Ontogeny of Hepatic and Renal Systemic Clearance Pathways in Infants Part I
- First Online:
- Cite this article as:
- Alcorn, J. & McNamara, P.J. Clin Pharmacokinet (2002) 41: 959. doi:10.2165/00003088-200241120-00003
- 461 Views
Dramatic developmental changes in the physiological and biochemical processes that govern drug pharmacokinetics and pharmacodynamics occur during the first year of life. These changes may have significant consequences for the way infants respond to and dealwith drugs. The ontogenesis of systemic clearance mechanisms is probably the most critical determinant of a pharmacological response in the developing infant. In recent years, advances in molecular techniques and an increased availability of fetal and infant tissues have afforded enhanced insight into the ontogeny of clearance mechanisms. Information from these studies is reviewed to highlight the dynamic and complex nature of developmental changes in clearance mechanisms in infants during the first year of life.
Hepatic and renal elimination mechanisms constitute the two principal clearance pathways of the developing infant. Drug metabolising enzyme activity is primarily responsible for the hepatic clearance of many drugs. In general, when compared with adult activity levels normalised to amount of hepatic microsomal protein, hepatic cytochrome P450-mediated metabolism and the phase II reactions of glucuronidation, glutathione conjugation and acetylation are deficient in the neonate, but sulfate conjugation is an efficient pathway at birth. Parturition triggers the dramatic development of drug metabolising enzymes, and each enzyme demonstrates an independent rate and pattern of maturation. Marked inter-individual variability is associated with their developmental expression, making the ontogenesis of hepatic metabolism a highly variable process. By the first year of life, most enzymes have matured to adult activity levels.
When compared with adult values, renal clearance mechanisms are compromised at birth. Dramatic increases in renal function occur in the ensuing postpar-tum period, and by 6 months of age glomerular filtration rate normalised to bodyweight has approached adult values. Maturation of renal tubular functions exhibits a more protracted time course of development, resulting in a glomerulotubular imbalance. This imbalance exists until adult renal tubule function values are approached by 1 year of age. The ontogeny of hepatic biliary and renal tubular transport processes and their impact on the elimination of drugs remain largely unknown.
The summary of the current understanding of the ontogeny of individual pathways of hepatic and renal elimination presented in this review should serve as a basis for the continued accruement of age-specific information concerning the ontogeny of clearance mechanisms in infants. Such information can only help to improve the pharmacotherapeutic management of paediatric patients.
The dearth of pharmacological data specific for the infant population poses a significant obstacle to the successful pharmacotherapeutic management of such patients. Until recently, moral, ethical and legal issues had seriously impeded rigorous scientific investigation into paediatric drug pharmacokinetics, such that most marketed drugs merely contain disclaimers for their use in these populations. In the absence of age-specific pharmacological data, infant dosage regimens become based upon information available only in adult patient populations. Despite their prudence and conservatism, empirical approaches to pharmacotherapy still produce suboptimal, serious and even lethal consequences of drug exposure in infants. Such adverse outcomes exist because physiological and biochemical maturation of pharmacokinetic and pharmacodynamic characteristics alter the pharmacology and potential toxicity of drugs in infants as compared with adult patients.
Systemic clearance constitutes a critical pharmacokinetic parameter in the determination of a pharmacological response. Yet the exact impact of clearance pathway ontogenesis on an infant’s capacity to eliminate drugs remains largely unknown. With recent improvements in molecular technology, more precise information on the development of individual clearance pathways has become available. Part I of this review provides a comprehensive and descriptive account of the ontogenesis of hepatic and renal elimination mechanisms, principally based upon recently available molecular and in vitro enzyme activity data for hepatic metabolic clearance mechanisms and in vivo probe substrate data for renal clearance due to glomerular filtration. Such descriptions highlight the dynamic nature of these developmental processes during early postnatal life and how these processes may affect drug elimination and pharmacological response in infants. In general, hepatic and renal clearance mechanisms are underdeveloped and inefficient in the neonate, a condition further exacerbated in the premature infant. These mechanisms undergo dramatic developmental changes following birth and demonstrate different rates and patterns of maturation to adult values. Discussion herein of the ontogeny of such hepatic and renal clearance mechanisms is based upon hepatic microsomal activity data normalised to the amount (mg) of microsomal protein content and renal glomerular filtration rate (GFR) normalised to bodyweight (kg) for both infant and adult data. Such background information should allow the reader to draw his/her own conclusions concerning the risk of adverse drug exposure.
Based upon the in vitro enzyme kinetic data and in vivo probe substrate data, part II of the review proposes a tentative mathematical model to serve as a foundation for the development of a more comprehensive model that may predict infant clearance in the first 6 months of life.
1. Literature Search Methods
The following review is not intended to be exhaustive. It principally highlights the recently available in vitro mRNA, protein and enzyme activity data for individual hepatic drug metabolising enzymes and probe substrate data for the determination of GFR in the developing infant. When clinical studies reported systemic clearance values for relatively narrow infant age group ranges, the data were incorporated into the review article.
The review contains information obtained from a number of different sources. An Internet search using the Google search engine (http://www.google.com) and keywords such as ‘ontogeny of clearance’, ‘cytochrome P450’ and ‘drug transporters’ revealed several important sites offering general information on cytochrome P450 (CYP) enzymes, drug transporters and general paediatric pharmacological information. Particularly relevant sites provide general information on CYP enzymes, general information on CYP enzymes with a table of age-dependent changes in liver microsomal CYP enzymes and phase II enzymes and associated references, a human CYP enzyme metabolism database and a database for all drug transporters cloned to date. An online book provides an excellent overview of the ontogeny of CYP enzymes and important issues related to pharmacology and pharmacotherapy in the paediatric population.
Additionally, a literature search using the PubMed database of bibliographic information (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi) provided the bulk of the information cited. Keywords included various combinations of ‘ontogeny’, ‘development’, the specific enzyme in question (for example, cytochrome P450, glucuronosyltransferase, sulfotransferase), ‘clearance’, ‘infant’ or ‘neonate’, ‘fetus’, ‘transporter’ and keywords for the physiological determinants of clearance (for example, protein binding, hepatic blood flow, GFR). Standard neonatology and paediatric textbooks and review papers were also accessed.
2.1 Systemic Clearance
Extrahepatic metabolism and intestinal and other routes of elimination will not be discussed, but they may have notable importance for some drugs.
2.2 Oral Clearance
F becomes an important determinant of infant systemic exposure after oral drug administration or inadvertent exposure to drugs or environmental pollutants via the oral route (i.e. breast milk). The ontogenesis of factors governing the rate and extent of oral drug absorption may have a tremendous impact on drug bioavailability during infant development. Anatomical and physiological differences between infants and adults may result in differences in both the rate and extent of drug absorption in infants.[7–11] Low bioavailability is typically the result of first-pass effects. Literature data provides evidence of a postnatal maturation of hepatic drug metabolising enzyme and transporter activities, and it is probable that the activity of proteins involved in the metabolism and transport of drugs in the gastrointestinal tract also undergoes a postnatal maturation. Given the paucity of data on the ontogeny of drug metabolising enzymes and transporters in both the gastrointestinal tract and liver, any conclusion regarding drug bioavailability in infants is necessarily limited to broad generalisations. In general, F will probably decrease with postnatal age as the anatomical and physiological characteristics of systems governing absorption and first-pass effects mature with infant development. The lack of available data precludes detailed discussion of the impact of drug metabolising enzyme and transporter ontogenesis on first-pass effects in the infant, and this topic remains beyond the scope of this review.
3. Determinants of Hepatic Clearance
3.1 Hepatic Blood Flow
Hepatic blood flow governs the hepatic clearance of drugs that exhibit high extraction ratios. Few studies have assessed hepatic blood flow in the infant. Studies in lambs suggest high fetal hepatic blood flow, constituting 30% of total cardiac output. Fetal hepatic blood flow consists of highly oxygenated umbilical vein flow (70%), hepatic arterial flow (10%) and portal vein flow (25%). In the adult, portal vein flow (75%) and hepatic arterial flow (25%) comprise total hepatic blood flow. As well, a functionally patent ductus venosus shunts a significant portion of the umbilical vein blood past the liver parenchyma directly to the inferior vena cava in the fetus. The ductus venosus closes within the first week of life. Birth results in the abrupt cessation of umbilical venous blood flow, a 4-fold reduction in hepatic blood flow, a 3-fold decrease in oxygen delivery to the liver, and a marked enhancement in portal venous blood flow and an increase flow through the ductus venosus within the first 10 hours of life.[18–20]
Although these dramatic hepatic circulatory changes and changes in hepatic oxygen tension may affect hepatic function during the immediate postpartum period, little difference in the systemic clearance of high clearance drugs such as lidocaine was reported between newborn and adult patients. Since hepatic blood flow governs the hepatic clearance of lidocaine, similar systemic clearance values suggest that infant hepatic blood flow may be comparable to adult values. No reports exist to suggest that hepatic blood flow limits the efficiency of drug elimination in the infant. Furthermore, immature drug metabolising pathways or hepatic transport mechanisms, rather than hepatic blood flow, may limit the clearance of most drugs in the infant. Hepatobiliary mechanisms probably have the greatest impact on hepatic drug elimination in the developing infant.
