maternal undernutrition in early gestation alters molecular regulation of the...

Maternal Undernutrition in Early Gestation Alters MolecularRegulation of the Hypothalamic-Pituitary-Adrenal Axis inthe Ovine Fetus

P. Hawkins,* M. A. Hanson* and S. G. Matthews²*Departments of Obstetrics and Gynaecology and Physiology, University College London, London, UK.

²Departments of Physiology and Obstetrics and Gynaecology, University of Toronto, Toronto, Ontario, Canada.

Key words: nutrition, hypothalamus, pituitary, fetal sheep, programming.

Abstract

We have demonstrated previously that plasma adrenocorticotropin hormone and cortisol responsesto exogenous and endogenous stimuli are reduced in fetuses of mildly undernourished ewes. In thepresent study, we examined the molecular regulation of fetal hypothalamic-pituitary-adrenal (HPA)axis function at 127±130 days gestation (dGA) following 15% reduction in maternal nutrition between0 and 70 dGA. Using in situ hybridization, we found that corticotropin releasing hormone (CRH)mRNA expression in the hypothalamic paraventricular nucleus (PVN) was lower in fetuses fromnutrient restricted ewes than in controls. Restricted fetuses also had greater levels of mRNAencoding preproenkephalin (PENK) and magnocellular arginine vasopressin (AVP) in the PVN.Expression of oxytocin mRNA and parvocellular AVP mRNA in the PVN and pro-opiomelanocortinmRNA in the pituitary were unchanged. Glucocorticoid receptor mRNA expression was unalteredat the PVN, but was reduced (>40%) in the anterior pituitary of restricted fetuses. Northern blotanalysis demonstrated that levels of adrenal P450scc mRNA and P450C17 mRNA were notdifferent between the groups. We conclude that the reduction in HPA function reported previouslyis mediated, at least in part, by a decrease in expression of CRH mRNA and increase in PENKmRNA in the PVN.

Epidemiological studies suggest that impaired fetal develop-ment is associated with increased risk of adult disease (1).Fetal development may be altered by environmental in¯u-ences acting during pregnancy. Maternal undernutrition hasbeen suggested as an adverse environmental in¯uence whichcould alter development of fetal organ systems, producingpermanent physiological changes, and increasing the risk oflater disease (1). Studies in rats (2, 3), guinea-pigs (4) andsheep (5) have shown that even modest maternal under-nutrition can alter development of cardiovascular andneuroendocrine systems in the offspring, and lead tohypertension in postnatal life.

Adverse intrauterine in¯uences have been linked to altereddevelopment of other systems, including the hypothalamic-pituitary-adrenal (HPA) axis. In humans, impaired fetaldevelopment, indicated by reduced size at birth, is associatedwith greater rates of urinary glucocorticoid excretion inchildren (6), elevated basal plasma cortisol concentrations (7)

and greater adrenocortical responsiveness to adrenocortico-tropin hormone (ACTH) in adults (8). In fetal sheep, surgicalrestriction of placental size produces a decrease in the level ofpro-opiomelanocortin (POMC) mRNA in the fetal pituitary(9). Changes in fetal HPA axis function could have importantconsequences for fetal organ maturation during the ®nalweeks of pregnancy (10, 11). Postnatally, altered function mayaffect the ability of offspring to mount appropriate stressresponses and to preserve a stable internal environment.Altered HPA axis function could also affect cardiovascularregulation, as it is known that glucocorticoids can producehypertension in sheep (12).

ACTH is synthesized in the ovine fetal pituitary gland fromthe precursor peptide POMC (13). Its release is under thecontrol of stimulatory inputs from the hypothalamus, includ-ing corticotropin releasing hormone (CRH) and argininevasopressin (AVP). These neuropeptides are released fromhypothalamic neurones primarily in the paraventricular

Correspondence to: Dr S. G. Matthews, Department of Physiology, Faculty of Medicine, University of Toronto, Medical Sciences Building,

1 King's College Circle, Toronto, Ontario, M5S 1A8, Canada (e-mail: [email protected]).

Journal of Neuroendocrinology, 2001, Vol. 13, 855±861

# 2001 Blackwell Science Ltd

nucleus (PVN), into the hypophyseal-portal circulation toregulate pituitary corticotroph function (14). Other inputsinclude preproenkephalin (PENK) (15) and oxytocin (16).Negative-feedback inhibition of pituitary ACTH secretion ismediated by circulating glucocorticoids which can act atglucocorticoid receptors (GR) at both the pituitary andhypothalamus (17, 18). At the fetal adrenal gland, ACTHstimulates steroidogenesis and secretion of cortisol (14), aprocess which is regulated, in part, by levels of the steroido-genic enzymes, cholesterol side-chain cleavage (P450scc) and17a-hydroxylase (P450C17) (19).

We have demonstrated previously that mild maternalundernutrition during early pregnancy alters developmentof the fetal HPA axis, producing a reduction of plasma ACTHand cortisol responses to both exogenous (20) and endogenous(21) stimuli in late gestation. The aim of the present investiga-tion was to examine if the reduced HPA responses weremediated by changes in the molecular mechanisms controllingfetal HPA axis function. Here, we examine the effects of mildmaternal undernutrition in early pregnancy on the expressionof mRNA encoding CRH, AVP, PENK, oxytocin and GR inthe hypothalamic PVN, POMC and GR in the pituitarygland, and P450scc and P450C17 at the adrenal gland, in fetalsheep at 127±130 days gestation (dGA). We have focusedour study on investigation of mRNA levels, since extensiveliterature exists on the molecular development of theseneuropeptide systems in the fetal sheep hypothalamus andpituitary (13, 15, 16, 22±24). The involved nature of theseexperiments and the limited sample size precluded us frommeasuring corresponding protein levels. However, there isgood evidence that close correlations exist between mRNAlevels and their corresponding neuropeptides in the fetal sheephypothalamus and pituitary (13, 22, 25).

