The advance of perinatal medicine has improved the survival of extremely premature babies, thereby creating a new and heterogeneous patient group with limited information on appropriate treatment regimens. The developing fetus and neonate have traditionally been ignored populations with regard to safety studies of drugs, making medication during pregnancy and in newborns a significant safety concern. Recent initiatives of the Food and Drug Administration and European Medicines Agency have been passed with the objective of expanding the safe pharmacological treatment options in these patients. There is a consensus that neonates should be included in clinical trials. Prior to these trials, drug leads are tested in toxicity and pharmacology studies, as governed by several guidelines summarized in the multidisciplinary International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use M3 (R2). Pharmacology studies must be performed in the major organ systems: cardiovascular, respiratory, and central nervous system. The chicken embryo and fetus have features that make the chicken a convenient animal model for nonclinical safety studies in which effects on all of these organ systems can be tested. The developing chicken is inexpensive, accessible, and nutritionally self-sufficient with a short incubation time and is ideal for drug-screening purposes. Other high-throughput models have been implemented. However, many of these have limitations, including difficulty in mimicking natural tissue architecture and function (human stem cells) and obvious differences from mammals regarding the respiratory organ system and certain aspects of central nervous system development (Caenorhabditis elegans, zebrafish).This minireview outlines the potential and limitations of the developing chicken as an additional model for the early exploratory phase of development of new pharmaceuticals.
Prior to the onset of clinical trials, new drug candidates are vigorously assessed in nonclinical studies comprising mandatory toxicological, pharmacokinetic, and pharmacological testing. The multidisciplinary International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) guideline M3 (R2), “Guidance on Nonclinical Safety Studies for the Conduct of Human Clinical Trials and Marketing Authorization for Pharmaceuticals,” (ICH, 2009) gives an overview over the toxicological and pharmacological studies that are required for a drug lead to proceed into clinical trials and refers to specific safety guidelines for further details. Estimates show that only about 10% of agents entering clinical trials ultimately receive marketing approval. Lack of safety contributes to 30% of failures, whereas lack of efficacy is the other major cause of attrition (Kola and Landis, 2004). The fact that several drug candidates progressed well into clinical trials before rare and potentially lethal adverse effects were detected led to an extension of nonclinical trials and the foundation of safety pharmacology as a discipline in the late 1990s (Pugsley et al., 2008). One of the ICH M3 (R2) safety guidelines referenced is S7A, “Safety Pharmacology Studies for Human Pharmaceuticals,” (ICH, 2009) which defines a core battery of vital organs for which the impact of new drugs needs to be assessed. The core battery comprises the central nervous system (CNS), cardiovascular system, and respiratory system. Traditionally, these studies have been performed on selected drug candidates late in the nonclinical phase and continued during first-in-human trials. In an attempt to reduce costs and accelerate development, there has been a shift toward performing safety pharmacology studies earlier in the discovery phase (Hamdam et al., 2013). Safety pharmacology endpoints may also be incorporated into other nonclinical toxicological and kinetic studies. For early screening purposes in the exploratory phase of drug development, it is thus beneficial to have efficient and inexpensive model systems suitable for high throughput in which functional tests, biochemical markers, and morphology can be assessed. This will increase the likeliness of a drug succeeding in later phases of nonclinical studies. A number of early high-throughput models have been proposed to improve productivity and efficiency in the early nonclinical phase, including Caenorhabditis elegans, zebrafish (Danio rerio), and human stem cells. However, animal models phylogenetically closer to mammals with well characterized CNS, cardiovascular, and respiratory systems are needed to better predict alteration in organ function relevant for humans. Established animal models in safety pharmacology may include pigs, monkeys, primates, and dogs, models that require substantial care and tend to stir public opinion (Pugsley et al., 2008). It would therefore be advantageous to develop additional models that, in the early exploratory phase, could help reduce the number of drug leads that later need to be tested in the established animal models. Ideally, model systems should be adjustable with respect to parameters such as organ function and age (Jacobs, 2006). The developing chicken has several advantageous features for drug discovery. Here, we summarize areas within pharmaceutical research where the use of chicken embryos is already established, and we suggest an extended use of the developing chicken as a model system for the CNS and respiratory system, two of the major organ systems for which safety needs to be assessed. The chicken embryo as a cardiovascular model system was recently discussed by Kain et al. (2014).
