ELSEVIER Early Human Development 49 (1997) 7--17

Fetal and maternal plasma insulin-like growth factors and binding proteins in pregnancies with

appropriate or retarded fetal growth

Robert Holmes”, Raul Montemagnob, Jennifer Jones”, Michael Preece”, Charles Rodeckb, Peter Soothill”‘*

*Fetal Medicine Research Unit, St. Michael’s Hospital. Southwell Street, Bristol BS2 $EG, LrK ‘Department of Obstetrics and Gymecology, University College London Medical School. 86-96

Chenies Mews, London WClE 6HX, UK “Institute of Child Health, 30 Guildford Street, London WC1 NIEH, (JK

Received 12 July 1996: revised 2 December 1996: accepted 18 December 1996

Abstract

A prospective observational study of 104 women was performed to study whether the insulin-like growth factor (IGF) system in pregnancy before labour is associated with reduced fetal growth. Fetal blood was obtained by cordocentesis for prenatal diagnosis or at elective caesarean delivery and a maternal sample was also obtained. IGF-1 and IGF-2 and their binding proteins -1 and -3 were measured by RIA. The 35 cases were smaller than -2S.D.s by ultrasound abdominal circumference and birthweight and were subdivided into fetal growth retardation (FGR, n = 20) and small for gestational age (SGA, n = 15) by Doppler velocimetry and neonatal outcome. Controls (n = 69) were normally grown.

Control maternal IGF-1 (r = 0.65, P < 0.0001) and IGFBP-3 (r = 0.46, P = 0.001) increased with advancing gestational age. In FGR cases, maternal IGF-1 was low (P = 0.0001) and IGFBP-1 was high (P = 0.03) and maternal IGF-2 was low in SGA (P = 0.005). In the SGA fetus, IGF-2 was low (P = 0.0009) and IGFBP-3 (P = 0.02) was high. In FGR. IGFBP-1 (P < 0.0001) and IGFBP-3 (P = 0.002) were both elevated.

These data do not support the hypothesis that fetal IGF-1 deficiency is a common cause 01 FGR. Elevated binding proteins may lead to a relative deficiency of free IGF but changes in binding proteins may be secondary to metabolic changes. In FGR, maternal IGF-1 was low with high binding proteins, so this system may be important in controlling placental transfer. 0 1997 Elsevier Science Ireland Ltd.

*Corresponding author. Tel.: + 44 I 17 9285277; fax: + 44 117 9285683

03783782/97/$17.00 0 1997 Elsevier Science Ireland Ltd. All rights reserved PII SO378-37X2(97)01867-7

8 R. Holmes et al. I Early Human Development 49 (1997) 7- 17

Keywords: Maternal; Fetal growth retardation; Insulin-like growth factor

1. Introduction

Individuals born with a low birthweight for gestational age are at much greater risk of a wide range of medical problems and at present there is no effective therapy apart from premature delivery. Therefore, understanding the mechanisms controlling fetal growth and the causes of fetal growth retardation would be extremely important if this led to therapeutic options. Small fetal size is not a diagnosis but a physical sign and can reflect many different pathologies such as fetal infection, genetic abnor- malities and impairment of nutrient supply across the placenta. When any pathologi- cal cause is found then the term fetal growth retardation (FGR) is used. In this study the term FGR was specifically restricted to cases of impaired growth due to utero-placental insufficiency, all other pathologies having been excluded. Frequently no clear pathology is identified and the fetus is described as being small for gestational age (SGA) rather then FGR. These fetuses may be considered normal small, growing to their genetic growth potential.

Animal studies, such as embryo transplant and cross breeding experiments, have shown that, unlike postnatal growth, the growth of the fetus is controlled predomi- nately by the uterine environment (‘maternal constraint’), and not fetal or paternal genetic factors [ 11. Evidence that environmental factors control fetal growth in human pregnancy comes from dizygotic twin pregnancies in which fetal size is reduced compared to sibling singletons, probably because of limitations in placental perfusion or nutrient transfer.