Methods

Animals and dietary manipulation

Before conception, time-mated Welsh Mountain ewes of uniform age, weightand body condition score were randomly assigned to either the control or thenutrient restricted group. The animals used in this study were housed inindividual pens to permit speci®c regulation of individual nutritional intake.The ¯oors of all pens were covered with wood shavings, and the animals werefed a complete pelleted diet which was regulated depending on the protocol.Animals were allowed free access to water. The diet consisted of barley, wheat,cooked cereal meal, micronized full fat soya, grass meal, molasses, choppedstraw, calcium carbonate, dicalcium phosphate, salt and a sheep vitamin/mineral supplement. It provided 10.81 mJ/kg metabolizable energy, 149.8 g/kgcrude protein and contained 88.4% dry matter. Rations were allocated basedupon recommendations made by an advisory manual prepared by the AFRCTechnical Committee on responses to nutrients (26). The diet rations wereadjusted according to bodyweight, condition score and stage of gestation.Nutrient intake was regulated by reducing the amount of the recommendeddaily ration. All components of the diet were reduced by the same degree. Allprocedures were approved by the Home Of®ce and were conducted inaccordance with the Animals (Scienti®c Procedures) Act 1986.

Control ewes were fed 100% of their recommended nutritional require-ments for the whole of gestation. Restricted animals received 85% of theirrecommended nutrient requirements from the time of conception until day 70of gestation, and 100% of their requirements thereafter. Thus, restricted ewesreceived 15% less compared to controls for the ®rst 70 days of gestation only.This manipulation resulted in reduced bodyweight during the period ofrestriction, but the bodyweight of ewes returned to normal soon after theend of the restriction period (20).

The present study was performed at 127±130 dGA, which allows direct com-parison with the data from our previous studies (20, 21). These comparisons

are based on the assumption that similar changes in fetal HPA axis functionoccurred in both studies. Due to the possible confounding effects of surgicalstress, fetuses were not chronically catheterized in the present study, andthus it was not possible to make measurements of plasma ACTH andcortisol concentrations.

Tissue collection

At 127±130 dGA, control (control; n=5) and nutrient restricted (restricted;n=5) ewes were killed by maternal i.v. injection of pentobarbitone (40 mlEuthatal i.v.; RhoÃne MeÂrieux, Harlow, Essex, UK). Fetuses were removedand the hypothalamus and pituitary gland were dissected rapidly and frozenon dry ice. The fetal adrenal glands were also collected and frozen usingliquid nitrogen. All tissues were stored at x80uC prior to analysis. Levels anddistribution of mRNA for CRH, AVP, GR, oxytocin and PENK in the hypo-thalamic PVN, and POMC and GR in the pituitary gland, weredetermined by in situ hybridization. Levels of P450scc mRNA and P450C17

mRNA at the adrenal gland were determined by Northern blot analysis,which has been shown previously to be an effective method of measuringsteroidogenic enzymes in the fetal adrenal gland (27). Also, speci®c regionalanalysis (i.e. in situ hybridization) was not required because steroidogenicenzymes are distributed throughout the adrenal cortex.

In situ hybridization

The technique for in situ hybridization has been described previously (22).Brie¯y, coronal cryosections (12 mm) were mounted onto (poly)-L-lysine(Sigma Chemical Company, St Louis, MO, USA)-coated slides and post®xedin paraformaldehyde (4%). The oligonucleotide probes were labelled usingterminal deoxynucleotidyl transferase (Gibco BRL, Burlington, Ontario,Canada) and [35S]deoxyadenosine 5k-(thio) triphosphate (1300 Ci/mmol; NEN,Du Pont Canada Inc., Mississauga, Ontario, Canada) to a speci®c activityof 1.0r109 c.p.m./mg. The labelled probe was used at a concentration of1.0r103 c.p.m./ml. Probe in hybridization buffer (200 ml) was applied to eachslide and incubated overnight in a moist chamber at 42uC. Sections were thenwashed [1rSSC at room temperature (30min) and 1rSSC at 55uC (30min)],dehydrated, air-dried, and exposed to X-ray ®lm (Biomax, Eastman Kodak,Rochester, NY, USA). The antisense CRH, AVP, GR, PENK, POMC andoxytocin deoxyribonucleotide probes were 45 bases and have been usedextensively in previous studies (13, 15±17, 22).

Northern analysis

The methods for Northern analysis have been described previously (17, 28).Brie¯y, total RNA was extracted using Trizol (Gibco BRL). The integrity ofthe RNA was assessed after electrophoresis on a 1% (w/v) agarose gel, stainingwith ethidium bromide. Only intact RNA samples were used. Electrophoresisof hydrolysed RNA (30 mg) was performed on agarose (1% w/v), formaldehydegels, and was followed by capillary transfer to a nylon membrane (GibcoBRL). The RNA was covalently cross-linked to the membrane by ultravioletlight (GS Gene Linker, Bio-Rad, Hercules, CA, USA).

