The literature on the long-term effects of nutritional deficiencies in early life is reviewed. The severity and duration of the deficiency, the stage of the children’s development, the biological condition of the children and the socio-cultural context may all modify the effect. There is substantial evidence that reduced breast-feeding, small-for-gestational-age birth weight, Fe and I deficiency, and protein-energy malnutrition (PEM) are associated with long-term deficits in cognition and school achievement. However, all these conditions are associated with poverty and poor health, which may account for the association. It is difficult to establish that the long-term relationship is causal, as it requires a randomized treatment trial with long-term follow-up. Such studies are only available for I deficiency in utero and early childhood PEM. Results from these studies indicate that I deficiency has a long-term effect and PEM probably has a long-term effect.

Fe plays a critical role in the immune system and defence against infection. However, many aspects of the way in which Fe influences these processes at the molecular and cellular level are unclear. The present review summarizes the role of Fe in lymphocyte activation and proliferation, and discusses how Fe is handled by macrophages.

Food intake is a regulated system. Afferent signals provide information to the central nervous system, which is the centre for the control of satiety or food seeking. Such signals can begin even before food is ingested through visual, auditory and olfactory stimuli. One of the recent interesting findings is the demonstration that there are selective fatty acid taste receptors on the tongue of rodents. The suppression of food intake by essential fatty acids infused into the stomach and the suppression of electrical signals in taste buds reflect activation of a K rectifier channel (K 1.5). In animals that become fat eating a high-fat diet the suppression of this current by linoleic acid is less than that in animals that are resistant to obesity induced by dietary fat. Inhibition of fatty acid oxidation with either mercaptoacetate (which blocks acetyl-CoA dehydrogenase) or methyl­palmoxirate will increase food intake. When animals have a choice of food, mercaptoacetate stimulates the intake of protein and carbohydrate, but not fat. Afferent gut signals also signal satiety. The first of these gut signals to be identified was cholecystokinin (CCK). When CCK acts on CCK-A receptors in the gastrointestinal tract, food intake is suppressed. These signals are transmitted by the vagus nerve to the nucleus tractus solitarius and thence to higher centres including the lateral parabrachial nucleus, amygdala, and other sites. Rats that lack the CCK-A receptor become obese, but transgenic mice lacking CCK-A receptors do not become obese. CCK inhibits food intake in human subjects. Enterostatin, the pentapeptide produced when pancreatic colipase is cleaved in the gut, has been shown to reduce food intake. This peptide differs in its action from CCK by selectively reducing fat intake. Enterostatin reduces hunger ratings in human subjects. Bombesin and its human analogue, gastrin inhibitory peptide (also gastrin-insulin peptide), reduce food intake in obese and lean subjects. Animals lacking bombesin-3 receptor become obese, suggesting that this peptide may also be important. Circulating glucose concentrations show a dip before the onset of most meals in human subjects and rodents. When the glucose dip is prevented, the next meal is delayed. The dip in glucose is preceded by a rise in insulin, and stimulating insulin release will decrease circulating glucose and lead to food intake. Pyruvate and lactate inhibit food intake differently in animals that become obese compared with lean animals. Leptin released from fat cells is an important peripheral signal from fat stores which modulates food intake. Leptin deficiency or leptin receptor defects produce massive obesity. This peptide signals a variety of central mechanisms by acting on receptors in the arcuate nucleus and hypothalamus. Pancreatic hormones including glucagon, amylin and pancreatic polypeptide reduce food intake. Four pituitary peptides also modify food intake. Vasopressin decreases feeding. In contrast, injections of desacetyl melanocyte-stimulating hormone, growth hormone and prolactin are associated with increased food intake. Finally, there are a group of miscellaneous peptides that modulate feeding. β-Casomorphin, a heptapeptide produced during the hydrolysis of casein, stimulates food intake in experimental animals. In contrast, the other peptides in this group, including calcitonin, apolipoprotein A-IV, the cyclized form of histidyl-proline, several cytokines and thyrotropin-releasing hormone, all decrease food intake. Many of these peptides act on gastrointestinal or hepatic receptors that relay messages to the brain via the afferent vagus nerve. As a group they provide a number of leads for potential drug development.