In postnatal life, growth hormone (GH) plays a key role in regulating cellular proliferation and tissue growth. It has direct effects but also indirect actions mediated by insulin-like growth factors (IGFs) [2]. GH does not have a major role in controlling fetal growth because children with a congenital deficiency of GH have only a slight reduction in birth length but normal birthweight [3]. It is possible, therefore, that the IGFs may directly control fetal growth.

The IGFs are pro-insulin-like polypeptides which markedly stimulate cell division and differentiation and are present in fetal organs from early development. Messenger RNAs for the IGFs are found in many fetal tissues suggesting local production and so paracrine and/or autocrine actions [4]. However, the concentration of IGF messenger RNA is very high in the fetal liver and circulating levels of IGFs may reflect liver excretion and an endocrine role. The majority of IGFs in blood and tissues are bound to specific carrier proteins which extend the half-life and delay IGF clearance. Six binding proteins (IGFBPs) have been identified in humans [5]; IGFBP-3 binds the bulk of IGF in adult human serum and the smaller IGFBPl may help delivery of IGF to the two IGF receptors. Both receptors have a wide distribution in fetal tissues and the placenta.

The most pressing evidence for the importance of the IGFs in fetal growth and development comes from gene knockout experiments in mice. Allelic disruption of

R. Holmes et al. I Early Human Development 4Y (1997) 7- I7

the IGF genes, either Igfl or Igf2, leads to profound FGR in mice with birth size reduced by 40% [6,7]. In addition, null mutants in knockout experiments involving the mouse gene for the type 1 IGF receptor (Igflr), which mediates the fetal effects of both IGFs, die at birth with a 55% reduction in size which is a greater effect than seen with Igfl or Igf2 knockouts [8]. Recently, umbilical venous blood concentrations of IGF-1 have been shown to be low in small new-borns, especially in those with reduced growth velocity in utero [9].

The maternal and fetal IGF systems seem to be independent since there is no correlation between maternal and umbilical cord serum IGF- 1 levels [lo] and no passage of radiolabelled IGF in animal studies [ 111. Evidence that the maternal lGF system may relate to fetal growth comes from the demonstration that maternal constraint of fetal growth is removed by increasing maternal IGF-1 levels of pregnant mice by infusion or selective breeding [12]. Furthermore, administration of IGF- 1 results in increased placental amino acid uptake and production of mRNA for B placental transporter molecule (GLUT- 1) [ 131.

We therefore wished to study before labour whether the IGF system in either the mother or fetus is associated with reduced fetal growth by comparing IGFs and IGFBPs in maternal and fetal plasma of FGR and SGA pregnancies with appro- priately grown for gestational age (AGA) controls.

2. Materials and methods

We studied 87 pregnant women presenting to the Fetal Medicine Unit at University College London Hospital for prenatal diagnosis of a variety of pathologies between 18 and 40 weeks’ gestation. A detailed ultrasound scan for measurement of fetal size and detection of abnormalities was performed and fetal blood sampling from the umbilical vein was undertaken by ultrasound guided cordocentesis as described previously [14]. The vessel sampled was identified as umbilical vein by observing ultrasonically the direction of the turbulence produced by a saline injection and the fetal origin and purity of the sample was confirmed by analysis with a Coulter T 540 haematology analyser (Coulter Ltd, Luton, Bedfordshire, UK). Twenty millilitres of maternal venous blood was obtained at the same time. Ethical Committee approval and patient consent was obtained to collect the extra maternal and fetal blood required for the study and so paired samples were available.

