Human Factor IX Transgenic Sheep Produced by Transfer of Nuclei from Transfected Fetal Fibroblasts

See allHide authors and affiliations

Science  19 Dec 1997:
Vol. 278, Issue 5346, pp. 2130-2133
DOI: 10.1126/science.278.5346.2130


Ovine primary fetal fibroblasts were cotransfected with a neomycin resistance marker gene (neo) and a human coagulation factor IX genomic construct designed for expression of the encoded protein in sheep milk. Two cloned transfectants and a population of neomycin (G418)–resistant cells were used as donors for nuclear transfer to enucleated oocytes. Six transgenic lambs were liveborn: Three produced from cloned cells contained factor IX and neo transgenes, whereas three produced from the uncloned population contained the marker gene only. Somatic cells can therefore be subjected to genetic manipulation in vitro and produce viable animals by nuclear transfer. Production of transgenic sheep by nuclear transfer requires fewer than half the animals needed for pronuclear microinjection.

Microinjection of DNA into the pronuclei of fertilized oocytes has been the only practical means of producing transgenic livestock since the method was established in 1985 (1). However, only a small proportion (∼5%) of animals integrate the transgene DNA into their genome (2,3). In addition, because the timing and site of integration are random, many transgenic lines do not provide sufficiently high levels of transgene expression or germline transmission. The consequent inefficient use of animals and associated high costs are a major drawback to pronuclear microinjection.

In mice, embryonic stem cells provide an alternative to pronuclear microinjection as a means of transferring exogenous DNA to the germline of an animal and allow precise genetic modifications by gene targeting (4, 5). However, despite considerable efforts, embryonic stem cells capable of contributing to the germline of any livestock species have not been isolated (6-11).

Recently, viable sheep have been produced by transfer of nuclei from a variety of somatic cell types cultured in vitro (12-14). We now demonstrate that nuclear transfer from stably transfected somatic cells provides a cell-mediated method for producing transgenic livestock.

We have used a transgene designed to express human clotting factor IX (FIX) protein in the milk of sheep. FIX plays an essential role in blood coagulation, and its deficiency results in hemophilia B (15). This disease is currently treated with FIX derived mainly from human plasma. Recombinant FIX produced in milk would provide an alternative source at lower cost and free of the potential infectious risks associated with products derived from human blood.

The transgene construct, pMIX1 (16), comprises the human FIX gene, containing the entire coding region (17), linked to the ovine β-lactoglobulin (BLG) gene promoter, which has been previously shown to provide a high level of transgene expression in ovine mammary glands (18). Analysis of pMIX1 expression in transgenic mice showed that seven of seven female founders expressed FIX in their milk (19). The level of expression in two animals (125 μg/ml) exceeded that achieved in previous studies (20, 21), indicating that pMIX1 is functional and suitable for introduction into sheep.

Primary strains of ovine cells, termed PDFF (Poll Dorset fetal fibroblast) 1 to 7, were derived from seven day-35 fetuses from the specific pathogen-free flock at PPL Therapeutics (22). Sex analysis of each cell strain by the polymerase chain reaction (PCR) (23) revealed PDFF5 to be male and the other six to be female.

Trial experiments indicated that both PDFF2 and PDFF5 cells could be readily transfected with a lacZ reporter gene with the use of the cationic lipid reagent Lipofectamine. PDFF2 cells at passage 1, after 3 days in culture, were cotransfected with pMIX1 DNA and the selectable marker construct PGKneo, and stable transfectants were selected with G418. Because the effects of drug selection and growth as single-cell clones on the ability of cells to support nuclear transfer were unknown, cells were then treated in two ways: One group was grown at high density under G418 selection and then cryopreserved as a pool for nuclear transfer. The other group was plated at low density under G418 selection, and cloned transfectants were grown from isolated colonies (22). A total of 24 clones was isolated, of which 21 were expanded for analysis of genomic DNA. Ten clones were found to contain pMIX1 by DNA hybridization analysis (24).

Untransfected PDFF2 cells cultured to passage 19 over a period of 80 days exhibited a modal chromosome number of 54, the euploid ovine chromosomal complement. The chromosome number of the four most rapidly growing pMIX1-transfected clones (PDFF2-12, -13, -31, -38) was determined at passage 6 or 7, after an average of 40 days in culture, and that of the uncloned PDFF2 pool was determined at passage 5, after 19 days in culture. Each clone and the pool showed a modal chromosome number of 54, indicating the absence of gross chromosomal instability during culture and drug selection.