At its most fundamental, cancer is a genetic disease resulting from inherited or acquired mutations in tumour suppressor genes and proto-oncogenes. Environmental factors, including ingested food components, interact with genetic inheritance to determine individual cancer risk. There is growing evidence that the immune system exerts selective pressure during neoplastic development. Tumour cells that evade this immunosurveillance because they are non-antigenic or because they defend themselves successfully against immune attack have a survival advantage. Effective chemopreventative agents will include dietary components that enhance the immune system’s ability to identify transformed cells and to target them for apoptosis.

Zn is an essential trace element for all organisms. In human subjects body growth and development is strictly dependent on Zn. The nervous, reproductive and immune systems are particularly influenced by Zn deficiency, as well as by increased levels of Zn. The relationship between Zn and the immune system is complex, since there are four different types of influence associated with Zn. (1) The dietary intake and the resorption of Zn depends on the composition of the diet and also on age and disease status. (2) Zn is a cofactor in more than 300 enzymes influencing various organ functions having a secondary effect on the immune system. (3) Direct effects of Zn on the production, maturation and function of leucocytes. (4) Zn influences the function of immunostimulants used in the experimental systems. Here we summarize all four types of influence on the immune function. Nutritional aspects of Zn, the physiology of Zn, the influence of Zn on enzymes and cellular functions, direct effects of Zn on leucocytes at the cellular and molecular level, Zn-altered function of immunostimulants and the therapeutic use of Zn will be discussed in detail.

The hypothalamus is the focus of many peripheral signals and neural pathways that control energy homeostasis and body weight. Emphasis has moved away from anatomical concepts of ‘feeding’ and ‘satiety’ centres to the specific neurotransmitters that modulate feeding behaviour and energy expenditure. We have chosen three examples to illustrate the physiological roles of hypothalamic neurotransmitters and their potential as targets for the development of new drugs to treat obesity and other nutritional disorders. Neuropeptide Y (NPY) is expressed by neurones of the hypothalamic arcuate nucleus (ARC) that project to important appetite-regulating nuclei, including the paraventricular nucleus (PVN). NPY injected into the PVN is the most potent central appetite stimulant known, and also inhibits thermogenesis; repeated administration rapidly induces obesity. The ARC NPY neurones are stimulated by starvation, probably mediated by falls in circulating leptin and insulin (which both inhibit these neurones), and contribute to the increased hunger in this and other conditions of energy deficit. They therefore act homeostatically to correct negative energy balance. ARC NPY neurones also mediate hyperphagia and obesity in the ob/ob and db/db mice and fa/fa rat, in which leptin inhibition is lost through mutations affecting leptin or its receptor. Antagonists of the Y5 receptor (currently thought to be the NPY ‘feeding’ receptor) have anti-obesity effects. Melanocortin-4 receptors (MC4-R) are expressed in various hypothalamic regions, including the ventromedial nucleus and ARC. Activation of MC4-R by agonists such as α-melanocyte-stimulating hormone (a cleavage product of pro-opiomelanocortin which is expressed in ARC neurones) inhibits feeding and causes weight loss. Conversely, MC4-R antagonists such as ‘agouti’ protein and agouti gene-related peptide (AGRP) stimulate feeding and cause obesity. Ectopic expression of agouti in the hypothalamus leads to obesity in the AVY mouse, while AGRP is co-expressed by NPY neurones in the ARC. Synthetic MC4-R agonists may ultimately find use as anti-obesity drugs in human subjects Orexins-A and -B, derived from prepro-orexin, are expressed in specific neurones of the lateral hypothalamic area (LHA). Orexin-A injected centrally stimulates eating and prepro-orexin mRNA is up regulated by fasting and hypoglycaemia. The LHA is important in receiving sensory signals from the gut and liver, and in sensing glucose, and orexin neurones may be involved in stimulating feeding in response to falls in plasma glucose.

