Observations that IGF is produced and acts locally in multiple tissues raise important questions about the biological significance of the major pool of IGF present in the circulation. Does it represent a pool of endocrine IGF en route to the tissues or conversely growth factor produced in excess of autocrine/paracrine requirements undergoing elimination? The primary objective of this thesis was to examine the kinetics and distribution of circulating IGF in sheep with a view to determining tissue destinations and thereby potential functions of the blood borne hormone. The IGFBP play a central role in facilitating IGF action. Characterization studies of the IGFBP and an examination of their physiology and potential involvement in IGF transport are also important parts of this thesis. Such studies are necessary because potential therapeutic uses of IGF will depend on systemic administration and endocrine action. Early work involved structural/functional characterization of a batch of recombinant methionyl insulin-like growth factor-I (N-Met IGF-I) designated for this project. The peptide was heterogenous on reversed phase chromatography eluting as two major peaks of approximate abundance 1:2. These each had the amino acid constitution expected of N-Met IGF-I and were carefully characterized in a range of binding and biological assays. Whereas the early eluting peptide demonstrated much reduced activity in each assay system, the second peak proved equipotent to a highly purified ovine plasma IGF-I preparation and was chosen for the investigative work of this thesis. The early eluting peptide may represent a variant with mismatched disulphides. Initial characterization of IGF binding activity in ovine tissue fluids was performed by competitive IGF tracer binding techniques together with size exclusion chromatography (SEC) and IGF-I affinity chromatography. Binding proteins (BP) of >200, 150 and 40-50 kDa were revealed in these studies and shown to be widely distributed in body fluids. Thus the >200 kDa binding protein, which is IGF-II specific, was identified in adult sheep plasma, colostrum, follicular fluid and fetal sheep plasma, and may be the ovine equivalent of the soluble type 2 IGF receptor. A 150 kDa binding protein complex, of mixed specificity for IGF-I and II, was also identified. In addition to vascular fluids, the 150 kDa complex was identified in mammay lymph, follicular fluid and, as a minor component, in vitreous humor. Binding proteins of 40-50 kDa were revealed in every fluid tested and multiple variants identified with distinct specificities for the IGF peptides. The BP 'make-up' of fluids and of 150 kDa and 40-50 kDa pools isolated by preliminary SEC was latter examined by IGF ligand blot analysis. Analysis of plasma 150 kDa pools revealed only the characteristic doublet of IGFBP-3 at 40-43 kDa, whereas the 40-50 kDa pool was heterogeneous containing IGFBP-3 together with smaller bands of 35, 30 and 23 kDa which may be the ovine equivalents of IGFBP-2, BP-4 and possibly BP-1. In support of the tracer binding data, IGFBP-3 was also identified in mammary lymph as were the smaller species. In an extension of the in vitro IGF tracer binding/SEC approach, kinetics of IGF equilibration with plasma binding sites was examined. Binding was found to be time and temperature dependent, reversible, dose responsive and relatively specific for the IGF peptides. Observations of special interest include a biphasic pattern of IGF-I equilibration with plasma, consistent with formation of the ternary 150 kDa complex of IGFBP-3, IGF and ALS, and evidence of relatively slow dissociation of IGF/BP complexes, suggesting that if release of IGF is required for full expression of IGF bioactivity in vivo, then specific processes may be involved. Avidity of isolated IGFBP complexes for Con A and heparin affinity adsorbents was also examined. The data indicate that the IGFBP belong to a relatively select group of proteins with high affinity for the glycosaminoglycan heparin suggesting roles for these proteins at the level of the capillary endothelium and/or extra-cellular matrix. Metabolic clearance of IGF-I and II was examined following intravenous (iv) bolus injection of the growth factors as radioiodinated tracer preparations. Tracer administration was followed by a rapid initial phase of clearance associated with tracer mixing in the vascular pool followed by intermediate and longer phases which appear to be direct consequences of interaction with and between the BP and to some extent accumulation of tracer degradation products in the circulation. Metabolic clearance of tracer complexed to the major molecular weight pools of BP was examined following SEC of sequential plasma samples. Average half-lives for IGF-I and II complexed to the 150 KDa and 40-50 kDa pools of carrier protein were established (150 kDa; 545±25 min, 325±30 min; 40-50 kDa, 34±2 min, 9.6±1.8 min, (mean±S.E.M., IGF-I and II respectively)) and compared to free IGF-I (t1/2 <5 min). Rapid clearance of free compared to bound IGF illustrates the central role of the IGFBP in maintaining IGF in the circulation and controlling tissue distribution. Whereas binding of IGF-II to different BP at 40-50 kDa (eg. IGFBP-2) may explain its shorter half-life compared with IGF-I, evidence suggests that IGF-I and II bind to the same carrier at 150 kDa. The observed difference in half-life of the 150 kDa complex is therefore suggestive of different metabolic handling of the BP depending on which of the IGFs is bound. The more rapid clearance of IGF-II complexed to the 150 kDa and 40-50 kDa carriers compared to IGf-I contributed to a more rapid clearance overall and is reflected in calculated metabolic clearance rates for IGF-I and II (IGF-I, 3.9 ml/min; IGF-II, 7.8 ml/min). Considering plasma IGF-II is significantly higher than IGF-I in post-natal sheep, a substantially greater secretion rate for IGF-II would be required to maintain plasma IGF-II in the face of the greater clearance rate. The secretion rate for IGF-II was estimated at ~ 1.6 nmol/min in the current study, some 8-fold greater than IGF-I. Clearance of IGF-I from plasma was associated with the appearance of radioactivity in lymph. Chromatography indicated that tracer in lymph was not degraded but retained its BP activity eluting on SEC