The chicken has been an important research tool in vertebrate developmental biology because the accessibility of the chick embryo allows embryonic manipulations that are not possible in a mammalian system. Fertilized chicken eggs are easily maintained in humidified chambers, and the chicken embryo floats on the nutritious egg yolk early in development, making it relatively easy to manipulate. The accessibility and ease of chicken embryonic manipulation has led to many experimental insights about developmental processes. The discovery that quail cells exhibit a different nuclear appearance than chicken cells has been exploited by transplanting quail cells into chicken embryos and following the implanted cell fate during development (Le Douarin,1973a,b). The quail:chick chimera model has also led to many discoveries about neural development (Kinutani and Le Dourain,1985; Kinutani et al.,1986). Although transgenic mice have been available to the research community for more than 20 years, transgenic chicken technology has lagged behind their mammalian counterparts. Because the poultry industry in the United States is a multi-billion dollar industry, there may eventually be commercial interest in using transgenic technology to improve the production of meat and eggs. The most likely use of transgenic technology in poultry production is to impart disease resistance, which has become a common practice in transgenic plants. In addition, the domestic hen is an ideal candidate for the efficient production of therapeutic proteins for the pharmaceutical industry because of the low cost of feeding a hen, the naturally sterile environment of the egg, the large amount of protein that may be produced per egg, and the large number of eggs that are produced per hen per year (Ivarie,2003). However, few researchers have attempted to create a transgenic chicken as a model for the developmental biology research community. Recently, lines of transgenic chickens that express bacterial beta-galactosidase have been created at North Carolina State University by using the SNTZ retroviral vector originally developed by Takashi Mikawa (Mikawa et al.,1991,1992). Generating transgenic chickens is a very complex and laborious endeavor, but there may be future opportunities to use a Cre/Lox system or green fluorescence protein to create new transgenic avian models for the research community. The objective of this manuscript is to provide a concise overview about the methods, the attempts, and the current status of technologies to generate transgenic birds.
EARLY DEVELOPMENT AND FERTILIZATION
Construction of an avian egg begins with the ovulation of a mature ovum from the ovary. Sperm are stored in the sperm host gland of the female, and fertilization happens when the ovum enters the infundibulum of the oviduct. The fertilized egg enters the magnum, which secretes albumen. The ovum next enters the isthmus where the outer and inner shell membranes are deposited to prepare for egg shell formation in the shell gland. It takes approximately 3.5 to 4 hr for the ovum to move from the infundibulum into the shell gland, where the egg remains for the ∼20 hr it takes to form an egg shell. The first cleavage division occurs upon entry of the ovum into the shell gland. Over the 20 hr it takes to form an egg shell, cell divisions continue, and at lay, the embryo comprises 50,000–60,000 cells on the surface of the yolk mass (Spratt and Haas,1960). A technical difficulty of using transgenic technology near the time of fertilization is obtaining/manipulating the embryos because there are many complex physiological phenomena happening that are meant to nourish and protect the embryo from the outside environment. Furthermore, every time one-cell eggs are harvested, a mature hen must be killed, making it difficult to obtain a sufficient number of eggs to successfully complete any procedures aimed at generating transgenic birds.
GERM CELL OVERVIEW
All methods of producing transgenic poultry rely on techniques designed to insert novel genetic material into cells that will give rise to germ cells. In general, most techniques will produce a rooster or hen with mosaic transgene insertion in the germline (i.e., only a subset of cells in the gonad will carry the transgene). Once a sperm or egg carrying the transgene unites with a wild-type cell, the resulting offspring will contain the transgene in all cells of the body. Therefore, efficient targeting of the germ line is essential in the production of transgenic poultry.
Development of the germ line begins with the formation of Primordial Germ Cells (PGCs) in the embryo (reviewed in Wentworth and Wentworth,2000; D'Costa et al.,2001). It has been suggested that the chicken germline is predetermined by maternally derived germ plasm, because the chicken homolog of the RNA binding protein vasa is detectable from the first cleavage division of fertilized eggs and in definitive PGCs (Tsunekawa et al.,2000), but the predetermination model of avian germ cell development remains to be demonstrated. Definitive PGCs are found in an extraembryonic region termed the germinal crescent. Subsequently, the migration of PGCs from the germinal crescent to the gonadal ridge occurs (1) passively through the intra- and extraembryonic circulation (Fujimoto et al.,1976a,b), and (2) actively when the blood-borne PGCs migrate into the germinal epithelium by means of the dorsal mesentery (Fig. 1).