Exposure of Pharmaceuticals to the Developing Embryo, Fetus, and Neonate and the Use of Juvenile Animal Models
The developing embryo, fetus, and neonate have previously represented ignored populations with respect to drug safety testing. However, the U.S. Food and Drug Administration and European Medicines Agency have recognized the need for increased knowledge of medicines in the pediatric population. The Best Pharmaceuticals for Children Act, effective in the United States since 2007, and the Pediatric Regulation in the European Union (EU) of 2006 (EC 1901/2006 and 1902/2006; EU 2006a, 2006b) were both passed with the objective to facilitate development of new drugs with pediatric indications and to improve the availability of information on the use of medicines for children. A principal goal is to include children in clinical trials. These trials should follow ICH guideline E11, “Clinical Investigation of Medicinal Products in the Pediatric Populations.” (ICH, 2000b). According to ICH M3 (R2) “Nonclinical Safety Studies for the Conduct of Human Clinical Trials and Marketing Authorization for Pharmaceuticals,” (ICH, 2009) section 12, the core battery of safety pharmacology tests used in adult animals and safety data from adult human exposure are usually available at the onset of pediatric clinical trials. The guidelines recognize the potential differences in safety profiles of drugs in adult versus those in young patients, and the use of juvenile animal models is discussed. ICH E11 states, “The need for juvenile animal studies should be considered on a case-by-case basis and be based on developmental toxicology concerns.” Apart from animal welfare and ethics concerns, the reason behind the lack of a strict requirement for studies in juvenile animals might be that results are difficult to extrapolate and, in some cases, have been found to be of little relevance to humans (Anderson et al., 2009). However, the interest in juvenile animal models in drug development is increasing (Baldrick, 2010). ICH is currently working (as of June 2015) on a new guideline S11, “Nonclinical Safety Testing in Support of Development of Pediatric Medicines,” to provide guidance, for example, on the design of studies in juvenile animals. This topic was endorsed by the ICH Steering Committee in November 2014, and a completed guideline is anticipated in 2018 (ICH, 2014). This highlights the importance of studies related to pediatric pharmacology.
The progression of neonatal medicine has made it possible to save infants born prior to gestational week 23 despite extreme prematurity (Blencowe et al., 2012). This has created an entirely new patient population for which increased knowledge of pharmacology is strongly needed. Neonatal therapeutic issues are often related to the cardiovascular system, the respiratory system, and the CNS, and there is an urgent demand for pharmacotherapy for conditions such as epilepsy and bronchopulmonary dysplasia in the pediatric population. This highlights the need for established prenatal and neonatal animal models in nonclinical studies of drugs that may potentially be administered to humans (Giacoia et al., 2012).
The Established Chicken Embryo Model
The Revival of a Well Established Model.
The need for juvenile and perinatal animal models prompts us to propose the in ovo chicken embryo model as a valuable addition to early safety studies. The chicken embryo and fetus constitute a model system with a long and fascinating history. Since Aristotle first started dissecting chicken embryos around the year 350 BCE, this model system has contributed substantially to a foundation for our understanding of human development and has also contributed important insight into immunology, cancer, and other areas of human disease (Stern, 2005; Datar and Bhonde, 2011; Kain et al., 2014).
Several features of the chicken embryo and fetus make it a desirable animal model: it is phylogenetically closer to mammals than, for example, zebrafish and the nematode worm, it has a convenient size, a short incubation time of 21 days, and is readily accessible throughout in ovo or ex ovo incubation. Development can be followed easily (Dohle et al., 2009), which is convenient for imaging and microsurgical procedures (Yalcin et al., 2010). Furthermore, the egg is nutritionally self-sufficient, and the embryo and fetus develop normally at 37–39°C and 45–55% air humidity without the need for specialized facilities and equipment.
Animal Welfare and Ethics with Regards to the Chicken Embryo Model.
Nonclinical testing should be carried out without unnecessary use of animals and other resources. Unanesthetized animals are preferred, and the number of animals in each experimental group should not exceed what is scientifically justified (ICH S7A; ICH, 2000a). All use of laboratory animals is subject to strict regulations with certain national variations. Guidelines regarding animal ethics and best practice aim to ensure minimal pain, suffering, and distress for the animal.