complexed to high molecular weight BP. The data illustrate that blood borne IGF is distributed into the extra vascular space and may therefore be available to the tissues. This contention is supported by observations that relatively little radioactivity (<20% in the course of these experiments) was cleared from plasma into urine suggesting that plasma IGF is not principally an elimination form. Similarly no other significant sites of elimination were identified. Questions of how physiological control of the BP may influence tissue distribution of IGF were investigated in the next major experimental section of this thesis. In the first study the influence of nutritional manipulation and GH treatment of growing lambs on the molecular distribution of IGF immunoreactivity in plasma was examined using a new IGF-I RIA in conjunction with SEC and saturation analysis for the estimation of the BP. Total plasma IGF-I was found to increase with nutritional intake (P<0.01) and with GH treatment (0.25 mg/kg body weight/d; P<0.001) but only on the higher intakes. Molecular size fractionation revealed IGF-I immunoreactivity in 150 kDa and 40-50 kDa binding fractions. 150 kDa bound IGF-I was increased on the higher plane of nutrition(P<0.05) and by GH treatment (P<0.001) but again, only at higher levels of nutrition. By contrast no change in 40-50 kDa bound IGF-I was observed with treatment. Unbound IGF-I was also identified in sheep plasma (2-5% of total) but demonstrated only slight changes in relation to treatment. Saturation analysis was an analytical approach chosen to estimate total binding capacity (TBC) and relative saturation of the binding protein pools. Evidence suggests that in ovine plasma constituents of the 150 kDa complex are available in excess of endogenous IGF (P<0.001). Relative saturation of this species did not change with treatment despite the observed differences in 150 kDa bound IGF-I. The data suggest that components of the 150 kDa complex were themselves responsive to treatment. By contrast large differences in saturation of the 40-50 kDa species were observed (P<0.001) despite little treatment dependent change in bound IGF-I. Binding capacity of the 40-50 kDa fraction was elevated at low levels of nutrition and suppressed on the higher feed intake resulting in near saturation. The data indicate complex regulation of the IGFBP in sheep. IGF-I, elevated in response to higher nutritional intake and by CH treatment was mostly distributed into the 150 kDa complex; paradoxically the species which most effectively maintains IGF in the circulation. Thus in conditions presumably conducive with growth related processes (high GH, high nutrition) access of circulating IGF to the tissues is apparently most restricted. This evidence is difficult to reconcile with the view that 150kDa bound IGF represents a pool of endocrine IGF en route to tissue sites of action. Galactopoietic effects of GH in lactating ruminants appear to be exerted in the absence of a mammary GH receptor and are associated with increased plasma and mammary IGF-I content. Thus it has been proposed that blood borne IGF, acting in the classical endocrine fashion may be the mediator of GHs lactogenic effects. Consequently the lactating sheep surgically prepared by the catheterization of efferent mammary lymph may be a useful model for examining questions of IGF/BP physiology. In a further study, plasma and efferent mammary lymph concentrations of IGF-I were determined in lactating ewes before and after treatment with GH (10 mg/d) for 3 days. Analysis of paired plasma/lymph samples revealed that the capillary endothelium constitutes a barrier to the passage of macro molecules which reduces the concentration of IGF in lymph to ~ 35% plasma. A key observation from the current study was the GH dependence of mammary lymph IGF-I. Thus, GH was found to increase mammary lymph IGF-I concentrations by a proportionately greater amount than the increase in plasma IGF-I (P<0.01). The increase in lymph IGF-I resulted from an increase in the concentration of IGF associated with both high molecular weight (150 kDa) and low molecular weight (40-50 kDa) binding fractions. However, the data indicate a proportionately greater increase in 40-50 kDa bound IGF-I in lymph compared with plasma suggesting that treatment either induces a selective redistribution of plasma 40-50 kDa IGF and BP into the mammary gland or, alternatively, treatment increases intra-mammary production of these factors. Ligand blot analysis of mammary lymph revealed IGFBP-3 and -2 as the major constituents of this fluid. IGFBP-2 declined in lymph with GH treatment whereas IGFBP-3 appeared to increase. Additionally, saturation analysis indicated that a substantial proportion of lymph IGFBP-3 was present in the 'free', uncomplexed form. Consequently observations that the total binding capacity (TBC) of the lymph 40-50 kDa fraction increased with treatment, would appear to result from an increase in IGFBP-3. Total binding capacity of the lymph 40-50 kDa binding fraction was found to increase by a proportionately greater amount that its plasma equivalent. Thus, if the lactogenic effects of GH are mediated by IGF distributed from blood into the mammary gland, the mechanism by which it is transferred would appear to involve BP of the 40-50 kDa pool and in particular IGFBP-3. In the final experimental section of this thesis a novel system was employed to examine tissue distribution and destinations of blood borne IGF-I. This involved intravenous infusion of the N-Met analog of IGF-I together with specific immunologic detection. For this an IGF antibody was employed which recognises the recombinant N-Met variant but demonstrates minimal cross-reactivity with any of a range of other IGF peptides including ovine plasma IGF-I and recombinant authentic sequence IGF-I. This antibody can therefore recognise the N-Met variant against a background of authentic endogenous IGF-I and was usefully applied to examining tissue destinations of the N-Met variant following iv infusion. Plasma N-Met IGF-I rose to plateau concentrations of ~150 ng/ml during iv infusion (Infusion rate, 8 µg/kg/h). Analysis revealed the N-Met variant was distributed on plasma BP in much the same proportion as endogenous IGF. N-Met IGF-I immunoreactivity was identified in