Production of Recombinant Therapeutic Proteins in the Milk of Transgenic Animals

August 1, 2006
Yann Echelard

Yann Echelard, PhD is vice president of research and development at GTC Biotherapeutics, Inc.

Harry M. Meade

Harry M. Meade, PhD, is senior vice president of research and development at GTC Biotherapeutics, Inc.

Carol A. Ziomek

Carol A. Ziomek, PhD is vice president for development at GTC Biotherapeutics, Inc.

BioPharm International, BioPharm International-08-01-2006, Volume 19, Issue 8

Capital investments in production plants represent a significant portion of the cost of new recombinant drugs


Transgenic milk production offers a cost-effective system for the manufacturing of large amounts of complex proteins. Specifically, commercialization is near for recombinant human antithrombin (rhAT) expressed in transgenic dairy goats. The product received a positive opinion from the Committee for Medicinal Products for Human Use of the European Medicine Agency.

This article reviews the reasons why transgenic milk is a cost-effective system. Also reviewed is the earlier research on targeting heterologous proteins to the mammary glands of many different animals. The final section describes the process by which goats express rhAT in their milk at approximately 2 g/L. The human AT purified from milk is structurally indistinguishable from human plasma–derived AT with the exception of carbohydrates. Clinical studies are ongoing on the prevention of deep-vein thrombosis.

The use of recombinant proteins has steadily increased during the last two decades. A large number of human proteins and potential therapeutic targets and their development for therapeutic uses have been identified. Clinical applications often require large amounts of highly purified molecules, for multiple or chronic treatments. The development of very efficient expression systems has been the key to the full exploitation of the recombinant technology. Thanks to a careful integration of molecular biology, large animal embryology and protein chemistry, transgenic milk production offers a cost-effective system for the manufacturing of large amounts of complex proteins.

Recombinant human antithrombin (rhAT, commercial name ATryn) is the most advanced of the transgenic milk-derived compounds. After several years in clinical development, it has recently received a positive opinion from the CHMP (Committee for Medicinal Products for Human Use) of the EMEA (European Medicine Agency) for its market application authorization for the prophylaxis of venous thromboembolism in surgery of patients with congenital antithrombin deficiency. This was the first positive opinion by a regulatory agency for a transgenically produced biopharmaceutical (from either plant or animal sources).

This article traces the development of protein technology from the microbe to the mammary glands. It will close with specific information on rhAT production and the status of clinical tests.

Various risk minimization measures have been instituted to protect this highly controlled closed donor goat population.


The first microbial bioreactors, in particular Escherichia coli and Saccharomyces cerevisiae, were found to be satisfactory for the production of simple polypeptides such as insulin and human growth hormone. However, microbial bioreactors were found to be unsuitable for proteins with complex post-translational modifications or intricate folding requirements, such as the coagulation factors, or monoclonal antibodies. This led to the development of large-scale mammalian cell culture, for example, the use of Chinese Hamster Ovary (CHO) cell bioreactors.

These technologies permitted the development of numerous monoclonal antibodies, cytokines, and other complex bioactive biomolecules. However, there are proteins that, due to a combination of complex structure and large therapeutic dosing, have until now eluded recombinant production using traditional bacterial and cell culture bioreactors. For example, commercial recombinant production of complex molecules, such as antithrombin and alpha1-antitrypsin, has not yet been achieved in microbial or mammalian cell derived bioreactors. The only source is human plasma because of the high dose needed.

Even human serum albumin, the therapeutic protein used in largest amounts (>400 tons, worldwide), the use of the recombinant form, produced in Saccharomyces cerevisiae, is limited to excipient applications. These are within the practical production capacity of this system but far too small for high-volume therapeutic indication (volume replacement).

Capital investments in production plants represent a significant portion of the development cost of new recombinant drugs. Also, the inherent risk associated with the regulatory approval process is a stimulus for the development of flexible and inexpensive approaches for the manufacture of therapeutic proteins. Milk-specific production offers a way to lessen the bite.