The membrane was prehybridized (42uC, 6 h) in a buffer containingformamide (50% v/v), SDS (7% w/v), 4rSSPE and denatured salmon sperm(0.5 mg/ml). The membranes were then hybridized overnight with oligo-nucleotide probes encoding steroidogenic enzymes. The difference in loadingand transfer among different lanes was assessed by hybridizing the blots withan oligonucleotide encoding 18S ribosomal RNA. The 45-mer oligonucleotideprobes were synthesized based on the DNA sequences for ovine P450C17

and P450scc mRNAs (27). The probes were 5k end labelled with [P32]ATP(3000 Ci/mmol; DuPont NEN, Mississauga, Ontario, Canada) using T4polynucleotide kinase (Promega, Madison, WI, USA). Probes were added(1.0±1.5r106 c.p.m./ml buffer) to the membranes, hybridized overnight,and then washed at the highest stringency of 1rSSC (42uC). Blots werethen exposed to X-ray ®lm (XAR 5, Kodak). Results are expressed as a ratioof adrenal enzyme mRNA to 18S rRNA. Northern analysis of total RNAextracted from fetal sheep liver was used as a negative control and demon-strated no hybridization when processed under the same conditions as theadrenal tissue (not illustrated).

Data analysis

In each analysis, sections from control and restricted ewes were processedsimultaneously to allow direct comparison between groups. The relativeoptical density of the signal on autoradiographic ®lm was quanti®ed, after

856 Nutrition and fetal HPA axis development

# 2001 Blackwell Science Ltd, Journal of Neuroendocrinology, 13, 855±861

subtraction of background values, using a computerized image analysis system

(Imaging Research Inc., St Catherines, Ontario, Canada). The value obtained

represents an average density over the area measured. For the in situ

hybridization experiments, sections were exposed together with 14C standards

(American Radiochemical Company, St Louis, MO, USA) to ensure analysis

in the linear region of the autoradiographic ®lm. Comparison between groups

was performed using the values from ®ve to 10 sections per animal.For pituitary POMC mRNA and GR mRNA, the level of the pituitary at

which analysis was undertaken was consistent throughout as described

previously (13, 17). Because POMC mRNA was regionally distributed within

the anterior pituitary, analysis of the superior (region around the pars

intermedia) and inferior (region at the base of the anterior pituitary) zones was

performed. Analysis of GR mRNA was performed using the entire anterior

pituitary as there was uniform distribution (17).For hypothalamic CRH mRNA and AVP mRNA analysis, the hypotha-

lamic PVN was split coronally into three regions: ventral, central, and dorsal

(22). For CRH mRNA, which is localized in parvocellular neurones, the entire

cross section of PVN was analysed at each level. For data presentation, the

values from each region were averaged to give mean values for the whole PVN.

AVP mRNA is contained in parvocellular and magnocellular ®elds of the

ovine PVN (22), and thus these areas were analysed separately. Such analysis

was possible, in as much as the extreme lateral ®elds of the PVN represent

almost exclusively magnocellular neurones, whereas the medial area adjacent

to the third ventricle is predominantly parvocellular (22). However, complete

separation of the magnocellular and parvocellular regions is not possible

at this time. AVP mRNA levels in parvocellular and magnocellular ®elds

within each animal were combined to give a total value at the ventral, central

and dorsal regions. For data presentation, the values for AVP mRNA in the

ventral, central and dorsal regions were averaged to give mean values for

the whole PVN for magnocellular, parvocellular and total AVP mRNA. For

GR mRNA, PENK mRNA and oxytocin mRNA, the PVN was split coronally

into two regions: ventral and dorsal, and analysed as for CRH mRNA.

Separate analysis of oxytocin mRNA in magnocellular and parvocellular

®elds was not performed as examination demonstrated that oxytocin

mRNA-containing magnocellular neurones were localized in parvocellular

®elds of the PVN, as reported previously (16). Oxytocin mRNA and GR

mRNA levels are presented for the whole PVN. PENK mRNA levels are

presented for the separate ventral and dorsal regions in addition to a mean

for the whole PVN. All values are presented as meanstSEM. Where

displayed, `n' indicates the number of fetuses in each group. Differences

between control and restricted fetuses were analysed by Student's unpaired

t-test. P<0.05 was considered statistically signi®cant.

Results

Hypothalamus

When the PVN was analysed as a whole, CRH mRNA levelswere signi®cantly lower in restricted fetuses (n=5) than incontrol fetuses (n=5) (P<0.05) (Fig. 1). There were nodifferences between the groups for total AVP mRNA orparvocellular AVP mRNA expression (Fig. 2). However,when magnocellular AVP mRNA was analysed separately, itwas found that levels were signi®cantly greater in restrictedfetuses than in control fetuses (P<0.05) (Fig. 2). PENKmRNA levels were also signi®cantly greater in restrictedfetuses than in control fetuses in the dorsal region of the PVN(P<0.05) (Fig. 3). Total and ventral PENK mRNA expres-

sion did not differ between the groups (Fig. 3). The relativeoptical density of oxytocin mRNA was 2.61t0.14 in controlfetuses and 2.58t0.21 in restricted fetuses, and did not differbetween the groups (not illustrated). Total GR mRNAexpression at the PVN was also similar in control andrestricted fetuses (Fig. 4).