mammary lymph providing further evidence supporting the contention that blood borne IGF is distributed outside the vascular space. At the end of the infusion N-Met IGF-I was identified in all tissues examined contributing between 35% (in kidney) and 62% (spleen) to total IGF. Major differences in morphological distribution of blood derived (N-Met IGF) were revealed by autoradiography in post infusion tissue slices. Thus the N-Met IGF was found to contribute relatively little to total localizable IGF-I immunoreactivity in connective tissue elements of the samples examined (muscle and mammary) but, by contrast, blood derived (N-Met) IGF-I constituted ~85% of total IGF immunoreactivity in other tissues. In particular, these include the metabolically active regions of muscle and mammary (fibre and epithelium respectively). The evidence suggests therefore that fibre and epithelium may be targets for blood derived ‘endocrine' IGF. Differences in the abundance of blood derived IGF between tissues may relate to the accessibility (vascularization or capillary permeability) of specific tissues or, alternatively, to local rates of production and turnover of IGF in tissues. Thus the contribution of blood derived IGF to total localizable IGF may be expected to be less in tissues which actively synthesize IGF such as those in which autocrine/paracrine modes of IGF action are operative. Examples from the current study would be connective tissues of muscle and mammary. Conversely blood derived IGF would be expected to represent a greater proportion of total IGF in tissue targets for endocrine IGF. Support for the current data was obtained from IGF-I mRNA in situ hybridization studies performed on human fetal tissue (437). Stromal elements of muscle tissue (perimysium, epimysium) were identified as active sites of transcription as opposed to muscle fibre where message could not be detected. Thus the evidence suggests that the N-Met infusion model is a useful technique for delineating tissue targets for circulating (endocrine) IGF. It is now widely accepted that the primary actions of IGF on growth and development occur via autocrine/paracrine mechanisms close to its site of production. Nonetheless such arguments do not exclude the possibility of classical endocrine roles for the major pool of IGF present in the circulation. The primary thrust of this thesis has been to examine kinetics and tissue distribution of this pool of IGF. The data confirm the availability of blood borne IGF to extra vascular tissues and appear to indicate that it is distributed into the tissues on a selective basis and under physiological control. It may, therefore, be available to selected tissues to fill specific endocrine functions. / Whole document restricted, but available by request, use the feedback form to request access.

Whole document restricted, see Access Instructions file below for details of how to access the print copy. / A dramatic increase in referrals of infants with non-synostotic positional plagiocephaly and brachycephaly has occurred since the adoption of supine sleep position recommendations to prevent sudden infant death syndrome (SIDS). Repeated positioning of the soft infant skull on firm surfaces is postulated to cause flattening of infant heads. There are concerns that parents who are worried about their infants' head shapes may reject SIDS prevention guidelines. This thesis was undertaken to provide greater understanding of the determinants, prevalence and natural history of non-synostotic aberrant head shapes. It includes a literature review that summarises the historical background, anatomy and skull growth of the infant cranium, clinical characteristics of non-synostotic plagiocephaly and brachycephaly, cephalometry methods, prevalence, risk factors, prevention, treatment and outcomes. Three studies were conducted. Firstly, a case control study was undertaken to investigate a range of possible risk factors. One hundred cases from plagiocephaly clinics and 94 community controls were administered a questionnaire covering obstetric, sociodemographic, infant, and infant care factors. Infants with plagiocephaly were significantly more likely to be male, firstborn, premature, supine sleeping, less active, to have a preferential head orientation at 6 weeks, to have a developmental delay, and to have a less educated mother. They were likely not to have had the head position varied when being put down to sleep in the first 6 weeks, and to have had less than 5 minutes a day of prone play time in the first 6 weeks. Next, a study of a new digital photographic technique that was developed to measure infant head shape used 31 plagiocephaly cases and 29 controls. The method, named HeadsUpTM, used an elastic band to define the head shape in the fronto-occipital plane. A digital camera recorded the head shape, and custom-written software quantified measurements from the photos. Compared with the conventional flexicurve measuring method, the HeadsUpTM method was more acceptable to both mothers and infants and less variable on measures of plagiocephaly and brachycephaly. Although it is recognised that head shape is a continuum from perfectly symmetrical to severely asymmetrical, thresholds were calculated to allow dichotomisation between "normal" and abnormal. The cut-off for cephalic index [(head width/head length) * l00] was identified as 93%, while the cut-off for the oblique cranial length ratio (the ratio of the longer cross-diagonal length to the shorter cross-diagonal length) was identified to be 106%. Beyond these points, brachycephaly and plagiocephaly, respectively, are deemed to exist. Finally, a prospective cohort study of 200 newborn infants combined the methods from both previous studies to further enlarge on determinants, prevalence and natural history of the condition. Ninety-one percent of the children were followed to two years of age. Using the cut-off points determined in the photo study, it was seen that 29.5% of the cohort developed plagiocephaly or brachycephaly at some time during the first 8 months, after which there were no new cases. Prevalence of plagiocephaly at 6 weeks was 16.0%, increasing to 19.7% at 4 months, then reducing to 9.2%, 6.8% and 3.3% at 8, 12, and 24 months respectively. Risk factor analysis was conducted on the 6-week and 