We have proposed that induction of quiescence in nuclear donor cells by serum deprivation is necessary for successful nuclear transfer (12). After 5 days of culture in medium with a reduced serum content (0.5%), immunofluorescence detection of proliferating-cell nuclear antigen (PCNA), which is an indicator of active DNA replication, showed that none of the cells analyzed was in S phase, consistent with cell cycle arrest (25). Restoration of serum content to 10% reversed this effect and cell growth resumed.

Four cell types were used as nuclear donors: untransfected male PDFF5 cells, pooled female PDFF2 transfectants, and two transfected clones, PDFF2-12 and PDFF2-13, which contained >10 and ∼5 copies of the pMIX1 transgene, respectively. Transfer of nuclei from each cell type into enucleated oocytes derived from Scottish Blackface ewes was performed as previously described (12, 13).

Live lambs were obtained from all four cell types (Table1). As expected, animals derived from PDFF5 cells were male and those from PDFF2 cells were female. The efficiency of nuclear transfer, expressed as the number of liveborn lambs obtained per 100 reconstructed embryos, varied from 0.89% for PDFF2-13 to 2.25% for PDFF2-12. This efficiency is similar to the value (1.35%) that we obtained previously for nonmanipulated fetal fibroblasts from another breed of sheep (BLWF1) (13).

Table 1

Results of nuclear transfer. Nuclear transfer was performed as described previously (12, 13). All cells were exposed to a reduced serum concentration (0.5%) for 5 days before use as nuclear donors. PDFF5 cells were used for nuclear transfer at passage 2 or 3, PDFF2 transfected pools at passage 5 to 7, and transfected clones PDFF2-12 and PDFF-13 at passage 7 to 9. Liveborn lambs were defined as those with a heartbeat and able to breathe unassisted at birth.

View this table:

Pregnancies resulting from embryo transfer were determined by ultrasound scan at about 60 days after estrus, and development was subsequently monitored at regular intervals. Of the original 14 fetuses, 7 were liveborn, as defined by heartbeat and unassisted breathing (Table 2). Postmortem examination of aborted fetuses and dead lambs did not reveal any common factor as a cause of death.

Table 2

Characteristics of nuclear transfer–derived lambs. Outcomes of 11 pregnancies resulting from nuclear transfer of PDFF donor cells. When judged necessary, labor was induced by injection of dexamethasone at day 153 of gestation; when required, cesarean section (CS) was performed 24 to 52 hours later. The average duration of gestation for the Poll Dorset flock at PPL Therapeutics is 145 days.

View this table:

All animals derived from PDFF cells exhibited a prolonged gestation, and, with the exception of animals 7LL5 to 7LL8, labor was induced artificially. Delayed delivery was likely the cause of death of lamb 7LL9. Subsequently, all surrogate ewes were induced at day 153, and, if necessary, cesarean section was performed. Three of 11 pregnancies were twin pregnancies. In two instances (7LL6 and 7LL7 and 7LL10 and 7LL11), the death of one fetus in late pregnancy probably resulted in the death of the sibling.

The birth weight of nuclear transfer–derived lambs whose gestation exceeded 145 days ranged from 3.0 to 8.7 kg, with a mean of 3.7 kg for twins and 5.9 kg for single pregnancies. Poll Dorset lambs in the PPL Therapeutics New Zealand–derived flock have mean weights of 3.75 kg for twins and 5.1 kg for single pregnancies. However, comparison is complicated by the fact that nuclear transfer–derived lambs were gestated in Scottish Blackface surrogate mothers. All animals from PDFF2 cells had an undershot lower jaw that did not interfere with their well-being. This characteristic is a genetic trait that occurs sporadically in the Poll Dorset breed and is considered to be unrelated to nuclear transfer. The PDFF5 lambs did not show this feature.

DNA from nuclear transfer–derived lambs was analyzed for the presence of pMIX1 and PGKneo transgenes (Fig. 1). All fetuses and animals derived from the transfected PDFF2 cells were transgenic. The three animals derived from the PDFF2 pool (7LL8, -9, -12) contained the selectable marker gene but lacked the FIX transgene (Fig. 1, A and B). Fetuses and lambs derived from the cell clones PDFF2-12 (7LL10, -14, -15, -16) and PDFF2-13 (7LL13) contained both the FIX transgene (Fig. 1B) and PGKneo.

