AU711144B2 - Materials and methods for management of hyperacute rejection in human xenotransplantation - Google Patents

Description

P00011 Regulation 3.2

AUSTRALIA

Patents Act, 1990

ORIGINAL

COMPLETE SPECIFICATION STANDARD PATENT 0O C C TO BE COMPLETED BY THE APPLICANT C r

C

P.S.

C

CC C NAME OF APPLICANT: BRESAGEN LIMITED AND ST VINCENT'S HOSPITAL (MELBOURNE) LIMITED ACTUAL INVENTORS: ADDRESS FOR SERVICE: INVENTION TITLE: MARTIN J PEARSE; ROBERT J CRAWFORD; ALLAN J ROBINS; ANTHONY J F d'APICE Peter Maxwell Associates Level 6 Pitt Street SYDNEY NSW 2000 MATERIALS AND METHODS FOR MANAGEMENT OF HYPERACUTE REJECTION IN HUMAN

XENOTRANSPLANTATION

DETAILS OF ASSOCIATED PROVISIONAL APPLICATION NO(S): NIL The following statement is a full description of this invention including the best method of performing it known to me:- MATERIALS AND METHODS FOR MANAGEMENT OF HYPERACUTE REJECTION IN HUMAN XENOTRANSPLANTATION Field of the Invention This invention relates generally to the field of xenotransplantation. In particular this invention relates to methods and materials for reduction or elimination of the hyperacute rejection response in humans. More particularly, this invention relates to methods and materials for generating non-human organs lacking or having reduced ca-1, 3 galactosyltransferase activity.

10 Background of the Invention It is widely acknowledged that there is an acute, worldwide shortage of human organs for transplantation. This is in spite of legislative changes and education programs to increase awareness of the problem. In the United States, for example, there is an estimated annual shortfall of approximately 18,000 kidneys/year. Similarly, in Australia in 1992, only 41 of renal patients awaiting transplantation received transplants. In Japan the imbalance between supply and demand is even greater due to "religious prohibitions on the use of organs from cadaveric donors.

The benefits of transplantation can be seen by comparing the rehabilitation rates of transplant patients with those of dialysis patients. In Australia and New Zealand, the majority of transplant patients are capable of full time work or school with a further 10% in part time work, while only 7% are unfit for work. In 2 contrast, 23% of dialysis patients are capable of full time work or school, with 15% involved in part time work and 20% unfit for work. The remainder are "retired." Fifteenth Report of the Australia and New Zealand Dialysis and Transplant Registry (ANZDATA), Queen Elizabeth Hospital, Woodville, APS Disney, ed.

(1992).

The direct financial cost of dialysis in Australia and New Zealand is approximately $A45,000/patient/year.

In addition, indirect costs due to unemployment and sickness are higher in dialysis patients and the social costs are considerable. Transplantation engenders an expense of approximately $A30,000/patient in the first year and $A14,000/patient/year thereafter. These 15 statistics indicate that a) transplantation is the optimal therapy for end stage renal failure; b) there is an undersupply of donor kidneys; and c) present strategies aimed at increasing the transplant rate have been less than successful. There are, in addition, serious shortages of other transplantable organs including hearts, livers, lungs and pancreases.

The use of xenografts (transplants between species) is one option for overcoming the short supply of human organs for transplantation. Non-viable, non- 25 antigenic xenografts are commonly used in vascular reconstruction (bovine arteries) and in cardiac surgery (porcine cardiac valves). However, despite their occasional use in the past, immunological barriers have prevented the common use of viable xenografts. Between 1964 and 1991 a total of 27 non-human primate to human organ xenografts was reported; the longest reported patient survival was 9 months. Two liver transplants from baboon to human were recently performed in anticipation that modern immunosuppressive therapies could cope with the severe rejection problems likely to L~jJ~UL~V" rLVYL;UL3 It:t-ly L 3 occur in xenotransplantation. To date, the course of one of these patients has been reported, and in this case rejection was not the direct cause of death. Starzl et al., Baboon-to-Human Liver Transplantation. Lancet 341: 65-71 (1993). This clinical experience indicates that a) non-human organs can function and support human life; b) rejection episodes can be reversed by conventional antirejection therapy; and c) the mechanisms of rejection are similar, in principle, to those in allograft rejection.

It is unlikely that primates will be a satisfactory source of organs for xenotransplantation.

Most are endangered species, breed slowly in the wild and poorly in captivity. The baboon is an exception to these generalizations, but other disadvantages limit the 15 usefulness of this species. Baboons have single pregnancies, long gestation times, are difficult and expensive to maintain and may be infected with or carry organisms, particularly viruses, that are pathogenic in humans. For hearts and kidneys where organ size may be a consideration, the smaller primates are unsatisfactory as donors to human adults. Finally, the use of primates is likely to arouse considerable opposition from the public.

These difficulties have led to renewed interest in the use of non-primate species as organ donors for human S. 25 patients. The pig is a widely acknowledged choice for xenotransplantation into humans. The pig erythrocyte diameter (6.5gm) and, by implication, its capillary size, are similar to humans, facilitating connection of xenografts to the human circulatory system. The pig breeds well in captivity, has a short gestation time and produces large litters. In addition, pigs can be bred and maintained in low pathogen facilities, can be reared to any size and do not arouse the level of public reaction associated with primates.

The immunological barriers to use of pig organs in human patients include a) an immediate severe ("hyperacute") rejection phenomenon that develops in minutes to hours after transplantation, and b) a proposed acute rejection that develops in days to weeks. Once the hyperacute rejection phenomenon has been overcome, it is expected that normal acute rejection would ensue. This form of rejection is thought to be similar to that experienced with allografts (transplants between individuals of the same species) and should be amenable to normal immunosuppressive therapies.

Both preformed "natural antibodies" (xenoantibodies) and complement regulating factors in human serum are thought to be involved in the process of hyperacute rejection. Hyperacute rejection is thought to i: be initiated when xenoantibodies bind to epitopes on the endothelium of a 0: donor organ, activating the classical complement pathway.

The present invention provides a DNA construct when used for inactivating an a-1, 3 galactosyltransferase gene, comprising a DNA sequence for porcine a-1, 3 galactosyltransferase, wherein said DNA sequence carries a disruption of said DNA sequence such that functional a- 1, 3 galactosyltransferase is not encoded by said DNA sequence.

As used herein, the term 3 galactosyltransferase gene" S 20 includes the exons encoding or potentially encoding a-1, 3 galactosyltransferase, introns contiguous with such exons, and regulatory elements associated with such exons and introns.

Preferably, the disruption is within exon 4, exon 7, exon 8 or exon 9 of the porcine a-1, 3 galactosyltransferase gene.

The said DNA sequence preferably comprises a selectable marker, including but not limited to the neoR gene, and the hygromycin resistance (hyg

R

gene.

30/7/99 The said selectable marker may be bordered at both ends by FRT DNA elements, with stop codons for each of the three reading frames being inserted 3' to the said selectable marker.

Presence of the FRT elements allows the selectable marker to b deleted from the targeted cell, and the stop codons ensure that the a-1, 3 galactosyltransferase gene remains inactivated following deletion of the selectable marker.

The invention further includes a DNA construct when used for inactivating an a-1, 3 galactosyltransferase gene, comprising a DNA 10 sequence for murine a- 1, 3 galactosyltransferase, wherein said DNA sequence carries a disruption of said DNA sequence such that functional a- 1, 3 galactosyltransferase is not encoded by said DNA sequence.

Preferably, the disruption is within exon 4, exon 7, exon 8 or exon 9 of the murine a-1, 3 galactosyltransferase gene.

15 The said DNA sequence preferably comprises a selectable marker, including but not limited to the newR gene, and the hyg gene.

The said selectable marker may be bordered at both ends by FRT DNA elements, with stop codons for each of the three reading frames being inserted 3' to the said selectable marker.

Presence of the FRT elements allows the selectable marker to be deleted from the targeted cell, and the stop codons ensure that a- 1, 3 galactosyltransferase gene remains inactivated following deletion of the selectable marker.

The invention also includes methods for generating a mammalian totipotent cell having at lease one inactivated (non-functional) a-1, 3 galactosyltransferase allele, where the totipotent cell is derived from a 30/7/99 6 mammalian species in which alleles for the a- 1, 3 galactosyltransferase gene normally are present and functional.

A "functional" allele is capable of being transcribed and translated to produce a polypeptide having an activity the same as or substantially similar to the native a-1, 3 galactosyltransferase.

The methods include providing a plurality of cells characterized as totipotent cells of the aforementioned mammalian species, introducing into the totipotent cells a nucleic acid construct effective for inactivating the a- 1, 3 galactosyltransferase gene, said DNA construction comprising of DNA 10 sequence for a- 1, 3 galactosyltransferase, wherein said DNA sequence carries a disruption of said DNA sequence such that functional a- 1,3 galactosyltransferase is not encoded by said DNA sequence, and then identifying a totipotent cell or mitotic descendants of a totipotent cell having at least one inactivated a- 1,3 galactosyltransferase allele.

15 The totipotent cells can include, without limitation, embryonic stem (ES) cells, primordial germ cells (PGC's) and fertilized or unfertilized eggs.

The cells can be taken from a variety of mammalian species in which alleles for the a- 1, 3 galactosyltransferase gene are present and functional, including, without limitation, murine and porcine species.

The invention further includes methods for generating a non-human mammal lacking a functional a- 1, 3 galactosyltransferase gene, where the mammal belongs to a species having a functional a- 1, 3 galactosyltransferase gene.

The methods include providing a mammalian totipotent cell having at least one inactivated a- 1, 3 galactosyltransferase allele, where the totipotent cell is derived from the aforementioned mammalian species having a functional a- 1, 3 galactosyltransferase gene, manipulating the totipotent cell such that mitotic descendants of the cell constitute all or 30/7/99 part of a developing embryo, allowing the embryo to develop to term, recovering a neonate individual derived from the embryo, and raising and breeding the neonate to obtain a mammal homozygous for inactivated a-1, 3 galactosyltransferase alleles, a mammal in which both of the a-1, 3 galactosyltransferase alleles are inactivated) and lacking the GAL epitope as determined by failure of somatic cells of said mammal to bind anti-GAL antibodies and IB4 lectin and resistance of said cells to lysis by human serum.

The totipotent cells can include, without limitation, ES cells, PGC's 10 and eggs.

The cells can be taken from a variety of mammalian species in which the alleles for the a-1, 3 galactosyltransferase gene are present and functional, including without limitation murine and porcine species.

ES cells, PGC's and eggs are manipulated in various ways such that their mitotic descendants are found in a developing embryo. These manipulations can include, without limitation, injection into a blastocyst or morula, co-culture with a zona pellucida-disrupted morula, and fusion with an enucleated zygote. Cells injected into a blastocyst- or morula- stage embryo become incorporated into the inner cell mass of the blastocyst embryo, giving rise to various differentiated cell types of the resulting embryo, including in some cases germ cells.

The embryo derived from such manipulations is a chimera composed of normal embryonic cells as well as mitotic descendants of the introduced ES cells or PGC's or eggs.

Alternatively, chimeric embryos can be obtained by co-culturing at least one ES cell or PGC or egg with a morula embryo in which the zona pellucida is sufficiently disrupted to allow direct contact between the ES cell/PGC/egg and at least one cell of the morula.

8 The zona pellucida-disrupted embryo may be an embryo that is completely free of the zona pellucida.

Finally, the genome of an ES cell or PGC or egg can be incorporated into an embryo by fusing the ES cell/PGC/egg with an enucleated zygote.

Such a procedure is capable of generating a non-chimeric embryo, i.e. an embryo in which all nuclei are mitotic descendants of the fused ES cell/PGC/egg nucleus. The resulting embryos are implanted in a recipient female, or surrogate mother, and allowed to develop to term.

When eggs, as opposed to ES cells or PGC's, are directly injected with a nucleic acid construct effective for inactivating the a-1, 3 galactosyltransferase gene, the eggs can be manipulated to form an embryo by implanting into a recipient female.

As used herein, "high stringency conditions" are those hybridization conditions generally understood by the skilled artisan to reflect standard conditions of high stringency as set out in widely recognized protocols for nucleic acid hybridization. See, Sambrook et al, Molecular Cloning: A Laboratory Manual (2nd Edition), Cold Spring Harbor Laboratory Press (1989), pp. 1.101 1.104; 9.47 9.58 and 11.45 11.57. Generally, these conditions reflect at least one wash of the hybridization membrane in 0.05x to 0.5x SSC with 0. 1% SDS at 65 0 C, or washing conditions of equivalent stringency.

BRIEF DESCRIPTION OF THE DRAWINGS FIGURE 1 is a graphical representation of fluorescence intensity following immunofluorescent staining of porcine aortic endothelial cells with anti- GAL antibody alone or with anti-GAL antibody that was preincubated with selected saccharides.