Pituitary

Pituitary POMC mRNA expression is illustrated in Table 1.POMC mRNA levels did not differ between control andrestricted fetuses in any area of the anterior lobe or in the

TotalC

RH

mR

NA

(R

OD

)

(B)

1.0

0

0.5

1.5

2.0

2.5

C

R

(A)

C R S

FIG. 1. (A) Representative autoradiographs of coronal sections of theparaventricular nucleus (PVN) showing expression of corticotropinreleasing hormone (CRH) mRNA in control (C) and nutrient restricted(R) fetuses. (S) An adjacent section hybridized with a sense oligonucleo-tide probe (no signal was detected). (B) Densitometric analysis of CRHmRNA levels in the PVN of control (%, n=5) and nutrient restricted (&,n=5) fetuses following in situ hybridization. Values are meantSEMrelative optical density (ROD). *P<0.05 (unpaired t-test) control versusrestricted. CRH mRNA levels were signi®cantly lower in restricted fetusescompared to control fetuses.

Total

AV

P m

RN

A (

RO

D)

15

0

5

20C

R

10

Magnocellular

15

0

5

20

10

Parvocellular

15

0

5

20

10

FIG. 2. Densitometric analysis of arginine vasopressin (AVP) mRNAin the paraventricular nucleus (PVN) of control (C) (%, n=5) andrestricted (R) (&, n=5) fetuses following in situ hybridization. Valuesare meantSEM relative optical density (ROD). *P<0.05 (unpairedt-test) control versus retrsicted. Magnocellular AVP mRNA levelswere signi®cantly greater in restricted fetuses compared to controlfetuses.

Nutrition and fetal HPA axis development 857


intermediate lobe. Levels of POMC mRNA were higher inthe intermediate lobe than in the anterior lobe in both C andR fetuses as reported previously (28). In contrast to the

hypothalamus, expression of GR mRNA in the anterior lobe

of the pituitary was signi®cantly lower in restricted fetusesthan in control fetuses (Fig. 5).

Adrenal

Northern blot analysis identi®ed single transcripts for P450scc

(size 1.9 kb) and P450C17 (size 1.7 kb) in adrenal tissue fromall groups of fetuses. Expression of P450scc mRNA appearedto be lower in restricted fetuses than in control fetuses;however, this was not signi®cant (P<0.06) (Fig. 6). There wasno difference in expression of P450C17 mRNA betweencontrol and restricted fetuses (Fig. 6).

Discussion

We have reported previously that fetal plasma ACTH andcortisol responses to stimulation by exogenous CRH+AVPor ACTH (20), or acute isocapnic hypoxaemia (21) arereduced following a modest level of maternal undernutritionin early gestation. These data suggested that pituitary andadrenal responsiveness was blunted in fetuses of under-nourished ewes. The major ®ndings of the present study arethat, at the fetal PVN, the same level of maternal under-nutrition produced a reduction of parvocellular CRH mRNA

Total

PE

NK

mR

NA

(R

OD

)

2.5

3.0

C

R

Ventral Dorsal

3.5

FIG. 3. Densitometric analysis of preproenkephalin (PENK) mRNA in theparaventricular nucleus (PVN) of control (C) (%, n=5) and restricted (R)(&, n=5) fetuses following in situ hybridization. Values are meantSEMrelative optical density (ROD). *P<0.05 (unpaired t-test) control versusretrsicted. PENK mRNA levels in the dorsal region of the PVN weresigni®cantly greater in restricted fetuses compared to control fetuses.

Total

GR

mR

NA

(R

OD

)

0

1

C

R

3

2

FIG. 4. Densitometric analysis of glucocorticoid receptor (GR) mRNA inthe paraventricular nucleus (PVN) of control (C) (%, n=5) and restricted(R) (&, n=5) fetuses. Values are meantSEM relative optical density(ROD).

Total

GR

mR

NA

(R

OD

)

1.0

0

0.5

1.5

C

R

C R

FIG. 5. (A) Representative autoradiographs of coronal sections of theanterior pituitary showing expression of glucocorticoid receptor (GR)mRNA in control (C) and nutrient restricted (R) fetuses. (B)Densitometric analysis of pituitary GR mRNA in the anterior lobe ofcontrol (%, n=5) and restricted (&, n=5) fetuses following in situhybridization. Values are meantSEM relative optical density (ROD).*P<0.05 (unpaired t-test) control versus restricted. GR mRNA levelswere signi®cantly lower in restricted fetuses compared to control fetuses.

TABLE 1. Densitometric Analysis of Pituitary Pro-Opiomelanocortin (POMC) mRNA in the Anterior Lobeand Intermediate Lobe in Control (n=5) and Restricted(n=5) Fetuses.

POMC mRNA (ROD)

Anterior lobe Intermediate lobe

Total Superior Inferior Total

Control 1.02t0.08 0.73t0.08 1.27t0.09 32.92t1.25Restricted 1.04t0.18 0.83t0.16 1.22t0.21 30.49t0.03

Values are meantSEM relative optical density (ROD).



expression, and elevation of PENK mRNA and magno-cellular AVP mRNA. In addition, pituitary GR mRNAexpression was reduced in fetuses of nutritionally challengedewes. At the adrenal gland, the expression of P450scc mRNAtended to be reduced in restricted fetuses; however, this justfailed to attain signi®cance. These data, together withprevious studies (20, 21), suggest roles for CRH and PENKat the PVN in mediating the reduction of fetal HPA axisresponsiveness in moderately nutrient restricted mothers.