Figure 1

DNA analysis of transfected clones and transgenic sheep. Genomic DNA was isolated from the blood of live animals or tongue samples from dead animals, digested with Bam HI and Eco RI, and subjected to Southern hybridization with either a 1.8-kb fragment of the BLG promoter or the neo gene. (A) Southern analysis of the uncloned pool of cells (PDFF2 pool), and two lambs (7LL8 and 7LL9) derived from them, for the presence of pMIX1 and PGKneo. (B) Assay for the presence of the pMIX1 transgene in lambs derived from the PDFF2 pool (7LL8 and 7LL12) and from the transfected clones PDFF2-12 (7LL10, 7LL14 to 7LL16) and PDFF2-13 (7LL13). PDFF5 cells were not transfected. The positions and sizes of fragments corresponding to the transgenes and the endogenous BLG gene are indicated. The lane marked λ is a 1-kb ladder of phage λ fragments from 3 to 12 kb.

Our approach has shown that cell-mediated transgenesis is possible in a mammal other than the mouse. The technique is still in the early stages of development and problems remain to be addressed—in particular, the lack of spontaneous parturition and the incidence of perinatal mortality. However, the mortality rate we observed (46%) was exacerbated by two twin pregnancies in which the death of one lamb in late gestation may have resulted in the loss of the other. The mortality rate for nontwin pregnancies was 28.6%, higher than that occurring after normal breeding (∼8%) but similar to that observed after nuclear transfer with embryonic blastomeres (5 to 40%) (26). Our data therefore do not suggest any correlation between lamb mortality and extended culture or genetic manipulation of the donor cell. Many types of manipulation of preimplantation embryos—for example, in vitro oocyte maturation and fertilization, in vitro culture, asynchronous embryo transfer, and progesterone treatment of the mother—have been shown to increase fetal morbidity and mortality (26, 27). An increased understanding of the interaction between the transplanted nucleus and the host cytoplasm and the relation between the early embryo and the maternal environment, together with improved culture systems, should increase the success of embryo production and manipulation in vitro.

The use of somatic cell donors for nuclear transfer in livestock offers many advantages over pronuclear microinjection. Since 1989, PPL Therapeutics has generated a substantial number of transgenic sheep by pronuclear microinjection. A total of 51.4 animals are required to produce one transgenic lamb by pronuclear microinjection, compared with 20.8 animals in the present study by nuclear transfer, values that differ by a factor of ∼2.5 (Table 3). The most important difference is that no recipients are wasted gestating nontransgenic lambs in the nuclear transfer technique.

Table 3

Comparison of the production of transgenic sheep by nuclear transfer or pronuclear microinjection.

View this table:

Gestation of large numbers of nontransgenic embryos represents a major source of inefficiency (28). Several schemes have been devised to identify transgenic embryos before embryo transfer, either by detection of the transgene in embryo biopsies by PCR (29) or by co-expression of a marker gene (30, 31). However, these methods, with the possible exception of that of Takada et al. (30), are restricted by the persistence of unintegrated DNA during the short time that embryos can be cultured before embryo transfer. In contrast, cells transfected in vitro can be analyzed extensively before effort is devoted to large animals. This advantage will be particularly important in instances in which microinjection is inefficient; for example, with large constructs such as yeast artificial chromosomes.

Delayed integration of microinjected DNA into the embryo genome often results in mosaic founder animals. The reduced rate of transgene transmission resulting from germline mosaicism can hinder or prevent the establishment of transgenic lines from potentially valuable founder animals. In contrast, animals produced by nuclear transfer are entirely transgenic.

Nuclear transfer allows the sex of transgenic animals to be predetermined and thus offers a further twofold increase in efficiency relative to pronuclear microinjection when the sex of the transgenic founder animal is critical. If, for example, the primary interest is the expression of human proteins in milk, the founder generation can be all females. Sheep with different random integrations of the transgene can be produced by nuclear transfer from independent cell clones and the milk analyzed. After a suitable clone has been identified, the corresponding stock of cells can be used to generate an “instant flock” by further nuclear transfer. Such a flock could be superior to those produced by conventional breeding as a source of proteins for human therapy because genetic identity would contribute to the consistency of the medicinal product.

The procedures of transfection, drug selection, and growth from single-cell clones described here are essentially the same as those required for gene targeting. The realistic prospect of targeted genetic manipulation in a livestock species should open a vast range of new applications and research possibilities.

  • Present address: PPL Therapeutics, Roslin, Midlothian, EH25 9PP, Scotland, UK.


View Abstract

Stay Connected to Science

Navigate This Article