3.2 Hepatic Transport Systems
The ontogeny of transport systems involved in the carrier-mediated hepatocellular uptake or biliary excretion of drugs in humans has received limited attention. Multiple transport systems have been identified in the adult liver.[22–24] These systems have overlapping substrate specificities and, furthermore, several transport proteins may mediate the uptake and/or excretion of an individual drug substrate. The developmental expression of transport processes may influence hepatic clearance by limiting biliary excretion and/or hepatocellular uptake and the rate of drug presentation to hepatic drug metabolising enzymes. Biliary excretion is important in the elimination of some drugs (e.g. anticancer drugs) and drug conjugates (glucuronides, glutathione conjugates), and in the excretion of endogenous bile acids and other organic anions (e.g. endogenous glucuronide conjugates).
Limited in vivo data suggest that hepatic excretory function is inefficient in infants. This assertion is supported by numerous animal studies that have shown deficient biliary excretion processes for a variety of compounds in infants.[26–30] Carrier-mediated hepatocellular uptake and efflux processes have been implicated as important mechanisms and potentially rate-limiting steps in the hepatic clearance of some drugs. As with biliary excretion, limited data exists concerning the postnatal development of these transport systems. Studies in animal models suggest carrier-mediated hepatocellular uptake is deficient in infants.[31–33]
Recently, several studies on the ontogeny of specific membrane transporters in animal models and humans have appeared in the literature.[34–41] In general, animal models have helped elucidate the roles of membrane transporters in drug absorption and disposition. However, few good animal models of postnatal development exist, and care should be exercised when extrapolating the developmental pattern and role of transporter expression in an animal model to human infants. The role of membrane transporters in hepatic clearance and to drug disposition as a whole is unknown, and the impact of the ontogeny of these processes awaits further investigation.
3.3 Hepatic Metabolism
Hepatic biotransformation is an important route of elimination for many drugs. In general, hepatic metabolism enhances the hydrophilicity of drugs to effect their renal elimination and abrogation of pharmacological/toxicological activity.
Hepatic drug metabolism may proceed via diverse pathways. These pathways are generally divided into phase I and phase II reactions, and often drug elimination requires both phases in a two-step sequential manner. In the adult liver, phase I and phase II reactions are well characterised, but much less information is available for infants. Phase I reactions are principally mediated by the CYP enzymes. These enzymes are located on the smooth endoplasmic reticulum and mediate the addition, formation or uncovering of a functional group on the drug molecule. Other phase I enzymes such as alcohol and aldehyde dehydrogenases, xanthine and amine oxidases, and esterases play a role in the hepatic metabolism of only select classes of drugs. Phase II reactions catalyse the conjugation of a water-soluble endogenous molecule to a drug compound and contribute significantly to the metabolism of both endogenous and exogenous compounds and their metabolites. The ontogeny of the important conjugation pathways (glucuronidation, sulfation and glutathione conjugation) in humans is not well characterised.
4. Approaches to Study Age-Related Changes in Hepatic Metabolism
Two general experimental approaches allow the study of age-related changes in drug biotransformation. The first involves the description of age-related changes in the systemic clearance of probe substrates in vivo. The second requires the assessment of age-related changes in drug metabolism following an in vitro analysis of immunoquantifiable enzyme levels using specific antibodies, mRNA levels using nucleic acid probes, and enzyme activity levels using specific probe substrates.
4.1 In Vivo Studies
Moral, ethical and technical concerns have limited the evaluation of the biochemical and functional maturity of a specific metabolic pathway in vivo. The measurement of metabolic enzyme activity requires serial blood and urine collections,[43,46,47] and few noninvasive methods (e.g. breath tests) exist to assess the developmental maturation of phase I or phase II enzyme function in vivo. Many paediatric clinical pharmacokinetic studies report average clearance values or half-lives of the drug investigated in infants of widely differing ages. Such studies ignore the dynamic and dramatic changes that occur in clearance mechanisms in the first months of life. Information from these studies generally lacks utility in the overall assessment of clearance capacity at defined infant age periods. As well, in vivo clearance data gathered from metabolic probe substrates must be interpreted cautiously, since their clearance may be determined additionally by renal clearance mechanisms and other phase I or II enzymes.
For example, midazolam undergoes hepatic oxidation principally by CYP3A4, which catalyses the hydroxylation of midazolam to the pharmacologically active 1α-hydroxymethyl metabolite.[48,49] The kidneys excrete 45 to 57% of an intravenous bolus dose of midazolam as the glucuronide conjugate of 1α-hydroxy-midazolam. Consequently, developmental maturation of CYP enzymes, glucuronide conjugating enzymes and renal elimination mechanisms principally determine the systemic clearance and pharmacological duration of action of midazolam in infants.
In general, the contribution of multiple pathways of elimination have resulted in conflicting data from studies that have attempted to assess the maturational stage of individual CYP enzymes using in vivo metabolic probes. For example, Vauzelle-Kervroedan et al. evaluated CYP3A4 maturation at birth using the ratio of 6β-hydroxycortisol to free cortisol (a specific CYP3A substrate). Newborn infant and adult ratios were similar, suggesting that CYP3A activity is fully developed in the newborn infant, but CYP3A activity declined with increasing age of the infant. This is consistent with the results of Nakamura et al., who reported that by 5 days of age the ratio decreased to one-third that of the first day after birth in term infants. Renal development and the importance of different CYP3A isozymes (see section 5.3.7) in the determination of this ratio may influence the in vivo determination of CYP3A4 activity. As well, factors other than enzyme activity may contribute to metabolic ratios of probe substrates. Consequently, in vivo probe substrates may lack sufficient specificity and may fall short in defining the developmental maturation of drug metabolising enzymes and its influence on infant systemic clearance.
4.2 In Vitro Studies
In the past several years, data have become available concerning protein, mRNA and activity levels of individual CYP enzymes in human fetal and infant hepatic tissues. Very limited data are available for the phase II enzymes important in drug metabolism. The ensuing discussion will focus on the ontogenesis of the various phase I and II pathways of hepatic metabolism based upon recently available in vitro data.
In the literature cited in this review article, the investigators ensured liver collection and storage (−70°C) within 1 hour of infant or fetal death (except for one report where samples were collected within 4 hours of death) and exclusion of livers from infants who received drugs known for their CYP enzyme inductive potential. In studies involving human hepatic drug metabolising enzymes, practical issues preclude the collection and processing of fresh liver samples and, often, liver tissue and hepatic microsomes are stored frozen for extended periods.
Several studies have examined the effects of freezing and thawing of human liver samples on the content and catalytic activities of drug metabolising enzymes.[54–57] Such studies indicate that rapid collection of post-mortem tissue is critical, since tissue hypoxia and autolysis results in a significant loss of mRNA content, hepatic metabolic enzyme content and enzyme activity within a few hours of death.[54–60] Several reports indicate that freezing storage of liver samples (−80°C) halts or slows the decline in CYP enzyme content and activities.[54–56] However, storage conditions of hepatic microsomes may influence CYP enzyme activity. Several authors report that CYP enzyme-dependent activities are very stable,[56,61,62] whereas Pearce et al. and Powis et al. reported a differential change in individual CYP enzyme activity, which suggests that CYP enzymes are not uniformly stable. Interestingly, in those studies, reduction in CYP enzyme activities with storage of liver tissues or hepatic microsomes did not result in an increased conversion of cytochrome P450 to the inactive cytochrome P420 form of the protein. Rather, the investigators postulated a decline in CYP enzyme activities due to contamination with other haem-containing proteins (such as haemoglobin) during tissue autolysis, and suggested that CYP enzyme protein content may not be an accurate indicator of CYP enzyme activity.[54,55] Despite the potential limitations of using post-mortem material and the effects of storage and processing on the expression of enzyme content and activity levels, human liver tissue and hepatic microsomes remain a widely used in vitro test for characterising phase I and phase II metabolism.[63,64]
5. Developmental Changes in Hepatic Metabolism
Although not readily apparent from table I, significant interindividual differences in the maturation of these enzymes exist, making the developmental expression of hepatic metabolic clearance a highly variable process. Although other factors may influence in vivo probe substrate clearances, the information in table I and table II demonstrates similarities in enzyme maturation between in vivo and in vitro probe substrate data. This suggests that in vitro enzyme activity data may provide estimates of in vivo clearance capacity. Part II of this review offers a tentative method to apply in vitro activity data to estimate infant systemic clearance at any age up to 6 months of life.
