Production of α-1,3-Galactosyltransferase Knockout Pigs by Nuclear Transfer Cloning

See allHide authors and affiliations

Science  08 Feb 2002:
Vol. 295, Issue 5557, pp. 1089-1092
DOI: 10.1126/science.1068228


The presence of galactose α-1,3-galactose residues on the surface of pig cells is a major obstacle to successful xenotransplantation. Here, we report the production of four live pigs in which one allele of the α-1,3-galactosyltransferase locus has been knocked out. These pigs were produced by nuclear transfer technology; clonal fetal fibroblast cell lines were used as nuclear donors for embryos reconstructed with enucleated pig oocytes.

Clinical transplantation has become one of the preferred treatments for end-stage organ failure since the introduction of chronic immunosuppressive drugs in the mid-1980s. One of the novel approaches to dealing with the limited supply of human organs is the use of alternative species as a source of organs (xenotransplantation). The pig is considered the primary alternative species because of ethical considerations, breeding characteristics, infectious disease concerns, and its compatible size and physiology (1).

A major barrier to progress in pig-to-primate organ transplantation is the presence of terminal α-1,3-galactosyl (Gal) epitopes on the surface of pig cells. Humans and Old World monkeys have lost the corresponding galactosyltransferase activity in the course of evolution and therefore produce preformed natural antibodies to the epitope, which are responsible for hyperacute rejection of porcine organs. The temporary removal of recipient antibodies to Gal through affinity adsorption and expression of complement regulators in transgenic pigs has allowed survival of pig organs beyond the hyperacute stage. However, returning antibody and residual complement activity are believed to be responsible for the acute and delayed damage that severely limit organ survival even in the presence of high levels of immunosuppressive drugs and other clinical intervention (2). Competitive inhibition of galactosyltransferase in α-1,2-fucosyl-transferase transgenic pigs has resulted in only partial reduction in epitope numbers (3). Similarly, attempts to block expression of Gal epitopes inN-acetylglucosaminyltransferase III transgenic pigs also resulted in partial reduction of Gal epitope number but failed to substantially extend graft survival in primate recipients (4). Given the large number of Gal epitopes present on pig cells (5), it seems unlikely that any dominant transgenic approach of this nature can provide sufficient protection from damage mediated by antibodies to Gal. In contrast, a genetic knockout of the α-1,3-galactosyltransferase (GGTA1) locus in pigs would provide permanent and complete protection.

Viable α-1,3-galactosyltransferase knockout mice have been produced by embryonic stem cell technology (6). The development of nuclear transfer technology has provided a means for locus-specific modification of large animals, as demonstrated by the production of viable sheep by means of in vitro targeted somatic cells (7). Successful cloning (8–11) and production of transgenic pigs by nuclear transfer of genetically modified somatic cells (12) have been reported. Attempts at targeting the GGTA1 locus in pigs (11) and sheep (13) have also been reported, but these failed to result in live birth of animals with the desired modification. In both cases, difficulties in obtaining viable targeted donor cell clones were encountered.

We chose to knock out the GGTA1 locus in a highly inbred, major histocompatibility complex–defined miniature pig line. Descendant from lines long used for xenotransplantation studies (14, 15), this line is an ideal size match for eventual use in clinical transplantation and has animals that consistently test negative for transmission of porcine endogenous retrovirus (PERV) to human cells in vitro (16). Cells were isolated from one male (F9) and three female (F3, F6, and F7) fetuses at day 37 of gestation for production of donor cell lines (17). A gene trap targeting vector, pGalGT, was used for homologous replacement of an endogenous GGTA1 allele (Fig. 1). The vector contains about 21 kb of homology to the GGTA1 locus, with the coding region upstream of the catalytic domain disrupted by insertion of a selection cassette consisting of a Bip internal ribosome entry site followed by sequences encoding G418 resistance. After transfection and 14 days of G418 selection, viable cell clones (18) were passaged in triplicate for further analysis and cryopreservation (19).

Figure 1

pGalGT targeting vector and genomic PCR assays for targeting (35).

A reverse transcription polymerase chain reaction (RT-PCR) was performed on crude cell lysates the day after passage with a forward primer from exon 7 (upstream of the 5′ end of the targeting vector) and a reverse primer from the selection cassette (20). Dot blot hybridization of the RT-PCR products with an exon 8 probe detected targeting in 22 of 159 clones analyzed.

The structure of the GGTA1 locus was analyzed in two overlapping PCR reactions (21). Clones with a targeted insertion of the cassette relative to vector external primer sites both upstream and downstream of the cassette, indicative of a replacement-type targeting event, were considered candidates for use in nuclear transfer. Of 17 clones analyzed, 8 were found to have undergone the desired recombination event, and one from each fetus (F3-C5, F6-C3, F7-H6, and F9-J7) was used for nuclear transfer.

Nuclear transfer was performed with the use of in vitro matured oocytes and, except for 4 of 28 embryo transfers, cryopreserved donor cells without further culture (22). Asynchronous embryo transfer—that is, transfer to a surrogate at an earlier stage of the estrus cycle than the embryos themselves—had previously been used with minimally manipulated (23), pronuclear microinjected (24), and nuclear transfer (NT)–derived embryos (10–12). The observed benefit of asynchronous transfer suggests that any manipulation may result in a delay in early embryonic development. Because the manipulations required for nuclear transfer are quite extensive, and because previous reports suggest that miniature swine embryos of the NIH strain used here may normally develop at a relatively slower rate (25), naturally cycling large white gilts that had displayed standing estrus but had not yet completed ovulation were used as surrogates (26). [For detailed information on all 28 embryo transfers, see (18).]