Here is the method to achieve milk-specific recombinant protein production. Fuse an expression vector, comprising a gene that is encoded for the human or humanized target protein with mammary gland-specific regulatory sequences, and then insert into the germline of the selected production species. When integrated, the milk-specific expression construct becomes a dominant genetic characteristic that is inherited by the progeny of the founder animal (Figure 1). This general strategy makes it possible to harness the ability of dairy animal mammary glands to produce large quantities of complex proteins.

Table 1. List of therapeutic proteins produced in the milk of transgenic animals that are currently in commercial development.

GTC Biotherapeutics and other have generated transgenic animal herds that yield large amounts of proteins as diverse as: human antithrombin (AT), alpha1-antitrypsin (AAT), C1 esterase inhibitor, fibrinogen, albumin, and monoclonal antibodies (Table 1). Technologies that permit the clinical-grade purification of recombinant therapeutic proteins from the milk of transgenic dairy animals have been developed and implemented.

Limitations of the transgenic expression systems are related to potential adverse effects of bioactive heterologous proteins on the health of the production animals and, to a lesser extent, to initial timelines. Although transgenic expression systems are able to perform complex post-translational modifications, such as γ-carboxylation, β-hydroxylation or N- and O-linked glycosylation, there are species- and tissue-specific characteristics for these modifications that may affect the appropriateness of a given system for the expression of specific proteins. This is also a challenge found with mammalian cell culture, microbial expression systems, or transgenic plants.


The targeting of heterologous proteins to the mammary gland of transgenic mice was independently reported by several groups during the late 1980s.1,2,3 These initial successes were followed by reports relating the generation of transgenic sheep, goats, cows, and pigs with milk-specific transgenes with the ultimate objective of producing recombinant proteins for clinical use (reviews by Clark,4 Meade et al.,5 Pollock et al.6 ).

The aim was to target recombinant proteins to the mammary gland of transgenic farm animals to solve problems associated with either microbial or animal cell expression systems. Bacteria often improperly fold complex proteins, leading to involved and expensive refolding processes, and both bacteria and yeast lack adequate post-translational modification machinery for mammalian-specific N- and O-linked glycosylation, γ-carboxylation, and proteolytic processing. Cell culture systems require high initial capital expenditures, lack scale-up (or down) flexibility, and use large volumes of culture media. On the other hand, transgenic livestock can be maintained and scaled-up in relatively inexpensive facilities, use animal feed as raw material, and can achieve impressive yields of recombinant proteins.

Figure 1. Schematic representation of the transgenic production process, using the production of rhAT in the milk of transgenic goats as an example.

The targeting of a recombinant protein to the milk of a transgenic animal (Figure 1) is achieved by first generating an expression vector containing the gene encoding the protein of interest fused to milk-specific regulatory elements. This transgene is then introduced in the germline of the chosen production species. Pronuclear microinjection of one-cell embryos (Figure 1) or, alternatively, transfection into a primary cell population suitable for somatic cell nuclear transfer (Figure 2) have both been used to generate transgenic founders.

Figure 2. Schematic representation of the somatic cell nuclear nuclear transfer process employed for the production of transgenic animals used for the production of recombinant proteins.

Following germline integration, mammary gland-specific transgenes are predictably inherited by the offspring of the founder animal. The expression level of the protein(s) of interest is variable. Concentrations surpassing 1 g/L are attained routinely and levels of up to 20 g/L have also been achieved. Expression levels are dependent on the mammary-specific regulatory sequences employed, the gene expressed, and the integration site of the transgene. Milk can easily be obtained using established large-scale technologies of the dairy industry, and is an excellent starting material from which recombinant therapeutic proteins can be purified. The choice of the production species is largely driven by the expected quantity of the therapeutic protein needed. There is usually a trade-off between milk yield and time to natural lactation. Another consideration may be a species-specific ability to perform specialized post-translational modifications more efficiently.

Figure 3. Timeline associated with the creation of a herd of transgenic goats producing recombinant proteins in their milk.