5.1 Hepatic Cytochrome P450 (CYP) Enzymes
Cytochrome P450 enzymes are a superfamily of haem-containing enzymes that catalyse the oxidative metabolism of a variety of lipophilic substrates. These enzymes are broadly classified into several families based upon sequence similarity. Only CYP enzyme families 1, 2, 3 and 4 play an important role in drug metabolism in the adult liver. All CYP enzymes have rather broad substrate specificities, which include both exogenous and endogenous compounds (steroids, bile acids, fatty acids, retinoids, biogenic amines, drugs and environmental pollutants). However, individual CYP enzymes have distinct and only partially overlapping substrate specificities. Constitutive and inducible forms exist, and several CYP enzymes express genetic polymorphisms, which may have important therapeutic implications in infants.[73,105,106] Such characteristics may have a significant impact on the interpretation of infant metabolic clearance.
5.2 Total Hepatic CYP Enzymes
CYP1A2, 2A6, 2B6, 2C8/9/19, 2D6, 2E1 and 3A4/5 comprise the principal drug metabolising CYP enzymes of the adult human liver and account for about 70% of total hepatic CYP content.[64,65,68] Figure 1 presents immunoquantifiable levels of these enzymes in adult livers (expressed as a percentage of total CYP content) and the percentage of drugs on the market metabolised by each specific CYP enzyme.[65,66,68] Immunoquantifiable levels of CYP enzyme do not necessarily correspond with its relative contribution to drug metabolism in the adult. For example, CYP3A accounts for about 30% of total CYP enzyme content in the liver and is responsible for the metabolism of about 50% of drugs on the market. On the other hand, CYP2D6 metabolises a significant percentage of marketed drugs (>30%), yet is found at low levels in the liver (about 2% of total). Figure 1 suggests that the developmental rate and pattern of CYP3A4, CYP2C (principally CYP2C8 and CYP2C9, which comprise 35 and 50% of total CYP2C, respectively) and CYP2D6 will have a significant impact on the infant’s capacity to eliminate drugs. CYP1A2, CYP2E1 and CYP2A6 are found at lower amounts in the liver and make only minor contributions to overall drug metabolism. As well, considerable interindividual variation in the hepatic content of each CYP enzyme exists (6-fold for CYP3A4 and 50-fold for CYP2D6). Such variation may confound estimations of hepatic clearance.
In general, CYP enzymes develop early in fetal life (table I). Electron microscopy studies have demonstrated early hepatocellular differentiation and development of the smooth endoplasmic reticulum by the 12th week of gestation, allowing for the development of functional CYP enzymes in the fetus.[16,108] This is consistent with literature reports demonstrating the metabolism of a variety of drugs in fetal hepatic microsomes. Total immunoquantifiable levels of CYP enzymes remain fairly stable at one-third the adult value throughout fetal life.[53,67,68,109,110] The early presence of functional CYP enzymes is not unexpected, given their role in maintaining the steady state levels of endogenous substrates involved in processes affecting homeostasis, growth and differentiation.
In the postnatal period, total hepatic CYP enzyme levels increase and reach adult levels by 1 year of age. If the assumption can be made that protein levels directly correlate with enzyme activity, then the drug metabolising capacity of the fetus and neonate is limited and efficient drug metabolism is not achieved until 1 year of age.
5.3 Individual CYP Enzymes
For most CYP enzymes, protein and activity levels in the fetus are low and parturition triggers their postnatal development.[17,107] This developmental increase in enzyme activity is consistent with literature reports of shortening half-lives and increases in in vivo clearances of drugs in developing infants (table II).[113–116]
fetal CYP enzymes;
early neonatal CYP enzymes; and
neonatal CYP enzymes.
Fetal CYP enzymes (CYP3A7 and CYP4A1) are expressed at high levels in the fetal liver and demonstrate extensive activity against endogenous substrates. A decline in their activity is observed in the postnatal period. Activities of early neonatal CYP enzymes (CYP2D6 and CYP2E1) generally surge within a few hours after birth, but activity is almost undetectable during fetal life. Activities of neonatal CYP enzymes (CYP3A4, CYP2C, CYP2B and CYP1A2) tend to rise only within the first few weeks after birth, exhibiting slower rates and onset of development as compared with early neonatal CYP enzymes.
Figure 2 represents a summary of the ontogeny of mRNA, protein, and activity levels of CYP enzymes expressed as a percentage of adult levels. Readily apparent in figure 2 is the lack of correlation between protein and mRNA levels and, in some cases, protein and activity levels during pre- and postnatal development. It is well known that gene transcription may be activated without concomitant increases in protein levels. Consequently, mRNA levels may precede increases in protein concentration in fetal and infant stages. Additionally, post-translational mechanisms may influence enzyme activity, resulting in low activity levels despite high amounts of immunoquantifiable protein. Attempting to assess systemic clearance on the basis of immunoquantifiable levels of enzyme or message levels may give erroneous ideas about the actual in vitro and in vivo ability of infants to eliminate drugs. A comparison of activity levels during defined stages of development provides a more direct means of assessing individual CYP enzyme function and an infant’s ability to eliminate drugs.
The following sections discuss briefly the ontogeny of the more important hepatic CYP enzymes according to the information supplied in table I, table II and figure 2. For a more in depth discussion, the reader should refer to Hakkola et al., Ring et al., Gow et al. and McCarver and Hines,[119,120] who present excellent reviews of the current understanding of the ontogeny of hepatic CYP enzymes during fetal and infant development, and de Wildt et al., who specifically discuss CYP3A ontogeny and disposition.
Although CYP1A2 is the major enzyme of the CYP1A subfamily in the liver (13% of total hepatic protein), the extrahepatic expression of CYP1A1 has toxicological importance due to its induction potential and ability to bioactivate procarcinogens. Limited information is available regarding the ontogenesis of extrahepatic CYP1A1. Several studies suggest an absence of quantifiable hepatic CYP1A2 mRNA and protein expression and negligible enzyme activity during fetal stages.[68,124–126] As well, CYP1A2 exhibits a protracted time course in its maturation and is the last major CYP enzyme to develop. The inability of the newborn and young infant to metabolise caffeine to paraxanthine, a CYP1A2 pathway, supports these observations.[70,112,127] This enzyme becomes readily detectable only by 1 to 3 months of age,[112,128] and adult levels may be approached only after 1 year of age.
CYP2A6 constitutes about 4% of total CYP enzyme content and has a limited role in drug metabolism (figure 1). Several studies have failed to demonstrate CYP2A6 activity in the fetal liver.[68,125,126] The expression of CYP2A6 during the postpartum period is largely unknown. One study found no significant differences in immunochemically detectable levels of CYP2A6 protein between infants greater and less than 1 year of age, suggesting that its development is complete by the first year of life.
CYP2B6 is detected at very low levels in human adult livers (figure 1). Fetal livers fail to express quantifiable mRNA and protein levels of this enzyme.[125,126] As with CYP2A6, the ontogeny of the CYP2B subfamily remains largely unknown. However, one study reported higher levels of expression of CYP2B6 enzyme in infants greater than 1 year of age as compared with younger infants, suggesting a delay in CYP2B6 development.
5.3.4 CYP2C Subfamily
The CYP2C subfamily is expressed at significant levels (20%) in the adult liver, consistent with its importance in drug metabolism (figure 1). This family of CYP enzymes contains several genes and alleles (i.e. CYP2C8, CYP2C9, CYP2C18 and CYP2C19), but CYP2C9 is the principal enzyme of the adult liver. Although low levels of mRNA have been detected in the fetal liver, fetal expression of CYP2C protein remains a matter of controversy. Some reports indicate a low level of protein expression,[125,126,128] whereas others failed to detect CYP2C protein.[67,68,75,110] Nevertheless, the fetal liver lacks any significant functional enzyme as evidenced by the lack of aryl hydrocarbon hydroxylase activity (a CYP2C-mediated reaction) in hepatic microsomes from fetuses and newborn infants.[110,131]
Within 1 day of birth, CYP2C mRNA levels rise significantly and reach adult values by 1 month of age.[53,67,128] Increases in CYP2C9 levels account largely for the increase in CYP2C mRNA levels postnatally. Similarly, 1 day after birth protein levels surge to reach one-third of adult levels by the first month,[53,67] without further apparent increases until 9 months to 1 year of age.[53,67] Livers from infants less than 8 days of age demonstrate limited CYP2C activity. Enzyme activity surges to 50% of adult values by 1 month of age, but declines slightly thereafter for the first year of life.
Clinical studies confirm the surge in activity levels observed in hepatic microsomes. Analysis of diazepam metabolites in the urine of infants administered intravenous diazepam demonstrated low metabolite levels in infants less than 2 days of age, but urinary metabolite levels increased markedly in infants greater than 1 week of age. After this time, metabolite levels remained stable in children up to 5 years of age.