The target for any modification of the avian genome is the germ cell because these cells ultimately deliver the transgene to the subsequent generations. Current targets for germ line modification include the mature oocyte/spermatozoa, the newly fertilized egg, and primordial germ cells during their establishment, migration, and colonization of the gonad.
A major difficulty that is encountered in using any methodology to create transgenic chickens by manipulating embryos from freshly laid eggs to introduce foreign genes into the germ line is achieving a sufficient hatchability to generate an acceptable number of G0 offspring to screen for the incorporation of the transgene. Historically, a window has been cut in the egg shell and sealed with an adhesive or shell membrane after embryonic manipulation. However, hatchability has been very low by using the classic and simplistic techniques (Kinutani and Le Dourain,1985; Petitte et al.,1990). Optimal hatchability of windowed unincubated eggs is important for the efficient production of transgenic poultry. Improvements in hatchability of manipulated freshly laid eggs have been achieved by grinding off the shell, covering the window with phosphate-buffered saline (PBS), cutting the shell membrane while it is under the PBS, injection of cells into the embryo, sealing the egg with shell membrane and Duco cement (Speksnijder and Ivarie,2000). The windowing procedure improved hatchability from approximately 6% to 30%. The windowing procedure also resulted in 23% to 35% hatchability when used with retroviral injections (Harvey et al.,2002). Improvements in hatchability have also been reported by storing the eggs with the pointed end down for 5 to 7 days before injecting cells through the blunt end of the egg (Bednarczyk et al.,2000). Another alternative to the windowing techniques is to use an ex ovo culture system of freshly laid eggs (Rowlett and Simkiss,1987; Perry,1988). Briefly, the contents of freshly laid eggs are placed in individual surrogate chicken egg shells, and the embryos are injected with a suitable vector. The surrogate chicken egg shell is covered with Saran-Wrap (Dow Chemical, Midland, MI) for 3 days. Subsequently, the chicken embryo is transferred to a surrogate turkey egg shell, which is covered with Handi-Wrap (Dow Chemical) until hatching (Petitte and Mozdziak,2002). The modifications of the original procedures of Perry (1988) and Rowlett and Simkiss (1987) by Borwornpinyo (2000) resulted in 20% to 35% hatchability when used in conjunction with retroviral injections. The hatchability was sufficient to generate G0 roosters carrying the lacZ gene in their semen (Mozdziak et al.,2003). Currently, the windowing and the ex ovo culture are the only ways to introduce foreign DNA into a chicken embryo, and have viable offspring.
Retroviral vectors have become a common gene transfer vehicle, and they can be used to introduce foreign genes into the avian germline. Retroviral gene transfer techniques have been used the most frequently to generate transgenic chickens (Bosselman et al.,1989a,b,1990; Harvey et al.,2002a,b; Harvey and Ivarie,2003; McGrew et al.,2003; Mozdziak et al.,2003), because high retroviral titers make it possible to transfer the gene of interest to many cells. Briefly, retroviruses have an RNA genome encased in a protein core containing integrase, reverse transcriptase, and protease. The viral envelope proteins bind to specific receptors on the host cells, and the viral particles are brought into the cellular cytoplasm by receptor-mediated endocytosis. The RNA is copied into cDNA, which is transported into the nucleus, and the cDNA is integrated into the host cell DNA through the action of an integrase on the viral long-terminal repeats. The integrated proviral DNA is replicated with the cell, and it is inherited in a Mendelian manner. The proviral DNA can be transcribed into viral RNA for the synthesis of proteins that include a polymerase (pol), group-associated antigens (gag), and envelope (env) proteins. These three classes of proteins assemble new viral cores and package the viral RNA genome. The new complex is transported to the host cell membrane, and it is transported outside of the cell (Fig. 2). In the case of replication-competent vectors, the viral structural genes and the packaging sequence are intact, allowing for continuous production of infectious particles that can infect other cells. Therefore, replication-competent retroviral vectors are likely easy to infect a sufficient number of germ cells to transfer the gene of interest to the second and subsequent generations.