4-month cases and controls. Significant determinants identified were: male gender, a limitation of head rotation, supine sleep position, more than 21 hours of supine-lying time a day at 6 weeks, head position when put down to sleep not varied in first 6 weeks, maternal reporting of low activity level at 4 months, and average-to-difficult temperament at 4 months. Although the cohort study showed that nearly all infants improved over time, there were a few persistent cases. Heads in the cohort were wider and shorter than those measured in infants during earlier decades when prone and side sleeping positions were the norm, highlighting a need for further research to provide age-specific norms for cephalic index in supine-sleeping infants. The photo study cases, recruited from plagiocephaly clinics, were in general of greater severity than the cohort cases. Further research is needed to allow early identification of infants who do not improve over time. Although supine sleeping is a risk factor for the development of plagiocephaly, this position is highly protective against SIDS and should be maintained. However, varying the head position at sleep, providing tummy time, limiting supine-lying time when awake, and awareness and treatment of head rotation problems may help to prevent the condition. These practices need to be confirmed in future studies of primary prevention.

Literature review Pertussis mortality and morbidity Mass immunisation was associated with a decrease in pertussis mortality and a profound reduction in pertussis incidence. Despite this pertussis remains prevalent. Infants account for the majority of pertussis deaths and hospitalisations. Immunisation Pertussis vaccines protect against disease rather than infection. Despite immunisation pertussis remains endemic. The efficacy of different whole cell and acellular pertussis vaccines varies considerably. There has only been a small increase in immunisation coverage in New Zealand over the past 25 years. Currently between 80% and 90% of New Zealand children receive the primary immunisation series. Other epidemiological features Bordetella pertussis is a highly infectious organism. Neither infection nor immunisation results in lifelong immunity. Pertussis affects all age groups. It is more severe in females than in males. The incidence has always been highest in infants and children but the reported incidence in adults is increasing. Pertussis epidemics occur at four yearly intervals. The epidemic periodicity has not been changed by immunisation. Risk factors for pertussis Contemporary case control studies from the United States have shown that exposure to someone outside of the home with pertussis increases the risk of introduction of pertussis into the home and that infants of adolescent mothers and of mothers with a preceding coughing illness are at increased risk of pertussis. Small sample size and imprecise measurement of immunisation status have compromised these studies. Other factors associated with an increased risk of pertussis in infants include younger age, low birth weight, the infant's immunisation status and household crowding. Prior to this current case control study there was no knowledge on the effect of infant characteristics, infant immunisation status, parental and household characteristics, or socioeconomic factors on the risk of pertussis in infants. Methods The pertussis mortality and hospital discharge statistics and notification data from 1872 to 2000 were reviewed. The characteristics of children hospitalised with pertussis during the 1995 to 1997 epidemic were described. Risk factors for pertussis in infants were determined using a case control study with two different control groups. A matched case-control design was used to compare infants with pertussis with well control infants from the community. An unmatched design was used to compare infants hospitalised with pertussis to infants hospitalised with other acute respiratory illnesses. Results Historical review of pertussis epidemiology Immunisation was associated with a significant decline in pertussis mortality rates in New Zealand. Pertussis incidence rates in New Zealand are five and 10 times higher than in the United Kingdom or the United States. New Zealand pertussis hospital discharge rates increased from 1920 to 1950, decreased from 1950 to 1970 and have been increasing since then. The severity of disease among those hospitalised in New Zealand is comparable to other developed countries. Case control study of risk factors for pertussis in infants In the community control sample factors associated with incomplete immunisation included poverty and household crowding, advice from a doctor that immunisations be delayed and the caregiver not having a record of the infant's immunisations. Primary and secondary pertussis in case households occurred in all age groups. Over half of the primary cases were infants. Factors associated with an increased risk of pertussis included incomplete immunisation of the infant, children five to nine years of age living in the household, household members with pertussis during the preceding two months and the family doctor advising that an immunisation be delayed. Preschool attendance by a household member was associated with a decreased risk of pertussis. Infants of low birth weight and infants with younger mothers were not at increased risk of pertussis. In a multivariate analysis, non-immunisation of other children in the household and the presence of someone in the household with clinical pertussis were associated with an increased risk of pertussis in infants. The associations between household members with cough and the risk of pertussis varied with the age of the household members and imply an age dependent disease modifying effect of immunisation. For many of the children in the study households it seems unlikely that any health professional knew whether or not they were fully immunised. Conclusions Immunisation reduced pertussis mortality in New Zealand. Pertussis hospitalisation rates are increasing despite improvements in the immunisation schedule. Sustained sub-optimal immunisation coverage appears to be the dominant reason for New Zealand’s excessive pertussis disease burden. Primary school aged children are important in household pertussis transmission.