CYP2D6 constitutes only 2% of total CYP content of the adult liver, yet is responsible for the metabolism of numerous drugs on the market (figure 1). The extensive contribution of CYP2D6 to drug metabolism has considerable significance, since it demonstrates an important genetic polymorphism (5 to 10% of Caucasians are poor metabolisers). Little literature evidence exists for functional expression of CYP2D6 during the fetal stage. Fetal livers expressed CYP2D6 mRNA after 12 weeks of gestation,[77,126] but several investigators failed to detect CYP2D6 protein and enzyme activity.[68,133] Interestingly, Treluyer et al. detected CYP2D6 protein in 30% of fetal livers examined, and a significant proportion of these positive samples were in fetal livers obtained after spontaneous and surgical abortions.
In the immediate postpartum period, CYP2D6 protein and activity begin to increase. Levels of mRNA increase to 2- to 3-fold of adult values in the first month of life, an increase that demonstrates considerable interindividual variability and no correlation with increases in protein levels.
Limited data are available for CYP2D6 developmental expression after 1 month of age. Protein levels were reported to be approximately two-thirds adult levels from livers 1 month to 5 years of age. However, another study reported no significant differences in protein levels of CYP2D6 in infants greater and less than 1 year of age, suggesting that CYP2D6 development is complete by 1 year of age.
CYP2E1 is moderately abundant (7%) in adult livers (figure 1). Despite its limited contribution to drug metabolism (figure 1) CYP2E1 metabolises several important compounds of toxicological significance (ethanol and cigarette smoke procarcinogens).
Fetal expression of CYP2E1 remains questionable. Several studies failed to detect fetal CYP2E1 mRNA and protein.[68,78,126,135,136] However, Carpenter et al. detected low levels of mRNA and activity in their fetal liver samples.
Postnatally, CYP2E1 protein and activity levels increase rapidly within the first 24 hours of birth without a significant increase in mRNA levels. By 1 month of age, increases in mRNA levels correlate with the increases in protein and activity. Activity, protein and mRNA levels rise gradually to reach adult levels by about 1 year of age.
5.3.7 CYP3A Subfamily
The CYP3A subfamily is quantitatively the most abundant CYP enzyme in adult livers (figure 1). In particular, CYP3A4 metabolises a wide variety of drug substrates and endogenous compounds and constitutes the most important CYP enzyme in drug metabolism (figure 1). Important members of this family include CYP3A4, CYP3A5 and CYP3A7. CYP3A4 is the principal adult liver enzyme, but low concentrations have been detected in fetal livers.[79,126] CYP3A5 demonstrates much lower and variable expression in adult and fetal livers.[140,141] Limited information is available concerning CYP3A5 hepatic expression in fetal and infant populations. In adult populations, approximately 30% of individuals express both CYP3A4 and CYP3A5, and in those individuals expressing CYP3A5 this enzyme may account for 50% of total CYP3A expression in hepatic tissue.[142,143] A similar percentage of CYP3A5-positive livers were noted in infant populations; protein levels surge within the first week of life and remain constant up to a year of life.
CYP3A5 generally exhibits a lower degree of catalytic capability and different regioselectivity as compared with CYP3A4.[142,144] However, higher catalytic activity has been noted for CYP3A5 with some substrates metabolised by both enzymes.[49,145] Possibly, the clinical pharmacokinetic variability in CYP3A substrates may relate to the expression pattern and differences in regioselectivity and catalytic capability of CYP3A5 as compared with CYP3A4.[49,143]
CYP3A7 represents the fetal form of CYP3A,[79,146–150] constituting about 30% of total fetal hepatic CYP enzyme content.[68,147] CYP3A4 and CYP3A7 exhibit high identity (95%) in the coding region of their nucleotide sequences, suggesting these enzymes share similar functional properties. Given the close similarity in the two enzymes, CYP3A7 and CYP3A4 may share overlapping substrate specificities, but important differences do exist. For instance, Ohmori et al. reported qualitatively differential catalytic capacities in the metabolism of endogenous (dehydroepiandrosterone 16α-hydroxylation, testosterone 6β-hydroxylation) and exogenous (carbamazepine 10,11-epoxidation, zonisamide reduction) substrates and inhibitory potential (triazolam, triacetyloleandomycin) between CYP3A4 and CYP3A7 in expression systems. Shimada et al. reported poor testosterone 6β-hydroxylation capacity of fetal liver microsomes but extensive metabolism in adult hepatic microsomes. 6β-Hydroxylation is an important pathway of CYP3A4-mediated testosterone metabolism in the adult, but is poorly catalysed by CYP3A7.
Expression of total hepatic CYP3A protein remains relatively constant throughout infant development. However, a developmental switch from CYP3A7 to CYP3A4 occurs during the early postnatal period, such that levels of CYP3A4 enzyme increase with a concomitant decrease in CYP3A7 enzyme levels. The pharmacokinetic consequences of this developmental switch remain unknown. An extensive qualitative and quantitative assessment of substrate specificities of CYP3A7 in comparison to CYP3A4 will help elucidate the impact of this postnatal switch from fetal CYP3A7 to CYP3A4 on drug metabolism in the infant.
Fetal livers express substantial CYP3A7 activity, which peaks by 1 week postpartum. Thereafter, CYP3A7 activity levels decline significantly during the first year of life and adult levels may be only about 10% that of fetal livers. Detection of CYP3A7 protein in adults remains equivocal[126,141,147,150] and Kuehl et al. suggest that residual expression of CYP3A7 may be due to a promoter region polymorphism. The fetal liver demonstrates limited CYP3A4 activity (about 10% of adult). Activity progressively increases postnatally to reach 30 to 40% of adult levels by 1 month of age and adult levels by 1 year. Fetal CYP3A4 mRNA levels are relatively low and increase rapidly following birth to plateau after 1 week of age. As with other CYP enzymes, little correlation exists between CYP3A4 mRNA and protein levels during the first months of life. Midazolam clearance in infant populations confirms the postnatal increase in CYP3A4 activity observed in vitro. In neonates, midazolam systemic clearance and production of one its primary metabolites, 1′-hydroxy-midazolam, is markedly reduced, especially in premature infants. However, by 3 months of age, midazolam clearance has increased almost 5-fold.
5.3.8 CYP4A Subfamily
5.4 Developmental Changes of the Liver
5.4.1 Anatomical and Functional Development of the Liver
The liver undergoes dramatic anatomical and physiological changes during prenatal and postnatal development. Ring et al. provide an excellent review on the anatomical and functional development of the fetal liver. As a brief summary, hepatobiliary morphogenesis and the organisation of the liver into defined acinar units occur in the first 10 weeks of gestation.[155,156] These changes are accompanied by the commencement of haemopoiesis, and dramatic increases in protein synthesis (due to increases in rough endoplasmic reticulum and Golgi apparatus) at 5 to 6 weeks gestation. By 10 weeks gestation, the smooth endoplasmic reticulum develops along with enhanced hepatic capacity for lipid and carbohydrate metabolism. Following the formation of the basic hepatic acinar structure in the first trimester, hepatocellular hyperplasia and hypertrophy result in linear liver growth through the remainder of the fetal period.[160,161]
At birth, the parenchymatous cells outnumber all other types of cells in the liver, but infant livers contain 20% fewer hepatocytes than do adult livers, and infant hepatocytes are only one-half the size of adult hepatocytes. Cellular hyperplasia and hypertrophy continue until the liver has reached full growth by young adulthood.
The acinus of the adult liver is divided into three functionally different zones. These functional differences may relate to adaptive responses to differences in the composition of sinusoidal blood as blood traverses from the periportal region to the pericentral region of the acinus.[165–167] Additionally, intrinsic alterations in hepatocyte genetic expression in each of the zonal regions contribute to the zonal functional differences of the acini. In the fetal liver, hepatocellular function is more homogenously distributed across the acinus and the zonal functional distinction of the acinus occurs only with postnatal development.[128,165]
In addition to the marked biochemical and morphological changes in the liver during the prenatal and postnatal period, marked haemodynamic changes occur during fetal (see Ring et al.) and postnatal life. At birth, with the cessation of umbilical blood flow, the infant liver becomes dependent upon its own immature and poorly oxygenated portal vein and hepatic arterial blood supply.[18,19] Dramatic changes in the architecture of both the portal venous and arterial blood vessel supplies occur during infant development.
In comparison to the adult, the newborn liver is anatomically and functionally immature. Both hepatocellular function and biliary epithelial function are generally deficient in the neonate.[169–171] Postnatal hepatic development is characterised by marked changes in the quantitative distribution of cells, cessation of hepatic haemopoiesis, changes in cell volume and an increase in enzyme activity and synthetic and metabolic capacity of hepatocytes and the biliary epithelium in general. These anatomical and functional changes may have a significant impact on both the quantitative and qualitative characteristics of drug elimination in the developing infant. How liver development and maturation affect hepatic drug elimination requires investigation.