Several factors have been found recently to have a significant impact on newborn bone mineral content (BMC) and developing fetal bone. Recently we showed that maternal vitamin D deficiency may affect fetal bone mineralization. Korean winter-born newborn infants had extremely low serum 25-hydroxyvitamin D (25-OHD), high serum cross-linked carboxy-terminal telopeptide of type I collagen (ICTP; a bone resorption marker), and markedly lower (8 %) total body BMC than summer-born newborn infants. Infant total body BMC was positively correlated with cord serum 25-OHD and inversely correlated with ICTP, which was also negatively correlated with vitamin D status. In three separate studies on North American neonates we found markedly lower (8–12 %) BMC in summer newborn infants compared with winter newborn infants, the opposite of the findings for Korean neonates. The major reason for the conflicting BMC results might be the markedly different maternal vitamin D status of the North American and Korean subjects. Recently, we found evidence of decreased bone formation rates in infants who were small-for-gestational age (SGA) compared with infants who were appropriate-for-gestational age; we reported reduced BMC, cord serum osteocalcin (a marker of bone formation) and 1,25-dihydroxyvitamin D (the active metabolite of vitamin D), but no alterations in indices of fetal bone collagen metabolism. In theory, reduced utero-placental blood flow in SGA infants may result in reduced transplacental mineral supply and reduced fetal bone formation. Infants of diabetic mothers (IDM) have low BMC at birth, and infant BMC correlated inversely with poor control of diabetes in the mother, specifically first trimester maternal mean capillary blood glucose concentration, implying that factors early in pregnancy might have an effect on fetal BMC. The low BMC in IDM may be related to the decreased transplacental mineral transfer. Cord serum ICTP concentrations were higher in IDM than in control subjects, implying increased intrauterine bone resorption. BMC is consistently increased with increasing body weight and length in infants. Race and gender differences in BMC appear in early life, but not at birth. Ethanol consumption and smoking by the mother during pregnancy affect fetal skeletal development.

Numerous case–control and prospective studies have identified elevated plasma homocysteine as a strong independent risk factor for cerebovascular, cardiovascular and peripheral vascular disease. Homocysteine is formed as a result of the breakdown of the dietary amino acid methionine. Once formed, homocysteine is either remethylated to methionine, or undergoes a trans-sulfuration reaction to form cysteine. The re-methylation of homocysteine to methionine is dependent on three B-vitamins, i.e. riboflavin, vitamin B12 and folate. The second pathway of homocysteine metabolism is the trans-sulfuration pathway which requires both vitamin B6 and riboflavin for its activity. Thus, up to four B-vitamins are required for intracellular homocysteine metabolism. Many studies have noted strong inverse relationships between homocysteine levels and the status of both vitamin B12 and folate. However, the relationship between vitamin B6 status and homocysteine is still uncertain. Similarly, numerous intervention studies have demonstrated effective lowering of homocysteine levels as a result of folate and vitamin B12 supplementation, while the homocysteine-lowering ability of vitamin B6 is unclear. Even though riboflavin plays a crucial role in both the trans-sulfuration and remethylation pathways of homocysteine metabolism, the relationship between riboflavin status and homocysteine levels has not been investigated. The exact mechanism that explains the vascular toxicity of elevated homocysteine levels is unknown at present, studies indicate that it is both atherogenic and thrombogenic. To date, no randomized clinical trial has demonstrated that lowering of homocysteine levels is beneficial in terms of reducing the prevalence of vascular disease. It is probable, however, that optimal B-vitamin status is important in the prevention of vascular disease.

Whilst a number of neuroendocrine afferent signals are implicated in body-weight homeostasis, the major efferent pathway is the sympathetic nervous system (SNS), which affects both energy expenditure and substrate utilization. Thyroid hormones and their interactions with the SNS may also have a role to play. Some of the variability in resting energy expenditure can be explained by differences in SNS activity, and β-blockade can reduce energy expenditure and diet-induced thermogenesis in Caucasians. Excess energy intake leads to SNS activation and increased diet-induced thermogenesis. A relationship has also been demonstrated between spontaneous physical activity and SNS activity. In many animal models the SNS activates brown adipose tissue thermogenesis, hence increasing diet-induced thermogenesis and dissipating excess energy as heat. This effect is mediated via β3-adrenoceptors and activation of an uncoupling protein unique to brown adipose tissue. Homologous proteins have been identified in human tissues and may play a role in human energy expenditure. How the SNS is implicated in this process is unclear at present. β3-Adrenoceptor polymorphism has been associated both with lower resting energy expenditure in some populations and with reduced autonomic nervous system activity. SNS effects on substrate cycling may also play a role. In the development of obesity the effects of the SNS in promoting lipolysis and fat oxidation are likely to be at least as important as its effects on thermogenesis. β-Blockade has relatively small effects on energy expenditure, but more pronounced effects on reducing lipid oxidation, so tending to favour fat storage and weight gain. Low lipid oxidation is a risk factor for weight gain, and there is some evidence that low basal sympathetic nerve activity in muscle is associated with this process. Overall, the relationship between SNS activity and obesity is complex, with evidence of low SNS activity occurring in some, but not all, studies.