FIGURE 2 shows the results of an experiment in which lysis of porcine aortic endothelial cells by human serum and by purified anti-GAL antibodies was determined using a 51 CR release assay.

FIGURE 3 depicts physiograph tracings of perfused rat heart contractions in the presence of human serum with or without selected saccharides.

FIGURE 4 is a comparison of the porcine a-1,3 galactosyltransferase cDNA sequence with the corresponding murine and bovine sequences. PGTCD porcine sequence. BOVGSTA bovine sequence. MUSGLYTNG murine sequence.

15 FIGURE 5 is a comparison of the porcine a-1,3 galactosyltransferase amino acid sequence with the corresponding murine and bovine amino acid sequences.

PGT porcine sequence. BGT bovine sequence. MGT murine sequence.

20 FIGURE 6 depicts the Sall restriction sites in four overlapping phage clones spanning a portion of the murine a-1,3 galactosyltransferase genomic region.

FIGURE 7 is a detailed restriction map of murine a-1,3 galactosyltransferase subclone FIGURE 8 is a detailed restriction map of murine a-1,3 galactosyltransferase subclone FIGURE 9 is a detailed restriction map of murine a-1,3 galactosyltransferase subclone paGT-S1l.

FIGURE 10 is a detailed restriction map of murine a-1,3 galactosyltransferase subclone paGT-S13.

FIGURE 11 is an additional detailed restriction map of murine a-1,3 galactosyltransferase subclone paGT- FIGURE 12 is an additional detailed restriction map of murine a-1,3 galactosyltransferase subclone paGT- FIGURE 13 is a diagram of a knockout construct carrying the 4.0 and 5.5kb Sail fragments from and paGT-S4.0, which flank the Exon 9 Sail site.

FIGURE 14 depicts the 8.3kb and 6.4kb BglII fragments that are diagnostic for the uninterrupted a-1,3 galactosyltransferase gene and the targeted (inactivated) a-1,3 galactosyltransferase gene, respectively, using the *probes identified in the text.

FIGURE 15 is a schematic representation of the generation of a knockout construct using the vector paGT- S5.5 as the starting vector.

15 FIGURE 16 sets out the nucleotide sequence of a neomycin resistance cassette used in the construction of a DNA construct for interrupting the a-1,3-GalT gene in mice.

FIGURE 17 is a diagram of one example of a final 20 knockout construct that has been sequenced to confirm the identity, copy number and orientation of the various inserts.

FIGURE 18 is a Southern blot of genomic DNA from various murine ES cell lines transformed with the knockout construct of Figure 16, probed to reveal the diagnostic fragments depicted in Figure 14.

FIGURE 19 depicts the "long" PCR products derived from wild type and interrupted a-1,3-GalT genes using the designated primers.

FIGURE 20 is a Southern blot of long PCR products obtained from wild type and knockout mice.

FIGURE 21 depicts the PCR products used for identification of the interrupted (targeted) galT locus.

FIGURE 22 shows PCR products generated from mice carrying interrupted (inactivated) GalT alleles.

11 FIGURE 23 depicts the PCR products expected from PCR analysis of cDNA generated from ac-1, 3-GalT mRNA in normal and knockout mice.

The ferrochelatase primers and PCR fragment represent a control demonstrating that cDNA synthesis had occurred.

FIGURE 24 shows the PCR fragments generated from cDNA obtained from RNA isolated from kidney heart and liver of a wild-type mouse a mouse heterozygous for the interrupted a-1, 3-GalT allele and a mouse homozygous for the interrupted a-1, 3-GalT allele FIGURE 25 is a graphical representation of the relative protection of 10 spleen cells, derived from GalT knockout mice, from lysis by human serum.

DETAILED DESCRIPTION ~Evidence presented herein establishes that a substantial portion of human pre-formed, anti-pig xenoantibodies recognize a specific terminal galactose linkage on the surface of pig endothelial cells. As demonstrated 1 5 in experiments carried out by the present inventors, it is possible to reduce the titers of preformed xenoantibodies by adsorption with immobilized 0. antigens containing the appropriate epitopes. This leads to reduction or elimination of cellular responses associated with the hyperacute rejection response.

Conversely, it is demonstrated to be possible to neutralize such antibodies by addition of appropriate carbohydrate antigens to human serum. In demonstrating the usefulness of these approaches, it was necessary to identify the relevant carbohydrate moieties and to demonstrate their efficacy in cultured cell systems and, importantly, in whole organs. As such, one approach to reducing or eliminating the hyperacute rejection response is identified as treatment of the recipient by eliminating or neutralizing the relevant antibody populations.

An alternative approach to xenotransplantation 15 would be elimination of the relevant epitope(s) in the donor organ. This could be accomplished, for example, by reducing or eliminating expression of the gene(s) encoding the metabolic machinery responsible for formation of the epitopes. The epitope defined by the a- 20 1,3 galactose linkage (termed the GAL epitope) is *generated by the enzyme UDP-galactose:-D-galactosyl-l,4- N-acetyl-D-glucosaminide a-1,3 galactosyl- transferase galactosyltransferase" or "a-l,3-GalT"). This enzyme transfers galactose to the terminal galactose residue of N-acetyllactosamine-type carbohydrate chains and lactosaminoglycans. The reaction catalyzed by a-1,3- GalT may be summarized as follows: UDP-Gal Galf-1,4-GlcNAc-R Gala-1,3-Galp-1,4- GlcNAc-R UDP The a-l,3-Gal T enzyme is found in most mammals, but is not present in Old World monkeys and humans. For purposes of xenotransplantation, it is significant that humans and Old World monkeys have naturally occurring xenoantibodies directed against the GAL epitope. The use of pig organs lacking the GAL epitope could reduce or eliminate the hyperacute rejection of such organs by human recipients. The utility of such an approach is buttressed by the present inventors' demonstration that the GAL epitope is, in fact, central to the hyperacute rejection phenomenon in cells and whole organs. One approach to obtaining such organs would be to generate pigs in which the gene encoding the a-l,3-GalT enzyme is "knocked out" by homologous recombination.

Role of the GAL Epitope in Hyperacute Relection 10 The present inventors have affinity purified antibodies directed against the GAL epitope (anti-GAL antibodies) from human serum. This was accomplished with affinity columns comprising the appropriate epitopes galactosyl-galactose or melibiose) attached to a solid phase. Total anti-GAL IgG and IgM were obtained in one set of experiments. In an alternative approach, anti-GAL IgG was obtained by passage of serum over an affinity column with specificity for all proteins except albumin and IgG. The wash-through from this column was 20 then applied to a galactosyl-galactose affinity column and purified anti-GAL IgG was collected as the eluate.

The obtained anti-GAL IgG can be further purified by passage over a protein G column, which specifically binds IgG but not other antibody isotypes. Conversely, the wash-through from the above-described columns can be used as sources of total anti-GAL (IgG IgM)-depleted serum or of anti-GAL IgG-depleted serum in further experiments.

Preferably, the anti-GAL antibody preparations are characterized for protein content, molecular weight and purity, and for antibody class and isotype.

To demonstrate the role of the GAL epitope in the hyperacute rejection response, it is necessary, first, to establish that IgG and IgM anti-GAL antibodies react with porcine cells and tissues. The present inventors investigated the binding of human anti-GAL antibodies to porcine cells and tissues using immunofluorescent staining. In this technique, selected human antibody preparations are reacted with intact porcine cells and then reacted with signal antibody comprising non-human anti-human IgG or IgM labeled with fluorescein isothiocyanate (FITC). Stained cells may be detected and quantified with a fluorescence-activated cell sorter (FACS) or other appropriate detection means. Other methods for detecting the presence of a bound antibody on a cell surface, for example through use of enzyme-labeled signal antibody reagents, are known to the skilled artisan.

Total anti-GAL (IgM and IgG), as well as 15 purified anti-GAL IgG, stained cells from a porcine epithelial cell line (PK 1 as well as cells from a porcine aortic endothelial cell line (PAE). Neither anti-GAL (total IgM IgG) antibody-depleted serum nor anti-GAL IgG-depleted serum gave detectable staining. To 20 further investigate the specificity of the response, it o* is desirable to determine whether or not reactivity of the antibodies with porcine cells can be diminished or eliminated by prior exposure to one or more molecules suspected of comprising the epitope(s) in question. In this regard, the present inventors have established that antibody binding is inhibited by galactose and by disaccharides having terminal galactose residues in the al configuration. Staining was not inhibited with sugars having a terminal galactose in a ~3-4 configuration.

These results demonstrate the specificity of the antibody binding and the ability of appropriate sugars to inhibit such binding.

Reactivity of anti-GAL antibodies with cultured pig cells was confirmed using tissue sections of pig organs. Again, using a fluorescent signal antibody system, staining was seen with total anti-GAL IgM IgG and with purified anti-GAL IgG but not with the anti-GAL antibody-depleted sera. Staining was particularly strong with kidney, heart and liver endothelium, with heart endocardium and with bile duct epithelium. The tissue binding was inhibited with melibiose but was not inhibited by other disaccharides not representative of the GAL epitope.

These data clearly indicate that the GAL epitope is expressed at high levels on the endothelial cells of arteries, veins and capillaries of porcine kidney, heart and liver. In a xenograft situation, the endothelial cells of these vessels come into direct contact with the anti-GAL antibodies in human serum. The above results 15 are consistent with evidence that binding of these antibodies (with attendant complement activation) is a :key component of the hyperacute rejection response.

To further investigate the specificities of naturally occurring xenoantibodies in human serum 20 directed against porcine antigens, the ability of human serum to cause agglutination of pig red blood cells was investigated. These studies revealed the presence of high levels of such antibodies in human serum. Moreover, sugars such as melibiose, stachyose, galactose and fucose, having terminal residues in the al-6 configuration, were found to inhibit agglutination in the pM to mM range. Sugars with other configurations were only inhibitory at very high doses, where the observed effects are likely due to simple changes in osmolarity or other non-specific mechanisms.

The above investigations establish a potential role for naturally occurring, human anti-GAL xenoantibodies in the complement-mediated destruction underlying hyperacute rejection. However, it is preferable to directly examine complement-mediated destruction of porcine cells in order to confirm the specificity of the GAL epitope and of anti-GAL antibodies in the process of lysis. To this end, the present inventors have examined the ability of human serum to cause lysis of porcine cells.

To investigate complement-mediated destruction of cells, it is necessary to employ one or more assays that provide quantitative data on cell lysis. Preferably, such assays measure a cell-sequestered component that is released into the medium upon complement-mediated cell lysis. Such experiments should control for involvement of complement in the induced lysis by employing both native complement proteins as well as heat-inactivated complement. The present inventors have used a 51 Cr- 15 release assay and a lactate dehydrogenase (LDH)-release assay to investigate the complement-mediated lysis of porcine epithelial and endothelial cells by human serum.

In the 51 Cr-release assay, porcine cells were pre-labeled with 51 Cr and then incubated in the presence of heat-inactivated human serum plus rabbit complement (PAE's) or human complement in non-heat-inactivated normal human serum (PK' Release of 51 Cr into the medium was measured with a gamma counter following addition of scintillation fluid. In the LDH-release 25 assay, cells were labeled with LDH as per the manufacturer's instructions (Promega, USA). Release of LDH into the medium was measured using an ELISA format, with absorbance read at 492nm. For both assays, the ability of various sugars to inhibit the complementinduced lysis was also tested.

Similar results were obtained with the two unrelated porcine cell lines, PAE and PK, using both types of assays. The results clearly demonstrate that naturally occurring xenoantibodies (NXAb's) are responsible for initiating the complement-induced lysis of porcine cells. The present inventors have also established that IgM and not IgG antibodies are responsible for the lysis in this system. Moreover, heat inactivation of the complement preparations prevented lysis, providing further evidence that lysis of the porcine cells is a complement-dependent phenomenon. The present inventors have also shown that melibiose, but not lactose, protects the porcine cells from lysis.

Importantly, the concentrations of sugar found to be effective in these studies covered the physiological range of blood sugar, about These results indicate that the anti-GAL NXAb's in normal human serum are primarily responsible for lysis of the porcine cells. As such, the binding of anti-GAL 15 NXAb's to the endothelial cells lining the blood vessels of a porcine xenograft, with attendant activation of the complement cascade, is likely to be a key component of the hyperacute rejection of porcine xenografts. This would also be the case with organs from other discordant species, such as rodents, sheep, cows and goats, all of which have active a-1,3-GalT genes in their genomes.

These conclusions are further supported in a whole-organ study performed by the present inventors.

For this study, isolated and perfused rat hearts were 25 used to further demonstrate the involvement of anti-GAL xenoantibodies in hyperacute rejection. Rat hearts were connected to a Langendorf perfusion apparatus, as described in Doring and Dehnert, The Isolated Perfused Heart According to Langendorf, Bionesstechnik-Verlag March GmbH, D7806, West Germany. The connected hearts were then stabilized by perfusion with a physiological buffer system, and perfused with the same buffer containing either melibiose or lactose (10mM). Human plasma was then added to a final concentration of 13% and 18 the effect of the added sugar on heart rate, strength of contraction and output were measured.