There is a consensus among scientists that the avian embryo gains the ability to experience pain at a certain point of development in ovo. The first sensory afferent nerves in chicken develop from embryonic day 4 (E4); however, synaptic connection with the dorsal horn is not established until E7, making perception of pain during the first trimester impossible (Eide and Glover, 1995, 1997; Rosenbruch, 1997). The development of tractus spinothalamicus is not fully characterized in the chicken embryo, so the exact time point at which the chicken embryo could begin experiencing pain is not known. At E13, the chicken fetus has developed a functional brain, and a few days prior to hatching, it is considered fully conscious (Animal Care and Use Committee California State Polytechnic University, 2012; Institutional Animal Care and Use Committee University of Louisville, 2012). Despite the uncertain onset of nociception and pain perception in the developing chicken, the American Veterinary Medical Association guidelines for euthanasia state that avian embryos that have attained >1/2 of the total incubation period have a sufficiently developed CNS for pain perception and should be euthanized by similar methods as used in neonates—for example, by anesthetic overdose, decapitation, or prolonged exposure to CO2. Eggs at <1/2 of the total incubation period may be destroyed by cooling (<4°C for 4 hours), freezing, or prolonged exposure to CO2. The EU legislation on the protection of animals used for scientific purposes (2010/63/EU; EU, 2010) defines embryos or fetuses developed to >2/3 of their total gestational period as neonates. Also included and protected by the requirements of this legislation are embryonic or fetal forms at a developmental stage less than 2/3 of their gestational length if they are subjected to experimentation during these two trimesters and allowed to live into the third trimester. This includes nonmammalian vertebrates, such as birds.
In accordance with the 3R principles of animal ethics (reduce, refine, replace), appropriate anesthesia of animals subjected to potentially painful procedures should be provided. The knowledge on nociception and distress during chicken embryonic development is relatively new and has led to new protocols for humane anesthesia of chicken embryos. These protocols have helped to refine the model to a higher ethical standard (Aleksandrowicz and Herr, 2015).
In mice and rats, litter size varies across breedings and cannot be accurately predicted in advance, whereas in general, a hen lays only one egg per day. To obtain a sample size of animals large enough for sufficient statistical power, additional matings are often required in mice and rats, causing distress for the mothers and potentially resulting in an unnecessarily large number of animals being handled, manipulated, and destroyed. In chicken hatcheries, however, the daily production of fertilized eggs is more predictable, making it easier to assign the appropriate number to experimental groups. Moreover, a broader genetic variation can be obtained, mimicking the clinical setting more closely. Additionally, the mother is excluded from experimentation. Thus, the use of chicken embryos instead of mammalian embryos obviously reduces the size of the experimental group, in line with international ethical guidelines. A further reduction is possible by monitoring effects on several major organ systems in each individual simultaneously. An important feature of the chicken embryo in this regard is that the development and maturation of the CNS and respiratory system overlap to a large extent temporally, as will be discussed in the following sections. It is thus possible to monitor several organ systems at the same time in one embryo—in line with the goal of reduction. In this respect, the chicken embryo model could serve as a replacement for other perinatal and juvenile animal models. However, the experimental use of embryonic stages is no longer considered an acceptable replacement for adult animals. Partial replacement by using, for example, tissue and cell culture, perfused organs, and tissue slices should be preferred instead of replacement with other live laboratory animals. All of these techniques are applicable to and described in the chicken.
Administration of Pharmaceuticals to the Developing Chicken.