Taste preferences, food choices and eating habits all change with age. The transition from childhood to adolescence and adult life is associated with reduced sweet taste preferences, lower sugars consumption and reduced energy density of the diet. Ageing is associated with elevated acceptance of bitter tastes, elevated preferences for vegetables and salad greens, and increased consumption of whole grains, vegetables and fruit. The age-associated drop in energy intakes is achieved through a reduction in the weight and volume of food consumed, as well as a reduction in the overall energy density of the diet. Energy density drops from a peak of 5 kJ (1.2 kcal)/g in adolescence and early adult life to a low of 3.1 kJ (0.75 kcal)/g for adult women aged 45–54 years. Older adults, particularly women, consume less fat and saturated fat and more fibre and vitamin C, suggesting a shift in consumption from snacks, sweets and desserts towards grains, vegetables and fruit. These changes in food preferences and eating habits are associated, on a population level, with a decline in preferences for sweet taste and with increased acceptance of bitter tastes. At present there are no data to show a causal relationship between age-related changes in sensory function and the selection of a more bulky energy-dilute diet. However, it is a plausible hypothesis that sensory factors mediate adjustments in energy density of foods at different life stages.

Substrates with immune-modulating actions have been identified among both macro- and micronutrients. Currently, the modes of action of individual immune-modulating substrates, and their effects on clinical outcomes, are being examined. At present, some enteral formulas are available for the clinical setting which are enriched with selected immune-modulating nutrients. The purpose of the present paper is to review the scientific rationale of enteral immunonutrition. The major aspects considered are mucosal barrier structure and function, cellular defence function and local or systemic inflammatory response. It is notable that in critical illness the mucosal barrier and cellular defence are impaired and a reinforcement with enteral immunonutrition is desirable, while local or systemic inflammatory response should be down regulated by nutritional interventions. The results available from clinical trials are conflicting. Meta-analyses of recent trials show improvements such as reduced risk of infection, fewer days on a ventilator, and reduced length of intensive care unit and hospital stay. Thus, a grade A recommendation was proclaimed for the clinical use of enteral immune-modulating diets. Improvement in outcome was only seen when critical amounts of the immune-modulating formula were tolerated in patients classified as being malnourished. However, in other patients with severe sepsis, shock and organ failure, no benefit or even disadvantages from immunonutrition were reported. In such severe conditions we hypothesize that systemic inflammation might be undesirably intensified by arginine and unsaturated fatty acids, directly affecting cellular defence and inflammatory response. We therefore recommend that in patients suffering from systemic inflammatory response syndrome great caution should be exercised when immune-enhancing substrates are involved which may aggravate systemic inflammation.

Retinoic acid (RA) is the bioactive metabolite of vitamin A (retinol) which acts on cells to establish or change the pattern of gene activity. Retinol is converted to RA by the action of two types of enzyme, retinol dehydrogenases and retinal dehydrogenases. In the nucleus RA acts as a ligand to activate two families of transcription factors, the RA receptors (RAR) and the retinoid X receptors (RXR) which heterodimerize and bind to the upstream sequences of RA-responsive genes. Thus, in addition to the well-established experimental paradigm of depriving animals of vitamin A to determine the role of RA in embryonic and post-embryonic development, molecular biology has provided us with two additional methodologies: knockout the enzymes or the RAR and RXR in the mouse embryo. The distribution of the enzymes and receptors, and recent experiments to determine the endogenous distribution of RA in the embryo are described here, as well as the effects on the embryo of knocking out the enzymes and receptors. In addition, recent studies using the classical vitamin A-deprivation technique are described, as they have provided novel insights into the regions of the embryo which crucially require RA, and the gene pathways involved in their development. Finally, the post-embryonic or regenerating systems in which RA plays a part are described, i.e. the regenerating limb, lung regeneration, hair cell regeneration in the ear and spinal cord regeneration in the adult.