These results demonstrate in a whole-organ system that: 1) Perfusion with unmodified human plasma causes rapid loss of function.

2) Perfusion of a rat heart with human plasma in the presence of melibiose, which competes for binding with the anti-GAL antibodies, prolongs heart survival and output. Lactose, however, which does not compete for binding with the anti-GAL antibodies, does not prolong heart survival.

3) Perfusion of a rat heart with anti-GAL antibody-depleted plasma prolongs heart survival and ~15 output.

4) If purified anti-GAL antibodies are added back to anti-GAL antibody-depleted plasma, the heart rapidly loses function The present inventors' experiments with cultured *oo 20 cells, tissues and whole organs provide important confirmation that anti-GAL antibodies are a critical element in the hyperacute rejection response. Moreover, *the disclosed results point to various approaches that be employed to eliminate or reduce the hyperacute rejection of xenogeneic mammalian organs by humans.

For example, the intravenous administration of the specific disaccharide galactose a 1-3 galactose will block the naturally occurring anti-GAL antibodies of all classes and prevent them binding to their specific epitopes on the surface of the endothelial cells of the xenograft, thus preventing them from initiating or participating in hyperacute rejection. The present inventors' results indicate that the concentration of galactose a 1-3 galactose required to achieve this effect

I

is in a physiologically tolerated range. The experiments also indicate that various other carbohydrates can be substituted for the specific disaccharide. These include the monosaccharide galactose and various other di-, trior tetra-saccharides in which there is a terminal a galactose linked to the next sugar via position 1. These other sugars include, but are not limited to, melibiose and stachyose.

Likewise, prior to xenotransplantation, all or a substantial portion of total IgM (that is, IgM of all specificities) can be removed from the serum of the patient by extracorporeal immunoabsorption.

Alternatively, anti-GAL antibodies of all classes can be removed by extracorporeal immunoabsorption. Most 15 preferably, the patient's pre-formed natural anti-GAL IgM antibodies can be removed. In this way, many or most of the primary immunological agents of the hyperacute response are eliminated, resulting in reduction or elimination of the response following xenotransplantation.

The a-1,3-GalT Gene as a Target for Suppressing the GAL Epitope The present inventors have succeeded in cloning the entire coding region of the porcine a-1,3-GalT gene.

This is desirable for full exploitation of the gene in genetic engineering of pigs for purposes of human xenotransplantation. Previous attempts to obtain the entire coding region of the porcine gene have, to the knowledge of the inventors, failed to generate the coding regions. See, Dabkowski et al., Transplant.

Proc. 25: 2921 (1993). The present inventors have employed a PCR-based approach to generate the full sequence. In designing the primers and experimental conditions required to obtain the 5' and 3' regions of the gene, the present inventors overcame significant theoretical and practical obstacles to success.

Primers were selected on the basis of careful analysis of published sequences for the murine, bovine and human c-1,3-GalT genes, the only published sequences available for this purpose. The present inventors' analysis revealed that in the reported sequence of the bovine cDNA, exon 3 (which is in the region) is missing. This had not been reported in the literature. Thus, in order to find appropriate regions for deriving useful primer sequences, the mouse and :-bovine sequences had to be realigned. Even with the o .i appropriate realignment, however, only one island of about 20 base pairs (bp) in the 5' untranslated region 15 displayed the desired homology (19 out of 20 bp) for design of a PCR oligonucleotide. The fact that the untranslated regions of the mouse and bovine genes do not *"seem substantially related even upon optimal alignment would not be considered unusual by the ordinary skilled artisan. This is because the 5' untranslated regions are often not well conserved between species. As such, the natural inclination would be to perform a less-thanexhaustive analysis and to conclude that design of PCR oligonucleotides based on homology from this region was S" 25 unlikely to be successful.

In the downstream 3'-untranslated region, the homology is less than obvious again. Various insertions and deletions had to be made in order to obtain proper alignment of the mouse and bovine sequences. Moreover, to obtain a region of appropriate homology for design of PCR oligonucleotides, it was necessary to select a region approximately 200 bp downstream of the stop codon.

Finally, to get the 5' and 3' primers to work properly, the present inventors found it necessary to drop the annealing temperature by 9 0 C. These technical and theoretical hurdles to successful use of a PCR-based approach were overcome by the present inventors and allowed the entire coding sequence to be determined.

Analysis of the nucleotide sequence indicates that a counterpart to murine exon 3 in the 5' untranslated region is not found in the porcine gene. The porcine sequence is similar to the bovine sequence in this regard. Analysis of the amino acid sequence demonstrates that the structure of the porcine a-l,3-GalT is similar to that of other glycosyltransferases, and in particular is closely related to bovine and murine a-1,3-GalTs. In each of these enzymes a short cytoplasmic amino-terminal domain of about 6 residues precedes a hydrophobic membrane-anchoring domain (extending from residues 7 to o 15 22). The stem region, which serves as a flexible tether, and the catalytic domain, which catalyses the synthesis of a-1,3-GAL linkages, are located in the lumen of the Golgi and extend from amino acid 23 to the carboxyl terminus at amino acid 371. The precise boundary between 20 the stem and catalytic domains is not well-defined.

Based on the suggested characteristics of the stem region, it appears to be the least conserved region and is rich in glycine and proline residues. Paulson and Colley, J. Biol. Chem. 264: 17615 (1989); Joziasse et al., J. Biol. Chem. 267: 5534 (1992). The stem/catalytic boundary may occur around amino acid To generate constructs for inactivating genes by homologous recombination, the gene is preferably interrupted within an appropriate coding exon by insertion of an additional DNA fragment. Upon analysis of the full-length porcine nucleic acid sequence, the present inventors have identified exons 4, 7, 8 and 9 as preferred locations for disruption of the gene by homologous recombination. However, identification of these exons as preferred sites should not be construed as limiting the scope of the present invention, as interruptions in exons 5 and 6 may be useful in particular cell types or in situations where less-thancomplete inhibition of a-l,3-GalT gene expression is desired. Moreover, regulatory elements associated with the coding sequence may also present useful targets for inactivation.

In a preferred embodiment, a Sall site located within exon 9 of the mouse a-1,3-GalT gene at codons 221- 222 is chosen as the site for disruption of the murine coding sequence. For disruption of the porcine sequence, it is noted that the amino acids encoded by the corresponding porcine nucleotides are conserved, although the Sall site is not. In a preferred embodiment for 15 inactivation of the porcine gene, a Sall site is engineered into the corresponding location of the pig sequence for convenient construction of a knockout sequence. Sall cuts only rarely in genomic DNA. Since multiple restriction sites can be a problem in manipulating large fragments of DNA, the presence of a Sall site in the exon is very useful since it is not likely that other Sall sites will be present at other locations in the knockout constructs.

A gene coding for a selectable marker is generally used to interrupt the targeted exon site by homologous recombination. Preferably, the selectable marker is flanked by sequences homologous to the sequences flanking the desired insertion site. Thomas and Capecchi, Cell 51: 503-12 (1987); Capecchi, Trends in Genetics 5: 70-76 (1989). It is not necessary for the flanking sequences to be immediately adjacent to the desired insertion site.

The gene imparting resistance to the antibiotic G418 (a neomycin derivative) frequently is used, although other antibiotic resistance markers hygromycin) also may be employed. Other selection systems include negativeselection markers such as the thymidine kinase (TK) gene from herpes simplex. Any selectable marker suitable for inclusion in a knockout vector is within the scope of the present invention.

However, it is possible that in some circumstances it will not be desirable to have an expressed antibiotic resistance gene incorporated into the cells of a transplanted organ. Therefore, in a preferred embodiment, one or more genetic elements are included in the knockout construct that permit the antibiotic resistance gene to be excised once the construct has undergone homologous recombination with the a-1,3-GalT gene.

The FLP/FRT recombinase system from yeast 15 represents one such set of genetic elements. o'Gorman et al., Science 251, 1351-1355 (1991). FLP recombinase is a protein of approximately 45 kD molecular weight. It is encoded by the FLP gene of the 2 micron plasmid of the yeast Saccharomyces cerevisiae. The protein acts by binding to the FLP Recombinase Target site, or FRT; the core region of the FRT is a DNA sequence of approximately 34 bp. FLP can mediate several kinds of recombination reactions including excision, insertion and inversion, depending on the relative orientations of flanking FRT sites. If a region of DNA is flanked by direct repeats of the FRT, FLP will act to excise the intervening

DNA,

leaving only a single FRT. FLP has been shown to function in a wide range of systems, including in the cultured mammalian cell lines CV-1 and F9, O'Gorman et al., Science 251: 1351 (1991), and in mouse ES cells, Jung et al., Science 259: 984 (1993).

Targeted cells carrying a genomic copy of an antibiotic resistance gene flanked by direct repeats of the FRT are supplied with FLP recominase by 1) introduction into cells of partially purified FLP protein by electroporation, or 2) transfection with expression plasmids containing the FLP gene. In this way, the antibiotic resistance gene is deleted by action of the FLP recombinase, and cells are generated that contain the inactivated a-l,3-GalT gene and are free of the exogenous antibiotic resistance gene.

Due to the relative infrequency of homologous recombination in targeted cells, most such cells will carry only one inactivated allele of the target gene.

That is, the great majority of cells taken through a single round of transformation with an appropriate knockout construct will be heterozygotes. As used herein, the term "transformed" is defined as introduction of exogenous DNA into the target cell by any means known to the skilled artisan. These methods of introduction can include, without limitation, transfection, microinjection, infection (with, for example, retroviralbased vectors), electroporation and microballistics. The term "transformed," unless otherwise indicated, is not intended herein to indicate alterations in cell behavior and growth patterns accompanying immortalization, density-independent growth, malignant transformation or similar acquired states in culture.

Although heterozygous cells can be used in the methods of the present invention, various manipulations can be employed to generate homozygous cells in culture.

For example, homozygous cells can be generated by performing a second homologous recombination procedure on cells heterozygous for the inactivated allele. If the knockout construct used in the initial transformation carried the neoR gene, a second construct may be employed in a second round of transformation in which the neoR gene is replaced with a gene conferring resistance to a separate antibiotic hygromycin). Cells resistant to both G418 and hygromycin can be screened by Southern blots in order to detect any "double knockouts" homozygotes). Both antibiotic resistance genes can be removed subsequently in a single procedure using FLP recombinase. By maintaining selection with G418, this approach ensures that the second construct does not simply replace the previously knocked-out allele, leaving the cells heterozygous.

Alternatively, the neoR gene can be deleted from an original heterozygous cell using FLP recombinase and a second knockout procedure conducted using the original neoR gene-containing construct. Double knockouts could be detected by Southern analysis. The newly introduced neo R gene then could be deleted by FLP recombinase. This alternative approach does not allow one to direct the knockout construct specifically to the non-inactivated allele. Nevertheless, screening of appropriate numbers of targeted cells can lead to identification of cells *"homozygous for the inactivated locus.

Cellular Vehicles for Incorporation of Knockout Constructs 20 To create animals having a particular gene inactivated in all cells, it is necessary to introduce a knockout construct into the germ cells (sperm or eggs, ooo*e the "germ line") of the desired species. Genes or p. p pother DNA sequences can be introduced into the pronuclei of fertilized eggs by microinjection. Following pronuclear fusion, the developing embryo may carry the introduced gene in all its somatic and germ cells since the zygote is the mitotic progenitor of all cells in the embryo. Since targeted insertion of a knockout construct is a relatively rare event, it is desirable to generate and screen a large number of animals when employing such an approach. Because of this, it can be advantageous to work with the large cell populations and selection criteria that are characteristic of cultured cell systems. However, for production of knockout animals from an initial population of cultured cells, it is necessary that a cultured cell containing the desired knockout construct be capable of generating a whole animal. This is generally accomplished by placing the cell into a developing embryo environment of some sort.

Cells capable of giving rise to at least several differentiated cell types are hereinafter termed "pluripotent" cells. Pluripotent cells capable of giving rise to all cell types of an embryo, including germ cells, are hereinafter termed "totipotent" cells.

Totipotent murine cell lines (embryonic stem, or "ES" cells) have been isolated by culture of cells derived from very young embryos (blastocysts). Such cells are ee" 15 capable, upon incorporation into an embryo, of differentiating into all cell types, including germ cells, and can be employed to generate animals lacking a functional a-1,3-GalT gene. That is, cultured ES cells can be transformed with a knockout construct and cells selected in which the a-1,3-GalT gene is inactivated through insertion of the construct within, for example, an appropriate exon. In fact, ES cell lines have been derived for both mice and pigs. See, Robertson, Embryo-Derived Stem Cell Lines. In: Teratocarcinomas and Embryonic Stem Cells: A Practical Approach (E.J.

Robertson, IRL Press, Oxford (1987); PCT Publication No. WO/90/03432; PCT Publication No.