In 35 of the 87 pregnancies the indication for prenatal diagnosis was small fetal size and all had an abdominal circumference (AC) measured by ultrasound which was smaller than - 2 standard deviations (S.D.) for gestational age (mean AC - 3.7 SD., range - 2.0 to - 7.44). Doppler velocimetry studies (Toshiba Sonolayer SSD-220’) and fetal blood gases (Radiometer ABL 520, Crawley. West Sussex, UK) were measured. None of these fetuses had any structural abnormality on ultrasound or postnatal assessment, the karyotype was subsequently reported as normal and fetal infection was not suspected either pre- or post-natally. These fetuses were therefore considered to be either small secondary to utero-placental insufficiency (fetal growth retardation- FGR) or entirely normal small for gestational age fetuses (SGA) growing

10 R. Holmes et al. I Early Human Development 49 (1997) 7- 17

to their genetic potential. Before the IGF results were known, these 35 cases were classified into FGR and SGA. All cases with abnormal umbilical/fetal Doppler studies or antenatal umbilical venous gases were assigned to the FGR group. Where these tests were normal or not available, one of us (PWS) reviewed all available information including antenatal (ultrasound growth velocity, amniotic fluid volume, fetal heart rate tracings), intrapartum (cardiotocograph recordings, fetal blood sampling, cord gases) and post-natal (hypoglycaemia, feeding, temperature regula- tion, intensive care requirements) to make a retrospective diagnosis. In this way, 20 fetuses were classified as FGR and 15 fetuses were classified as SGA.

The other 52 pregnancies in which fetal and maternal venous blood was taken for prenatal diagnosis acted as controls. In all these cases the fetus was appropriate size for gestational age and was subsequently shown not to be affected by the condition under investigation. Examples included choroid plexus cysts when the chromosomes were normal and maternal allo-immunisation when the fetus was antigen negative. Although some of these fetuses had markers of chromosomal abnormality, no fetus had a structural anomaly associated with abnormal growth. Such a group of fetuses was as ‘normal’ as is obtainable for a control population considering the invasive nature of fetal blood sampling. It was considered appropriate that maternal control samples should be obtained from the same pregnancies in order to maintain consistency with the paired maternal and fetal sample collection in the FGR and SGA groups. None of these mothers had evidence of glucose intolerance.

The control group was enlarged by 17 umbilical cord venous blood samples obtained at elective caesarean section at term performed for maternal indications (excluding diabetes mellitus) or abnormal fetal presentation. The maternal venous blood was obtained prior to pre-operative fasting. All the control fetuses were normally grown both by the abdominal circumference measurement and the birth- weight being larger than -2S.D.s. The pooling of two different groups of pregnancies to derive the control population was considered valid because no statistical differ- ences in IGF results were detected between them.

All samples were placed into heparinised tubes, centrifuged and the plasma was stored at - 70°C until analysis. IGF-1 and IGF-2 concentrations in plasma were measured by specific RIA after acid ethanol extraction to remove endogenous binding proteins. For IGF-1, a mouse monoclonal antibody (BPL-M23) was used as described by Morrell et al [15] with a sensitivity of 13.9 rig/ml, intra-assay coefficients of variation (CVs) of 9.0% at 35 rig/ml and 4.7% at 698 rig/ml and inter-assay CVs of 10.5% at 75 rig/ml and 5.1% at 698 rig/ml. The assay was unaffected by the addition of IGFBP-1, IGFBP-2 or IGFBP-3 at concentrations of 1000 rig/ml prior to extraction of binding proteins. IGF-2 was measured by the method of Blum et al. [16] with sensitivity of 0.18 rig/ml, an intra-assay CV of 3.9% at 520 rig/ml and inter-assay CV of 10.2% at 520 rig/ml. The assay did not cross react with IGFBP-2 (at 2 ,uglml), IGFBP-3 (at 3 pg/ml), pro-insulin (at 2 pg/ml), insulin (at 4.3 ,uglml) and growth hormone (at 2 ,uglml).

IGFBP-1 was measured by the method of Povoa et al. [17] with a sensitivity of 1 rig/ml and intra- and inter-assay CVs of 4.0% at 35 rig/ml and 6.2% at 55 rig/ml respectively. There was no cross reactivity with the addition of 1 pug/tube of

R. Holmes et al. I Early Human Development 49 (19971 7- 17

Table 1 Maternal plasma results in control, SGA and FGR pregnancies

Assays Maternal controls (n = 69)

Maternal SGA (n = 15)

Maternal FGR (n = 20)

IGF-I (rig/ml)

IGF-2 (n&/ml)

IGFBP-1 (nglml)

IGFBP-3 (rig/ml)

r = 0.65, P < 0.0001

486 (344-595)

133 (x6-206)

r = 0.46, P = 0.001

F = 2.48, NS Fig. 1 Low F= 17.1. P = o.ooo1 Fig. I

336 (277-394) 384 (336-441). 4S u = 134, Y = 0.005 166 (SO-256), 197 (130-307 1 NS Fig. 2 u = 429.