5.4.2 Developmental Distribution of CYP Enzymes in Hepatocytes
Hepatocellular distribution and expression of CYP enzymes may change during neonatal development. To illustrate, enzymes of the CYP3A subfamily are homogeneously and abundantly expressed in the fetal liver. During postnatal development, hepatocellular expression of CYP3A protein shifts to centrilobular areas of the hepatic acinus, and, by adulthood, CYP3A expression is restricted principally to the centrilobular and midzonal hepatocytes with minimal expression in the periportal hepatocytes. CYP2C subfamily protein expression in the fetal liver is negligible, but a dramatic postnatal increase of CYP2C enzymes is observed in all hepatocytes during the first weeks of life. As with the CYP3A subfamily, expression of CYP2C enzymes progressively concentrates in the centrilobular and midzonal hepatocytes with maturation to adulthood. Additionally, CYP1A2 protein is detected only after several weeks postpartum, with CYP1A2 enzyme developing principally in the centrilobular hepatocytes.
The mechanisms underlying alterations in hepatocellular CYP enzyme distribution with infant development remain unknown, but may affect the apparent CYP enzyme activity observed in vivo. Any differences in the efficiency of metabolic elimination of drugs between the infant and the adult may relate not only to differences in the levels of CYP expression in the liver but also to the distribution pattern of these enzymes in the hepatic acinus.
5.5 Development of Phase II Reactions
Phase II enzymes metabolise a variety of endogenous compounds. Hence, these enzymes function in the regulation and homeostasis of physiological and biochemical processes important in growth, development and cellular function. Since substrate availability and enzyme activity may be limiting during the postnatal period, drugs may inhibit the endogenous function of phase II enzymes and cause poorly defined and subtle toxic effects in the developing infant. A serious potential exists for clinically significant adverse effects following long-term drug exposures in the developing fetus and infant.
5.5.1 Glucuronide Conjugation
Uridine 5′-diphosphate (UDP)-glucuronosyltransferases (UGT) catalyse the conjugation of glucuronic acid to their substrates. More than 18 different enzymes, divided into two families, UGT1A and UGT2B, according to sequence homologies, have been identified in humans. Members of the UGT1A family result from alternative splicing of a single gene transcript.[184,185] In contrast, separate genes give rise to the various members of the UGT2 family.
Numerous drugs, drug metabolites and endogenous compounds undergo metabolism and detoxification by UGT enzymes. Typical substrates of the UGT1A family include simple phenols, estrogens, bilirubin, primary amines and opioids, whereas UGT2 enzymes catalyse the glucuronidation of opioids, bile acids and endogenous steroids. Individual UGT enzymes demonstrate broad and overlapping substrate specificities. Consequently, several different UGT enzymes may participate in the metabolism of a single substrate, but their contribution will vary quantitatively to the overall metabolism of that substrate. For instance, UGT1A10 demonstrates 5- to 10-fold less activity than UGT1A9 for phenolic substrates,[182,190] but may glucuronidate a number of steroids with remarkably greater efficiency than UGT1A9.[182,190] Consequently, differences in the levels of expression of individual UGT enzymes may account for the marked variability reported in the glucuronidation capacity of infants.
In general, the glucuronidation capacity of infants is deficient such that drug elimination is prolonged and a risk for drug accumulation is possible.[191–193] This is consistent with immunoreactivity studies that have detected the presence of low levels of hepatic UGT enzymes during the fetal and early postnatal period.[174,175,194–196] Birth triggers the increase in UGT enzyme expression and, by 3 months of age, a full complement of UGT enzymes is present but at reduced levels (about 25% of adult levels). In vivo pharmacokinetic studies using morphine as an UGT enzyme substrate further support these findings, as premature and full term infants demonstrate reduced and variable morphine clearances due to immaturity of the glucuronidation pathways.
The ontogenesis of only a few UGT enzymes has been examined. The fetal liver expresses low immunoquantifiable levels of UGT1A1 protein and activity as compared with the adult. Postnatally, UGT1A1 activity and protein levels dramatically increase in parallel until adult levels are reached around 3 to 6 months of age. The UGT2B17 enzyme, which is responsible for the metabolism of androgenic steroids, catalyses testosterone glucuronidation at 3 and 13% of adult levels in fetal and infant liver microsomes, respectively. This suggests that UGT2B17 undergoes developmental increases in expression postnatally. Fetal liver expression of UGT1A6 mRNA was demonstrated at 20 weeks of gestation. Interestingly, fetal and adult liver microsomes catalyse serotonin glucuronidation at similar levels, and UGT1A6 was recently implicated in the metabolism of this neurotransmitter. Finally, morphine is a suggested probe substrate of UGT2B7. Fetal hepatic microsomes at 15 to 27 weeks of gestation catalysed morphine glucuronidation with 10 to 20% the efficiency of adult hepatic microsomes.[199,200] Clinical studies suggest that morphine glucuronidation is deficient in young infants and matures to adult values by 6 to 30 months of age, depending upon whether morphine clearance was corrected according to bodyweight or an allometric model.[96,201,202]
5.5.2 Sulfate Conjugation
Sulfate conjugation represents another important phase II pathway in the metabolism and elimination of drugs and endogenous compounds. Sulfotransferases are a family of cytosolic proteins that catalyse the conjugation of inorganic sulfate, derived from the donor molecule 3′-phosphoadenosine-5′-sulfophosphate (PAPS), with compounds containing hydroxyl functional groups.[173,177] Each sulfotransferase enzyme exhibits different substrate specificities which overlap with those of the UGT enzymes. The number of different sulfotransferase enzymes is unknown, but they can be loosely categorised into catechol sulfotransferases and phenol sulfotransferases, which catalyse the sulfation of dopamine (SULT1A3) and p-nitrophenol (SULT1A1), respectively.[177,180,204–206] Catechol sulfotransferases develop earlier in fetal life as compared with the phenol forms. Richard et al. demonstrated significant levels of SULT1A3 protein and activity during early fetal development, with a substantial decrease in the late fetal/early neonatal period and its absence in the adult liver. As well, SULT1A1 activity and protein were higher in fetal tissues as compared with infant livers.
In general, the fetus expresses significant sulfotransferase activity,[176,206,207] but the extent of activity is substrate dependent. For example, fetal liver sulfation of ritodrine exhibited about 1.5-fold greater activity than adult livers. On the other hand, fetal 2-naphthol sulfation was only one-third of adult levels, and demonstrated marked variability in activity (up to two orders of magnitude difference). Sulfotransferases in fetal, infant and adult liver cytosols failed to demonstrate significant differences in activity towards hydroxylamine and hydroxamic acid model compounds, suggesting a lack of age-dependent expression of sulfotransferase enzymes for these compounds. Sulfotransferases involved in dehydroepiandrosterone sulfation demonstrate low levels of hepatic expression in the first 25 weeks of fetal life, but rapidly increase in later gestational periods with adult levels being reached by the early postnatal period.
As with the UGT enzymes, accurate interpretation of in vitro and in vivo clinical data awaits the identification of individual sulfotransferase enzymes and specific probe substrates. Nevertheless, significant hepatic sulfotransferase activity is expressed in the fetus and infant, which suggests that sulfate conjugation is an efficient pathway in infants. For some drugs that undergo extensive glucuronidation in the adult, sulfation may become the predominant pathway in the fetus and infant, since sulfotransferase and UGT enzymes exhibit overlapping substrate specificities. For example, ritodrine, a β2-adrenoceptor agonist used in the management of preterm labour, is equally metabolised both by sulfation and glucuronidation in the adult. Extensive sulfate conjugation of ritodrine in infants resulted in efficient elimination of the drug despite low levels of the glucuronide conjugate as compared with the adult. Consequently, sulfation may play an important role in homeostasis and detoxification in the fetus and newborn infant, since other conjugation pathways (such as UGT) are not well developed until the postnatal period.
5.5.3 Glutathione Conjugation
Glutathione-S-transferases (GST) represent a family of proteins responsible for the conjugation of glutathione with a wide variety of electrophilic or reactive lipophilic and alkylating agents. In addition to their ability to metabolise drugs, these enzymes offer a protective function by catalysing the covalent binding of glutathione to alkylating chemicals. Up to five different families of GST with overlapping substrate specificities exist.[209,210] At least three genetic loci (GST1, GST2, GST3) give rise to the mu, alpha, theta and pi classes of this family of enzymes.[211–213] A complex pattern of tissue-specific and time-dependent expression of these enzymes has been observed,[179,214,215] but the liver expresses the greatest amount of GST protein.