Transgenic mice have mainly been used for the testing of expression constructs prior to or concomitant with the generation of larger founder transgenic animals. This model allows the relatively inexpensive and rapid evaluation and optimization of transgene constructs and has proven crucial to the development of milk expression technology. The model allows the definition of regulatory sequences that efficiently target expression of heterologous genes to the mammary gland.7 Obviously, the very limited milk yield from transgenic mice restricts expression of recombinant proteins to small amounts. But this can be sufficient to obtain meaningful data on the protein of interest. As an example, it was possible to purify enough Malaria antigen MSP142 from transgenic mouse milk to test for immune protection in a primate model.8

The generation of transgenic rabbits by pronuclear microinjection is straightforward and inexpensive. Relative to ruminants, rabbits have a short gestation interval that allows up to eight lactations per year. However, only 1.5 L of milk can be obtained per lactation, and this limits the value of this expression system to products with a commercial scale in the low-kilogram range;9–12 Labor-intensive milking and high husbandry costs could become prohibitive for larger quantities of purified proteins.

Recombinant protein production in the milk of transgenic sows has been reported for human Protein C,13 factor VIII,14 and factor IX.15 Lactating sows can yield a surprising amount of milk (100–200 L) and it has been reported that the porcine mammary gland cells can carry out the complex post-translational modifications (γ-carboxylation, proteolytic processing) on factor IX and Protein C at rates higher than those encountered with mammalian cell and transgenic mouse milk systems.16

Transgenic ruminants are obvious candidates for targeting expression of recombinant proteins to the mammary gland. Thousands of years of patient genetic selection have yielded breeds of sheep, goats, and cattle that can produce prodigious quantities of milk. The first published report of production of therapeutic proteins in the milk of transgenic dairy farm animals was the targeting of factor IX and alpha1-antitrypsin to the milk of transgenic ewes.17 Other proteins such as fibrinogen and factor VIII have also been expressed in the mammary gland of transgenic sheep.

Transgenic dairy goats, with an average milk output per doe on the order of 600 to 800 L per natural lactation, have shown to be well adapted to the production of therapeutic proteins. The timeline from initiation of transgene transfer to natural lactation of resulting transgenic does is 16 to 18 months for goats (Figure 3). A large number of production females can be easily generated from a transgenic male using artificial insemination or embryo transfer techniques. Relatively small herds of a few hundred transgenic does can then easily yield several hundred kilograms of purified product per year. This level of production can meet the manufacturing needs of several factors traditionally derived from plasma fractionation and for a large number of recombinant antibodies currently in development.6

Dairy cows have a yearly milk output in the range of 10,000 L. Consequently, with concentrations routinely achieved with most mammary gland-specific proteins, yields of tens of kilograms of recombinant proteins can be produced by one lactating transgenic cow. In addition, embryo culture and transfer technologies are well established for cattle breeds, allowing efficient generation of transgenic cows by somatic cell nuclear transfer. However, it takes almost three years from the onset of transgene transfer to obtain milk from a cow's natural lactation. The tremendous scale-up potential offered by transgenic cattle may compensate for this drawback, especially for indications that necessitate large quantities of protein.


Now let us show this applied to a special product. The recombinant production of AT presented numerous challenges. Antithrombin is a complex glycoprotein carrying 4 N-linked glycosylation sites and 3 disulfide bonds. These characteristics, which are crucial for the functions of AT, precluded the use of microbial bioreactors for its recombinant production. In addition, the therapeutic use of AT calls for large amounts, often grams, of purified protein per course of treatment. This ruled out the use of standard mammalian cell culture bioreactors, because production costs with this approach would be prohibitive.

Expression in the milk of transgenic dairy goats was employed. The promoter region of the goat beta-casein gene was linked to hAT cDNA. This transgene was introduced into the chromosomes of goat embryos, which were then transferred to surrogate mothers. The resulting goats carrying this transgene produce the gene product, rhAT, in their milk. Transgenic offspring from the line selected for commercial development consistently express rhAT in their milk at approximately 2 g/L.18 Expression levels of up to 10 g/L were observed in other lines that were not developed further because of timing issues.