P = 0.03 Fig. 2 F = 2.39, NS F-0.21. NS

In maternal plasma, control IGF-I and IGFBP-3 were normally distributed after log transformation and both significantly changed with gestational age assessed by regression analysis. Analysis of covariance F-test was used to adjust for the effect of gestational age. IGF-2 and IGFBP-1 could not be transformed to normal distribution and did not change with gestational age. Medians and inter-quartile ranges (25th-75th) are presented for these factors and the P-values are the results of the Mann-Whitney U-test. (NS = P -; 0.0.5).

IGFBP-2 or IGFBP-3. IGFBP-3 was measured by an immunoradiometric assay (Diagnostic Systems Laboratories, Webster, TX, USA) with a sensitivity of 0.5 rig/ml. Irma-assay CVs were 6.1% at 1000 rig/ml and 4.4% at 9800 rig/ml and inter-assay CVs were 4.6% at 3500 rig/ml and 3.8% at 11000 rig/ml. One microgram per tube of IGF- 1, IGF-2, IGFBP-1 or IGFBP-2 were detected at concentrations of <3 rig/ml.

2. I. Statistical analysis

The control data were first tested for normal distribution by the Shapiro-Wilk test. Much of the data were not normally distributed despite log transformation so the non-parametric Mann-Whitney u-test was used to compare groups. The median and inter-quartile ranges are presented in the Table 1 and Table 2. Two assays in the maternal control samples changed significantly with gestational age but both these were corrected to normal distribution after log transformation. Therefore, for these two factors log transformed data were compared to controls by analysis of covariance (F-test) to remove the effects of gestational age. The correlation coefficient (rf and P-value are presented in Table 1.

3. Results

In the control group after log transformation maternal serum IGF-1 rose with advancing gestational age (I = 0.65, n = 45, P < 0.0001) and the derived reference ranges are displayed in Fig. 1. There was also an increase in control log transformed maternal IGFBP-3 levels (Y = 0.46, n = 45, P = 0.002) but no other component of the maternal or fetal IGF system correlated significantly with gestational age. The

12 R. Holmes et al. I Early Human Development 49 (1997) 7-17

Table 2 Fetal plasma results in control, SGA and FGR pregnancies

Assays

IGF-1 (nglml) IGF-2 (nglml)

IGFBP-1 (rig/ml)

IGFBP-3 (rig/ml)

Fetal controls (n = 69)

63 (45-96) 140 (106-192)

80 (28-l 11)

766 (416-1003)

Fetal SGA (n = 15)

74 (58-118), NS 96 (71-IO@, u = 177, P = 0.0009 91 (46-190), NS Fig. 3

1014 (805-1207), U=609, P = 0.02

Fetal FGR (n = 20)

83 (62-92), NS 123 (84-408) NS

159 (105-468), U = 687, P < 0.0001 Fig. 3 1023 (773-1705), (I = 864, P = 0.002

In fetal plasma, all the data were not normally distributed despite transformation and no changes with gestational age were found. Medians and inter-quartile ranges (25th-75th) are presented and the P-values are the results of the Mann-Whitney U-test. (NS = P > 0.05).

pooling of two different groups of pregnancies to derive the control population was considered valid because no statistical differences were detected between them. The median and inter-quartile ranges of the maternal controls are shown in Table 1 and for the fetal controls in Table 2.