Fetal livers exhibit a more limited and dissimilar pattern of GST enzyme expression as compared with the adult.[179,210,216,217] A gradual increase in GST1 activity with a concomitant decline in GST3 activity levels is observed during the fetal stage. Both alpha- and pi-class GST enzymes demonstrate intense expression in hepatocytes between 16 and 24 weeks of gestation, but alpha-class GST enzymes predominate in infant and adult livers. By 6 months of age, hepatocellular pi-class GST enzymes are nonexistent[218,219] and remain present only in epithelial cells of the biliary canaliculi. The literature presents conflicting information about mu-class GST expression. Mera et al. demonstrated mu-class GST expression by 30 weeks gestational age in fetal livers, but Pacifici et al. found it to be lacking from fetal livers examined. In addition, theta-class GST enzymes are expressed in the adult liver but were not detected during the fetal stage.[213,214]
Individual GST enzymes mature at an independent rate and pattern of development. Quantitative and qualitative differences in GST enzyme expression profiles during development may modify the efficiency of GST substrate metabolism in the infant. For example, mu-class GST enzymes preferentially catalyse the conjugation of benzo[a]pyrene 4,5-oxide. An absence of mu-class GST enzymes in fetal livers resulted in average GST activities towards benzo[a]pyrene 4,5-oxide that were one-half adult levels.[179,216] In another study, despite similar activity towards chlorodinitrobenzene, preterm infant livers exhibited 60% greater activity towards chloramphenicol than fetal livers. These data suggest that different GST enzymes, which undergo different patterns of development, metabolise these two substrates. As well, fetal livers catalyse glutathione conjugation of styrene oxide with more than 7-fold less activity than adult livers.[222,223] Yet, fetal and adult liver exhibit similar GST enzyme activities against 1-chloro-2,4-dinitrobenzene.[224,225] Interestingly, Faulder et al. failed to find any significant developmental changes with respect to total GST activity.
The data on the developmental expression of GST enzymes remain confusing, but suggest that the GST enzyme developmental pattern is substrate-dependent and relatively well developed in the infant. The clinical significance of the quantitative and qualitative differences in the developmental expression of individual GST enzymes remains unknown, but these differences have implications for toxicity and carcinogenesis.
N-Acetyltransferases (NAT) have wide tissue distribution.[227–229] These enzymes catalyse the transfer of the acetyl moiety of acetyl-CoA to NAT substrates. Two human genes, NAT1 and NAT2, have been identified, and each enzyme form possesses different substrate specificities.[228–230] Several different allelic forms of NAT1 and NAT2 genes exist, but only NAT2 demonstrates polymorphic activity.[228–230] The genetic polymorphism of NAT2 enzymes leads to the wide interindividual variability reported in NAT activity and divides the general population almost equally into slow and rapid acetylators.[228,229]
The ontogenesis of hepatic NAT enzymes has received limited attention. The literature indicates a substrate-dependent developmental pattern of NAT activity, which suggests quantitative differences in the rate of NAT1 and NAT2 maturation in the fetus and infant. NAT activity develops during the first trimester and fetal livers are capable of catalysing the acetylation of several substrates specific for the different NAT enzymes,[227,231] even at 9 to 12 weeks gestation. However, the activities of NAT enzymes are significantly lower in the fetal liver as compared with the adult, and extrahepatic organs contribute significantly to the overall metabolism of NAT substrates in the fetus.
Several studies have investigated the postnatal maturation of acetylation status in infants. In general, the acetylation capacity of full-term infants is limited, which is further exacerbated in the premature infant. With respect to NAT2 activity, infants less than 1 year of age are generally (83%) slow acetylators.[102,233] An age-dependent alteration in phenotype distribution occurs, such that many infants who are genotypically fast acetylators become fast acetylators by 2 to 4 years of age (48% slow acetylators), depending on the substrate.[102,233,234] Substrate-dependent differences in the rate of maturation of acetylator phenotype suggest that different NAT2 enzymes are involved; these await characterisation. More investigation is required to evaluate the ontogenesis of NAT enzymes but, in general, acetylation status of the infant is deficient and requires at least 1 year of life to reflect adult levels.
5.6 Developmental Effects on Metabolite Profiles
Many drugs undergo multiple routes of metabolism. Given the enzyme-specific differences in the rates and patterns of conjugative and CYP enzyme ontogenesis, the relative contribution of different routes of metabolism may vary with infant age, causing changes in metabolic pathway predominance during infant development. This may lead to altered metabolite profiles between the infant and the adult, a significant concern when pharmacologically active or toxic metabolites are produced. The metabolism of theophylline presents a unique example of how differences in the rate and pattern of enzyme maturation may lead to a pharmacologically significant change in metabolite profiles.
Adults metabolise theophylline to 3-methylxanthine and uric acid derivatives, whereas full-term and premature infants principally metabolise theophylline to caffeine and eliminate theophylline unchanged in the urine,[235–237] since the N-demethylation pathway (CYP1A2) is essentially absent from newborn infants.[238,239] Newborn infants eliminate caffeine very slowly (preterm infants demonstrate a half-life of 36 to 144 hours) and significant accumulation of caffeine, with concomitant adverse effects, is possible.[235–240] With maturation of CYP1A2, infant clearance values begin to approach adult values by 4 to 5 months of age and metabolite patterns become more similar. Consequently, failure to consider the rate, pattern and extent of metabolic pathway maturation may lead to an adverse exposure outcome during the pharmacotherapeutic management of an infant.
Alternatively, despite an altered metabolic profile, a well-developed route of metabolism in an infant may allow for efficient drug elimination even though a different metabolic pathway acts as the principal route of elimination in the adult. For example, the highly active sulfotransferases in infants may efficiently eliminate drugs as sulfate conjugates despite their extensive glucuronidation in adult populations.[242–244] Both paracetamol (acetaminophen) and ritodrine undergo extensive glucuronide conjugation in adults but sulfate conjugation in infants. Although no age-related changes in overall elimination are observed, quantitative differences in metabolite levels, between infants and adults do exist.[245,246] The predominance of a specific elimination route may change at different stages of infant development. The possibility of an adverse outcome following an infant exposure to a drug whose primary path(s) of elimination is poorly developed will be dependent upon the availability of an alternative but efficient pathway to effectively eliminate the drug in the infant.
5.7 Interindividual Variability in the Developmental Expression of Hepatic Metabolism
Drug pharmacokinetics often exhibit quite substantial inter- and intraindividual variability due to genetic and/or environmental causes. Interindividual differences in the rates and patterns of the development of individual clearance pathways may largely explain the variations in rates of renal and hepatic elimination observed in infant patients. Phenytoin pharmacokinetics in the infant provides an excellent example. At birth, the ability of the full term infant to metabolise phenytoin is only 30% of the adult capacity. Plasma half-lives are prolonged and extensive variability in half-lives is observed in the first 3 days of life. With increasing postnatal age, both plasma half-life and the degree of interindividual variability decrease. Differences in the extent of CYP enzyme expression at birth contribute to the wide interindividual variability, but rapid maturation to adult values results in a quantitative reduction in half-lives and associated interindividual variability. Understanding the significance of this variability is essential when attempting to determine infant capacity to eliminate drugs.
Drug metabolising enzymes may undergo induction and/or inhibition upon exposure to certain drugs or endogenous compounds.[249,250] Enzyme induction may occur with relatively greater magnitude and speed in the infant as compared with the adult. Enzyme induction and inhibition following uterine or postnatal exposure to exogenous compounds will further exacerbate the variable rate and pattern of maturation of the drug metabolising enzymes. For example, uterine exposure to antiepileptic drugs results in enhanced aminopyrine N-demethylation in exposed infants. Induction of metabolic enzymes may explain the wide inter- and intrapatient variability observed in theophylline pharmacokinetics in infants. Induction agents such as barbiturates may alter the pattern of CYP2C development.[67,253,254] Neonates treated with diazepam, a CYP2C substrate, alone demonstrated prolonged half-lives for diazepam (31 ± 2 hours) as compared with infants treated concomitantly with barbiturates (18 ± 1 hours).[253,254] The possibility of previous or concurrent exposure to enzyme inducers/inhibitors requires consideration in any assessment of infant capacity to eliminate drugs.
Genetic polymorphisms in drug metabolism may result in enhanced, diminished or unaltered enzyme activities, leading to marked interindividual variability in drug pharmacokinetics. Several phase I and phase II enzymes, including CYP2D6, CYP2C9, CYP2C19, CYP2E1, UGT and NAT, exhibit genetic polymorphisms. Such polymorphisms may increase the risk of an adverse outcome to a chronic drug exposure in the infant if these polymorphisms compromise metabolic clearance and slow drug removal. Polymorphisms in drug metabolism and the potential for enzyme induction and/or inhibition add further complexity to the assessment of elimination capacity of an infant.