Perinatal asphyxia is thought to be a major cause of subsequent neurological deficits. Pathological studies suggest that many of these events occur before birth. However, the relationship between specific prenatal events and neurological outcome is not clear. This thesis tested the hypothesis that certain fetal factors play a role in determining the severity and distribution of neuronal loss following in utero asphyxia. Chronically instrumented fetal sheep at three different gestational ages; midgestation (90d), late-gestation (120-130d) or near term (> 135d) were subjected to either a single or repeated insult. The insult consisted of an episode of either systemic asphyxia (umbilical cord occlusion) or cerebral ischaemia (carotid artery occlusion). The fetal parietal cortical electroencephalogram (EEG), cortical impedance (CI) indicating intracellular edema, blood pressure (MAP), electrocardiogram (ECG) and frequent fetal blood gases and metabolites were measured. Three days after the insult, histopathological analysis or immunohistochemistry was performed to determine neuronal loss and specific neurotransmitters respectively. Transient (10min) occlusion of the umbilical cord in late-gestation fetuses, resulted in severe fetal asphyxia, hypotension (24±5mmHg, p<0.01), bradycardia (72±14bpm, p<0.001), depressed EEG activity (-17±2dB, p<0.001) and intracellular edema. The intracellular edema resolved within 27±6min, whereas the EEG activity was depressed for 5±2h, despite rapid recovery of pO2. Neither seizures or infarction were observed. The degree of hypotension, increase in CI, lactate and recovery of post-asphyxial EEG intensity were more marked in 135d fetuses compared with the midgestation fetus (p<0.01). Neuronal loss, which was only observed in the older group, was predominantly in the hippocampus and associated with the severity of hypotension during occlusion. Repeated episodes of cerebral ischaemia, altered the distribution of neuronal loss compared with single insults, inducing damage mainly in the striatum. The frequency of the insults determined the severity of the damage. Similarly, recurrent episodes of fetal asphyxia induced predominantly striatal neuronal loss. Each occlusion resulted in fetal hypoxia and bradycardia accompanied by increased T/QRS ratio as noted on the ECG. Progressively severe hypotension and lactic acidosis developed during successive occlusions. The EEG was depressed and CI increased with each occlusion. After the asphyxial episodes, blood pressure and heart rate returned to normal, while the T wave was inverted for 310±60min. The EEG remained depressed for 90±10min and intermittent seizures developed at 3.3±0.6h after the last occlusion. The extent of neuronal loss correlated with the degree of hypotension, increase in T/QRS ratio, duration of post-asphyxial EEG depression and number of seizures. Immunohistochemical analysis showed loss of striatal GABAergic projection neurons. These findings demonstrate that certain prenatal factors, such as neurological maturation, pattern of the insult and cardiovascular instability can influence neuronal outcome following fetal asphyxia. An isolated brief episode of asphyxia can lead to selective hippocampal neuronal loss, while repeated insults induce predominantly striatal damage. These distributions of neuronal loss may be associated with postnatal sequelae such as learning disorders and cerebral palsy.

Lactose malabsorption (LM) in children was diagnosed by an elevated breath hydrogen (BH) level following a 10g lactose load. A portable gas chromatograph and a semiconductor detector, designed and constructed for use in the Pacific Islands is described. Following verification on known malabsorber and. normal subjects, the technique was used to determine the prevalence of LM in Europeans at Auckland and Rarotonga, and in Samoans at two locations in Auckland and two locations in Western Samoa. The prevalence of LM in Europeans was significantly (p<0.01) higher at Rarotonga than at Auckland. For Samoans, the prevalence of LM was significantly (p<0.01) higher in Western Samoa than at Auckland. The prevalence of LM was very highly significantly (p<0.001) related to race. Each child tested for LM filled in a questionaire to determine attitude, consumption of and perceived intolerance to milk, milk biscuits and ice cream. Lactose malabsorption was significantly (p<0.05) correlated to milk consumption and to attitude to dairy products, but not to sex, age, and perceived intolerance. The consumption of dairy products was very highly significantly (p<0.0001) correlated to attitude, and highly significantly (p<0.001) correlated to location and perceived intolerance. There was no significant correlation between consumption and race, sex or age. Perceived intolerance to individual dairy products was significantly correlated to attitude to milk (p<0.0001), milk biscuits (p<0.02) and ice cream (p<0.001). Perceived intolerance was not related to age, sex, race, location or the actual symptoms following the consumption of 10g lactose. The unexpected finding of increased LM in the Pacific Islands, was investigated further by studying the LM status of the Medical Team during a visit to western Samoa, and by performing a microbiological survey of water quality. It was found that half of the Medical Team 3/6, became malabsorbers during the week spent in Western Samoa. On returning to New Zealand it was shown that lactase levels took 3 months to normalise. The water supply in Western Samoa was shown to contain very high levels of coliform bacteria. The currently held hypothesis that genetic factors are solely involved in the onset of LM, was not supported. The evidence from the survey supported environmental factors are also involved in adult onset LM. The hypothesis suggesting that dietary lactose was a requirement for retaining elevated lactase levels, was tested using Galactosemic and Phenyl Ketonuria patients. None of the patients had developed LM although they had been on a low lactose diet for years, hence the theory was not supported. The BH method proved highly successful in diagnosing LM with many of the children actually enjoying it.