The rhAT protein is isolated from the milk of the transgenic females and conventionally purified using tangential flow filtration, heparin affinity chromatography, nanofiltration, anion exchange chromatography, and hydrophobic interaction chromatography, with a yield of greater than 50% (Figure 4). The human AT purified from transgenic goat's milk is structurally indistinguishable from human plasma-derived AT (hpAT) with the exception of the carbohydrates. The main glycosylation differences observed for rhAT were the presence of fucose and GalNAc, a higher level of mannose, and a lower level of galactose and sialic acid. There was also substitution of 40-50% of the N-acetyl neuraminic acid with N-glycolyl-neuraminic acid.18 The terminal sialic acid in the rhAT contained the same 2-6 linkage found in hpAT.

Figure 4. Schematic representation of the process used to purify ATryn from the milk of transgenic goats.

Several independent laboratories have determined that differences in glycosylation of AT do not affect the intrinsic rate constant of the uncatalyzed or heparin catalyzed inhibition of thrombin, indicating that the carbohydrate chains solely affect heparin binding and not heparin activation or proteinase binding functions. Thus, glycosylation does not impact the major biological activity of AT, which is thrombin inhibition, but explains the differences in affinity for heparin and in pharmacokinetics.

The manufacturing process for rhAT has been validated for its viral and prion removal capacity. The rhAT viral validation studies demonstrated that a significant virus reduction of >8.5 to >25.3 log10 (roughly 300 million fold to septillion fold) was accomplished across the distinctly different modes of the rhAT process.19 All GTC goats are certified free of scrapie in the United States Department of Agriculture (USDA) Scrapie Flock Certification Program and various risk minimization measures have been instituted to protect this highly controlled closed donor goat population. The rhAT purification process has been validated for its ability to reduce scrapie contamination over a 100 billion fold (>11.3 log10 scrapie removal).


Ten human clinical studies have been undertaken with rhAT, and several other studies are ongoing. Two clinical indications were pursued:

  • Heparin resistance in patients undergoing cardiac surgery involving CPB20,21

  • Prevention of deep vein thrombosis (DVT) in patients who have a hereditary deficiency of AT and who are in high risk situations such as delivery or surgery 22-24

In all the human studies completed, rhAT has not generated significant adverse events and has met the primary clinical endpoints. All studies demonstrated that rhAT was well tolerated in these patient populations.

A European regulatory filing was submitted in January 2004, for the use of rhAT in the prophylaxis of DVT in Hereditary AT deficient patients in a high-risk situation. On June 2, 2006 the CHMP of the EMEA adopted a positive opinion on the market authorization application for Atryn.25 The CHMP has recommended that ATryn be granted market authorization for the prophylaxis of venous thromboembolism in surgery of patients with congenital antithrombin deficiency. ATryn may be given in association with heparin or low molecular weight heparin in these situations.

Upon approval, expected about three months after the positive opinion, ATryn will be the first antithrombin product approved for use in all 25 countries of the European Union. ATryn will also be the only available recombinant antithrombin product that is not derived from the human blood supply. Furthermore, GTC Biotherapeutics anticipates using the results from both the completed study reviewed by the CHMP and an ongoing pivotal Phase 3 study to prepare a Biologics License Application for the FDA. The results of the pivotal Phase 3 study will also be submitted for consideration by the CHMP for expansion of the use of ATryn in Europe to prevent deep vein thromboses and thromboembolisms in women with hereditary antithrombin deficiencies who are undergoing childbirth.

The development of a recombinant option for antithrombin will provide a safe and reliable supply of this important factor and will facilitate the resumption of clinical trials aimed at acquired deficiencies of antithrombin, such as cardiovascular surgery, severe burns, and severe sepsis. For example, LEO Pharma, in partnership with GTC Biotherapeutics, has also begun development of ATryn in Europe as a potential treatment for disseminated intravascular coagulation, or DIC, associated with severe sepsis. DIC occurs in an estimated 220,000 severe sepsis cases in the European Union each year. Approximately 50% are fatal, representing a major unmet medical need of significant interest in critical care. Fulfilling the AT needs associated with the DIC clinical indication, if these development activities lead to an approval, would require several hundred kilograms of purified rhAT. This can be easily accomplished by scaling-up the existing herd of rhAT-producing transgenic goats.