However, transgenic poultry constructed using a replication-competent retrovirus will also continually shed virus into the environment, which is an unacceptable situation. However, replication-defective retroviral vectors that do not contain the pol, gag, or env genes necessary to generate new infectious particles alleviate any environmental issues. Replication-defective vectors are capable of infecting host cells, but the provirus will not generate new infectious particles. The development of retroviral packaging cell lines that are capable of providing the gag, pol, and env proteins to assemble the viral construct containing the gene of interest was a major advancement in generating replication-incompetent vectors. Replication-defective vectors allow for reasonably large exogenous genes (up to ∼10 kb), because they are missing the viral genes necessary for replication. New retroviral vectors are being developed based upon new strategies that include pseudotyping that allows for optimal infectivity (Mizuarai et al.,2001). A common problem with some retroviral vectors used in transgenic technology has been gene silencing. A new approach that does not appear to be affected by gene silencing has been to use lentiviral vectors, which are a type of retrovirus that can infect both dividing and quiescent cells (Pfeifer et al.,2002). The preintegration complex of lentiviral vectors can penetrate the intact membrane of the nucleus of the target cell, and lentiviral vectors have been used to develop transgenic mice (Lois et al.,2002). In addition, lentiviral vectors have been used to generate lines of transgenic chickens, but expression was tissue-specific, which may be related to the (CMV) promoter used in the vectors (McGrew et al.,2003).
RETROVIRAL HISTORICAL PERSPECTIVE
The first successful development of transgenic chickens was reported by Salter et al. (1986,1987) who used replication competent reticuloendotheliosis virus (REV) and avian leukosis virus. The retroviral vectors were injected near the blastoderm resulting in 25% of the G0 males being mosaic for the transgene, and they transmitted the provirus in the germline at a rate of 1 to 11%. Chen et al. (1990) reported that a modified form of the Rous sarcoma virus containing bovine growth hormone can enter the chicken germ line. In other studies, transgenic chickens have been produced by using a replication-defective REV vector that carries an inserted growth hormone gene (Bosselman et al.,1989a,b,1990). The investigators injected approximately 2,599 embryos to generate 33 males that carried that vector sequences in their semen. The G0 birds transmitted the vector to the G1 progeny at a rate between 2% and 8% (Bosselman et al.,1989a). Transgenic birds carrying the lacZ gene have been produced by infecting primordial germ cells with a replication defective spleen necrosis-derived retroviral vector encoding the lacZ gene and transplanting the transformed primordial germ cells into early recipient embryos (Vick et al.,1993). However, the previous investigators (Vick et al.,1993) never demonstrated beta-galactosidase expression in the offspring. Transgenic chickens carrying the lacZ gene were also produced using a replication-defective avian leukosis based retroviral vector (Thoraval et al.,1995). Unfortunately, beta-galactosidase expression was only noted in cultures of embryonic fibroblasts from G2 progeny, and expression was not reported in the entire embryo. Recently, Mizuarai et al. (2001) created transgenic quail by using a pseudotyped VSV-G (vesicular stomatitis virus-glycoprotein) vector, but the quail failed to adequately express the reporter gene. More recently, other groups have been successful at creating transgenic chickens using retroviral systems. First, Harvey et al. (2002a,b) demonstrated that they created a transgenic chicken that expressed beta-lactamase in the egg white and that when the line was bred to homozygosity; expression was greater in the homozygotes than in the heterozygotes (Harvey and Ivarie,2003). The work of Harvey et al. (2002a,b) demonstrated that a chicken can be engineered to express a foreign protein in the egg showing the feasibility of the hen as a bioreactor for the production of therapeutic proteins. Second, a new chicken model for developmental biology research has been currently engineered (Mozdziak et al.,2003) by using a retroviral vector system that has been previously shown to have high expression in the chicken (Mikawa et al.,1991,1992). These transgenic chickens are a new model system for developmental biology research (Mozdziak et al.,2003) that is an improvement over the quail:chick chimera, because the chickens carry the lacZ gene and express beta- galactosidase, making it possible to track cells into a wild-type chick embryo (Giamario and Mozdziak, unpublished observations).