The physical isolation of the chicken embryo and fetus from the mother is an advantage in studies of mechanisms of pharmaceuticals. Drug administration is easily controlled, and both primary and secondary pharmacology, off-target effects, and toxicity can be tested. However, the chicken model does have limitations in this respect. As a rule, the route of drug administration in an animal model should appropriately mimic the intended therapeutic route of administration in humans. Historically, however, studies of acute toxicity have been performed by parenteral administration in addition to the intended clinical administration, as discussed in ICH M3(R2; ICH, 2009). In the developing chicken, various drug delivery systems and formulations have been evaluated, and topical administration onto the chorioallantoic membrane as well as systemic administration intravenously or via injections into the yolk sac, amnion, or peritoneum have been tested (Vargas et al., 2007). Drug inhalation in the developing chicken can only be done in the postpipping stage after E19, whereas oral drug administration can be performed after hatching. As in all nonmammalian animal models, challenges arise when attempting to translate results to a similar setting in mammals. For example, drug metabolism in the mother and the placental barrier in mammals may, depending on properties of the pharmaceutical, protect the fetus against exposure, a feature which cannot be mimicked correctly in avian models. Further, there is no excretion from the egg until hatching, a feature that may cause a prolonged effect of a single exposure. In studies of pharmacokinetics, the chorioallantoic membrane has raised interest in testing drug delivery systems and has been suggested as a valuable tool in nonclinical studies, thoroughly reviewed by Vargas et al. (2007). One has to be aware of the possible differences in pharmacokinetics in chickens compared with humans (Nowak-Sliwinska et al., 2014).
Established Use of the Chicken Model in Nonclinical Safety Studies.
Avian embryos are established models in teratology and developmental toxicology. Such issues are addressed by the ICH guideline S5 (R2) “Detection of Toxicity to Reproduction for Medicinal Products and Toxicity to Male Fertility.” (ICH, 2000c) Several apoptosis assays (activated caspase-3, terminal deoxynucleotidyl transferase–mediated digoxigenin-deoxyuridine nick-end labeling assay) in addition to echocardiography and morphologic changes in organs have been used to test effects of toxicants (Smith et al., 2012). The chicken embryo has also been used to study the effects of retinoids that alter patterns of gene expression, neuronal differentiation, and axonal projections. In situ hybridization, RNA sequencing, immunostaining, and retrograde labeling techniques have been used to assess such effects (Forehand et al., 1998).
A challenge in nonclinical studies is poor translation of results between species. Retinoids are known teratogens that also affect the nervous system in the human embryo (Adams, 2010). Here, there is good concordance between the chicken and other animal models. In general, proper cross-species translation of results from animal models requires knowledge of comparative biology and careful selection of relevant and translatable endpoints before concluding. Observation of similar results in animal models of different species increases the likeliness of the results being translatable to humans.
As another example from the area of developmental toxicology, the effect of mammalian cardiac inward rectifying potassium channel blockers has been shown to produce digit malformations in developing rats. Knowledge of the comparative biology of expression of the human Ether-à-go-go-Related Gene (hERG) channel was essential for translation of these results into a human drug-development context. This channel is important in the chicken embryo as well (Krogh-Madsen et al., 2005; Danielsson et al., 2007), which makes it possible to study primary or off-target effects of pharmaceuticals on this channel in the developing chicken.
One area of drug development where the chicken model has been more firmly implemented is in ocular toxicology testing. The chicken embryo is an established model for the study of retinal development (Vergara and Canto-Soler, 2012). The chicken has large eyes during development and at birth, and retinal cells can be cultured in vitro (Daston et al., 1991). The isolated chicken eye is an accepted method for toxicology testing in both the U.S. and the EU (Dholakiya and Barile, 2013). Toxicity is assessed by a qualitative examination of opacity of the cornea, swelling, and damage measured by fluorescein staining (OECD, 2009).
Expanding the Use of the Chicken Embryo Model
The Developing Chicken as a CNS Model.
Conditions affecting the nervous system are common in neonates; one typical sequela of dysregulated brain development is the occurrence of seizures, which requires treatment with antiepileptic drugs (Giacoia et al., 2012). During CNS development, cell division, cell death, and cell differentiation are all tightly regulated (Marzban et al., 2014). To avoid administering neuroactive drugs that may alter CNS development, safety studies are recommended to investigate key functional domains of the developing CNS (Himmel, 2008). These studies should be performed in the early nonclinical phase of drug development.