The reduction of CRH mRNA demonstrated in the presentstudy could have important functional consequences for fetalHPA axis regulation. Studies in well-fed animals have shownthat CRH is one of the primary endogenous factors whichstimulates synthesis and secretion of ACTH from the fetalpituitary (14). In the fetal sheep, CRH mRNA and immuno-reactive (ir)-CRH can be detected at the PVN by 50±60 dGA(19, 22, 29, 30), and CRH containing cell bodies, with axonsprojecting from the PVN to the median eminence, have beenreported to be present at 90 dGA (31). Exogenous CRH canalso activate pituitary function and cause increases in plasmaACTH concentration prior to 100 dGA (32, 33). Levels ofCRH mRNA and ir-CRH at the PVN increase during lategestation (22, 24, 34) in parallel with the increases inpituitary POMC mRNA expression (13, 24) and plasmaACTH concentrations (35). Furthermore, CRH mRNAexpression in the PVN increase rapidly during fetal hypoxia(22). Thus, CRH probably functions as a regulator of fetalpituitary activity from an early age, and appears to have arole in maturation of the pituitary corticotroph (36).

A decrease in CRH mRNA expression in the PVN, asoccurs following maternal undernutrition in the presentstudy, is consistent with the reduction in stimulated fetalHPA axis activity that we demonstrated following CRH/AVP,ACTH and hypoxic challenge (20, 21). This effect does notappear to be mediated through a decrease in pituitary POMCmRNA expression, but could have involved changes inPOMC translation or processing mechanisms, or changes inACTH secretion. A recent study in the rat has demonstratedthat maternal undernutrition during pregnancy results inneonatal offspring that exhibit reduced CRH mRNA expres-sion and decreased plasma ACTH concentrations, but nochange in pituitary POMC mRNA (37).

The reduction in CRH mRNA expression in the PVNcould result from an increase in glucocorticoid-mediated

negative-feedback. CRH neurones in the PVN of the fetalsheep are highly sensitive to glucocorticoid negative-feedback(18, 22, 38). In the present study, GR mRNA expression inthe PVN was not signi®cantly different between the groups,suggesting that sensitivity to circulating glucocorticoids wasnot elevated at this level of the axis. However, changes in GRbinding in the PVN, although not measured in the presentstudy, may have been modi®ed in nutrient-restricted animals.It is also possible that glucocorticoid feedback sensitivity isaltered at the hippocampus in this group of animals, leading toincreased inhibition of CRH production in the PVN. Thefetal sheep hippocampus expresses high levels of GR in thesecond half of gestation (18), and is known to tonically inhibitpituitary-adrenocortical activity (39).

CRH-mediated activation of pituitary function could alsohave been altered by changes in the numbers of pituitaryCRH receptors. Investigation of CRH binding to anteriorpituitary membrane preparations has demonstrated that thenumber of binding sites increases progressively from 65 to70 dGA to reach a maximum at 125±130 dGA (40), a patternwhich follows a similar time-course to changes in pituitaryCRH responsiveness (35). Thus, a reduction in pituitary CRHreceptor number could have contributed to the decrease inpituitary responsiveness demonstrated previously (20, 21).

As for CRH, expression of PENK mRNA has been detec-ted in parvocellular ®elds of the fetal PVN by 60 dGA, and it isthought that the two peptides may coexist in a proportion ofparvocellular neurones (15, 41). Expression of PENK mRNAincreases initially during gestation in parallel with levels ofCRH mRNA (15). However, expression then falls in the weekprior to term. At the hypothalamus, the majority of PENK isprocessed to methionine enkephalin (Met-Enk) (42), and fetalplasma concentrations of Met-Enk increase during gestation(43, 44). In the rat, it has been proposed that Met-Enk couldhave a dual effect on HPA axis function, stimulating pituitaryACTH secretion, but inhibiting CRH release at the medianeminence (45, 46). Thus, the decrease in PENK mRNA expres-sion at the PVN that occurs near term in fetal sheep maycontribute to a maintenance of CRH release and pituitaryactivation (15). In the present study, PENK mRNA expres-sion was elevated in the dorsal region of the PVN in fetuses ofnutritionally restricted ewes. This may result in an inhibitionof release of CRH from the hypothalamus. This effect couldhave important consequences for fetal HPA axis function andmay be an additional factor which contributes to the reduc-tion of fetal HPA axis activity (20, 21). The differences in thehypothalamic CRH and PENK mRNA pro®les in the fetusesfrom restricted mothers (i.e. higher PENK mRNA, lowerCRH mRNA), may be suggestive of developmental delayin these fetuses. However, subsequent studies indicate thatthere is no difference in gestation length between control andrestricted fetuses.

Maternal undernutrition also led to a small but signi®cantincrease in magnocellular AVP mRNA levels in the fetalPVN. Previous studies have shown that magnocellular AVPneurones are rather inert with respect to stimulus (i.e.dehydration)-induced changes in AVP mRNA in the sheep(47), and so relatively small changes in AVP mRNA, asdescribed in the present study, may be of physiologicalsigni®cance. In parallel studies, we have shown that, following

P45

0 SC

C m

RN

A/1

8S r

RN

A

0

0.5

C

R

2.0

1.0

1.5

P45

0 C17

mR

NA

/18S

rR

NA

0

1

3

2

FIG. 6. Ratios of P450scc mRNA and P450C17 mRNA to 18S rRNA inadrenal glands of control (C) (%, n=5) and restricted (R) (&, n=5)fetuses. Values are meantSEM.



maternal undernutrition, arterial blood pressure is lower infetal life, but becomes elevated postnatally (5, 48).