In newborn infants, birth asphyxia represents the predominant cause of brain injury. These infants will later exhibit neurodevelopmental disabilities or a more major cerebral palsy. Prevention of adverse outcomes requires an understanding of the way in which these deficits develop. Endogenous protective mechanisms arising from the insult have opened new insights in neuroprotective strategies. Neurotrophic factors such as IGF-1 and its N-terminal tripeptide GPE have been shown to confer some neuroprotection after HI injury in the adult rodent. In the P21 rat brain after moderate HI injury, exogenous intracerebral and intraperitoneal injections of GPE (30μg and 300μg respectively) were neuroprotective in the hippocampus and lateral cortex possibly through binding to glia as detected by autoradiography of 3H-GPE. In the preterm infant the mechanisms of white matter injury remain to be clearly elucidated. To mimic the pattern of diffuse cerebral injury of the very preterm infant, a transient moderate focal HI injury has been applied on the immature P3 rat. This new model showed a significant reduction in the lateral cortical volume with reduction and alteration of the myelination pattern in the cortical white matter (WM) at P21. These cortical alterations result from neuro-axonal damage 24h after the insult as shown with Fluoro-Jade B staining and β-APP accumulation. In addition activated astrocytes from 24h after HI up to P21 were present. This model should enable us to elucidate some of the pathogenic mechanisms involved in diffuse WM injury. Brain damage in the developing brain has two components: 1) the pattern and mechanisms of injury are correlated with the stage of development at the time of injury; 2) it will influence subsequent brain development.

Ovine placental lactogen (oPL) is secreted by the placenta into both the fetal and maternal compartments. Its biological function(s) during pregnancy and the mechanisms involved are still unclear. A purification procedure was developed for oPL from sheep placental cotyledons of late gestation. Four procedures were attempted to obtain homogeneous oPL. Recoveries of oPL and total protein were measured throughout the several procedures using a specific oPL RIA and the Bradford protein estimation respectively. The third and fourth procedures resulted in homogeneous oPL and a partial amino acid sequence was obtained from the fourth procedure. In the successful procedures, the placental tissue was extracted with 0.1 M ammonium bicarbonate pH 8.5. A pH precipitation of the soluble fraction was performed, followed by 60% saturation with ammonium sulphate. Further separation steps involved chromatogaphic procedures. Carboxymethylcellulose (CM32) cation exchange was performed batchwise at pH 5.6. Subsequently chromatofocusing was performed to elute proteins in order of their isoelectric points. This was carried out using a pH gadient of 0.9 to 6.0. The final chromatographic step was reverse-phase high performance liquid chromatography (RP-HPLC) using a C4 column. To obtain homogeneous oPL in the third procedure, the partially purified oPL was subjected to SDS polyacrylamide gel electrophoresis and the separated proteins were transferred to nitrocellulose membrane. The homogeneous oPL was eluted from the membrane, however, sequencing was unsuccessful. It was assumed that the N terminal of oPL was blocked. Homogeneous oPL was obtained in the fourth procedure by electrophoretic elution from the Hunkapiller gel system performed at 4°C. The oPL was digested with trypsin, the fragments were separated by RP-HPLC chromatography and two peptides were sequenced. Peptide 1: F D E Q Y G Q G I Peptide 2: Y I N C H T Several strategies were attempted to provide more homogeneous oPL to enable more sequencing. The partially purified oPL fractions from each of these attempts were pooled and electrophoresed on an SDS polyacrylamide gel. The section of acrylamide containing the oPL band was homogenised and a trypsin digest was performed. The digested oPL was separated from the gel pieces, filtered through a Sep-Pak filter and the fragments were separated by RP-HPLC. The yield of oPL was low, but sufficient homogeneous oPL was obtained to provide a partial amino acid sequence from tryptic peptides. A further two peptides provided sequences. Peptide 3: (L) A G E M V N R F D E Q Y G Q G I Peptide 4: (L) Q P G K C Q I P L Q S L F Collaborators from Genentech Inc (San Francisco USA) used partially purified oPL produced from the present study and also obtained homogeneous oPL (Colosi et al., 1989). Complementary DNA clones of oPL were isolated and expressed in mammalian cells by recombinant DNA techniques (Colosi et al., 1989). These clones were sequenced, demonstrating that the full sequence of oPL consists of 198 amino acids preceded by a 38 amino acid sequence signal. Recombinant oPL was generated by Colosi et al. (1989) which provided sufficient material to perform physiological studies in vivo. The somatogenic effects of recombinant oPL were investigated in the growth hormone (GH) deficient dwarf rat and compared to identical doses of recombinant bovine GH (bGH) in 3 independent studies. Both oPL and bGH treatments resulted in an increase (p<0.05) in body weight gain compared to that in saline treated controls, with oPL treatment being more potent than bGH (p<0.05). In promoting linear growth, oPL was more potent (p<0.05) than bGH in some instances. Nitrogen content of dry carcass matter was increased with oPL treatment compared to saline (p<0.05), with a nonsignificant increase in bGH treated animals. Carcass fat was similarly reduced by both oPL and bGH treatment (p<0.05) compared to saline. Serum insulin-like growth