94/26884. Generally these cells lines must be propagated in a medium containing a differentiation-inhibiting factor (DIF) to prevent spontaneous differentiation and loss of mitotic capability. Leukemia Inhibitory Factor (LIF) is particularly useful as a DIF. Other DIF's useful for prevention of ES cell differentiation include, without limitation, Oncostatin M (Gearing and Bruce, The New Biologist 4: 61-65 (1992); personal communication

M

from A. Smith), interleukin 6 (IL-6) with soluble IL-6 receptor (sIL-6R) (Taga et al., Cell 58: 573-81 (1989); personal communication from A. Smith), and ciliary neurotropic factor (CNTF) (Conover et al., Development 19: 559-65 (1993). Other known cytokines may also function as appropriate DIF's, alone or in combination with other DIF's.

As a useful advance in maintenance of ES cells in an undifferentiated state, the present inventors have identified a novel variant of LIF. In contrast to the previously identified forms of LIF which are extracellular, this new form of LIF (hereinafter

T-LIF)

i" s intracellularly localized. The transcript was cloned from murine ES cells using the RACE technique, Frohman et i 15 al., Proc. Natl. Acad. Sci. USA 85: 8998-9002 (1988), and subjected to sequence analysis. Analysis of the obtained nucleic acid sequence and deduced amino acid sequence indicates that T-LIF is a truncated form of the LIF sequence previously reported in the literature.

Expression of the T-LIF nucleic acid in an appropriate host cell yields a 17 kD protein that is unglycosylated.

S. This protein is useful for inhibiting differentiation of murine ES cells in culture. The protein is expected to have a similar activity with porcine cells, since murine D-LIF is effective at inhibiting both murine and porcine ES cell differentiation. The present inventors have also determined the sequence of the human form of T-LIF.

To generate a knockout animal, ES cells having at least one inactivated a-1,3-GalT allele are identified and incorporated into a developing embryo. This can be accomplished through injection into the blastocyst cavity of a murine blastocyst-stage embryo, by injection into a morula-stage embryo, by co-culture of ES cells with a morula-stage embryo, or through fusion of the ES cell with an enucleated zygote. The resulting embryo is raised to sexual maturity and bred in order to obtain animals, all of whose cells (including germ cells) carry the inactivated a-1,3-GalT allele. If the original ES cell was heterozygous for the inactivated a-l,3-GalT allele, several of these animals must be bred with each other in order to generate animals homozygous for the inactivated allele.

Although direct microinjection of DNA into eggs does not generate the large numbers of recombination events obtained through transfecting large numbers of cultured cells, nevertheless direct injection of eggs can be a useful approach since this avoids the special manipulations (see above) required to turn a cultured cell into an animal. This is because fertilized eggs 15 are, of course, quintessentially "totipotent" i.e., *000 capable of developing into an adult without further substantive manipulation other than implantation into a surrogate mother. To enhance the probability of S* homologous recombination when eggs are directly injected with knockout constructs, it is useful to incorporate at least about 8 kb of homologous DNA into the targeting construct. In addition, it is also useful to prepare the knockout constructs from isogenic DNA. For example, for injection of porcine eggs, it is useful to prepare the constructs from DNA isolated from the boar whose sperm are employed to fertilize the eggs used for injection.

Embryos derived from microinjected eggs can be screened for homologous recombination events in several ways. For example, if the GalT gene is interrupted by a coding region that produces a detectable fluorescent) gene product, then the injected eggs are cultured to the blastocyst stage and analyzed for presence of the indicator polypeptide. Embryos with fluorescing cells, for example, are then implanted into a surrogate mother and allowed to develop to term.

Alternatively, injected eggs are allowed to develop and the resulting piglets analyzed by polymerase chain reaction (PCR) or reverse transcription PCR (RT/PCR) for evidence of homologous recombination.

Characterization of Knockout Animals Animals having either one (heterozygous) or two (homozygous) inactivated GalT genes are characterized to confirm the expected alterations in gene expression and phenotypic effect. For example, GalT mRNA should be absent from homozygous knockout animals. This can be confirmed, for example, with reverse transcription

PCR

(RT-PCR) using appropriate GalT-specific primers. In addition, various tests can be performed to evaluate 15 expression of the GAL epitope in homozygous knockout animals. For example, anti-GAL antibodies and IB 4 Lectin (which has an exclusive affinity for terminal a-Dgalactosyl residues) can be used in various assay or immunohistological formats to test for the presence of the GAL epitope in an array of tissues. As another indication of GAL epitope status, lysis of cells by human serum can be tested through use of a 51 chromium release assay.

.EXAMPLE 1 S 25 Affinity Purification of Human Anti-GAL Antibodies Anti-GAL antibodies were purified from normal heat inactivated AB serum (from CS1, Parkville, Victoria, Australia) using the following sets of procedures.

A. Preparation of total anti-GAL (IgG+IqM) antibodies The following procedures are performed at 4'C.

1. Desalt 15-30ml serum (in 3ml batches) by passage through a pre-equilibrated (20ml application buffer:

K

2

HPO

4 30mM NaC1, pH 8) Econo Pac 10DG (Bio-Rad, Richmond, USA) column. Alternatively, for large scale preparations, desalt by dialysis exhaustively against application buffer.

2. Wash column with 4ml aliquots of application buffer. Collect and pool column eluates.

3. Apply pooled desalted serum to a pre-equilibrated application buffer) Synsorb 115 (galactosylgalactose; Chembiomed, Alberta, Canada) or Melibiose-Agarose (Sigma) affinity column (5ml-50 ml depending on the yield required).

4. Collect run-through (partially anti-GAL-depleted) and reapply to column. Repeat process 3 times to ensure complete removal of anti-GAL antibodies. The washthrough from the 3rd passage through the Synsorb column is collected and the volume adjusted to the original volume of the serum with phosphate-buffered saline

(PBS)

r: pH 7 +0.05% azide. This is used as a source of anti-GAL antibody-depleted serum.

5. Wash column with PBS pH 8 until the eluate is protein free 280nm=0).

6. Elute anti-GAL antibodies with 3.5M KSCN, pH Collect 4ml fractions, determine the O.D. 280 and pool peak fractions (usually 1-6).

7. Concentrate anti-GAL antibodies using ultrafiltration cones (Amicon, Danvers, USA). Add 7ml of S 25 the pooled fractions containing anti-GAL antibodies to spin cone and centrifuge (2,000 RPM, 10min, Refill cone and recentrifuge until volume is reduced to 8. To dilute the KSCN, adjust vol. to 7ml with PBS and centrifuge (2,000 RPM, 10min, Repeat process a further 10 times.

9. Remove sample containing anti-GAL antibodies from cone using plastic pipette; rinse cone with PBS pH7 +0.05% azide.

B. Preparation of IqG anti-GAL antibodies The following procedures are performed at 4'C.

1. Desalt 15-30 ml serum (in 3ml batches) by passage through a pre-equilibrated (20ml application buffer) Econo Pac 10DG (Bio-Rad, Richmond, USA) column.

Alternatively for large scale preparations desalt by dialysis exhaustively against application buffer.

2. Wash column with 4ml aliquots of application buffer. Collect and pool column eluates.

3. Apply desalted serum to a pre-equilibrated application buffer) Affi-Blue column (Bio-Rad, Richmond, USA) (Affi-Blue binds all proteins except albumin and IgG).

4. Wash column with 20ml application buffer to elute 15 IgG enriched fraction.

5. Apply IgG enriched fraction to a pre-equilibrated application buffer, pH 8.0) Synsorb 115 (galactosyl-galactose; Chembiomed, Alberta, Canada) affinity column 6. Collect run-through and reapply to column. Repeat process 3 times to ensure complete removal of anti-GAL antibodies. The wash-through from the 3rd passage through the Synsorb column is collected and the volume adjusted to the original volume of the serum with PBS pH 25 7 +0.05% azide. This is used as a source of control anti-GAL-depleted IgG.

In some cases anti-GAL IgG was further purified using a protein G column, which efficiently binds IgG but not other antibody isotypes. IgG was then eluted from the protein G column using glycine pH 2.4.

All anti-GAL antibody preparations were analyzed for the following: a. Protein content was determined using the Bradford colorimetric method (Bradford, M.M 1976, Anal. Biochem. 72:248-254), using purified human IgG as the standard.

b. Molecular weight and purity were determined using polyacrylamide gel electrophoresis according to method described by Laemli, Nature (London) 227: 680 (1970), and protein was detected in the gels by silver staining using standard kit reagents (Amersham,

UK).

c. Antibody class and isotype were determined by radial 15 immunodiffusion using standard techniques as set out in Rose et al. Manual of Clinical Laboratory Immunology, American Society for Microbiology, 20 Washington, D.C. IgG anti-GAL preparations were found to contain all subclasses, with IgG2 predominating.

EXAMPLE 2 Reactivity of IqG and IqM Anti-GAL Antibodies and Depleted Serum with Porcine Cells and Tissues I. CELLS Reactivity of IgG and IgM anti-GAL antibodies was assessed using either porcine aortic endothelial cells (prepared by the inventors as described below) or porcine epithelial cell line LLC PK 1

(PK

1 obtained from the American Type Culture Collection (ATCC), Accession No.

CRL1392.

A. Isolation and culture of porcine aortic endothelial cells (PAE's) Pigs were blood typed (using human typing reagents) to identify "O-type" pigs, i.e, pigs unreactive with antibodies to A or B human red blood cell antigens.

Aortas were excised from "O-type" pigs, then transported from the abattoir to the laboratory on ice. PAE's were isolated by collagenase treatment as described by Gimbrone et al., J. Cell Biol. 60: 673-84 (1974). PAE's were cultured in RPMI medium containing 10% fetal calf serum (FCS), supplemented with 150Ag/ml endothelial cell supplement (Sigma) and 50Ag/ml heparin (Sigma). The 15 cells were identified as endothelial cells by their typical cobblestone morphology and by their immunoreactivity with Factor VIII antibodies, as identified using immunofluorescence. In all the assays described below, the PAE's were used between the 8th and 12th passages.

o B.Tissue Culture: Maintenance of PK-1 and PAE cell lines All tissue culture was performed in a laminar flow hood, using appropriate tissue culture sterile technique.

All tissue culture reagents, unless otherwise indicated, were purchased from CSL, Melbourne, Australia. Media were constituted as follows: PK-1 Culture Medium: DMEM (Cytosystems, Castle Hill, Australia) 500ml FCS (CSL, Melbourne, Australia) 3 Glutamine (200mM) (Cytosystems) Hepes (IM) (CSL) 7.5ml Penicillin (CSL) 0.5ml (10 5 U/ml final) Streptomycin (CSL) 0.5ml (10 g/ml final) PAE Culture Medium: RPMI (CSL) FCS (CSL) Endothelial cell supplement (3mg/ml) (Sigma) Heparin (10mg/ml) (CSL) Endothelial cell supplement was purchased from Sigma Chem. Co. (St. Louis, Missouri, USA) as a lyophilized powder, resuspended in sterile HBBS, and 3ml aliquots stored at 4'C.

Heparin (Sigma, Missouri, USA) dissolved in PBS (10mg/ml) filter sterilized 1 (0.22um) Hanks Buffer purchased from Cytosystems The following general procedures were used in propagating the cell lines.

1) Pour off old medium 20 2) Rinse cells twice with sterile PBS 3) Add 3ml of TED (0.05 M trypsin, 0.53 M EDTA, Gibco,

NY,USA)

4) Incubate 10 min. in CO 2 incubator at 37 0

C

Add 7ml RPMI with 10% FCS 25 6) Resuspend cells and transfer to a sterile tube 7) Centrifuge for 5min at 1200 rpm, discard supernatant 8) Resuspend cells in RPMI with 10% Newborn Bovine Serum (NBS) and repeat centrifugation 9) Resuspend cells in 1ml DMEM (PK-1's) or RPMI (PAE's) (with additives, as described above).

Add 10ml medium and the appropriate volume of cell suspension to achieve the desired dilution for each 75cm 2 tissue culture flask, and return to humidified

CO

2 incubator.

C. Antibody staining and FACS analysis 1) Add 2ml TED to a 75cm 2 culture flask containing PK-1 or PAE's, and incubate at room temperature for min.

2) Add RPMI plus 10% FCS (5ml) to neutralize trypsin.

3) Pellet cells by centrifugation (700g, 5 min, 4 C).

4) Wash cells by resuspension and centrifugation in Hanks Buffer (x2).

5) Pellet cells by centrifugation (700g, 5 min, 4'C).

6) Resuspend cell pellet in Hanks buffer containing purified anti-GAL antibodies, GAL-depleted serum or GAL-depleted IgG and incubate 20 at 4'C for 60 min. All

C..

antibodies were used undiluted, or diluted 1:2 or 1:4 in Hanks buffer.

7) Add Iml of Hanks Buffer, pellet cells by centrifugation and aspirate off supernatant.