Levels of GR mRNA in the anterior pituitary weresigni®cantly lower in fetuses of undernourished ewes than innormally-fed controls. A functional role for the pituitary inmediating negative-feedback control of ACTH release hasbeen shown in fetal sheep in which the hypothalamus andpituitary are disconnected surgically. Following this pro-cedure, stimulated, but not basal, ACTH release can beinhibited by cortisol infusion (49). Thus, it has been suggestedthat the pituitary is important for feedback control ofstimulated ACTH secretion, but that regulation of basalACTH production is mediated by higher centres such as thehypothalamus (49). Levels of GR mRNA at the anteriorpituitary are at relatively low levels through most of lategestation, but increase at term (17). It is suggested that thispattern results in a low degree of feedback during late gesta-tion, allowing maintained secretion of ACTH and cortisol.The reduction of GR mRNA expression in the present studywould be expected to result in a decrease in negative-feedbackeffects at the pituitary and, consequently, in greater ACTHrelease following stimulation. However, in our previousinvestigation we found that stimulated ACTH responses werereduced following maternal undernutrition (20, 21). Thus, theeffect of this change in pituitary GR mRNA expression is notclear. However, it is not known whether the decrease in pitui-tary GR mRNA expression was localized to corticotrophs, orif it re¯ected changes within other cells types at the pituitary

Maternal nutrient restriction, identical to that in the presentstudy, resulted in reduced fetal adrenocortical responsesto exogenous ACTH administration, suggesting a reductionin adrenal sensitivity (20). During fetal development, thereare three phases of adrenal responsiveness, with high activ-ity during early (40±50 dGA) and late gestation, but lowresponsiveness in mid-gestation (90±100 dGA) (14, 50±52).This pattern is mirrored by changes in levels of mRNA andimmunoreactive protein for P450scc and P450C17 at theadrenal cortex (19, 53). Thus, these enzymes are consideredto be important determinants of steroid synthesis andadrenocortical responsiveness. In contrast, the adrenal hypo-responsive period occurs during early neonatal life in therat (54). In the present study, levels of mRNA for P450scc orP450C17 did not differ signi®cantly between control andrestricted fetuses but there was a tendency for P450scc mRNAlevels to be lower in restricted fetuses.

In conclusion, a modest level of maternal undernutritionin early pregnancy can alter molecular regulation of thefetal HPA axis in late gestation. In previous studies, wedemonstrated that fetuses of nutritionally restricted ewes hadreduced plasma ACTH and cortisol responses to exogenousand endogenous stimuli. Our data suggest that these effectswere mediated by a decrease in CRH mRNA and an increasein PENK mRNA expression at the PVN. This is particularlystriking, as the level of maternal undernutrition used does notproduce any overt effects on fetal blood gas status, fetal bodyweight or organ size (21). The changes in molecular regulationmay represent a permanent resetting of HPA axis controlmechanisms, which could have consequences for HPA axisfunction in postnatal life.

Acknowledgements

This work was funded by The Wellcome Trust and Medical Research Council.P. H. was supported by a Bogue Research Fellowship from University CollegeLondon. Studies undertaken in Toronto (S. G. M.), were supported, in part, bythe Natural Sciences and Engineering Research Council of Canada.

Accepted 2 July 2001

References

1 Barker DJP. Mothers Babies and Health in Later Life. Edinburgh:Churchill Livingstone, 1998.

2 Langley-Evans SC, Phillips GJ, Benediktsson R, Gardner DS,Edwards CRW, Jackson AA, Seckl JR. Protein intake in pregnancy,placental glucocorticoid metabolism and the programming of hyper-tension in the rat. Placenta 1996; 17: 169±172.

3 Ozaki T, Nishina H, Hawkins P, Crowe C, Poston L, Hanson MA.Isolated systemic resistance vessel function in hypertensive male ratoffspring of mildly nutritionally restricted dams. J Physiol 2001;513: 118P.

4 Lingas R, Matthews SG. A short period of maternal nutrient restrictionin late gestation modi®es pituitary-adrenal function in adult guinea pigoffspring. Neuroendocrinology, 2001; 73: 302±311.

5 Hawkins P, Crowe C, Calder NA, Saito T, Ozaki T, Stratford LL,Noakes DE, Hanson MA. Cardiovascular development in late gestationfetal sheep and young lambs following modest maternal nutrientrestriction in early gestation. J Physiol 1997; 505: 18P.

6 Clark PM, Hindmarsh PC, Shiell AW, Law CM, Honour JW,Barker DJP. Size at birth and adrenocortical function in childhood.Clin Endocrinol (Oxf) 1996; 45: 721±726.

7 Phillips DI, Barker DJ, Fall CH, Seckl JR, Whorwood CB, Wood PJ,Walker BR. Elevated plasma cortisol concentrations: a link between lowbirthweight and the insulin resistance syndrome. J Clin Endocrinol Metab1998; 83: 757±760.

8 Reynolds RM, Walker BR, Syddall HE, Andrew R, Wood PJ,Whorwood CB, Phillips DI. Altered control of cortisol secretion inadult men with low birth weight and cardiovascular risk factors. J Clin

Endocrinol Metabol 2001; 86: 245±250.9 Phillips ID, Simonetta G, Owens JA, Robinson JS, Clarke IJ,

McMillen IC. Placental restriction alters the functional developmentof the pituitary-adrenal axis in the sheep fetus during late gestation.Pediatr Res 1996; 40: 861±866.