factor I (IGF-I) concentrations were increased significantly (p<0.05) by both oPL and bGH treatments, with a significantly greater effect of oPL suggested in one study. No increase in hepatic IGF-I mRNA was evident with either treatment, suggesting that the increase in serum IGF-I is due to posttranscriptional mechanisms. The expression of IGF binding protein 3 (IGFBP-3) hepatic mRNA was increased (p<0.05) with bGH treatment compared to that after saline treatment, but was unaffected by opL treatment, indicating regulation by GH at the transcriptional level. The binding of [125I]bGH to hepatic membrane preparations demonstrated no difference in specific binding compared to that in saline controls. However, [125I]oPL specific binding was greater in oPL treated animals (p<0.05). Animals treated with bGH had reduced (p<0.05) hepatic GH receptor mRNA compared to saline controls, but oPL treatment had no effect. Thus, oPL is a potent anabolic and lipolytic agent in the dwarf rat, exerting greater somatogenic effects on some parameters than bGH. The studies in this thesis have described biochemical and biological characterisitics of oPL. The amino acid sequence of oPL is more closely related to prolactin (PRL) than to GH (Colosi et al., 1989). However, oPL has potent somatogenic activities in the GH deficient dwarf rat. Our data suggest differences in receptor binding and effects on GH receptor and IGFBP-3 expression with these two treatments, raising the possibility of actions through different pathways or differential effects at the GH receptor level. These results do not fully resolve whether GH and PRL exert all effects through a single receptor or whether there is a separate PL receptor.

The introduction to this thesis is a literature review. Kramer, in a study commissioned by WHO, reviewed the literature prior to 1985 on low birthweight. This is extended, mainly in respect to infants who are small for gestational age with emphasis on important findings in relation to birthweight since that time. Work in New Zealand on birthweight is also summarised. The literature is also reviewed in respect to the mechanisms in the pathway between the placenta and the fetus, and in respect to recent work suggesting a link between birthweight and disease in adult life. This thesis examines factors that influence birthweight and placental weight. Birthweight for gestational age percentile curves for the New Zealand population were firstly defined. small for gestational age (SGA) infants could then be categorised. The thesis considers two sources of data, the first a cross-sectional sample of the New Zealand population from 1987 to 1990 (the control subjects of the New Zealand cot Death study, a national case-control study on sudden infant death syndrome), and the second a hospital population in Auckland (National Womens Hospital (l992)). These two datasets are investigated to determine factors that influence birthweight in a univariate situation and then in the multivariate situation. Independent variables are considered using a priori categorisations and where appropriate Quantile-Quantile (Q-Q) derived categorisations determined by producing plots of the quantiles of cases versus controls. A number of variables under the headings of socio-demographic, lifestyle, genetic, obstetric and nutrition are examined and found to be associated with the outcomes of interest at the univariate level. After controlling in multivariate analyses a number of variables are found to be no longer significant, however some show strong relationships. The variable relating to smoking in both datasets shows the greatest detrimental effect on the outcomes considered in respect to birthweight. This confirms that in New Zealand, as in other places in the world, smoking has significant consequences on birthweight. The data is also investigated for the timing of insult to the fetus from smoking, and is found to be most important during pregnancy. comparison of the results comparing those obtained using a binary outcome for SGA, and those obtained using birthweight continuously, show relatively consistent results. The odds ratios and the decreases in birthweight obtained from both datasets show a relatively linear relationship between the two. An examination into whether a distinct group of individuals exists in respect to having large placentae for birthweight, indentified an artefact in the dataset relating to recording of placental weight for twins. After removal of twins from the dataset, examination of factors that influence placental weight showed that the factors that influence placental weight are not the same as those that influence birthweight. In particular smoking is found not to influence placental weight, and haemoglobin, which has no influence on birthweight, is found to be inversely associated with placental weight. other factors such as parity are found to influence placental weight in the same proportion in which birthweight is affected. In conclusion this thesis shows that factors investigated in New Zealand are consistent with findings in the international literature in relation to birthweight. The results on factors that influence placental weight add to the international literature on a topic on which little work has been carried out. The results of this thesis point to areas where future research needs to be carried out, in particular in relation to maternal nutrition during pregnancy and maternal energy expenditure during pregnancy. There is also a need for further research into the relationships of factors on placental weight and the ratio of birthweight to placental weight, and how these relationships affect health outcomes in childhood and adult life.