Other recombinant proteins expressed in the milk of transgenic animals are currently in development, including human albumin, human growth hormone, C1-esterase inhibitor, alpha1-antitrypsin, as well as monoclonal antibodies. It is likely that in the coming years several regulatory filings in various jurisdictions will be submitted with some of these products, realizing the promise of transgenic technology in offering a safe cost-efficient alternative for the production of complex recombinant proteins.


The authors are deeply indebted to their many colleagues at GTC Biotherapeutics and Genzyme Corporation who have dedicated years to the development of transgenic technology and to the ATryn filings.

Yann Echelard, PhD, is vice president of research and development at GTC Biotherapeutics, Inc. 5 the Mountain Rd, Framingham, MA 01701, 508.370.5420, fax 508.370.5104,

Carol A. Ziomek, PhD, is vice president for development at GTC Biotherapeutics, Inc. 175 Crossing Blvd., Framingham MA 01702, 508.370.5421, fax 508.370.5266,

Harry M. Meade, PhD, is senior vice president of research and development at GTC Biotherapeutics, Inc., 175 Crossing Blvd., Framingham MA 01702, 508.370.5256, fax 508.370.5266,


1. Simons JP, McClenaghan M, Clark AJ. Alteration of the quality of milk by expression of sheep β-lactoglobulin in transgenic mice. Nature 1987; 328:530-532.

2. Gordon K, Lee E, Vitale JA, Smith AE, Westphal H, Hennighausen L. Production of human tissue plasminogen activator in transgenic mouse milk. Bio/Technology 1987; 5:1183-1187.

3. Meade H, Gates L, Lacy E, Lonberg N. Bovine αS1-casein gene sequences direct expression of active human urokinase in mouse milk Bio/Technology 1990; 8:443-446.

4. Clark, AJ.The mammary gland as a bioreactor: expression, processing, and production of recombinant proteins. J Mammary Gland Biol Neopl 1998; 3:337-350.

5. Meade HM, Echelard Y, Ziomek CA, Young MW, Harvey M, Cole ES, Groet S, Smith TE, Curling JM. Expression of recombinant proteins in the milk of transgenic animals. In: Fernandez JM and Hoeffler JP editors. Gene expression systems: Using nature for the art of expression. San Diego: Academic Press; 1998 p. 399-427.

6. Pollock DP, Kutzko JP, Birck-Wilson E, Williams JL, Echelard Y, Meade HM. Transgenic milk as a method for the production of recombinant antibodies. J. Immunol. Methods 1999; 137:147-157.

7. Maga EA and Murray JD. Mammary gland expression of transgenes and the potential for altering the properties of milk. Bio/Technology 1995; 13:1452-1457.

8. Stowers AW, Chen LH, Zhang Y, Kennedy, MC, Zou L, Lambert L, Rice TJ, Kaslow DC, Saul A, Long CA, Meade HM, Miller LH. A recombinant vaccine expressed in the milk of transgenic mice protects Aotus monkeys from a lethal challenge with Plasmodium falciparum. Proc Natl Acad Sci USA 2002; 99:339-344.

9. Stinnackre, MG, Massoud M, Viglietta C, Houdebine LM.The preparation of recombinant proteins from mouse and rabbit milk for biomedical and pharmaceutical studies. In: Houdebine LM editor. Transgenic animals, generation and use. Amsterdam: Harwood Academic Publishers; 1997. p. 461-463.

10. Korhonen VP, Tolvanen M, Hyttinen JM, Uusi-Oukari M, Sinervirta R, Alhonen L, Jauhiainen M, Janne OA, Janne J. Expression of bovine beta-lactoglobulin/human erythropoietin fusion protein in the milk of transgenic mice and rabbits. Eur J Biochem 1997; 245:482-489.

11. Stromqvist M, Houdebine LM, Andersson JO, Edlund A, Johansson T, Viglietta C, Puissant C, Hansson L. Recombinant human extracellular superoxide dismutase produced in milk of trans-genic rabbits. Transgenic Res 1997; 6:271-278.