As a major model in developmental biology, the chicken embryo has also played a pivotal role in the study of neurodevelopment. Famously, the neural crest was first discovered in the chicken embryo by Wilhelm His in 1868. Rita Levi-Montalcini shared the Nobel Prize in Medicine or Physiology in 1986 for her discovery of nerve growth factor, and her work was primarily performed in the chicken embryo. Her mentor, Viktor Hamburger, together with Howard Hamilton described the 46 stages of chick development, a staging system that is still used today (Fig. 1), and performed seminal studies of the development of motility and behavior as well as trophic interactions (Hamburger and Hamilton, 1951). Several aspects of the stages of neuronal development in chickens are characterized. As an example, the efferent motor neurons develop early in chickens. This gives the embryo the ability to move spontaneously from E3.5, when early movements start in the neck before spreading to the trunk. At E6.5, motility in the legs appears (Hamburger and Balaban, 1963; Bekoff, 1976). Nicole LeDouarin invented and developed the technique of quail-chick chimeras to launch a wide-reaching series of fate-mapping studies that have helped elucidate the fates and developmental roles of neural crest cells (Dupin et al., 1998; Le Douarin, 2004; Le Douarin et al., 2008; Le Douarin and Dieterlen-Lievre, 2013). Viral approaches for genetic fate mapping were introduced in the late 1980s (Price et al., 1987; Gray et al., 1988; Sanes, 1989). More recently, the chicken embryo has been used to study neural induction and early patterning (Le Douarin, 2004; Teillet et al., 2008). Fluorescent and other tracers have been developed to map the structural and functional development of axon projections and neural circuits (Glover et al., 1986; O'Donovan et al., 1994). The tradition of utilizing the chicken embryo in neurodevelopmental studies has continued to progress to embrace molecular genetic approaches, including in ovo electroporation for gain- and loss-of-function studies (Nakamura et al., 2004) and a genetic toolbox of DNA constructs and viral vectors for targeted tracing and functional manipulation of specific neuronal subtypes (Hadas et al., 2014). The chicken embryo and fetus are accessible for in ovo kinematic, electrophysiological, and noninvasive imaging studies of neurophysiological and behavioral development (Bekoff, 1992; Ryu and Bradley, 2009; Balaban et al., 2012). Last, the chicken embryo has been used as a recipient for xenotypic grafting studies of human neural stem and progenitor cells (Glover et al., 2009; Boulland et al., 2010).
Given its proud history in the field of neurodevelopment, it is surprising that, when it comes to nonclinical neuropharmacological studies, the chicken model is not yet routinely used.
An area where the advantages of the chicken embryo model have been highlighted is developmental endocrinology (De Groef et al., 2008). It is of particular benefit that the only endocrine influence the embryo receives from its mother is through the egg yolk. This facilitates external manipulation of the endocrine system and the study of whether pharmaceuticals and environmental chemicals cause endocrine disruption (Mathisen et al., 2013). Tight regulation of the endocrine system is crucial for correct development of the CNS. The chicken embryo can be a valuable tool in early screening of drug candidates to determine whether they alter endocrine profiles.
Epilepsy is one of the most serious conditions affecting human neonates. A homozygous strain of epileptic chickens, the Fayoumi strain, has been used in epilepsy research for over 30 years (Nunoya et al., 1983). Epileptic chickens show differences in electroencephalography compared with normal chickens from E17 (Guy et al., 1995). They have been used for magnetic resonance imaging (Fabene and Sbarbati, 2004) and evaluation of anticonvulsant drugs (Loscher, 1984; Johnson et al., 1985) and N-methyl-D-aspartate receptor (NMDA) antagonists (Pedder et al., 1990). Studies have been carried out to examine how generalized seizures alter the development of the chicken brain (Gong et al., 2003).
For studies of neural development in chicken, the cerebellar cortex is an important model system (Nieuwenhuys, 1967; Solecki et al., 2006). The structure of the cerebellar cortex was first described in 1911 by Ramon y Cajal, who found it to be so conserved among species that he proposed it to be a “law of biology” (Sultan and Glickstein, 2007). The cerebellum is considered an appropriate model that reflects well the general aspects of CNS development. Cerebellar granule neurons are numerous, constitute the most homogenous neuron population in the brain, are easily accessible, and may be purified and grown in vitro. They undergo all key stages of neural development and thus provide a model of choice for studying mechanisms of neuronal proliferation, apoptosis and survival, differentiation, and migration (Contestabile, 2002). A key feature of cerebellar development is the migration of granule neurons through layers of the cerebellar cortex—from the external granule layer to the internal granule layer. To a large extent, this occurs postnatally in some mammals, including humans. Alterations in migration may be indicative of impaired neurodevelopment and could be monitored by histologic methods in safety studies to assess whether a drug candidate is likely to be safe for children or during pregnancy. Over time, cerebellar granule neurons in vitro differentiate and acquire the characteristics of mature neurons, express NMDA receptors, and rapidly develop excitotoxicity in culture (Contestabile, 2002). Exposure of these cultured neurons to a drug candidate can determine at an early stage of preclinical testing whether the drug is toxic to neurons and, for example, initiates apoptosis (Jacobs et al., 2006). An important aim in drug development is to discover detrimental effects as early as possible (Pugsley et al., 2008). This is of particular importance when testing new antiepileptic drugs that interfere with glutamate (as NMDA antagonists) or GABA signaling, or to discover off-target effects of other potential drugs.