Median and inter-quartile ranges of the FGR and SGA groups as well as differences from the controls are summarised in Table 1 and Table 2. In the maternal samples from the FGR cases but not the SGA cases, IGF-1 was low (F = 17.1, P = 0.0001, Fig. 1). Maternal IGF-2 levels were lower than controls in SGA cases

t;; z 0 ,X / I I , ( I 1 I I I

15 20 25 30 35 40

Gestational age (weeks)

Fig. 1. IGF-1 levels in plasma from mothers with small fetuses (X = FGR, Cl = SGA). Data are superimposed upon the back transformed mean and 95% confidence intervals for the control population. Individual control values are not presented. IGF-1 was significantly low in the mothers of FGR (F = 17.1, P = 0.0001) but not SGA fetuses.

R. Holmes et al. 1 Early Human Developmerzt 49 (19971 7- I7

800

E p 600

7 4 400

5 z 200

i z O

-........-....-.... . . .._..~...______......~............................... - .l...........l... -... I i / T x i

cl

15 20 25 30 35 40

Gestational age (weeks)

Fig. 2. IGFBP-1 in plasma from mothers with small fetuses (X = FGR, c1 = SGA) plotted against median and inter-quartile range (25th-75th) of the controls. IGF’BP-I was significantly high in mothers of FGR (U = 429, P = 0.03) hut not SGA fetuses.

(U = 134, P = 0.005) but just not significantly low in FGR cases (U = 260, P = 0.06). Maternal IGPBP-1 was high in FGR cases (U = 429, P = 0.03) but not in SGA cases (Fig. 2). No differences in maternal IGFBP-3 were found in either FGR or SGA compared to controls.

In the fetal samples IGF-1 was not significantly different to controls in either FGR or SGA cases and IGF-2 was low only in SGA fetuses (U = 177, P = 0.0009). Fetal IGFBP-1 (U = 687, P < 0.0001, Fig. 3) and IGFBP-3 (U = 864, P = 0.002) were

800

600

.----- -..- I- ..~...I._..._.. -..- ..__..._...__....__.........~~.................... I 1 X

c

25 35

Gestational age (weeks)

Fig. 3. IGFBP-1 in fetal plasma (X = FGR, 0 = SGA) plotted against the median and inter-quartile range (2Sth-75th) of the controls. IGFBP-I was significantly high in FGR (U = 6X7. P < 0.0001) but not SGA.

14 R. Holmes et al. I Early Human Development 49 (1997) 7-17

high in FGR compared with the control group but only IGFBP-3 was elevated in SGA ((I = 609, P = 0.02).

4. Discussion

The IGF results from mothers of AGA fetuses were consistent with previous reports using RIA techniques [ 181. The techniques used provide more numerical results than the semi-quantitative Western ligand blotting methods and not only were the absolute levels very similar to previous studies but the same changes with gestational age (increase in plasma IGF-1 and IGFBP-3 but stable IGF-2 and IGFBP-1) were found [ 19-211. This consistency with previous publications strongly supports that the sampling techniques, storage and analytical techniques were appropriate.

No correlation with gestational age was seen in IGF-1, IGF-2, IGFBP-1 or IGFBP-3 in plasma from AGA fetuses. This contrasts with some previous reports [22], most of which measured cord blood at delivery. Correlations with gestational age in cord blood samples are always difficult to interpret because of the possible effects of labour (such as the rise in IGFBP-1 reported by Hills et al. [23]) particularly in the pre-term group when obstetric complications are more common. However, our results differ from the other major study of cordocentesis IGF data, Lassarre et al. [24], who reported a marked rise in both IGF-1 and IGF-2 after the 33rd week until term to mean values at term of 150 and 725 rig/ml respectively. The rise they reported at the end of pregnancy is surprising since it took the levels much above results from other groups at normal term delivery.

The data in this study were consistent with the hypothesis that the maternal IGF and IGFBP system controls the placental partition of nutrients and so maternal constraint. Maternal IGF-I was highly significantly reduced and IGPBP-I elevated in FGR but not SGA pregnancies and although these findings have not been consistently found in previous studies this is likely to be because they did not classify the data by the cause of growth retardation. Maternal administration of IGF-I in mice and rats has been shown to promote fetal [ 121 and placental growth and both IGF- 1 and IGF-2 can regulate nutrient transport. Giudice et al. [25] suggested that maternal IGFBP-1 and IGFBP-3 act in a reciprocal fashion; the former, secreted predominately by the decidua inhibiting the actions of IGF and the latter making IGF available for placental growth. Our findings of elevated maternal IGFBP-1 in the absence of any change of IGFBP-3 in FGR pregnancies are consistent with this.