8) Resuspend pellet in FITC-labelled sheep-antihuman IgG Fab2 or IgM Fab2 (Silenus, Hawthorn, Australia) diluted 1:80 in Hanks buffer.

9) Incubate for 30 min. at 4'C.

Wash three times with Hanks buffer; resuspend pellet from final wash in Hanks buffer.

11) Analyze stained samples using a FACScan II (Becton Dickinson) according to the manufacturer's instructions.

The specificity of the anti-GAL antibody binding to porcine cells was determined by examining the ability of sugars of various structures to inhibit antibody binding. In these competition studies the anti-GAL antibodies were pre-incubated with sugar (0.1M) at 37'C for 30 min before adding to the cells.

D. Results Using immunofluorescence it was found that total anti-GAL (IgM IgG) and purified anti-GAL IgG stained both PK- 1 and PAE's cells. On the other hand, neither the total anti-GAL antibody-depleted serum nor the anti- GAL IgG-depleted serum gave detectable staining over background. The staining with anti-GAL IgM and/or IgG was inhibited with purified galactose and with disaccharides having terminal galactose residues in the al-configuration such as melibiose (6-O-a-Dgalactopyranosyl- D-glucose) and stachyose (a-D-Gal-[l- 6 Staining was not inhibited with sugars such as lactose 4 -O--D-galatopyranosyl-a-Dglucose), which has a terminal galactose residue, but in a P1->4 configuration. The results of one such experiment are represented in Figure 1. PAE's were stained with anti-GAL antibody alone (GAL:PBS) or with anti-GAL antibody that had been pre-incubated with either melibiose (GAL:MELIBIOSE), galactose (GAL:GALACTOSE) or lactose (GAL:LACTOSE). Anti-GAL antibody staining was approximately 10 fold less in the samples containing melibiose and galactose, but was not affected significantly by lactose.

II. TISSUES A. Methods Pig kidney was fixed in formalin and dehydrated before embedding in Paraplast. Pig heart and liver were fixed in paraformaldehyde-lysine-periodate fixative and snap frozen in O.C.T. embedding compound (10.24% w/w polyvinyl alcohol, 4.26% w/w polyethylene glycol, 85.50% w/w nonreactive ingredients; Tissue Tek®, Miles, Inc., Elkhart, Indiana, USA). Four pm-thick sections of pig heart and liver and 2 pm-thick sections of kidney were incubated with purified anti-GAL antibodies (undiluted, 1:2 and 1:4) for 60 min. and then incubated with a **fluorescein isothiocyanate (FITC)-conjugated sheep antihuman immunoglobulin F(ab') fragment (Silenus 15 Laboratories, Hawthorn, Australia) (1:100) for 30 min. or a peroxidase-conjugated rabbit anti-human IgG (Dakopatts, Glostrup, Denmark) (1:50) for 60 min. Control sections were analyzed for autofluorescence, with the secondary antibody alone, or with the anti-GAL-depleted IgG or normal pig serum as the primary antibody. No staining was detected. The specificity of the anti-GAL antibodies was tested by pre-incubating sections of pig renal cortex with a variety of sugars, including melibiose, lactose, Ssucrose and glucose at 0.1M.

B. Results As with the analyses performed on the pig cells using immunofluorescence, total anti-GAL IgM IgG, purified anti-GAL IgG, but not the anti-GAL IgM and/or IgG-depleted sera, stained all pig tissues examined. The individual staining parameters varied from organ to organ as set out below: 1111111~111111 Immunostaining of Pig Tissues with Anti-GAL Antibodies: Tissue Anti-GAL Reactivity Staining Intensity Kidney Proximal and distal convoluted tubules Variable Endothelium: Intertubular sinusoids Variable Endothelium: Arteries and veins Strong Glomerular capillaries Variable Heart Endothelium: Arteries, veins, capillaries Strong Endocardium Strong Myocardium Perinuclear Liver Small Bile Ducts (lining cells) Strong Endothelium: Arteries, veins Strong Intertubular sinusoids Negative The specificity of the binding of anti-GAL antibodies was tested on sections of pig renal cortex by 15 inhibition with 0.1 M melibiose, lactose, sucrose and glucose. Reactivity of the anti-GAL antibodies with proximal tubule brush borders was reduced to near background by preincubation of antibody with melibiose, but was not inhibited by the other saccharides.

S 20 EXAMPLE 3 Hemaqqlutination of Pig RBC by Human Serum: Sugar Inhibition Studies The methods used to investigate the hemagglutination of pig red blood cells (RBC's) by human serum was adapted from the methods described by Galili, J.Exp. Med. 160: 1579-81 (1984) and Severson, Immunol.

96: 785-789 (1966).

I. METHODS A. Media/Solution Preparation 1. Human Serum Albumin (HSA) (CSL, Melbourne, Australia) (5mg/ml) was dissolved in PBS, filter sterilized, and stored at 4'C.

2. Preparation of sugars: 1M stock solutions of sugar were prepared by dissolving the amount indicated in 100ml of PBS. Sodium azide was added and solutions stored at 4'C.

a-Lactose (4-0-3-D-galactopyranosyl-a-D-glucose 3 6.Og D(+)galactose 18.0g Stachyose 66.6g Melibiose 6 -O-a-D-galactopyranosyl- D-glucose) 34.2 g "Sucrose (a-D-Glucopyranosyl P-D-fructofuranoside) 34.2 g D-(+)-Glucose 18.0 g a-D-(+)-Fucose 6 -Deoxy-D-galactopyranose) 16.4 g All sugars were purchased from Sigma (St. Louis, Missouri, USA). Sugar solutions were diluted in PBS to the appropriate concentration as required.

B. Preparation of pig RBC'S 1. Heparinised pig blood (Animal Resources, Clayton, 25 Australia) is centrifuged at 800 RPM for 10min to pellet the RBC.

2. The RBC pellet is washed by resuspension in PBS (10ml) and recentrifugation (repeated 3 times). After the final wash, the RBC pellet is resuspended in

PBS.

3. A 0.5% solution of RBC's is prepared by adding 50ul RBC solution (from step 2, above) to ml PBS containing 0.5g/100 ml of HSA.

C. Preparation of 96-well microtitre plates (Titretek, USA) 1. Add 25ul of PBS to each well.

2. Add 25ul of pooled human AB serum (CSL, Melbourne, Australia) to column 1 and serially dilute by removing from column 1 and adding to column 2, then repeating by sequentially removing and adding from and to each well 15 across the plate, finally discarding 25ul from column 11 and adding no serum to column 12.

3. Add 25ul of sugar solution S 20 to each row in decreasing concentrations down rows. No sugar solution is added to the final row.

4. Incubate at 4'C overnight and then at 37'C for 30 min.

Add 50ul of 0.5% pig RBC to each well; vortex and incubate at room temperature for 2 hours.

Determine agglutination visually.

II. RESULTS Human serum caused the agglutination of pig RBC' s at a titre of between 1/32-1/64, which is consistent with the presence of high levels of naturally occurring 41 xenoantibody (NXAb) in human serum. To examine the specificity of the NXAb response, sugar inhibition studies were performed. Sugars such as melibiose, stachyose, galactose and fucose which have terminal galactose residues in the rl-6 configuration were found to inhibit agglutination in the gM to mM range. Sugars with other structures, such as lactose and sucrose, were only inhibitory when very high concentrations were used.

At these high concentrations, the observed effects are most probably non-specific, due, for example, to changes in osmolarity. Results are summarized below: Pig RBC Hemagglutination by Human Serum: Sugar Inhibition Sugar Linkage Inhibitory Concentration 1 Melibiose Gal al-6Glc 5x10- 4

M

Stachyose Gal al-6Gal 2x10-3M Galactose 2x10-M Fucose 6-Deoxy-a-L-Gal lxl0 3

M

Lactose Galpl-4-Glc 10-1M Sucrose a-D-Glc-P-D-Fruc >10-1M 20 EXAMPLE 4 Inhibition of Human Serum-Induced Lysis of Porcine Cells by Sugars The ability of human serum to cause the lysis of porcine cells was examined using both pig epithelial

(PK

1 and aortic endothelial (PAE's) cells, the isolation and culture of which is described in Example 2. Cell lysis was determined using either the 51 Chromium release assay as described by Cerottini and Brunner, Nature New Biol. 237:272, 1972 or the Cytotox LDH release assay according to the manufacturer's instructions (Promega,

USA).

I. METHODS A. 5 1 CR Release Assay 1. Preparation of Cells: a) Trypsinize a confluent flask of cells.

On average, approximately 3 x 106 PAE's and approximately 3 x 107 PK cells are obtained per 10 ml flask. About 1 x 105 cells are required for each well in the 51 CR Release Assay.

b) Wash cells 4 times in 10 ml RPMI (no FCS); spin 1200 rpm for 5 min.

c) Resuspend cells in 100 Al RPMI (with heat-inactivated FCS; see below).

2. Labelling Cells with 51 CR: a) Combine in a 10 ml tube: Cells in 195 il RPMI/10% FCS (heat inactivated); 5 pl 51 CR (120 ACi).

b) Incubate at 370C for 2 hr.

c) Add 2 ml RPMI/10% FCS (heat inactivated).

d) Centrifuge cells through a layer of FCS (heat inactivated) to remove excess label.

e) Gently overlay the labelled cells onto a 4 ml cushion of FCS using a Pasteur pipette.

20 f) Centrifuge at 700g for 5 min. at 4 0

C.

g) Remove supernatant taking care not to disturb the cell pellet.

h) Resuspend pellet in RPMI/10% FCS (heat inactivated) at about 3 x 10 7 cells/ml.

3. Assay Conditions: a) For PAE' s, rabbit complement was used as the complement source, since the 51 CR-release assay was not sufficiently sensitive to detect lysis when human complement, a less "active" source, was used. In contrast, with the LDH assay, which is significantly more sensitive, normal human serum (NHS) was used as the source of complement.

b) To each test well of a 96-well V bottom plate, add: 100 gl labelled cells 43 of final) 10-50 gl NHS (heat inactivated) (5-25% of final) Complement: PAE's: 50 Al absorbed rabbit complement (25% final) final) PK 1 10-40 gl NHS (5-25% of 50 ~l antibody (total anti-GAL (IgG IgM, anti-GAL IgG, anti-GAL antibody-depleted serum, or anti-GAL antibody-depleted IgG) c) Adjust volume to 200 pl with FCS (heat inactivated) if required d) Incubate plates at 37 0 C for 3 hr.

e) Centrifuge plates at 1000 rpm for 15 min to pellet cells *0 f) Remove 100 il of supernatant from each well and transfer to a gamma counter tube g) Add 3 ml scintillation fluid and measure 51 CR release using a gamma counter (Packard 20 Instrument Company, Illinois,

USA)

(To determine maximum release, add 100 AL 8% Triton X-100 made up in RPMI/10% FCS (heat inactivated) to 100 gl labelled cells) (Note: Each reaction is set up in quadruplicate) 4. Calculation of Lysis: Lysis Experimental cpm Spontaneous Release c x 100 Max. Release cpm Spontaneous Release cpm Sugar Inhibition of Complement-Induced Cell Cytotoxicity: In a 96-well test plate, mix the following: 50 1l labelled cells 50 pl complement (PAE's: pig spleen cell absorbed complement;

PK

1 s: NHS) 10-1 to 10- M)x Al sugar (final concentration of sugar: y gl NHS (heat inactivated) final concentration make volume to 200 41 with RPMI 44 Plate Layout: Plate 1 Plate 2 10% 15% Rows: 1-4 5-8 1-4 5-8 Columns: I. Spontaneous Release 2. Maximum Release 3. Melibiose 4. Lactose B. LDH Release Assay 1. General Procedures: a) Prepare cells as for 51 CR Release "assay, and labeled with LDH as per the manufacturer's S. instructions (Cytotox non-radioactive LDH release assay, Promega, USA) S 15 b) To each well of a 96-well plate add (each reaction set up in quadruplicate): 25 il labeled cells 5-20 il NHS -20 x0l 1 sugar (final concentration of sugar: 10-1 to 10 3

M)

RPMI/10% FCS (heat inactivated), to total volume of 100 Al c) Incubate plates at 37 0 C for 3 hr.

d) Centrifuge plates at 1500 rpm for 5 min.

25 e) Remove 50 L1 supernatant from each well (taking care not to remove any cells) and transfer to ELISA plate containing 50 Al substrate mix (prepared according to manufacturer's instructions f) Cover tray and incubate in the dark at room temperature for 30 min.

g) Add 50 pl stop solution to each well using multichannel pipette h) Read absorbance at 492 nm.