10 Fowden AL. Endocrine regulation of fetal growth. Reprod Fertil Dev1995; 7: 351±363.

11 Liggins GC. The role of cortisol in preparing the fetus for birth. ReprodFertil Dev 1994; 6: 141±150.

12 Tangalakis K, Lumbers ER, Moritz KM, Towstoless M, Wintour EM.Effect of cortisol on blood pressure and vascular reactivity in the ovinefetus. Exp Physiol 1992; 77: 709±717.

13 Matthews SG, Han X, Lu F, Challis JRG. Developmental changes inthe distribution of pro-opiomelanocortin and prolactin mRNA in hepituitary of the ovine fetus and lamb. J Mol Endocrinol 1994; 13: 175±185.

14 Challis JRG, Matthews SG, Gibb W, Lye SJ. Endocrine and paracrineregulationof birth at term, and preterm. Endocr Rev 2000; 21: 514±550.

15 Matthews SG, Challis JR. Developmental regulation of preproenkepha-lin mRNA in the ovine paraventricular nucleus: effects of stress andglucocorticoids. Dev Brain Res 1995; 86: 259±267.

16 Matthews SG. Hypothalamic oxytocin in the developing ovinefetus: interaction with pituitary-adrenocortical function. Brain Res1999; 820: 92±100.

17 Matthews SG, Yang K, Challis JRG. Changes in glucocorticoid receptormRNA in the developing ovine pituitary and the effects of exogenouscortisol. J Endocrinol 1995; 144: 483±490.

18 Andrews MH, Matthews SG. Regulation of glucocorticoid receptormRNA and heat shock protein 70 mRNA in the developing sheep brain.Brain Res 2000; 878: 174±182.

19 Tangalakis K, Coghlan JP, Connell J, Crawford R, Darling P,Hammond VE, Haralambidis J, Penschow J, Wintour EM. Tissuedistribution and levels of gene expression of three steroid hydro-xylases in ovine fetal adrenal glands. Acta Endocrinol Copenh 1989;120: 225±232.



20 Hawkins P, Steyn C, McGarrigle HH, Saito T, Ozaki T, Stratford LL,Noakes DE, Hanson MA. Effect of maternal nutrient restrictionin early gestation development of the hypothalamic-pituitary-adrenalaxis in fetal sheep at 0.8±0.9 of gestation. J Endocrinol 1999; 163:553±561.

21 Hawkins P, Steyn C, McGarrigle HH, Saito T, Ozaki T, Stratford LL,Noakes DE, Hanson MA. Effect of maternal nutrient restriction inearly gestation on responses of the hypothalamic-pituitary-adrenal axisto acute isocapnic hypoxaemia in late gestation fetal sheep. Exp Physiol2000; 85: 85±96.

22 Matthews SG, Challis JRG. Regulation of CRH and AVP mRNA in thedeveloping ovine hypothalamus: Effects of stress and glucocorticoids.Am J Physiol Endocrinol Metab 1995; 268: E1096±E1107.

23 Myers DA, Ding XY, Nathanielsz PW. Effect of fetal adrenalectomyon messenger ribonucleic acid for proopiomelanocortin in the anteriorpituitary and for corticotropin-releasing hormone in the paraventricularnucleus of the ovine fetus. Endocrinology 1991; 128: 2985±2991.

24 Myers DA, Myers TR, Grober MS, Nathanielsz PW. Levels ofcorticotropin-releasing hormone messenger ribonucleic acid (mRNA)in the hypothalamic paraventricular nucleus and proopiomelanocortinmRNA in the anterior pituitary during late gestation in fetal sheep.Endocrinology 1993; 132: 2109±2116.

25 Saoud CJ, Wood CE. Ontogeny and molecular weight of immunoreactivearginine vasopressin and corticotropin-releasing factor in the ovine fetalhypothalamus. Peptides 1996; 17: 55±61.

26 AFRC. Energy and Protein Requirements of Ruminants. An advisorymanual prepared by the AFRC technical committee on responses tonutrients. Wallingford, UK: CAB International, 1993.

27 Braems GA, Han VK, Challis JR. Gestational age-dependent changes inthe levels of mRNAs encoding cortisol biosynthetic enzymes and IGF-IIin the adrenal gland of fetal sheep during prolonged hypoxemia.J Endocrinol 1998; 159: 257±264.

28 Yang K, Hammond GL, Challis JR. Characterization of an ovineglucocorticoid receptor cDNA and developmentalchanges in its mRNAlevels in the fetal sheep hypothalamus, pituitary gland and adrenal. J Mol

Endocrinol 1992; 8: 173±180.29 Watabe T, Levidiotis ML, Old®eld B, Wintour EM. Ontogeny of

corticotrophin-releasing factor (CRF) in the ovine fetal hypothalamus:use of multiple CRF antibodies. J Endocrinol 1991; 129: 335±341.

30 Marcos P, Corio M, CovenÄas R, Tramu G. Neurotensin and tyrosinehydroxylase in the tuberoinfundibular system of the guinea pig hypo-thalamus with special emphasis to lactation status. Peptides 1996;17: 139±146.

31 Levidiotis M, Old®eld B, Wintour EM. Corticotrophin-releasing factorand arginine vasopressin ®bre projections to the median eminence of fetalsheep. Neuroendocrinology 1987; 46: 453±456.

32 MacIsaac RJ, Congiu M, Levidiotis M, McDougall JG, Wintour EM.In vivo regulation of adrenocorticotrophin secretion in the immatureovine fetus. Modulation by ovine corticotrophin releasing hormone andarginine vasopressin. J Dev Physiol 1989; 12: 41±47.