As pointed out by Ramon y Cajal, the structure of the cerebellar cortex is highly conserved among species, which makes histologic examination of alterations in the cortex relevant to a similar setting in humans. However, there are some differences. The vermis (a medial expansion of the cerebellum) is only seen in mammalian species (Butts et al., 2014), and the migration of granule neurons exhibits different patterns in different species. In the human embryo, the external granule layer reaches its peak thickness by gestational week 25, and the process of migration continues well into the first postnatal year. In the chicken cerebellum, the migration of cerebellar granule neurons occurs prenatally (Aden et al., 2008, 2011; Volpe, 2009; Mathisen et al., 2013), in contrast to the situation in the rodent cerebellum where the external granule layer is formed postnatally, and granule neurons migrate between postnatal days 4 and 15 (Fujita, 1967). In fish, the migration of granule neurons follows yet a different pattern from that of mammals and chicken or is absent altogether (Butts et al., 2014). That the migration of cerebellar granule neurons in chicken resembles the human counterpart more closely than what is seen, for example, in zebrafish makes the chicken embryo cerebellum a more appropriate model than the fish cerebellum for high-throughput screening in safety studies of the CNS. That granule neuron migration in the chicken occurs at a fetal stage is another advantage of the chicken embryo model that facilitates the study of insults to cerebellar development. This has been used to assess the neurodevelopmental effects of glucocorticoids (Aden et al., 2008, 2011), a clinically relevant neonatal issue (Roberts and Dalziel, 2006; Brownfoot et al., 2013). The same model has been used to address developmental disruption caused by bisphenol A, with results in concordance with those reported in a mouse model (Mathisen et al., 2013). Histopathology remains the gold standard (Sasseville et al., 2014) for assessing detrimental effects on the developing CNS in chickens (Table 1). In addition, there are several biomarkers that could be useful for early screening purposes, although few are firmly established. The transcription factor PAX6 is important for cerebellar development and migration (Engelkamp et al., 1999), and perturbations of PAX6 expression have been linked to autism (Umeda et al., 2010). Also, the role of PAX6 has been studied and found to be important for brain development in the chicken embryo (Mathisen et al., 2013). Other markers that have been suggested and partly implemented in studies of neurotoxicity are Calbindin D28k and microtubule-associated protein-2 (Haworth et al., 2006). Calbindin D28k is a calcium-binding protein highly expressed in the cerebellum across various species, has been studied in chickens (Hunziker and Schrickel, 1988; Koszka et al., 1991; Marzban et al., 2010; Flace et al., 2014), and is associated with autism (Whitney et al., 2008) and hereditary ataxia (Koeppen et al., 2013) in humans. Microtubule-associated protein-2 has roles in neuronal growth and plasticity (Johnson and Jope, 1992), has been studied in chickens (Pinkas et al., 2015), and has been shown to be important in human traumatic brain injury (Mondello et al., 2012) and Rett syndrome (Johnston et al., 2001). Therefore, these are examples of markers that could serve as endpoints in safety studies with results relevant for conditions seen in humans. Other relevant neuronal markers are listed in Table 1. We thus argue that the chicken cerebellum is a valuable tool for the study of neurodevelopment that could successfully be implemented in early nonclinical studies of new pharmaceuticals.
The Developing Chicken as a Lung Model.