2. Controls: complement) a) Spontaneous release (no antibody or 25 gl labeled cells 75 p1 RPMI/10% FCS (heat inactivated) b) Maximum release 25 il labeled cells 50 ~l 16% Triton X-100 25 pl RPMI/10% FCS (heat inactivated) 3. Calculation of Lysis: Lysis 10 Experimental release -(Spontaneous release cpm sugar cpm) x 100 Maximum release (Spontaneous release cpm sugar cpm) a

S

4. Experimental Design: Plate 1 15 columns: l.spontaneous release 2. maximum release 3. 5% serum 4. 10% serum 5. 25% serum 20 6. RF10 alone Rows: 1-4: cells no sugar 5-8: no cells no sugar Plate 2 Plate 3 Plate 4 Plate 5 Plates 6-9 Columns: 1-2 3-4 5-6 7-8 9-10 11-12 melibiose galactose lactose sucrose same as plates 2-5 but no cells added Sugar Conc.

1 x 10- 1

M

5 x 10- 2

M

1 x 10- 2

M

5 x 10- 3

M

2 x 10- 3

M

1 x 10 3

M

Rows: 1-2 3-4 5-6 7-8 0% serum 5% serum 10% serum 25% serum Preparation of Pig Spleen-Absorbed Rabbit Complement: a) Cut pig spleen (obtained from local abattoir) into small pieces and prepare a single-cell suspension by passage through a fine metal sieve b) Pellet cells by centrifugation at 700g, 7 min.

at 4 0

C

c) Resuspend cell pellet in RPMI/10% FCS and repeat centrifugation d) Resuspend in RPMI/10% dimethylsulfoxide

(DMSO)

e) Count cells and store frozen aliquots (3 x 109 cells/aliquot) use one aliquot for each absorption f) For absorption, thaw and centrifuge at 600g, 15 min. at 4 0 C and remove the supernatant containing the

DMSO

g) Wash two times with RPMI/10% FCS (10 ml) h) Resuspend the cell pellet in rabbit complement; mix (rotary wheel) 2 hr. at 4 0

C

20 i) Centrifuge 600g, 5 min. at 4 0 C and remove the supernatant containing the rabbit complement; store at 4 0

C

II. RESULTS Comparable results were obtained with both cell types (PAE's and PKl's) using both lysis assays. The results of a typical lysis experiment are represented in Figure 2, in which the lysis of PAE's by human serum and by purified anti-GAL antibodies was determined using the release assay. Comparable results were also obtained with PK 1 cells using the 51 CR release assay and with both cell lines using the LDH release assay. The results of these assays can be summarized as follows: 1. Xenoantibodies (NXAb) in human serum in the presence of complement are capable of lysing porcine cells. Lysis increases with increasing concentrations of serum.

2. Pre-absorption of NHS with pig spleen cells (which removes the NXAb): No lysis.

3. Use of heat-inactivated complement: No lysis.

4. Use of NHS depleted of anti-GAL antibodies: No lysis.

Use of purified total anti-GAL antibodies (IgG IgM): Lysis.

6. Use of purified anti-GAL IgG: No lysis.

7. Use of purified total anti-GAL antibodies S" (IgG IgM) and dithiothreitol (DTT): No lysis. (DTT is a reducing agent that disrupts the multimeric structure S 15 of IgM antibodies without affecting IgG.) Together these results demonstrate that the anti- GAL antibodies are responsible for the observed lysis.

Purified anti-GAL IgG and DTT-treated total (IgG IgM) anti-GAL antibodies failed to elicit lysis, indicating that IgM, but not IgG, antibodies are causative agents in this system. Preliminary attempts to verify this observation using purified IgM prepared either in crude form by euglobulin fractionation or by a-IgM affinity chromatography were unsuccessful. The inventors believe this reflects inactivation of the IgM during preparation, rather than a true reflection of the capacity of anti-GAL IgM to cause lysis of porcine cells, heat inactivation of the complement prevented lysis, indicating that lysis of porcine cells is a complement-dependent phenomenon.

The effect of adding the disaccharide sugars melibiose (Gal al- 6 Gal) and lactose (Gal fl- 4 Glu) on the lysis of PAE's by human serum was assessed using the Cytotox non-radioactive LDH release assay. PAE's were incubated in the presence of 50% human serum as the

S.

S

S.

*5 S

S.

*SS*

S

S S *5 9* 5@

S.

source of xenoantibody and complement, together with various concentrations of each sugar (1mM to 100mM).

Under these conditions, melibiose, which has the Gal al- 6 Gal configuration, but not lactose, which has the terminal Gal moiety by in a pl- 4 configuration, protected the pig cells from lysis.

EXAMPLE Inhibition of Human Serum-Induced Damage to Rat Hearts by Sugars The Langendorf isolated perfused ex vivo heart model was used to further demonstrate the involvement of anti-GAL xenoantibodies in hyperacute rejection.

I. METHODS A. Preparation and storage of Human Plasma 1. Centrifuge fresh human blood at 3000 rpm, 10 min., 4 0 C to remove red blood cells (RBC's) 2. Remove the plasma 3. Centrifuge the plasma at 10,000 rpm, min. 4 0 C to remove any remaining cells; decant the plasma 4. Add 2.5 ml of 0.1M EDTA pH 7.30 for every 50 ml of plasma 5. Store 50.ml aliquots at -70 0

C

6. For heat-inactivated plasma, heat at 56°C for 60 min., then centrifuge at 2,500 rpm for min.

B. Assessment of Complement Activity Before being used in the ex vivo model, both heat inactivated and control plasma was tested for complement activity. Classical complement activity was determined by hemolysis using sensitized sheep RBC' s as described by Harrison and Lachman, In: Weir et al. Handbook of Experimental Immunology and Immunochemistry, 4th Ed., Blackwell scientific Publications (1986). Alternative complement pathway activity was determined using the rabbit hemolytic assay as described by Serrais et al., J.

Immunol. Meth. 140: 93-100 (1991). The assay was performed in buffer containing EGTA and MgCl 2 The EGTA chelates the Ca thus inhibiting the classical pathway.

The Mg is required for activation and assembly of CdbBb, the alternative pathway C3 convertase.

C. Preparation of Plasma for Heart Perfusions Plasma prepared from different blood packs is thawed at 37 0 C, pooled and filtered (100 pm steel mesh, 8.0 pm and 4.5 pm Millipore filters, sequentially).

CaCl2 is added at 0.58 mg/ml plasma, and the plasma kept on ice until ready for perfusion.

D.

D. Ex Vivo Isolated Perfused Rodent Heart Model 1. Anesthetize rats with Nembutal (1 1l sodiun 15 pentobarbitone (60 mg/ml)/g body weight) and mice with ether.

2. Surgically expose the heart and inject heparin (Porcine Mucous, 10,000 U/ml) into the femoral vein (rats: 0.3 ml injected).

20 3. Remove heart and place in ice-cold Krebs- Henseleit buffer containing heparin (0.2 ml/50 ml buffer.

Krebs-Henseleit buffer: 119 mM NaCI S- 25 mM NaHCO 3 4.6 mM KC1 1.2 mM MgSO 4 7H20 1.3 mM CaC12 2H 2 0 1.2 mM KH 2

PO

4 11 mM glucose 0.25%

BSA

Adjust to pH 7.4; store at 4. Connect aorta to the canula of the Langendorf perfusion apparatus and tie firmly. The apparatus was assembled by the present inventors according to experimental requirements of the Langendorf heart model as described in Doring Dehnerrt, The Isolated Perfused Heart According to Langendorf, Bionesstechnik-Verlag March GmbH, D7806, West Germany.

Perfuse with Krebs-Henseleit buffer (made fresh each day), which is gassed continuously with carbogen (95% 02, 5% CO2) at a pressure of 100 mmHg, at 37 0

C.

6. Attach a hook, connected to a transducer (Physiograph MK-111-S, Narco Bio-Systems) to the apex of the heart.

7. Perfuse heart for 20 min. with Krebs- Henseleit buffer to enable heart to stabilize (reservoir volume: 270 ml).

8. Add plasma (pre-warmed to 37 0 C) as follows: at 20 min. add 10 ml plasma plasma) at 25 min. add 10 ml plasma 9 plasma) at 30 min. add 10 ml plasma 13 plasma) 15 9. Monitor heart for a further 30 min. and record heart flow and contraction rate.

E. Sugar perfusion 1. Stabilize heart in Krebs Henseleit buffer for 30 min. as described above.

20 2. Add 2.5 ml of 1.08 M stock sugar solution to reservoir; total volume 270 ml; final sugar concentration 3. Allow heart to restabilize for 10 min, then add plasma (control or heat inactivated) as per the schedule described above.

4. Record heart beat and flow rate.

F. Large-Scale Preparation of anti-GAL antibody- Depleted Plasma (all manipulations are performed at 4 0

C)

1. Start with 200 ml freshly prepared human plasma; 100 ml is subject to depletion; 100 ml is used as an untreated control from the sam patient drawn on the same day; store at 4 2. Filter the plasma sequentially through a 100 pm, 8 pm metal sieves and finally through a 0.45 pm Millipore filter; dilute to 1000 ml with PBS, pH 3. Concentrate to 200 ml using an Amicon spiral wound cartridge (removes salt).

4. Equilibrate melibiose sepharose column ml) with PBS, pH 8.0 (10 column volumes).

Passage the plasma through the melibiose sepharose column; collect the run-through and store at 70 0 C (=partially depleted plasma).

6. Wash column with PBS, pH 8.0 (10 column volumes) until the O.D. (280nm) of the eluate is approximately zero.

7. Combine the partially depleted plasma and the 15 eluate from the wash; concentrate to 200 ml (Amicon spiral concentrator).

8. Elute the anti-GAL antibody fraction with 4M guanidinium HC1 pH 6.4 (2 column volumes).

9. Regenerate the column with PBS (10 column volumes).

1 0. Repeat the entire process an additional two times, repassage plasma through the melibiose column, wash, elute the anti-GAL antibody fraction and regenerate column.

S. 25 11. For the anti-GAL antibody-depleted fraction: combine the eluate from the melibiose sepharose column with run-through from the final wash adjust the volume to 5 liters with Krebs Henseleit buffer and add EDTA to 10 mM; adjust pH to concentrate back to original volume (Amicon spiral concentrator); aliquot (35 ml) and store at -70 0

C

12. For the anti-Gal antibody fraction: combine the eluted anti-GAL antibody fractions, dilute to 5 liters with Krebs Henseleit buffer and add EDTA to 10 mM concentrate back to 10 ml (Amicon spiral concentrator); aliquot (1 ml) and store at -700C 13. The anti-GAL antibody-depleted fraction and the purified anti-GAl antibody fraction are tested for a) Anti-GAL reactivity: Use as primary reagents to stain porcine cells (PK' Detect staining as described in Example 2, above. Analyze stained samples using a FACScan II (Becton Dickinson), according to the manufacturer's instructions.

b) Protein content: Determine using the colorimetric method of Bradford, Anal. Biochem. 72: 15 248-54 (1976), with purified human IgG as the standard.

c) Electrolyte concentration: On the day of the perfusion, the anti-GAL antibody depleted plasma is also tested to determine the calcium, magnesium and potassium levels using an electrolyte autoanalyser (Olympus); the levels of each are adjusted to normal as required.

II. RESULTS Rat hearts were connected to the Langendorf apparatus and then stabilized by perfusion with Krebs Henseleit buffer for 10 min., and then a further 10 min.

with the same buffer containing either melibiose or lactose (10mM). Human plasma was then added in stages as described above to a final concentration of 13 and the effect of the added sugar on cardiac function was assessed. The parameters measured were heart rate, amplitude (strength) of contraction and output (Figure 3).

In the presence of human serum alone (lower trace), the heart essentially stopped beating within minutes. The same result was obtained if lactose was added. In the presence of melibiose (upper trace) or anti-GAL antibody-depleted plasma, however, the heart was able to maintain a strong beat. When the purified anti- GAL antibody was added back to the anti-GAL antibodydepleted plasma, the heart again stopped beating within minutes.

EXAMPLE 6 Characterization of the Porcine a-1,3-GalT Gene cDNA's encoding porcine a-l,3-GalT were generated by Polymerase Chain Reaction (PCR) technology. Total RNA of pig liver was isolated by homogenizing liver slices in 7M guanidinium thiocyanate, as described by Chomczynski Sacchi, Anal. Biochem 162, 156-159 (1987); Sambrook et 15 al., Molecular Cloning: A Laboratory Manual (2nd Edition), Cold Spring Harbor Laboratory Press (1989).

Sixteen pg of the RNA, together with lpg oligo dT primer, were heat denatured for 5 minutes at 65 0 C prior to being transcribed into cDNA using avian myeloblastosis virus 20 (AMV) reverse transcriptase in a 100al reaction carried *out at 37 0 C for 90 minutes. Three A1 of the cDNA synthesis reaction was used in the subsequent

PCR

amplifications. General procedures used for generation of cDNA are provided in Sambrook et al (1989), supra.

Primers for PCR were synthesized using phosphoramidite technology, on an Applied Biosystems

DNA

synthesizer. The sequence of the PCR primers was based on identifying conserved regions within the published sequences for murine and bovine a-l,3-GalT genes.