A more complete interpretation of the data requires consideration of an IGFBP-3 specific protease [25] which dissociates the 150 kDa complex, thereby reducing binding affinity for IGF-1 and possibly increasing its biological activity [29]. However, protease activity is increased in pregnancy and by the third trimester IGFBP-3 may be totally dissociated. Landford et al. [30] demonstrated increased protease activity in the maternal serum of pregnancies with FGR due to utero-

R. Holmes et al. / Early Human Development 49 (1997) 7-17 is

placental insufficiency compared to SGA and normally grown pregnancies but there appeared to be no remaining complexed IGFBP-3 in any group. This would render quantification of protease activity irrelevant but this needs confirmation. IGFBP-1 is known to undergo phosphorylation and this alters the binding affinity for IGF-1 [3 11. To make further inferences about the role of the increased maternal IGFBP-1 in FGR a study needs to be made of any changes in phosphorylation accompanying this rise.

The alternative explanation for the data is that the maternal changes are secondary to the placental disease. It is possible that placental hormones such as human placental lactogen may stimulate IGF-I production or abnormal placentation may secondarily influence the maternal IGF system. Another possibility is that changes in maternal carbohydrate metabolism including insulin, pro-insulin and glucose resulting from nutritional deprivation or reduced placental hormones may lead to increased maternal IGFBP-I levels.

The hypothesis that reduced fetal growth may sometimes be due to fetal IGF-1 deficiency was not supported by the data because IGF-1 levels were normal. This is at variance with the cord blood data from Spencer et al. [9] and Verhaeghe et al. 1221, but again the uncertain effects of delivery may be important here and cases in this study were much more severely growth retarded than in their studies. However, IGF-2 was low in SGA and the binding proteins were at high concentrations (which has been shown by Taylor et al. [26] to inhibit IGF action in vitro) so ‘relative’ IGF deficiency remains a remote possibility. We consider this unlikely as a biologically important mechanism of poor fetal growth because the changes in the binding proteins were more marked in the FGR than the SGA group and FGR is associated with numerous metabolic consequences including hypoglycaemia [27]. The binding protein changes are, therefore, most likely to be secondary to changes resulting from inadequate placental function.

The results contrast with the data of Lassarre et al. (1991) [24] who reported reduced IGF-1 (but not IGF-2) in small fetuses. Size is not a diagnosis however but a physical sign. It is, therefore, probable that they included a heterogeneous group of different pathologies with normal small fetuses. In this study fetal abnormality, both structural and genetic, as a cause of being small was excluded and those with normal placental function (‘normal’ small fetuses (SGA)) were classified differently from those with placental insufficiency (FGR). This was important because clear differ- ences between these two groups were demonstrated. Langford et al. [28] made such a distinction and found low fetal IGF-1 in FGR. Discrepancies with our data may reflect differences in classification into FGR and SGA.

A better understanding of the role of the maternal IGF system in exerting maternal constraint of fetal growth will require a comprehensive study of larger numbers of pregnancies with impaired growth of known cause. This should include study of the effect of IGFBP proteases, the type, distribution, and function of placental IGF receptors and assessment of maternal carbohydrate metabolism. At present, iatrogenic premature delivery is the only available treatment for FGR. Maternal IGF supple- mentation has already been demonstrated to be effective in animals and may ultimately lead to therapeutic possibilities in humans.

16 R. Holmes et al. I Early Human Development 49 (1997) 7-17

Acknowledgments

“We are very grateful to Pharmacia and Upjohn for Financial support of this research”.