5.8 Regulation of the Developmental Expression of Hepatic Metabolic Enzymes
The various phase I and phase II enzymes exhibit marked differences in their rates and patterns of maturation. In many cases, developmental changes in mRNA, protein and activity levels fail to correlate for an individual enzyme (table I and figure 2). Although the regulatory mechanisms of enzyme expression during ontogenesis remain largely unknown, the dramatic postnatal development of many enzymes suggests that parturition events play a primary role in expression and activation of these enzymes.[256,257] The pattern of CYP enzyme regulation may relate to the maturation of normal biochemical processes of the body, since many of these drug metabolising enzymes contribute to the metabolism and homeostasis of endogenous compounds.
Expression of CYPs and conjugative enzymes is probably controlled at the post-translational (i.e. substrate-dependent stabilisation, phosphorylation), translational (i.e. mRNA stability) and transcriptional (i.e. activation of gene expression or removal of an inhibitor) levels of regulation. In the adult liver, transcriptional control of constitutive enzyme expression has been postulated. Transcriptional regulation is still incompletely known, but the 5′-flanking regions of several human CYP enzymes have been determined and several putative transcription factor binding sites identified. Such sites include hepatic nuclear factor (HNF), CAAT/enhancer binding protein, AP-1 and glucocorticoid receptor (GRE) cis-acting elements.[79,117,258–260] Hence, transcriptional regulation during the maturation of these enzymes may be due to the developmental expression of such transcription factors, and different transcription factors may predominate at various stages of development to control enzyme expression.
Several studies have examined the developmental regulation of CYP2E1, the CYP enzyme important in ethanol metabolism. Early fetal livers (<18 weeks gestation) and late stage fetuses fail to express quantifiable levels of mRNA, protein or activity.[78,136] Marked differences in the methylation status of the 5′region of the CYP2E1 gene were noted in these studies, suggesting that methylation of specific 5′ residues may account for the lack of CYP2E1 transcription during fetal stages. Within the first hours of parturition, CYP2E1 protein and activity levels rise dramatically and reach maximum levels by 1 week of age. Stabilisation of CYP2E1 protein may explain the rapid and dramatic rise in enzyme protein and activity without concomitant increases in mRNA levels during the first week of life. From 1 month to 1 year, enzyme protein and activity increase in concordance with mRNA levels, suggesting transcriptional activation of CYP2E1 expression. Activation of CYP2E1 transcription may be controlled in part by HNF-1α,[261,262] and may correlate with decreased methylation of the CYP2E1 gene. Furthermore, infant and adult CYP2E1 may undergo enzyme induction due to post-transcriptional regulatory mechanisms (i.e. substrate-induced protein stabilisation).[117,263] Regulation of CYP2E1 maturation exemplifies the complexity of the regulatory processes likely to be involved in the developmental maturation of drug metabolising enzymes in the fetus and infant.
6. Renal Clearance
In general, glomerular filtration governs the renal elimination of many drugs. Only a limited number of drugs undergo active tubular secretion or reabsorption. At birth, anatomical and functional immaturity of the kidney limit glomerular and tubular functional capacity, which results in inefficient drug elimination and prolonged half-lives.[264–267] Rapid improvements in renal glomerular and tubular functions occur during the postnatal period, greatly enhancing renal drug elimination. Postnatal renal maturation is exemplified by the renal clearance of digoxin. Glomerular filtration and active tubular secretion processes govern digoxin clearance, and average renal clearances of 0.624, 1.97, 5.33 and 8.67 L/h/1.73m2 (10.4, 32.9, 88.9 and 144.44 ml/min/1.73m2) have been reported in premature infants, full term infants less than 1 week of age, 3-month-old infants and children 1.5 years of age, respectively.[268,269]
6.1 Anatomical Development of the Kidney
The developing kidney undergoes profound structural and functional changes that alter the elimination of drugs during fetal and infant development. Nephrons of the nascent kidney begin development at 8 weeks of gestation and continue to increase in number until 36 weeks gestation.[96,270–273] Morphological changes in the glomeruli of the nephrons between 34 and 36 weeks of gestation result in significant maturation of fetal glomerular function. At 36 weeks, nephrogenesis is complete and no new nephrons are formed.[270–273,275] Renal tubular growth contributes exclusively to the large increase in renal mass from 36 weeks gestation to adulthood.[270–273,275] Prior to 36 weeks gestation, the increase in GFR is due principally to increases in the number of glomeruli.[276–278] At 36 weeks of gestation with a full complement of nephrons, GFR is only about 5% of the adult value and renal blood flow is diminished. GFR reflects the stage of fetal development, and a positive correlation exists between the GFR and gestational age in newborn infants born between 27 and 43 weeks of age.[264,279] Consequently, gestational age has a marked impact on renal function, since incomplete nephrogenesis further compromises glomerular and tubular function in premature infants.[266,279]
Renal tubular development continues throughout fetal and infant maturation into adulthood. As compared with the adult, renal tubular function is severely compromised in the infant due to limited renal tubular size and mass, a condition further exacerbated in the premature infant. As well, insufficient peritubular blood flow, immaturity of the biochemical processes supplying energy, a limited ability to concentrate urine and lower urinary pH values further hinder tubular function in infants. Each of these processes undergo an independent rate of development.[8,266,281] The development of tubular function generally exhibits a more protracted time course than GFR, resulting in a functional glomerulotubular imbalance until renal maturation is complete by the first year of life.[8,270,280–282]
6.2 Functional Development of the Kidney
Upon examination of the information presented in table V, absolute clearance, which allows an explicit description of age-specific maturation of renal function, demonstrates generally low renal functions at birth, which undergo rapid development to adult levels within the first year of life. Similarly, renal functions normalised to body surface area demonstrate a rapid maturation to adult levels by 1 year of life. On the other hand, renal function normalised to bodyweight shows an irregular development and little maturational trend to adult values. Care should be exercised in the evaluation and interpretation of the information presented in table V.
6.2.1 Glomerular Filtration Rate
Vasoconstriction of the renal microvasculature and diminished renal blood flow contribute principally to the differences in GFR between the full term infant and the adult.[270,301,302] Parturition triggers a dramatic decrease in renal vascular resistance and enhancements in both cardiac output and renal blood flow (see table V).[286,303–307] These circulatory changes and possibly increases in glomerular permeability and filtering area[308,309] lead to a rapid rise in GFR during the early postnatal period.[266,310–313] At birth, renal blood flow is only 5 to 6% of cardiac output and reaches adult values (15 to 25%) by 1 year of age.[301,314,315] Additionally, effective renal plasma flow is typically determined following a clearance estimate with a low dose of p-aminohippurate (PAH). Table V indicates a rapid increase in renal plasma flow in the first year of life, which reaches adult values after 2 years of age.[286,287,302] However, PAH clearance underestimates renal plasma flow during the postnatal period. In adults, PAH is completely extracted by the kidney, making it an effective marker for the determination of renal plasma flow. In the first 3 months of life, the renal extraction of PAH is only 60%, and is 95% by 5 months of age.
Enhancements in renal blood flow and redistribution of blood flow within the kidney lead to a rapid rise in GFR during the early postnatal period. In full term infants, GFR determined by inulin or creatinine clearance is generally only 10 to 15 ml/min/m2 (2 to 4 ml/min),[265,266] but it undergoes rapid postnatal maturation.[287,291] By 1 to 2 weeks of age, GFR is 20 to 30 ml/min/m2[264,266] (8 to 20 ml/min) and adult levels (73 ml/min/m2 or 127 ml/min; inulin) are approached by 6 months of age. For drugs whose renal clearance is governed by GFR, the rapid improvement in efficiency of glomerular filtration leads to rapid enhancements in renal drug clearance and a diminished risk of significant drug accumulation.
At birth, a direct proportionality exists between gestational age and GFR in the full term infant. Furthermore, the increase in GFR correlates with postconceptual age rather than postnatal age.[265,266,279,290,317–320] This suggests that, in addition to postnatal changes in renal vasculature resistance and blood flow, the extent of renal anatomical maturation influences GFR development in the immediate postpartum period.[265,279,294] Consequently, premature infants on average exhibit much lower GFR values (5 to 10 ml/min/m2 or 0.7 to 2 ml/min), and incomplete nephrogenesis results in a slower pattern of postnatal development, since less dramatic increases in GFR occur during the first 1 to 2 weeks postpartum (10 to 12 ml/min/m2 or 2 to 4 ml/min) as compared with the full term infant.[11,265,279,290,291] After the first week of life, enhancements in GFR proceed at the same rate in preterm and full term infants, but even by 5 weeks of age the absolute value for GFR remains lower in preterm infants. This functional delay in the development of glomerular filtration efficiency is an important consideration in the estimation of the capacity of premature infants to eliminate drugs in the early postpartum period.