Joziazze et al., J. Biol. Chem 264: 14290-97 (1989); Joziazze et al., Biol. Chem 267: 5534-5541 (1992). All primers were synthesized with EcoRl linkers at the 5' end for ease of cloning. In the following listing of the primers used in the present study, nucleotide positions varying between bovine and murine sequences are singleunderlined; nucleotide positions varying between bovine and human sequences are double-underlined: Exon 2 primer (forward): 5' -GTGAATTCAGCCCTGCCTCCTTCTGCAG-3I (SEQ ID NO: 1) Designation: GTE2F 28-mer 1 difference b/w bovine murine no sequence available for human exon 2 Exon 4 primer (forward): 3

I

*(SEQ ID NO: 2) SDesignation: GTE4F 28-mer -no differences b/w bovine, murine human Exon 9 primer (reverse): 3 ID NO: 3) *Designation: GTE9R 28-mer 3 differences b/w bovine murine 1 difference b/w bovine human 3'-UJTR primer (reverse):

S'-GTGAATTCGAAATCACTGGGATTTACA-

3

I

(SEQ ID NO: 4) Designation: GT3UR 28-mer no differences b/w bovine murine no differences b/w bovine human Exon 9 primer (forward): 3

I

(SEQ ID NO: Designation: GTE9F 28-mer -no differences b/w bovine murine difulferences b/w bovine human PolyA primer (reverse): 5'-TTGAATTCTTTTTTTTTTTTV*N**-31 (SEQ ID NO: 6) V A or C or G; N A or C or G or T (primer includes all nucleotide variants for V and N) Designation: APATR 23-mer The PCR conditions used to generate porcine a-1,3-GalT cDNA fragments were as follows: 1) For GTE2F GTE9R and GTE4F GTE9R: heat to 94 0 C seconds); then proceed with 35 reiterations (cycles) of the following three steps: 94 0 C, 40 seconds, (2) 57 0 C, 50 seconds, and 72 0 C, 80 seconds.

2) For GTE9F GT3UR: heat to 940C (120 seconds); then 15 proceed with 35 cycles of: 94 0 C, 40 seconds, (2) 48 0 C, 45 seconds, and 72 0 C, 60 seconds.

The PCR fragments were subcloned into EcoR1restricted pBluescript II KS+ (Stratagene, Cat, 2 12206) and the DNA sequence was determined using the 20 chain termination method. The DNA sequence was assembled and analyzed using DNASIS-Mac v2.01 (Hitachi) The nucleotide sequence of porcine a-1,3-GalT

(SEQ

ID NO: 7) and the derived amino acid sequence (SEQ ID NO: 10) of the enzyme are shown in Figures 4 and 5. A single large open reading frame extends from the initiating methionine at nucleotide 91 to a stop codon located at nucleotide 1204. The sequence surrounding the putative initiating methionine conforms to the consensus eukaryotic initiation sequence. Kozak, Cell 44, 283-92 (1986).

The porcine cDNA sequence is compared to the corresponding murine (SEQ ID NO: 9) and bovine (SEQ ID NO: 8) sequences in Figure 4. The locations of introns withi4 the murine gene are also shown. Joziazze et al., J. Biol. Chem 267: 5534 (1992). This alignment demonstrates that exon 3, located within the untranslated region of the mouse gene, is not found in either the porcine or bovine cDNAs. The overall sequence identities between the coding sequences are as follows: a) pig compared to mouse:- 75.02% (exon 3 not considered) b) pig compared to bovine:- 85.15% The amino acid sequences of the porcine (SEQ ID NO: 10), murine (SEQ ID NO: 12) and bovine (SEQ ID NO: 11) a-1,3-GalT enzymes are depicted in Figure 5. The locations of introns are also shown, based on their positions within the mouse gene (Joziasse et al., 1992).

This alignment illustrates that the overall amino acid homologies are: 15 a) pig compared to mouse: 71.98% b) pig compared to bovine: 82.87% c) bovine compared to mouse: 73.72% EXAMPLE 7 Identification of Potential Sites to Interrupt the a-1-3-GalT Gene 20 The present inventors' choice of a site for interrupting the a-1,3-GalT gene has been influenced by several characteristics of the gene and its expression.

i In particular, several mRNAs for a-l,3-GalT have been detected in the mouse. Joziazze et al., J. Biol. Chem.

267: 5534 (1992). These mRNAs are products of alternative splicing events in which exons 5 and/or 6 may be deleted. Hence, these exons are not appropriate interruption sites in the mouse, since a transcript encoding a functional a-l,3-GalT enzyme presumably could be formed when exons 5 or 6 are spliced out. Moreover, the present inventors have isolated two different classes of a-1,3-GalT cDNA clones from the pig one that includes exon 5 and one with exon 5 deleted. It is possible that mRNA's with and without exon 6 are also formed by alternative splicing in the pig. Thus, for initial experiments the present inventors have not chosen these exons as sites for interruption.

Insertion of an interrupting-DNA fragment into exon 4 (which encodes the cytoplasmic

NH

2 -terminal domain and the membrane-anchoring domain; see Figure 5) would disturb production of a transcript encoding an active a- 1,3-GalT. Hence this exon is an appropriate site to disrupt the a-l,3-GalT gene. Similarly, exons 7 and 8, which encode the NH 2 -terminal region of the catalytic domain, are suitable disruption sites. Insertion of a interrupting DNA fragment within these exons would prevent the synthesis of an active catalytic domain.

A preferred site for interrupting the mouse gene 15 is located at a Sail site found within exon 9 of the mouse e-l,3-GalT gene, at codons 221 222 (see Figure This site is positioned 150 amino acids from the COOH-terminus, within the catalytic domain. The mouse gene within the present inventors' constructs for homologous recombination is interrupted at this Sail site. The amino acids encoded by nucleotides at this Sail site are conserved in the pig and bovine sequences, although the Sail site itself is not. Construction of a Sail site at this position in the pig gene by in 25 vitro mutagenesis) provides a useful construct to inactivate the gene.

EXAMPLE 8 Choice of a DNA Fragment to Interrupt the a-1,3-GalT Gene The present inventors have used both the neomycin 58 resistance (neoR) gene and the hygromycin resistance gene (hygR) to interrupt the a-1,3-GalT gene. In one set of "knockout" constructs the neoR and hygR genes are linked to the murine phosphoglycerate kinase (PGK) promoter (Adra et al., Gene 60: 65-74 (1987) and are both bordered by polylinker sequences that include restriction sites for EcoRV and ClaI.

In another construct, expression of the neoR gene is directed by an altered polyoma virus promoter (PMC1; Thomas and Cappechi, cell 51: 503-12 (1987)). In this construct the present inventors have addressed the problem of including an antibiotic resistance gene within the genome of transplant organs. That is, in some circumstances it may not be desirable to have genes 15 conferring resistance to antibiotics present in the organ to be transplanted. The FLP/FRT recombinase system of yeast has been used to eliminate the neoR gene from the '"sequence that interrupts the a-1,3-GalT gene.

In a construct of the present invention, the neoR gene is bordered at both the 5' and 3' ends by FRT DNA elements. In addition, stop codons for each of three reading frames have been inserted 3' to the neoR gene, and these stop codons, together with a single FRT *99999 sequence, will remain within the a-l,3-GalT gene after 25 the neoR gene has been excised by FLP. Targeted cells carrying a genomic copy of the neo gene flanked by direct repeats of the FRT could be supplied with FLP recombinase in two ways: 1) Introduction into cells of partially purified FLP protein: FLP protein (0.1 10 p.g) is introduced ("transfected") into approximately 107 cells using standard electroporation conditions. The cells are plated out into gelatinized tissue culture dishes in appropriate medium, at a sufficient dilution to result in individual colonies. Approximately 200 of these colonies are then picked for further analysis.

2) Transfection with plasmids containing the FLP gene: A plasmid containing the FLP gene under control of a promoter able to drive FLP expression, the human interferon-inducible 6-16 promoter, is constructed according to standard methods. Porter et al., EMBO J. 7: 85 (1988). Approximately 10 ,g of FLP expression plasmid is transfected into approximately 107 cells using standard electroporation conditions. With a plasmid containing the human 6-16 promoter, interferon is added at approximately 500 units/ml, in order to induce 15 expression of FLP. The cells are then treated as in above.

The procedure to knock out the a-1,3-GalT gene in ES cells using an FRT-containing construct is: a) electroporate the complete construct into ES cells b) select neoR cells, and identify those ES cells having an interrupted a-l,3-GalT gene c) delete the neoR gene using FLP recombinase, Sas described above; cells are tested for the excision 25 event as follows: First, samples of each selected cell line are tested for the absence of the neoR gene by treatment with the chemical G418. The cells will die in the presence of approximately 200 /g/ml G418 unless the neoR gene is still present in the genome. Cell lines that are G418 sensitive are then tested further to confirm that excision of neoR has occurred. This is done by Southern analysis or PCR analysis, both described in Sambrook et al. (1989). For Southern analysis, genomic DNA is isolated from the cells, digested with an appropriate restriction enzyme, subjected to agarose gel electrophoresis, and the digested DNA transferred to a membrane. The DNA is hybridized with a labeled probe, the label is detected with X-ray film or color development), and the pattern of bands indicates whether or not an excision event had occurred in the cell line.

For PCR analysis, genomic DNA is isolated from the cells and subjected to PCR reaction with suitable oligonucleotide primers.

d) following confirmation of neoR excision, the manipulated ES cells or PGC' s are used to generate chimeric animals.

C

EXAMPLE 9 Preparation of DNA Constructs to Interrupt the a-1.3-GalT 15 Gene in Mice Gene targeting (homologous recombination) is more "efficient if the cloned cDNA fragments used for targeting S: are isolated from the cell line which is used for the gene knockout the DNA is "isogeneic").

20 Accordingly, DNA was isolated from-the E14 ES cell line (Hooper et al., Nature 326: 292-95 (1987)) and used to construct a mouse genomic library. The DNA was digested partially with the restriction enzyme Sau 3A, and fragments 12 kb 20 kb in size were isolated by glycerol gradient fractionation. The size-fractionated DNA was ligated into the Bam H1 site of IEMBL3 (Sambrook et al.

1989, supra), and packaged in vitro to form lambda phage particles. The lambda library was plated by infection of E. coli strain PMC103 host cells (Doherty et al., Gene 124: 29-35 (1993)) at a density of 4x10 4 phage per plate.

A bovine cDNA clone, about 900 bp in length and containing a portion of the a-l,3-GalT gene corresponding to exons 7 9, was used to probe a total of 5.6x10 independent recombinant phage. Four overlapping clones containing a-1,3-GalT gene sequences were isolated and purified. The Sail restriction sites within these clones were mapped (Figure and the 4.0kb, 5.5kb, 11kb and 12kb SalI fragments from two of the clones (13 and were subcloned into pBlueScript KS+ (Stratagene) or pUBS (pUC19 carrying the pBlueScript KS+ polylinker) to facilitate further detailed mapping of restriction sites.

These four subclones (designated paGT-S4.0, paGTpaGT-S11 and paGT-S13) were mapped for restriction sites with restriction enzymes BamHI, EcoRI, HindIII, XbaI, XhoI, KpnI, SacI, SacII, EcoRV, PstI, SmaI, NotI and BglII. paGT-S4.0 and paGT-S5.5 were also checked for PvuI, PvuII, NdeI and SphI restriction sites. Detailed restriction maps of the 4 subclones were drawn from these 15 data (Figures 7-12).

On the basis of these maps a knockout strategy was conceived. Basically the strategy is to insert a *resistance gene (either neoR or hygR) into the SalI site which lies within Exon 9. The knockout construct carries the 4.0 and 5.5kb Sail fragments from paGT-S4.0 and paGT-S5.5 which flank the Exon 9 Sail site (Figure 13).

Screening for homologous recombination events then can be carried out using a DNA fragment representing the genomic region but lying outside the DNA included in the knockout 25 construct, outside the 9.5kb covered by and paGT-S5.5. A 0.7kb EcoRl/XmnI fragment from paGT-S11 is used to screen Southern blots of BglII digested

ES

cell DNA for homologous recombinant events. An 8.3kb band should appear on these Southerns when the uninterrupted al,3-GalT gene is probed with this EcoRl/XmnI fragment (Figure 14). Insertion of the neoR gene after a homologous recombination event will give rise to a 6.4kb band, due to the presence of a BglII site just flanking the Exon 9 Sail site within the knockout construct. Thus the presence of the 6.4kb band is diagnostic for a homologous recombination event.