A minimum of four viable embryos is required for establishment of pregnancy in pigs (27). Thus, we used two methods to increase the likelihood of establishing pregnancies with NT-derived embryos. Although pregnancy was established in five of seven surrogates receiving parthenogenetic “carrier” embryos, no live births resulted. Therefore, we transferred reconstructed embryos to a mated surrogate. In the one embryo transfer performed in this group, the surrogate (O212) was mated on the first day of standing estrus and received NT-derived embryos the same day. Although any fertilized embryos would theoretically be 43 to 55 hours behind development of the transferred NT-derived embryos, the actual in vivo development rate for NT-derived embryos is unknown. Early pig embryos have a lower rate of survival when present in a surrogate along with embryos at a slightly more advanced stage (25). Thus, an apparent embryonic asynchrony may be advantageous should NT-derived embryos develop at a slower rate than naturally fertilized embryos.

Seven piglets, four females and three males (Table 1), were delivered by cesarean section at term (28). Microsatellite analysis (29) revealed that six of six haplotypes for one female piglet (O212-2) were identical to that of the F7 fetal cell line from which knockout donor line F7-H6 was derived (18). Furthermore, three of six haplotypes of O212-2 were not compatible with mating of the surrogate. All other piglets had at least four haplotype mismatches with the F7 line and were compatible with mating of the surrogate.

Table 1

Pregnancies carried to term after transfer of embryos reconstructed with GGTA1 knockout cell lines.

View this table:

We performed 20 additional transfers of only NT-derived embryos to unmated surrogates. Pregnancy was confirmed by ultrasound in six of these surrogates, with two continuing to cesarean section at term. Two live piglets were delivered from surrogate O230 (Table 1), one of which (O230-2) died from respiratory distress syndrome shortly after delivery. Four live piglets were delivered from surrogate O226 (Table 1), again with one (O226-4) dying shortly after birth from respiratory distress syndrome. Microsatellite haplotypes of all six piglets from the two litters were identical to the F7 and F3 donor parental cell lines, respectively.

Genomic targeting analysis was performed on DNA samples from all NT-derived piglets, the untransfected F3 and F7 donor parental cell lines, and surrogates (Fig. 2). For all piglets except O230-2, analysis of both ends of the GGTA1 locus revealed the presence of one replacement-type targeted allele. Whether O230-2 was derived from an untargeted miniature swine cell in the F7-H6 donor line or had a GGTA1 rearrangement incompatible with detection by the targeting assays performed has yet to be determined.

Figure 2

Targeting analysis of NT-derived piglets, parental miniature swine fetal cell lines F3 and F7, and surrogate sows. SeeFig. 1 for a description of the assays. (A) Upstream genomic PCR analysis with primers F238 and R823. (B) Downstream genomic PCR analysis with primers F527 and GR2520. After transfer, digested reactions were probed with an oligonucleotide (Bip419) from the IRES portion of the selection cassette. The analysis of all offspring, with the exception of O230-2, is consistent with a replacement-type targeting event at one GGTA1 allele.

Table 2 presents a health summary for the seven NT-derived piglets. Four of the five piglets surviving beyond the immediate postpartum period remain healthy, with a normal growth rate for miniature swine. The fifth, O226-3, died suddenly at 17 days of age during a routine blood draw. Necropsy revealed a dilated right ventricle and thickening of the heart wall. Another animal, O230-1, has shown cardiac defects similar to those reported in NT-derived animals of other species (7, 12, 30, 31).

Table 2

Health summary of NT-derived miniature swine piglets. H, healthy; NG, normal growth; D, dead.

View this table:

A number of other abnormalities were noted at birth among surviving piglets, none of which appear to affect their overall health and well-being. Flexure tendon deformities similar to those reported here have been observed in previous NT-derived commercial strain pigs (32). It is unlikely that the abnormalities we have seen are related to the genetic modification, as there is not a consistent phenotype and only one allele has been targeted, but rather are the result of improperly reprogrammed epigenetic factors. With the exception of O212-2, the four surviving piglets were somewhat undersized, with birth weights of 450 to 650 g (strain average 860 g).

Under our growth and selection conditions, miniature swine fetal fibroblasts maintain a steady doubling time of about 24 hours. Clonal lines senesce after 30 to 32 days of culture on average. The ability to quickly select clonal lines for nuclear transfer is likely to be a requirement for introduction of other complex genetic alterations into the pig genome. The ability to use cryopreserved donor cells without further culture, demonstrated by two of our litters, is also advantageous, as it extends the number of potential donor lines available for use in nuclear transfer. Our efficiencies in producing NT-derived GGTA1 knockout animals are similar to those previously reported in which extensively cultured primary fetal cells were used as nuclear donors (10, 11), despite the nearly fourfold difference in adult size between the miniature swine strain modified here and the commercial oocyte donor and surrogate strains used. The ability to use readily available oocyte donors and surrogates in a nuclear transfer program is essential when modification of less commonly available animals is required.

The next step will be to create α-1,3-galactosyltransferase–null (homozygous knockout) pigs, either by breeding to a heterozygous male produced by nuclear transfer or by sequential nuclear transfer modification of cell lines produced from the four female pigs reported here. Because α-1,3-galactosyltransferase–null mice have already been produced (6), it is not anticipated that this genetic modification will be lethal in the null animals. We hope that α-1,3-galactosyltransferase–null pigs will not only eliminate hyperacute rejection but also ameliorate later rejection processes, and (in conjunction with clinically relevant immunosuppressive therapy) will permit long-term survival of transplanted porcine organs. At a minimum, the availability of galactosyltransferase-null pigs will allow a clearer evaluation of approaches currently in development aimed at overcoming potential delayed and chronic rejection mechanisms in porcine xenotransplantation.

  • * To whom correspondence should be addressed. E-mail: pratherR{at}


View Abstract

Stay Connected to Science

Navigate This Article