33 Hargrave BY, Rose JC. By 94 days of gestation plasma cortisol increasesblock ACTH response to hypotension in lamb fetuses. Am J Physiol 1985;249: E350±E354.

34 Currie IS, Brooks AN. Corticotrophin-releasing factors in the hypo-thalamus of the developing fetal sheep. J Dev Physiol 1992; 17: 241±246.

35 Norman LJ, Lye SJ, Wlodek ME, Challis JR. Changes in pituitaryresponses to synthetic ovine corticotrophin releasing factor in fetal sheep.Can J Physiol Pharmacol 1985; 63: 1398±1403.

36 Antolovich GC, McMillen IC, Robinson PM, Silver M, Young IR,Perry RA. The effect of hypothalamo-pituitary disconnection on thefunctional and morphologic development of the pituitary-adrenal axisin the fetal sheep in the last third of gestation. Neuroendocrinology 1991;54: 254±261.

37 Lesage J, Blondeau B, Grino M, Breant B, Dupouy JP. Maternalundernutrition during late gestation induces fetal overexposure toglucocorticoids and intrauterine growth retardation, and disturbs thehypothalamo-pituitary adrenal axis in the newborn rat. Endocrinology2001; 142: 1692±1702.

38 Myers DA, McDonald TJ, Dunn TG, Moss GE, Nathanielsz PW.Effect of implantation of dexamethasone adjacent to the paraventri-cular nucleus on messenger ribonucleic acid for corticotropin-releasinghormone and proopiomelanocortin during late gestation in fetal sheep.Endocrinology 1992; 130: 2167±2172.

39 McDonald TJ, Nijland MJ. Fornix transection increases plasmaACTH concentrations in response to hypotension in 120 day fetal sheep.Endocrinology 81st Meeting, Toronto, 2000: A881.

40 Lu F, Yang K, Challis JR. Characteristics and developmental changes ofcorticotrophin-releasing hormone-binding sites in the fetal sheep anteriorpituitary gland. J Endocrinol 1991; 130: 223±229.

41 Pretel S, Piekut D. Coexistence of corticotropin-releasing factor andenkephalin in the paraventricularnucleus of the rat. J Comp Neurol 1990;294: 192±201.

42 Almeida OF, Hassan AH, Harbuz MS, Linton EA, Lightman SL.Hypothalamic corticotropin-releasing hormone and opioid peptideneurons: functional changes after adrenalectomy and/or castration.Brain Res 1992; 571: 189±198.

43 Jones CT. Enkephalin peptides in fetal sheep, changes with gestation,origin and production by the placenta. J Dev Physiol 1992; 17: 15±20.

44 Simonetta G, Young IR, Coulter CL, Hey NJ, McMillen IC. Fetaladrenalectomy does not affect circulating enkephalins in the sheep fetusduring late gestation. Neuroendocrinology 1993; 57: 408±415.

45 Hashimoto K, Suemaru S, Ono N, Hattori T, Inoue H, Takao T,Sugawara M, Kageyama J, Ota Z. Dual effects of (D-Ala2,Met5)-enkephalinamide on CRF and ACTH secretion. Peptides 1987;8: 113±118.

46 Yajima F, Suda T, Tomori N, Sumitomo T, Nakagami Y, Ushiyama T,Demura H, Shizume K. Effects of opioid peptides on immunoreactivecorticotropin- releasing factor release from the rat hypothalamus in vitro.Life Sci 1986; 39: 181±186.

47 Matthews SG, Parrott RF, Sirinathsinghji DJS. Isolation- and dehydra-tion-induced changes in neuropeptide gene expression in the sheephypothalamus. J Mol Endocrinol 1993; 11: 181±189.

48 Hawkins P, Steyn C, Ozaki T, Saito T, Noakes DE, Hanson MA.Effect of maternal undernutrition in early gestation on ovine fetalblood pressure and cardiovascular re¯exes. Am J Physiol Regul Integrat

Comp Physiol 2000; 279: R340±R348.49 Ozolins IZ, Young IR, McMillen IC. Effect of cortisol infusion on

basal and corticotropin-releasing factor (CRF)-stimulated plasma ACTHconcentrations in the sheep fetus after surgical isolation of the pituitary.Endocrinology 1990; 127: 1833±1840.

50 Ozaki T, Hawkins P, Nishina H, Steyn C, Poston L, Hanson MA. Effectsof undernutrition in early pregnancy on systemic small artery function inlate-gestation fetal sheep. Am J Obstet Gynecol 2000; 183: 1301±1307.

51 Rose JC, Meis PJ, Urban RR, Greiss FC Jr. In vivo evidence forincreased adrenal sensitivity to adrenocorticotropin-(1-24) in the lambfetus late in gestation. Endocrinology 1982; 111: 80±85.

52 Wintour EM, Brown EH, Denton DA, Hardy KJ, McDougall JG,Oddie CJ, Whipp GT. The ontogeny and regulation of corticosteroidsecretion by the ovine foetal adrenal. Acta Endocrinol 1975; 79: 301±316.

53 Han X, Berdusco ET, Lu F, Challis JR. Immunolocalisationof P450 (C17) in the fetal sheep adrenal gland during gestationand in response to ACTH and glucocorticoid administration. EquineVet J 1997; 24 (Suppl.): 62±67.

54 Sapolsky RM, Meaney MJ. Maturation of the adrenocortical stressresponse: Neuroendocrine control mechanisms and the stress hypo-responsive period. Brain Res 1986; 396: 64±76.



maternal undernutrition in early gestation alters molecular regulation of the...

Documents