The common method for introducing genes into mammals is to microinject DNA into the pronucleus of a newly fertilized egg (Gordon et al.,1980; Brinster et al.,1981). However, direct DNA injection of freshly laid chicken eggs is much more technically difficult compared with mammals because a fertile freshly laid chicken egg contains approximately 50,000–60,000 cells when it is laid (Spratt and Haas,1960). Therefore, DNA microinjection has not been extensively attempted on freshly laid eggs, but on fertile embryos collected after killing a hen (Love et al.,1994). Microinjected zygotes were brought to hatch by using a three-step ex ovo culture system (Love et al.,1994; Sherman et al.,1998). Briefly, newly fertilized eggs, which are surrounded with a capsule of albumen, are removed from the magnum and cultured for approximately 18 to 24 hr with synthetic oviductal fluid. Next, the eggs were transferred to a surrogate egg shell that is sealed with no air space. After 2 to 4 days, the embryos are placed into a larger shell with an upper air space for the remainder of the incubation. By using microinjection of a DNA construct carrying the lacZ gene driven by a truncated egg white lysozyme promoter into the germinal disk of newly fertilized eggs, Love et al. (1994) reported the production of a single mosaic transgenic rooster harboring the lacZ gene that transmitted the beta-galactosidase gene at a 3.4% rate to its offspring. Mendelian inheritance of the gene was reported between the G1 and the G2 generations, showing that it was possible to create a transgenic chicken using DNA microinjection. Unfortunately, there was no report of beta-galactosidase expression in any of the transgenic offspring. In another report, microinjection of a plasmid carrying an active Drosophila transposable element mariner resulted in transgenic chickens carrying the mariner gene, which was transferred from the G1 to the G2 progeny in a predicted Mendelian ratio (Sherman et al.,1998). Although microinjection can successfully be used to create transgenic poultry, it generally has resulted in relatively small numbers of transgenic birds.
Another technology to create transgenic chickens is to target blastodermal cells, PGCs, or embryonic stem cells. Subsequently, the manipulated cells are implanted into a developing embryo to create a germ-line chimera. Working with these cells provides the ability to perform the gene insertion in vitro. The cell-based method of gene insertion is attractive, because the integration into the genome and expression of the transgene can be examined before the cells are implanted into the embryo, and it could allow the production of transgenics with targeted changes to the genome. The first step of cell-based transgenic production is to obtain cultures of cells that are capable of entering the germ line (i.e., primordial germ cells), culture them, transfect them with DNA constructs, and implant the transfected cells into an embryo. Subsequently, the G0 offspring are mated to generate G1 chickens that carry the inserted transgene throughout their body. This technical scheme represents the coordinated efforts of three technologies: the ability to culture cells capable of entering the germ line, the development of suitable DNA constructs, and the ability to produce germline chimeras. Petitte et al. (1990) produced the first chicken germline chimera. Subsequently, the production of avian germline chimeras has become a common practice (Petitte et al.,1990; Kagami et al.,1997; Bednarczyk et al.,2000; Speksnijder and Ivarie,2000; Li et al.,2002), methods to culture embryonic stem cells and chicken germ cells have been developed (Pain et al.,1996; Acloque et al.,2001; Petitte and Mozdziak,2002), and construction of DNA vectors has also become a common practice. Therefore, all the tools are available to use in vitro germ line modifications to create transgenic birds, but transgenic birds have not yet been accomplished through in vitro germline manipulation.
Lastly, sperm-mediated gene transfer has made some recent advances. Briefly, a DNA construct is bound to sperm, and the semen/sperm is used to inseminate a fertile female. The sperm that successfully fertilizes an egg carries the transgene of interest into the egg, the transgene becomes incorporated into the resulting embryo and hatched chicks. Recently, sperm-mediated transfer has been reported to be a reproducible way to make transgenic mammals (Lavitrano et al.,2002,2003). However, sperm-mediated gene transfer has not yet provided any convincing data to suggest that a transgene of interest can successfully be incorporated into the avian germline. Presently, retroviral and microinjection technology have been the only methods used to generate transgenic birds. Once a G0 bird is hatched, rearing the birds and establishing a line is not a trivial process.