References

[1] Walton, A and Hammond, J. (1938): Proceedings of the Royal Society of London 125, 311-335. [2] Sara, V.R. and Hall, C. (1990): Insulin-like growth factors and their binding proteins. Physiol. Rev.,

70, 591-614. [3] Gluckman, P.D., Gunn, A.J., Wray, A. et al. (1992): Congenital idiopathic growth hormone deficiency

is associated with prenatal and early postnatal growth failure. J. Pediatr., 121, 920-923. [4] Han, V.K.M., Lund, P.K., Lee, D.C. and D’Ercole, A.J. (1988): Expression of somatomedinlinsulin-

like growth factor messenger ribonucleic acids in the human fetus: identification, characterization, and tissue distribution. J. Clin. Endocrinol. Metab., 66, 422-429.

[5] Drop, S.L.S. (1992): Report on the nomenclature of the IGF binding proteins. Endocrinology, 130, 1736-1737.

[6] Liu, J.-P, Baker, J., Perkins, A.S., Robertson, E.J. and Efstratiadis, A. (1993): Mice carrying null mutations of the genes encoding insulin-like growth factor 1 (Igfl) and Type 1 IGF receptor (Igflr). Cell, 75, 59-72.

[7] DeChiara, T.M., Efstratiadis, A. and Robertson, E.J. (1990: A growth-deficiency phenotype in heterozygous mice carrying an insulin-like growth factor II gene disrupted by targeting. Nature, 345, 78-80.

[8] Baker, J., Liu, J.-P, Robertson, E.J. and Efstratiadis, A. (1993): Role of insulin-like growth factors in embryonic and postnatal growth. Cell, 75, 73-82.

[9] Spencer, J.A.D., Chang, T.C., Jones, J., Robson, SC. and Preece, M.A. (1995): Third trimester fetal growth and umbilical venous blood levels of IGF-1, IGFBP-1 and growth hormone at term. Arch. Dis. Child, 73, F87-F90.

[lo] Wang, H.S., Lim, J., English, J., Irvine, L. and Chard, T. (1991): The concentration of insulin-like growth factor-l and insulin-like growth factor binding protein-l in human umbilical cord serum at delivery: relation to fetal weight. J. Endocrinol., 129, 459-464.

[II] Davenport, M.L., Clemmons, D.R., Miles, MY., Camacho-Hubner, C., D’Ercole, A.J. and Under- wood, L.E. (1990): Regulation of serum insulin-like growth factor-l (IGF-1) and IGF binding proteins during rat pregnancy. Endocrinology, 127, 1278- 1286.

[12] Gluckman, PD., Morel, P.C.H., Ambler, G.R., Breier, B.H., Blair, H.T. and McCutcheon, S.N. (1992): Elevating maternal insulin-like growth factor-l in mice and rats alters the pattern of fetal growth by removing maternal constraint. J. Endocrinol., 134, Rl-R3.

[13] Bassett, N.S., Currie, M.J., Woodall, S. and Gluckman, PD. (1994): The effect of IGF-1 and bGH treatment on placental glut 1 gene expression. Proceedings of the Third International Symposium on insulin-like growth factors 1994 Growth Regulation, 4 Suppl 1: 1105, 72.

[14] Nicolaides, K.H., Soothill, P.W., Rodeck, C.H. and Campbell, S. (1986): Ultrasound-guided sampling of umbilical cord and placental blood to assess fetal wellbeing. Lancet, i, 1065-1067.

[15] Morrell, D.J., Dadi, H., More, J. et al. (1989): A monoclonal antibody to insulin-like growth factor-I: Characterization, use in radio-immunoassay and effect on biological activities of the growth factor. J. Mol. Endocrinol., 2, 201-206.

[16] Blum, W.F., Ranke, M.B. and Bierich, J.R. (1988): A specific radioimmunoassay for insulin-like growth factor-II: the interference of IGF binding proteins can be blocked by excess IGF-I. Acta Endocrinol., 118, 374-380.

[17] Povoa, G., Roovete, A. and Hall, K. (1984): Cross-reaction of serum somatomedin-binding protein in a radioimmunoassay developed for somatomedin-binding protein isolated from human amniotic fluid. Acta Endocrinol., 107, 563-570.