Over the past decade, a number of epidemiological studies have provided significant evidence that certain major adult noncommunicable diseases, such as hypertension, ischaemic heart disease and non-insulin dependent diabetes mellitus, may be associated with impaired fetal growth. This phenomenon has been termed "programming" which is essentially the term used for persisting changes in structure and function caused by undernutrition or other adverse influences acting during critical periods of early development. Programming has been used as the mechanistic basis to explain the long term sequelae of intrauterine growth retardation (IUGR). The mechanisms underlying the epidemiological observations remain to be elucidated and developed. While it is well established that severe maternal undernutrition during pregnancy leads to IUGR, there has been relatively little well defined animal studies of the somatotrophic axis and postnatal development of growth retarded offspring. The major objectives of this thesis were to establish a model in the rat of IUGR by nutritional restriction of the dam throughout gestation and to examine the effects of fetal growth retardation on endocrine, molecular and growth parameters during postnatal development. In addition, the development of an animal model for IUGR enabled well defined studies testing distinct hypotheses suggested by the epidemiological observations of professor David Barker and colleagues. Timed matings were performed in Wistar rats and dams were randomly assigned to one of two dietary treatment groups. A control group was fed ad libitum throughout pregnancy and a restricted group was fed 30% of ad libitum intake. Restricted fed dams were observed to lose a significant amount of body weight throughout gestation, due to undernutrition, but caught up to the ad libitum group during the lactating period. Maternal undernutrition significantly reduced fetal and placental weights without altering litter size. Postnatally, body weights of offspring from undernourished dams continued to be reduced until at least 18 weeks of age, although they were observed to be growing at the same rate as ad libitum offspring by 2 weeks of age. A cohort of animals from undernourished dams were maintained to measure blood pressure by tail cuff plethysmography. Offspring from undernourished dams were found to have significantly elevated systolic blood pressures from 18 weeks of age. This observation provides direct experimental support for the hypothesis, derived from human epidemiological studies, that the origin of adult hypertension may originate during fetal life as a result of exposure to a sub-optimal intrauterine environment. Parallel reductions in plasma IGF-I and hepatic IGF-I mRNA concentrations before 15 days of age were also observed in growth retarded offspring. Hepatic IGF-I transcription start sites within exon 1 and exon 2 were coordinately reduced with IUGR up to 15 days of age without changes in GHR and GHBP mRNA abundance. The lack of catch-up growth observed in the IUGR offspring despite normalization of their plasma IGF-I and IGF-I mRNA levels from 15 days of age may be due to a state of partial resistance to GH. This observation lead to a series of treatment studies in which neonatal and juvenile offspring from ad libitum and undernourished dams were treated with growth factors to investigate somatic growth responses as a measure for hormone sensitivity. In both treatment studies, ad libitum offspring from both age groups and juvenile IUGR offspring responded to GH treatment However, neonatal IUGR offspring did not exhibit any response to GH treatment.. Analysis of IGF-I gene expression in neonatal offspring showed that GH treatment elevated IGF-I Eb mRNA in ad libitum but not IUGR offspring. These results suggest a possible mechanism for transient GH resistance in that a post-receptor defect in GH action may contribute to the development of temporary postnatal GH resistance as a consequence of IUGR and fetal programming of IGF-I gene expression. In summary, the development of a model of IUGR in the rat using maternal undernutrition throughout gestation has enabled detailed investigation of nutritional regulation of the somatotrophic axis during fetal development and postnatal sequelae. The studies in this thesis have shown that the somatotrophic axis is markedly altered postnatally by nutritional restriction of the dam throughout gestation, leading to prolonged postnatal growth retardation and elevated blood pressure The mechanisms which lead to the induction of such fetal programming and whether these changes may contribute to the development of subsequent adult-onset disease remain to be addressed in future studies.

The developmental origins of health and disease hypothesis suggests that exposure to adverse prenatal environmental influences can determine an individual’s susceptibility to obesity in adult life. However, the specific causal mechanisms which underlie this hypothesis have yet to be identified. Focusing on the potential mechanistic role of the leptin endocrine axis, the main objective of this thesis was to investigate the long term effects of prenatal undernutrition and different levels of postnatal nutrition on leptin sensitivity and the development of diet-induced obesity (DIO) in the Wistar rat. A well established animal model of maternal undernutrition during pregnancy was used to induce prenatal undernutrition in experimental offspring. To investigate the interaction between prenatal nutrition and postnatal diet, and its effects on obesity development, female offspring were placed on three different diets: standard chow, a high fat diet or a calorie restricted diet. The effects of prenatal undernutrition and postnatal diet on leptin sensitivity were investigated, in adult offspring, by measuring the response to 14 days of peripheral leptin treatment. Changes in gene expression in the liver, retroperitoneal adipose tissue and soleus muscle were then characterised by custom microarray and quantitative real-time RT-PCR (QPCR) analysis. Adult female offspring exposed to prenatal undernutrition (UN offspring) were found to exhibit leptin resistance in adulthood, independent of postnatal DIO. This result demonstrates for the first time that exposure to prenatal undernutrition has a long term effect on adult leptin sensitivity. In UN offspring fed on a high-fat diet, leptin resistance significantly accelerated the development of DIO while in contrast, offspring maintained on calorie restriction remained lean. These findings suggest that prenatal nutrition can shape future susceptibility to DIO by altering postnatal leptin sensitivity. An analysis of gene expression suggests that prenatal undernutrition causes the development of peripheral tissue-specific leptin resistance, and may also further enhance an offspring’s susceptibility to DIO by altering the regulation of peripheral tissue lipogenesis, mitochondrial function, glucocorticoid metabolism and insulin sensitivity. In conclusion, these studies identify peripheral leptin resistance as a key mechanism that can influence postnatal susceptibility to DIO in female offspring exposed to prenatal undernutrition. Furthermore, the identification of specific changes in peripheral gene expression highlights four additional metabolic mechanisms which may also facilitate the development of DIO in leptin resistant UN offspring.