The full program of producing transgenic birds requires a significant investment in laboratory and animal facilities. At the very least, incubation facilities must be sufficient for the incubation and hatching of manipulated embryos. Second, facilities must be sufficient for the incubation and hatching of eggs from any test matings and the generation of replacement stock. Small egg incubators and hatchers are sufficient for the manipulated eggs. These incubators must have digital temperature control, digital humidity control, and they must provide for different angles of egg turning. Larger capacity commercial incubators are necessary for incubating the continuous supply of fertile eggs necessary to support a transgenic project, and also the eggs from the transgenic test-matings. All eggs must be transferred to separate hatching incubators to help maintain cleanliness in the bird operation and to allow for pedigreed hatching. Proper brooding and rearing of the hatched chickens requires a significant investment in capital equipment. Battery style brooders are convenient for raising birds from hatch to approximately 4 weeks of age. All birds must be identified by wing banding or neck tagging. Once a transgenic chicken is identified by a polymerase chain reaction screening procedure with semen or blood DNA as the template, sequentially numbered bands must be inserted on each wing to ensure that the birds will never be misidentified. After the birds outgrow the brooder, they must be reared in a light controlled facility providing 8 hr of light. When the birds reach 17 to 20 weeks of age, they must be moved to a production facility where they must receive between 14 and 18 hr of light to stimulate sexual maturity and efficient reproduction. Test mating is performed by means of artificial insemination, and pedigrees must be maintained for all offspring. For the most part, G0 and G1 males are the most valuable, because they likely required much effort to generate, and they can be mated with many females to generate large numbers of offspring. In summary, it is important to note that a successful transgenic chicken facility must have good incubation facilities, brooding facilities, growing facilities, and production facilities. Many universities do not have these chicken rearing facilities readily available, and simply rearing the birds is a major roadblock for transgenic chicken technology that cannot be overlooked by new investigators seeking to enter the avian transgenic field.
Lastly, it is important to consider the time involved to generate a successful transgenic line. If a retroviral vector is injected into a prospective G0 embryo before incubation, it will take approximately 21 days until the embryo hatches (3 weeks). After hatching, it takes approximately another 17–20 weeks for the birds to reach sexual maturity, and semen can be screened for the transgene (∼5–6 months from the start of the project). A major unpredictable obstacle in the generation of transgenic lines is generating the G1 offspring carrying the transgene, because G0 birds exhibit mosaic transgene expression in the gonad. It is generally more efficient to only test-mate male G0 birds, because a single male can inseminate many females to generate many offspring at a single time point for screening. Subsequently, it will be at least another 21 weeks until the progeny from the G1 generation are sexually mature, and at least another 3 to 4 weeks until there is sufficient progeny to demonstrate Mendelian inheritance from the G1 to the G2 generation (Fig. 3). Therefore, it takes at least (optimistic minimum time) 12 months to generate founder stock for transgenic lines, and concurrently, strong evidence that a transgenic bird has actually been successfully constructed.
SUMMARY AND FUTURE DIRECTIONS
Overall, generating transgenic chickens for developmental biology research holds significant promise for the future, because new transgenic models may provide insight into early embryonic events. Nevertheless, success of generating a transgenic model is dependent on the choice of vector/method used to deliver the transgene to the embryo. In fact, simply delivering the transgene to an avian embryo is much more complicated than in a mammalian system. There is also a challenge to achieve sufficient hatchability to generate a sufficient number of G0 transgenic birds. Subsequently, the investigators must have sufficient space to maintain flocks for test-mating the birds. Despite these challenges, the recent successes of Harvey et al. (2002a,b;2003), McGrew et al. (2003), and Mozdziak et al. (2003) show that transgenic chicken technology can be used to generate tools for biological research.
We thank Becca Barnes, Suparerk Borwornpinyo, Carol Giamario, Jeff Hall, Darell McCoy, Simone Pophal, and Yonghong Song for their contributions to our transgenic research program. We also thank Jennifer Petitte for assistance preparing the illustrations.