Table V provides, when data are available, gentamicin, creatinine, inulin and mannitol clearances as markers for GFR development in premature (<37 weeks) and full term (≥37 weeks) infants. Gentamicin clearances correlate well with GFR, since tubular secretion is a minor pathway of gentamicin renal elimination. Creatinine clearance, as a marker of GFR, may be somewhat inaccurate since creatinine is both filtered by the glomerulus and secreted by the tubules. As alluded to earlier, renal tubular development lags behind glomerular filtration efficiency and this glomerulotubular imbalance may affect infant creatinine clearances. As well, infant creatinine concentrations may reflect more closely maternal creatinine levels in the early postnatal period. Hence, high initial maternal creatinine loads may result in an overestimation of GFR in the early postnatal stage.[325–327] However, a recent study found a good correlation between serum creatinine levels and GFR, supporting creatinine clearance as an appropriate measure of GFR.[290,292,319] Inulin is often used as a marker of GFR, but reliable estimates of inulin clearance require a 24-hour continuous intravenous inulin infusion. Typically, mannitol clearance as a marker of GFR is determined following a single intravenous bolus injection.
6.2.2 Renal Tubular Function
Anatomical and functional immaturity of the renal tubules exists at birth. In general, incomplete anatomical development of renal tubules compromises both passive reabsorption[310,330] and active secretion and reabsorption processes[284,331,332] in the developing infant.[8,270,280,282,301] Increases in renal tubular growth and mass, maturation of renal tubular transport systems and redistribution of blood flow to the secretory areas of the kidney probably account for the enhancements in renal tubular transport function during infant development.[266,311,333] Renal tubular development is delayed somewhat compared with GFR, such that there exists a glomerulotubular imbalance until renal tubular maturation is complete at about 1 year of age.[8,10,282,301]
Both drugs and endogenous compounds may serve as substrates for renal tubular active transport systems. However, limited information is available on the development of drug carrier systems at the renal tubular epithelium. Functionally, the kidney exhibits a reduced capacity to excrete weak organic acids such as penicillins, sulfonamides, cephalosporins and phenolsulfophthalein.[284,331,332] For example, renal tubular secretion of high doses of PAH, a substrate for the predominant organic anion transport system of the kidney, was only 20 to 30% of adult values at birth and approached adult levels only by 7 to 8 months of age (see table V). These results are consistent with the in vivo elimination of furosemide, a PAH transport pathway substrate, where half-lives of 19.9 and 7.7 hours in premature and full term infants, respectively, were observed as compared with 0.5 hours in the adult. Other than in vivo descriptions of changes in tubular transport function, the literature is devoid of information concerning the ontogeny of individual renal transport proteins and their impact on renal elimination efficiency in human infants.
For drugs principally eliminated by the kidney, immature renal clearance mechanisms will result generally in the inefficient elimination of drugs and a prolongation of their time in the body (i.e. longer half-life). When renal tubular mechanisms are important in the elimination of a drug, the disproportional rate of development of glomerular filtration and tubular function (i.e. the glomerulotubular imbalance) may have variable and complex effects on the renal clearance of that drug. For example, infant renal clearance values for a given drug may exceed adult values, since a low GFR may be matched by a greater reduction in tubular reabsorption capacity (see equation 5). Stimulation of renal tubular transport systems following uterine or infant exposure to certain agents adds further complexity in the estimation of infant renal function.[281,336] Induction of transporter expression may enhance tubular transport functions and reduce the effects of glomerulotubular imbalance as compared with the uninduced infant. Finally, infant urinary pH values are generally lower than adult values. Urinary pH may influence the reabsorption of weak organic acids and bases, and differences in renal drug elimination may reflect the discrepancy in urinary pH values.
7. Drug Transporters
The literature lacks information concerning the ontogenesis of specific drug transporters in the infant. A few studies report the ontogeny of various transporters in animal models and the reader is referred to these references.[34–39] However, as with other renal and hepatic clearance mechanisms, differences in the ontogenesis of drug transport proteins probably exist between human infants and animal models. Apart from the multidrug resistance drug transporter, no studies were found on the ontogenesis of drug transporters in the human fetus and infant.
The multidrug resistance family of transporters is encoded by two genes, MDR1 and MDR3.[337,338] The MDR1 gene encodes for P-glycoprotein, an energy-dependent efflux pump that transports a wide range of amphiphilic hydrophobic compounds. In adult tissues, P-glycoprotein is expressed in epithelia of organs with excretory function (colon, kidney, liver, adrenal gland, placenta) and in blood-tissue barrier sites (blood, testes).[339–341] The expression pattern of P-glycoprotein suggests a role for this protein in the transport of drugs and endogenous compounds, and a role in barrier function and organ-specific drug sensitivity.
A few studies have evaluated the ontogenic expression of P-glycoprotein in the fetus. P-glycoprotein expression was detected in the 7-week embryo, and differences in tissue distribution between adult and fetus were observed. P-glycoprotein mRNA and protein were detected in the hepatic biliary tract and kidney tubules by weeks 11 to 14 of fetal life, in the gastrointestinal tract at 28 weeks, in the adrenal by 13 weeks and in the brain at 28 weeks gestation. Interestingly, Schumacher and Mollgard demonstrated P-glycoprotein expression in the endothelium of brain microvessels at 8 weeks of development. Expression of this transporter during infant development awaits further study.
8. Premature Infants
Most of the information presented in this review relates to the full term infant. Premature infants (gestational age <37 weeks) may exhibit markedly different pharmacokinetic characteristics as compared with the full term infant. In general, incomplete fetal development further exacerbates the anatomical and functional immaturity of the organs of elimination and other biochemical and physiological processes of the newborn infant. This will compromise further the capacity of the premature infant to eliminate drugs. Parturition triggers the development of many drug metabolising enzymes, but immaturity of other biochemical and physiological processes may impede the rate of metabolic enzyme maturation and metabolic clearance in the premature infant. Incomplete nephrogenesis and renal function will compromise the premature infant’s capacity for renal elimination of drugs as compared with the full term infant. Finally, premature infants present additional concerns such as the integrity of the blood-brain-barrier, integrity of the gastrointestinal tract and lower plasma protein binding of drugs, all processes that influence the extent of drug exposure in the infant.
An assumption implicit in the pharmacotherapeutic management of paediatric patients is that infants and adults exhibit comparable relationships between drug concentration and response. Few pharmacodynamic studies have been conducted in infants to substantiate this assumption. Age-specific differences in receptor concentrations, intrinsic activities and affinities for drugs may result in altered concentration-response relationships,[114,252,344] and large pharmacokinetic and pharmacodynamic variations involving drug receptor affinity and density have been reported. The pharmacokinetic/pharmacodynamic interface in the infant has received limited attention in the literature, partly because age-related pharmacodynamic differences can often not be distinguished from age-related pharmacokinetic differences in paediatric drug studies. An enhanced understanding of the developmental maturation of clearance mechanisms will help to identify pharmacodynamic differences of clinical relevance.
10. Implications of Developmental Changes in Clearance Mechanisms
In general, systemic clearance mechanisms in the developing infant are inefficient. Drug elimination is prolonged, plasma half-lives are increased and a potential exists for the accumulation of drug following chronic exposure. The extent to which drug elimination is compromised in infants depends upon the contribution of a particular pathway to its elimination, the relative maturity of this pathway and the availability of alternative well-developed routes of elimination. Immaturity of clearance mechanisms may lead to saturable elimination and non-linear, dose-dependent, pharmacokinetics, which may enhance the potential for an adverse drug reaction. A classic example is the widely cited ‘grey baby’ syndrome, which results from the reduced capacity of the infant to glucuronidate chloramphenicol. As well, metabolite profiles may change with development and a failure to recognise this property may result in an adverse outcome. The inability of the neonate to N-demethylate theophylline to 3-methylxanthine results in the enhanced formation of the methylated metabolite, caffeine. The neonate has a poor capacity to eliminate caffeine and this metabolite may accumulate to toxic concentrations.[238,239]
When little pharmacokinetic information is available, careful consideration of the developmental aspects of systemic clearance mechanisms is fundamental to the safe and effective design of dosage regimens in the developing infant. Additionally, the wide interindividual variability in clearance pathway maturation, the limited understanding of pharmacodynamics in the infant and the potential for adverse drug effects on general growth and development must engender prudent and diligent therapeutic outcome analyses as requisite components of the pharmacotherapeutic management of paediatric patients. A continued and more intensive effort to gain a clearer understanding of the pharmacokinetic and clinical consequences of the ontogenesis of systemic clearance mechanisms will help to establish more effective guidelines and specific recommendations for the use of drugs in infants.
Jane Alcorn was supported by the University of Kentucky Research Challenge Trust Fund. This work was supported in part by the National Institutes of Health grant HD37463.