Obesity, which is becoming one of the major health hazards in developed and developing societies, results from a long-term positive energy balance. Body-weight regulation and stability depend on an axis with three interrelated components: food intake, energy expenditure and adipogenesis, although there are still many unknown features concerning fuel homeostasis and energy balance. Biochemical processes are interconnected, and a separate consideration of each component is often useful for methodological purposes and to achieve a better understanding of the whole system. Thus, many different experimental approaches can be applied by using laboratory animals, cell culture or human subjects to unravel the molecular mechanisms which participate in body-weight regulation. Thus, both in vitro (cellular and subcellular models) and in vivo methods have dramatically increased our knowledge of weight control. Several strategies in obesity research are reported here, exploiting the opportunities of the molecular era as well as novel whole-body approaches, which will impact on the development of new targets for obesity management and prevention.

The application of a lifespan perspective in human biology in recent years has shown that a number of early environmental factors influencing human growth and development have long-term biological or psycho-social consequences. Human growth is characterized by prolonged infancy, an extended childhood phase and high rates of growth during the adolescent growth spurt. It is unlikely that these characteristics would have evolved without having advantages, and curtailments have the potential for disadvantage. The present paper examines the evidence for long-term physiological and economic consequences of growth retardation in children and adolescents. The emphasis is the biological and economic imperatives of survival, subsistence, reproduction and production rather than aspects of metabolic competence. Many of the consequences of growth retardation are determined by the direct effect on body size, but many other consequences arise from the conditions that cause the growth retardation. Catch up of retarded growth can occur, but does not usually do so because of the continued presence of the retarding agents. Basal metabolism and physical work capacity are usually commensurate with the size of the individual; mechanical efficiency of physical work is unchanged, but falls in activity levels may occur along with a reduction in the pace of activity. Growth retardation in childhood is associated with a higher disease and mortality risk in adulthood, with decreased productivity and employment and promotion prospects. Studies are showing that relative deprivation and the accumulation of socially patterned exposures are important in some societies. Height and growth retardation have proved invaluable in reflecting these factors, but the next generation of studies may require more discriminating indices.

I is required for the synthesis of thyroid hormones. These hormones, in turn, are required for brain development, which occurs during fetal and early postnatal life. The present paper reviews the impact of I deficiency (1) on thyroid function during pregnancy and in the neonate, and (2) on the intellectual development of infants and children. All extents of I deficiency (based on I intake (μg/d); mild 50–99, moderate 20–49, severe > 20) affect the thyroid function of the mother and neonate, and the mental development of the child. The damage increases with the extent of the deficiency, with overt endemic cretinism as the severest consequence. This syndrome combines irreversible mental retardation, neurological damage and thyroid failure. Maternal hypothyroxinaemia during early pregnancy is a key factor in the development of the neurological damage in the cretin. Se deficiency superimposed on I deficiency partly prevents the neurological damage, but precipitates severe hypothyroidism in cretins. I deficiency results in a global loss of 10–15 intellectual quotient points at a population level, and constitutes the world’s greatest single cause of preventable brain damage and mental retardation.

Breast-feeding policy tends to be an emotive issue. International agencies recommend exclusive breast-feeding for 4–6 months followed by continued partial breast-feeding into the second year of life in order to promote infant and child health and minimize the damage caused by the malnutrition–infection cycle. To what extent are these recommendations supported by the experimental evidence? Are they a simplification for emotional reasons or public health purposes? Breast-feeding is believed to benefit infants because breast milk contains the ideal mix of nutrients for infants, because it contains factors which promote development of the infant’s gut and immune system and which prevent pathogen invasion, and because exclusive breast-feeding prevents intake of pathogens in food or water. However, some apparently contradictory evidence exists. First, in environments which are not highly contaminated breast-fed infants tend to growth falter relative to those fed formula. Second, in such environments partial breast-feeding is not associated with significantly increased gut damage relative to exclusive breast-feeding, suggesting that active promotion of gut development by breast-feeding is more important than simple avoidance of pathogens from other foods. Third, many immune factors in breast milk are probably present primarily to protect the mother, not the infant. Finally, breast milk itself may contain bacteria or viruses. This problem has come to the fore with the human immunodeficiency virus epidemic, since it is clear breast-feeding is an important mode of mother-to-child transmission. The present review will examine these challenges to the basis of the international infant feeding recommendations and will suggest that the science does actually support the policy.