12. McKee C, Gibson A, Dalrymple M, Emslie L, Garner I, Cottingham I. Production of biologically active salmon calcitonin in the milk of transgenic rabbits. Nat Biotechnol 1998; 16:647-651.

13. Velander WH, Johnson JL, Page RL, Russell CG, Subramanian A, Wilkins TD, Gwazdauskas FC, Pittius C, Drohan WN. High level expression of a heterologous protein in the milk of transgenic swine using the cDNA encoding human protein. Proc Natl Acad Sci USA 1992; 89:12003-12007.

14. Paleyanda RK, Velander WH, Lee TK, Scandella DH, Gwazdauskas FC, Knight JW, Hoyer LW, Drohan WN, Lubon H. Transgenic pigs produce functional human factor VIII in milk. Nat Biotechnol 1997; 15:971-975.

15. Van Cott KE, Butler SP, Russell CG, Subramanian A, Lubon H, Gwazdauskas FC, Knight J, Drohan WN, Velander WH. Transgenic pigs as bio-reactors: a comparison of gamma-carboxylation of glutamic acid in recombinant human protein C and factor IX by the mammary gland. Genet Anal Biomol Eng 1999; 15:155-160.

16. Lubon H and Paleyanda RK. Vitamin K-dependent protein production in transgenic animals. Thromb Haemost 1997; 78:532-536.

17. Clark AJ, Bessos H, Bishop JO, Brown P, Harris S, Lathe R, McClenaghan M, Prowse C, Simons JP, Whitelaw CBA, Wilmut I. Expression of human anti-hemophilic factor IX in the milk of transgenic sheep. Bio/Technology 1989; 7:487-492.

18. Edmunds T, Van Patten SM, Pollock J, Hanson E, Bernasconi R, Higgins E, Manavalan P, Ziomek C, Meade H, McPherson JM, Cole ES. Transgenically produced human antithrombin: Structural and functional comparison to human plasma-derived antithrombin. Blood 1998; 91, 4561-4571.

19. Ziomek C, Kutzko J, Sherman L, Gavin W, Hendry C, Hayes M, Cole ES. Viral and prion safety of recombinant human antithrombin. Ann Hematol 2003; 82(Suppl 1):S95.

20. Avidan MS, Levy JH, van Aken H, Feneck RO, Latimer RD, Ott E, Martin E, Birnbaum DE, Bonfiglio J, Despotis GJ. A phase III, double-blind, placebo-controlled, multicenter study on the efficacy of recombinant human antithrombin in heparin-resistant patients scheduled to undergo cardiac surgery necessitating cardiopulmonary bypass. Anesthesiology 2005; 102:276-284.

21. Avidan MS, Levy JH, van Aken H, Feneck RO, Latimer RD, Ott E, Martin E, Birnbaum DE, Bonfiglio LJ, Kajdasz DK, Despotis GJ. Recombinant human antithrombin III restores heparin responsiveness and decreases activation of coagulation in heparin-resistant patients during cardiopulmonary bypass. J Thorac Cardiovasc Surg 2005; 130:107-113.

22. Tait RC, Morfini M, Walker ID, Makris M, Dolan G, Menon R, Noonan PK, Frieling J, Bonfiglio J. Pharmacokinetics of recombinant human antithrombin in hereditary antithrombin deficient patients. Blood 2002; 100: Abst 3970.

23. Tait RC, Konkle BA, Bauer KA, Bonfiglio J, Dolan G, Frieling J, Greist A, Holmes HE, Makris M, Morfini M, Noonan B, Streisand JB, Weinstein R, Walker ID, Echelard Y. Clinical application of recombinant human antithrombin (AT) in patients with hereditary deficiency. Ann Hematol 2003; 82:S84.

24. Konkle BA, Bauer, KA, Weinstein R, Greist A, Holmes HE, Bonfiglio J. Use of recombinant human antithrombin in patients with congenital antithrombin deficiency undergoing surgical procedures. Transfusion 2003; 43:390-394.

25. GTC Biotherapeutics. Press Release. 2006 June 2. Available at