R. Holmes et al. I Early Human Development 49 (I9973 7-l 7 i7

[ 181 Mirlesse, V., Frankenne, F., Alsat, E., Poncelet, M., Hennen, G. and Evain-Brian, D. (1993): Placental growth hormone levels in normal pregnancy and in pregnancies with intrauterine growth retardation. Pediatr. Res., 34, 439-442.

1191 Hall, K., Enberg, G., Hellem, E. et al. (1984): Somatomedin levels in pregnancy: longitudinal study in healthy subjects and patients with growth hormone deficiency. J. Clin. Endocrinol. Metab., 59. 587-594.

[20] Hall, K., Hansson, U., Lundin, G. et al. (1986): Serum levels of somatomedins and somatomedin- binding protein in pregnant women with type I or gestational diabetes and their infants. J Chn. Endocrinol. Metab., 63, 1300-1306.

[21] Wilson, D.M., Bennett, A., Adamson, G.D. et al. (1982): Somatomedins in pregnancy: I cross- sectional study of insulin-like growth factors I and II and somatomedin peptide content in normal human pregnancies. J. Clin. Endocrinol. Metab., 55, 858-865.

[22] Verhaeghe, J., Van Bree, R., Van Herck, E., Laureys, J., Bouillon, R. and Van Assche, F.A. (1993): C-peptide, insulin-like growth factors I and 2, and insulin-like growth factor binding protein-l in umbilical cord serum: Correlations with birth weight. Am. J. Obstet. Gynecol., 169, 89-97.

[23] Hills, F.A., Crawford, R., Harding, S. et al. (1994): The effects of labor on maternal and fetal levels of insulin-like growth factor binding protein-I. Am. J. Obstet. Gynecol., 17 I, 1292-1295.

[24] Lassarre, C., Hardouin, S., Daffos, F., Forestier, F., Frankenne, F. and Binoux, M. (1991): Serum insulin-like growth factors and insulin-like growth factor binding proteins in the human fetus, Relationships with growth in normal subjects and in subjects with intrauterine growth retardation. Pediatr. Res., 29, 219-225.

(251 Giudice, L.C., Farrell, E.M., Pham, H., Lamson, G. and Rosenfeld, R.G. (1990): Insulin-like growth factor binding proteins in maternal serum throughout gestation and in the puerperium: effects of a pregnancy-associated serum protease activity. J. Clin. Endocrinol. Metab., 71, 806-816.

[26] Taylor, A.M., Dunger, D.B., Preece, M.A. et al. (1990): The growth hormone independent insulin-like growth factor-l binding protein BP-28 is associated with serum insulin-like growth factor-l inhibitory bioactivity in adolescent insulin-dependent diabetics. Clin. Endocrinol., 32, 229-239.

[27] Economides, D.L. and Nicolaides, K.H. (1989): Blood glucose and oxygen tension levels m small for gestational age fetuses. Am. J. Obstet. Gynecol., 161, 10041GO8.

[28] Langford, K., Blum, W., Nicolaides, K., Jones, J., McGregor, A. and Miell, J. ( 1994): The pathophysiology of the insulin-like growth factor axis in fetal growth failure: a basis for programming by undernutrition. Eur. J. Clin. Invest., 24, 851-856.

1291 Lassarre, C. and Binoux, M. (1994): Insulin-like growth factor binding protein-3 is functionally altered in pregnancy plasma. Endocrinology, 134, 1254- 1262.

1301 Langford, KS., Nicolaides, K.H., Jones, J., Abbas, A., McGregor, A.M. and Miell, J.P. (1995): Serum insulin-like growth factor binding protein-3 (IGFBP-3) levels and IGFBP-3 protease activity in normal. abnormal and multiple pregnancy. J. Clin. Endocrinol. Metab., 80, 21-27.

[31] Jones, J.I., D’Ercole, A.J., Camacho-Hubner, C. and Clemmons, D.R. (1991): Phosphorylation of insulin-like growth factor (IGF) binding protein 1 in cell culture and in viva: effects on affinity foi IGF-I Proc. Nat]. Acad. Sci. USA, 88, 7481-7485.