In the 17th century, the lung was suggested to be a link between vital air particles in the atmosphere (later determined to be oxygen) and the blood (Partington, 1956). Since then, it has been acknowledged as a prerequisite for life, and the respiratory system is a natural component of the core battery of safety pharmacology studies (Murphy, 2002; Dhokarh et al., 2012). In nonclinical testing of drugs on laboratory animals, clinical observation alone to assess respiratory function is not adequate. Thus, several parameters of respiratory function, including monitoring both the pumping apparatus and the gas-exchange unit, should also be quantified using appropriate methods, as discussed by Murphy (2002, 2014). Important conditions in neonatology that affect the respiratory system include neonatal respiratory distress syndrome, apnoe of prematurity, and persistent bronchopulmonal dysplasia, and call for immediate treatment. Fetal hypoxia is an important concern during pregnancy and is often present in combination with premature birth and other conditions, such as chorioamnionitis, preeclampsia, and maternal diabetes. Premature birth occurs in >10% of all live births worldwide (Blencowe et al., 2012), and drugs that have minimal adverse effects in healthy and normally developed individuals can be life-threatening in individuals with a diseased or immature respiratory system. Both the pumping apparatus and the gas-exchange unit are essential for respiration. Mild respiratory depressant drugs may produce serious adverse effects if given to a patient with a ventilatory defect such as apnea, often seen in low-birth-weight neonates with an immature CNS. Drugs that mildly hamper lung development and maturation may be mortally dangerous to premature individuals with marginally developed lungs. Adverse effects on the respiratory system in newborn babies after drug administration to the mother during pregnancy may also occur. Selective serotonin reuptake inhibitors are associated with an increased risk of pulmonary hypertension in the newborns of women receiving this therapy during the last trimester of pregnancy (Moses-Kolko et al., 2005; Chambers et al., 2006). Administering drugs to newborn children and pregnant women is thus a major safety concern. There is a need for animal models to study whether pharmaceuticals affect the respiratory system in fetal and neonatal life. As conditions affecting the respiratory system can be severe, there is also a clinical need for new treatment options. Models in which lung development is well characterized can provide an important research tool to identify targets for future therapy.
The respiratory anatomy and physiology of birds differs from that of mammals, and the morphology of the developing chicken lung is well described (Maina, 2003a,b, 2004b). The parabronchial bird lungs are connected to air sacs that serve as reservoirs for oxygen-rich air and are used for respiration during exhaling. This results in a very efficient respiration that is highly useful, for example, during long flights. In chickens, the respiratory movements start around E18 of development. The human fetus, by contrast, already makes breathing-like movements from week 9, and the circulation of amniotic fluid into the airways is essential for sufficient lung development. Unlike mammals at birth, birds adapt to air breathing gradually, with an overlap of chorioallantoic membrane and lung-dependent gas exchange lasting from hours to even days during hatching. Partly because of these major anatomic and physiologic differences, the chicken has traditionally not been implemented as a model system in safety studies of respiratory function. However, the commercial broiler industry, which uses in ovo vaccination and drug treatments as strategies to defeat disease among flocks, has used monitoring of the respiratory system of the developing chicken in safety pharmacology testing of new vaccines. For example, effects on distal lung maturation with insufficient development of air capillaries have been found in an in ovo trial testing a vaccine against Newcastle disease virus (Dilaveris et al., 2007). Safety studies of respiratory function in adult and juvenile animals are usually monitored in head-out or whole-body plethysmographs. Whereas measuring respiratory function in a mammalian fetus requires invasive or indirect electrophysiology or imaging techniques (Greer, 2012), the same monitoring in the chicken fetus can be performed using noninvasive barometric methods, a real advantage of the chicken embryo model system (Szdzuy and Mortola, 2007).
Although not yet widely implemented in nonclinical pharmaceutical studies, the respiratory system of the chicken embryo has been used as a model for research purposes. Antenatal corticosteroids are routinely given to pregnant women when there is a threat of premature birth to accelerate lung maturation and prevent respiratory distress. However, after several decades, there is still no international consensus on an optimal regimen of corticosteroid treatment. This is still a highly relevant clinical problem, and experimental embryo animal model systems in which the effects of different corticosteroids on lung development can be monitored are crucial. Hylka and Doneen (1983) used the chicken embryo model to study the actions of steroid hormones on the fetal lung. This study used growth, cellular proliferation, and rate of synthesis of the major surfactant phospholipids as parameters of lung development and function. Further, effects on lung development of in ovo hypophysectomy and treatment with corticosterone or pituitary transplantation were assessed (Hylka and Doneen, 1983). Antenatal corticosteroid treatment is also associated with impaired cerebellar development in humans, a finding that has been reproduced in the chicken embryo (Aden et al., 2008, 2011). Thus, the chicken embryo model can be used advantageously to study the impact on two organ systems (Fig. 1) of pharmaceuticals clinically administered for a major human neonatal condition.