Cow’s milk-based infant formulas and cow’s milk consumption in childhood have been suggested to promote the development of type 1 diabetes mellitus and other immune-mediated or neurological diseases. Epidemiological studies in man have led to the hypothesis that introduction of cow’s milk-based infant formula within the first 3 months of life is associated with increased risk of type 1 diabetes mellitus. Furthermore, in animal models of type 1 diabetes mellitus, cow’s milk proteins have been proven to be ‘diabetogenic’. However, the issue seems far from being resolved. Several epidemiological studies and, more importantly, the first prospective trials did not show an association between early exposure to cow’s milk and type 1 diabetes mellitus. In animal models, cow’s milk proteins are modestly and variably diabetogenic, wheat or soyabean proteins in the diet cause higher rates of autoimmune diabetes. In both man and rodents there is increasing evidence that the gut-associated immune system plays a major role in disease development, probably because of disturbed oral tolerance mechanisms. Oral tolerance depends on immunological homeostasis and normal maturation of the gut. These factors are influenced by growth factors and cytokines from breast milk, normal bacterial colonization, infections and diet. All these factors have been proposed as risk factors for type 1 diabetes mellitus. Hence, cow’s milk proteins may provide mimicry epitopes relevant in autoimmunity, as well as destabilizing oral tolerance mechanisms by biologically active peptides. The concept of dietary regulation of autoimmunity does not apply only to cow’s milk protein, but also to other dietary proteins.

The ‘fetal origins’ hypothesis (Barker, 1995) would predict that the rising epidemic of diabetes and CHD in India would be due to poor intrauterine growth of the Indian babies. While this explanation may be valid to an extent, the higher prevalence of these disorders in urban compared with rural India (where birth weights are lower) would suggest a significant role for postnatal factors. In a cohort of 477 children born in the King Edward Memorial Hospital, Pune, we found that at 8 years of age current obesity strongly predicted insulin resistance. When this effect was allowed for, low birth weight was significantly associated with insulin-resistance variables and other cardiovascular risk factors. Children who were born small but had grown heavy (or tall) were the most insulin resistant and had the highest levels of cardiovascular risk factors. Accelerated growth in relation to mid parental height was similarly predictive. Poor intrauterine growth also predicted higher central adiposity at 8 years of age. We have also studied maternal nutrition and fetal growth in six villages near Pune. A newborn Indian baby is small (2650 g, SD score (SDS) –1.6 compared with an average white Caucasian baby born in the UK) and ‘thin’ (ponderal index 2.45 kg/m3, SDS –1.2), but has preserved its subcutaneous fat (subscapular skinfold thickness SDS –0.6). The thinness of the Indian babies is due to poor muscle and small abdominal viscera. We have proposed this composition as the ‘thrifty phenotype’ (Hales & Barker, 1992) of Indian babies. Maternal size and intake of certain food groups during pregnancy were important determinants of the baby’s phenotype. Thus, the small Indian babies are programmed to deposit fat from their intrauterine life. Exaggeration of this tendency in later life is associated with insulin-resistance syndrome. Control of the insulin-resistance epidemic in India might depend on improved intrauterine development and prevention of childhood obesity.

Insufficient polyamine intake could play a role in the induction of sensitization to dietary allergens. This proposal is based essentially on investigations made in sucking rats and in children. In sucking rats it has been established that oral administration of spermine can induce all the modifications occurring in the digestive tract at weaning. In the intestine events occur in two phases. The early event consists of desquamation of the epithelium resulting from an activation of apoptosis. The late event appears to involve an hormonal cascade in which adrenocorticotropic hormone, cytokines, bombesin and corticosterone are included. Observations in human subjects show that: (1) the spermine and spermidine concentrations are generally lower in infant formulas than in human breast milk. Mothers seem consistently to have relatively high or relatively low concentrations of spermine and spermidine in their milk. These individual variations may be due to diet, lifestyle or genetic background; (2) the probability of developing allergy can reach 80 % if the mean spermine concentration in the milk is lower than 2 nmol/ml milk. It is approximately 0 % if the mean spermine concentration is higher than 13 nmol/ml milk; (3) preliminary results show that the intestinal permeability to macromolecules differs in premature babies when they are fed on breast milk compared with infant formulas (J Senterre, J Rigo, G Forget, G Dandrifosse and N Romain, unpublished results). This difference does not seem to be present when powdered milk is supplemented with polyamines at the concentration found in breast milk; (4) spermine increases proliferation and differentiation of lymphocytes isolated from the tonsils of children.