Fetal hypoxia is another important concern in neonatal medicine, and a hypoxia model has been established in chickens. In a study from 2010, fertilized eggs were incubated at various levels of oxygenation to support the concept that fetal oxygen tension is a key signal in the regulation of the surfactant system. Disaturated phospholipids and vascular endothelial growth factor were used as parameters of lung function at different stages of chicken embryo development, and a critical window of vulnerability to oxygen tension was established (Been et al., 2010). Another relevant example is surfactant protein B, which is expressed in chicken lungs from E17 (Bjornstad et al., 2014). Qualitative and quantitative assessment of relevant biomarkers related to morphology and protein expression are thus established methods to study the chicken embryo lung, and have shown that, despite the phylogenetic distance, central aspects of lung development are similar at cellular and protein levels in humans and chickens. Hence, as cross-species comparative knowledge on relevant biomarker panels (Table 1) continues to grow, there is an expanding potential for translation from the developing chicken lung to the human context, despite the anatomic and physiologic differences.
Utilization of the chicken model has evolved over time in parallel with technological innovations. Sophisticated techniques combining structural and functional tracing witness this advance (Table 2). Positron emission tomography imaging (Balaban et al., 2012), for example, enables (among other things) far more effective monitoring of the uptake and elimination of radiolabeled drugs than before, without sacrificing the animal. Molecular targeting of cell populations combined with viral vector–mediated labeling and imaging techniques has created new possibilities for reliable and high-throughput structural and functional characterization of cells and cell populations, as recently shown in studies of the developing spinal cord of the chicken (Hadas et al., 2014). As genes can be delivered into the chicken embryo using in ovo electroporation or retroviral vectors (Luo and Redies, 2004; Sato, 2004), and gene expression can be knocked down using small interfering RNA (Sauka-Spengler and Barembaum, 2008), gain and loss of function may additionally be used to induce disease and alter receptor expression and enzymatic activity. For example, retroviral vectors have been used to generate mis-/overexpression of Wnt5a in the chicken embryo, demonstrating its role in coordinating lung development (Loscertales et al., 2008). Several protocols have been described for gene silencing by RNA interference in the nervous system of the developing chicken, both at very early stages of brain development (from E3 to E3.5) and at later stages in specific brain regions, such as the cerebellum (Andermatt and Stoeckli, 2014). For example, an important role for axonin-1/TAG-1 in the pathfinding of granule cell axons in the developing cerebellum was established using this approach (Baeriswyl and Stoeckli, 2008).
These new technologies will likely lead to the discovery of novel biomarker panels central to lung, cardiovascular, and CNS development. The chicken model will thus be a more comprehensive animal model with highly accurate gene targeting analogous to that provided by transgenic rodents. This expanded toolbox, now available for the chicken embryo model, increases its usefulness with regard to testing pharmaceuticals, as it is possible to link adverse effects to individual variability in, for example, receptor or enzyme profiles. Increased use of the chicken model will contribute to the documentation necessary for approval as a safety pharmacology model by the regulatory authorities in the future. With this in mind, the time has come for the chicken embryo model to leave the realms of academia and become an industrial worker.
Many features of the chicken embryo, including low cost, self-contained development, short incubation time, conveniently large size, accessibility, and availability of a large number of techniques for experimental manipulation and characterization, make it a desirable laboratory animal model. The chicken embryo has traditionally been used to address mechanisms of development and neonatal function, but has an expanding potential as an efficient and inexpensive screening model for drug development. The development of the CNS and the cardiovascular and respiratory systems is well characterized, as are the main differences from human development. We therefore propose that the chicken embryo should be included as a mainstay model for early nonclinical safety assessment.
The authors thank Anne Soleng at the Norwegian Medicines Agency for valuable discussion.
Wrote or contributed to the writing of the manuscript: Bjørnstad, Austdal, Roald, Glover, Paulsen.
- central nervous system
- European Union
- International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use
- Copyright © 2015 by The American Society for Pharmacology and Experimental Therapeutics