To carry out this strategy, the present inventors prepared a series of knockout constructs. The generation of one such construct is outlined in detail in Figure The vector paGT-S5.5, which carries the 5.5kb fragment immediately 3' to the Exon 9 SalI site, was chosen as the starting vector. paGT-S5.5 was digested with EcoRV and ClaI, generating a vector with a blunt end and a Clal compatible end. A 1.3kb fragment carrying the PMC1 promoter-driven neoR gene flanked by FRT sites was excised from plasmid pNeo2FRT (previously constructed by the present inventors) by digesting with BamHI, filling in the restriction site and then digesting with Clal to 15 generate a fragment with one blunt end and one ClaI compatible end. The nucleotide sequence of this 1.3kb fragment is provided in Figure 16 (SEQ ID NO: 13). This fragment was then ligated into the ClaI/EcoRV digested paGT-S5.5, the ligation mix transformed and colonies screened for recombinants. One colony was recovered that contained the NeoR fragment inserted into the EcoRV/ClaI of paGT-S5.5, based on the restriction pattern after digestion with diagnostic restriction enzymes Clal, o EcoRV, Xbal and EcoRI. This construct was designated S 25 PNeoaGT6.8.

pNeoaGT6.8 was digested with SmaI, generating a vector with blunt ends. The 4.0kb Sall fragment was excised from paGT-S4.0 and the ends filled. This fragment was then ligated into the SmaI digested paGT- S5.5, the ligation mix transformed and colonies screened for recombinants. Four colonies were recovered which contained the 4.0kb SalI fragment inserted into the SmaI sites of pNeoaGT6.8 with the 5' portion of Exon 9 lying near the 3' portion of the exon in the nearby Sail fragment. The identity and orientation of the insert was confirmed by the restriction pattern after digestion with diagnostic restriction enzymes XbaI, EcoRI, HindIII, BamHI, EcoRV and others. This construct was designated pNeoaGT10.8.

pNeoaGT10.8 was digested with Clal, generating a vector with Clal compatible ends. Two complementary oligomers were synthesized that, when annealed, generated a linker containing translation termination codons in all three reading frames and a BglII site. The linker has Clal compatible ends. The linker was ligated into the Clal digested pNeoaGT10.8, the ligation mix transformed and colonies screened for recombinants. Many colonies were recovered that contained the linker inserted into the Clal sites within pNeoaGT10.8 based on the 15 restriction pattern after digestion with diagnostic restriction enzymes BglII, Cla and BglII/NotI. This construct has been sequenced to confirm the identity, copy number and orientation of the insert. This construct is called pNeoaGT10.8B (Figure 17).

20 EXAMPLE ES Cells General Materials and Methods Working Conditions 'o"C Procedures for the isolation and culturing of all cell lines (embryonic stem, primordial germ and fetal fibroblast cell lines) require aseptic conditions to prevent growth of contaminating organisms: 1. All laboratory bench tops and equipment are wiped down with 70% ethanol prior to use.

2. All surgical instruments are autoclaved prior to use.

3. Water for media preparation and cleaning of glassware is of high quality Milli-Q water, prepared by passage through a Milli-Q ultrapure water system (Millipore).

64 4. Glassware is either dry-heat sterilized or autoclaved following extensive cleaning in Milli-Q water before use.

All tissue culture work is carried out under laminar flow conditions (Hepa filtered horizontal laminar flow workstation).

6. All media are filter sterilized (22pm disposable filter) prior to use.

7. Antibiotics are used to minimize the risk of bacterial contamination (Penicillin, Streptomycin and Gentamicin for bacteria; Nystatin for fungi).

Media/Solution Preparation DULBECCOS MODIFIED EAGLE MEDIUM (DMEM) 10.Og DMEM powder- Gibco 15 (the low-glucose or high-glucose formulation, with or without pyruvate, may be used; L-glutamine is included) liter Milli-Q-Water 3.7g NaHCO 3 Stir slowly until dissolved Adjust pH 7.2 Filter sterilize (following filter sterilization pH to rises to 7.4) SKeep at 4'C.

STO CELL MEDIUM 25 83.0 ml DMEM 15.0 ml 15% fetal bovine serum (FBS); batch tested before use ml Pen/Strep 1:100 ml Glutamine 1:100 (if needed) (see note below) Filter sterilize and keep at 4'C.

Note: Replenish complete medium (DMEM medium) (STO or ES) with glutamine.

*This step is only required if medium is older than 1 week 10 days, as the glutamine breaks down after this time.

ES CELL MEDIUM WITH OR WITHOUT LIF up to 100.0 ml DMEM 15.0 ml 15% FBS (batch tested before use; see below) ml (from 0.01M stock) P-mercaptoethanol (0.1 mM final concentration) 1.0 ml Pen Strep. 1:100 0 1.0 ml Glutamine 1:100 (if needed) ml Nystatin 1:100 0 2.5 ml Recombinant murine LIF (from 4x10 4 U/ml; 1000U/ml stock); activity-tested using LIF Assay (see below) 0.4 ml Gentamicin ml Nucleotides ml Non-essential amino acids PENICILLIN/STREPTOMYCIN ANTIBIOTIC SOLUTION (1:100) Commonwealth Serum Laboratories, Australia; Catalogue No. 05081901 Penicillin G 5000 U/ml Streptomycin Sulphate 5000 pg/ml.

.9*s MITOMYCIN-C

SOLUTION

S 20 2.0 mg Mitomycin-C (Sigma Chemical Co. ("Sigma"); Catalogue No. M0503) 200.0 ml STO Cell Medium Filter sterilize, divide into 20x 10 ml aliquot's and store at -20 0

C.

PHOSPHATE BUFFERED SALINE

(PBS)

For 100 ml Milli-Q Water: (Ca and Mg containing) (Ca and Mg free) NaCl 0.89 0.80 KC1 0.02 0.02

KH

2

PO

4 0.02 0.02 Na 2 HPO412H 2 0 0.289 1.115 CaCI 2 2H 2 0 .014 MgC12 6H20 0.01 Na pyruvate 0.0036 D-glucose 0.1 g Adjust to pH 7.4 and filter sterilize (Ca and Mg++ free PBS is purchased from ICN Cell Biology and Tissue Culture, Cat. No. 18-604-54) TRYPSIN/VERSENE (TV) WORKING SOLUTION (TV x 1) In PBS (Ca and Mg free): 0.25% trypsin (lyophilized) 0.04% EDTA or EGTA or: To 1 liter of milli-Q water add the following: Trypsin powder (Porcine, Difco) 2.5 g EDTA or EGTA 0.4 g NaCl 7.0 g Na 2 HPO412H 2 0 0.3 g

KH

2

PO

4 0.24 g KCl 0.37 g D-Glucose 1.0 g Tris 3.0 g Phenol red 1.0 ml 15 Adjust to pH 7.6, filter sterilize, aliquot and store frozen.

EGTA: Ethylene-glycol-bis(P-amino-ethyl ether)N,N,N',N- -tetra-acetic acid [Ethylene-bis(oxyethylenenitrilo)]tetraacetic acid 20 EDTA: Ethylenediaminetetraacetic Acid 5* Use either EDTA or EGTA. EGTA is preferred as it is less damaging to the ES/PGC cells.

GELATIN WORKING SOLUTION 0.1% gelatin in Milli-Q Water 25 Dissolve gelatin by heating to 60 0

C.

S" Filter sterilize when still warm.

To gelatinize tissue culture plates: 1. Cover dish with solution, leave 30 minutes 2. Aspirate gelatin and let dish air-dry.

NUCLEOSIDE STOCK SOLUTION Milli-Q Water 100 ml Adenosine (Sigma) 80 mg Guanosine (Sigma) 85 mg Cytidine (Sigma) 73 mg Uridine (Sigma) 73 mg Thymidine (Sigma) 24 mg 1. Dissolve by warming to 37°C.

2. Filter sterilize and aliquot while warm.

3. Store at 4°C or -200C.

4. Thawing of nucleotides for use in ES cell media nucleotides come out of solution upon thawing; Warm to 37 0 C to resolubilize before use.

NON-ESSENTIAL AMINO ACIDS (1:100) Commonwealth Serum Laboratories; Catalogue No. 09751301 100x concentrate for minimum essential medium (Eagle): ml is added to 100 ml ES Cell Medium) mq/10 ml milli-O Glycine L-Alanine 8.9 L-Asparagine

H

2 0 15.0 L-Aspartic Acid 13.3 L-Glutamic Acid 14.7 L-Proline 11.5 L-Serine 10.5 20 WHITTEN'S CULTURE MEDIUM KC1 0.0356

KH

2 PO4 0.0162 MgSO 4 7H 2 0 0.0294 NaC1 0.4 25 NaHCO 3 0.2106 Glucose 0.1 Na Pyruvate 0.0036 Ca Lactate 5H 2 0 0.0527 Na Lactate 0.2416 ml Milli-Q-H20 100 ml The solution is adjusted to a final milliosmolarity of 250-280 by addition of H 2 0 or NaCl.

Filter sterilize and store at 4 0 C for two weeks.

Working solution: 10 ml Whitten's medium Diagnostic division, Code No. 81-001-4) BSA fraction V (Miles Pentex, Kankakee, II., USA; 9 S to of 0 9 *9 S S

S.

S. S Filter Sterilize and equilibrate in 5%0 2 :5%CO2:90%

N

2 at 39.5°C, 95% humidity.

FBS BATCH TRIALS Batches of FBS vary in the ability to support growth of ES cells, and in the ability to maintain the undifferentiated state of such cells. The following procedure is used to identify suitable batches of FBS.

Use ES cells from between 2 20 passages: Day 1 Split ES colonies and plate into dishes without feeder cells but with LIF.

Incubate for 3 days.

Day 4 Trypsinise to detach colonies and cells.

Count cells and dispense into gelatinized 6cm dishes containing ES Cell Medium and LIF (no serum added) as follows: Dish Number Non-Inactivated Serum 1 2 3 4 6 7 8 9 10 11 12 No. Cells Batch FBS Control Serum (Batch Tested)

B

250 250 250 2000 2000 2000 5 mi 5 ml 5 ml 5 ml 5 ml 5 ml Inactivated Serum, as control 13 14 16 17 18 19 20 (56 0 C for 250 250 250 2000 2000 2000 15 min.) 5 ml 5 ml 5 ml 5 ml 5 ml 5 ml There are duplicate plates for each treatment.

69 Incubate low density dishes for 5 days Incubate high density dishes for 3 days Day 7 Fix high density cells and stain with hematoxylin.

Day 9 Fix low density cells and stain for alkaline phosphatase.

LIF ASSAY This procedure is used to assay the potency of Leukaemic Inhibitory Factor (LIF).

S. 6 S S *5 0 0O

S

S.

S.

*5 0 0 *5 0 090 S r 0

S.

r Day 1 Day 3/4 Day 6/7 Split one 10 cm dish of confluent STO cells into five dishes. Incubate for 2 3 days in STO medium.

When cells are confluent, replace medium with DMEM 10% FBS. Incubate for 3 days.

Collect conditioned medium (CM) and store at 4 0

C.

*Prepare low density ES cell cultures as described above.

Dish 20 1,2,3 4,5,6 7,8,9 10,11,12 13,14,15 25 16,17,18 No. Cells C.M. Medium 1000 U/ml

LIF

250 250 250 250 250 250 0.1 ml 0.25 ml 0.5 ml 1.0 ml 4.9 ml 4.75 ml 4.5 ml 4.0 ml Medium Presumed w/o LIF LIF Content 200 U/ml 500 U/ml 1000 U/ml 2000 U/ml 5 ml 5 ml There are triplicate plates for each treatment.

Fix and stain for alkaline phosphatase.

Preparation of Fibroblast Feeder Cell Layers Embryonic pluripotential cells are cultured in vitro on a layer of fetal fibroblast cells. The fibroblast cells provide a wide range of factors necessary for the growth of pluripotential embryonic cells growth factors, cytokines, factors that are essential for maintenance of ES cell pluripotency).

ISOLATION OF PORCINE FETAL FIBROBLASTS: 1. Remove developing porcine fetuses (preferably between days 16-30 of development) from uterus by aseptic dissection.

2. Remove skin layer from fetus.

3. Dissect out soft tissue avoiding developing viscera.

The white (fibroblast containing) tissue is found just under the skin layer.

4. Wash dissected tissue in PBS (Ca and Mg free).

Centrifuge at 1000 rpm for 5 min.

5. Remove supernatant.

6. Incubate tissue in Trypsin/Versene Working Solution 15 for 20 min.

S7. Dissociate cells by vigorously pipetting.

Centrifuge at 1000 rpm for 5 min.

8. Remove supernatant.

9. Resuspend cells in STO Cell Medium. Allow large cell-clumps to settle.

00oS 10. Plate out cells within supernatant large cell clumps are not included) onto gelatinized tissue culture plates. Incubate cells in an atmosphere of 5% CO 2 95% air (37.5'C, 95% humidity) until a 25 confluent layer of fibroblast cells appears days).

11. Passage of cells may be continued to increase cell numbers, or cells may be frozen or inactivated for further use.

CULTURE OF FETAL FIBROBLAST FEEDER LAYERS FROM FROZEN

STOCKS:

Several different types of mouse feeder (STO cells) and porcine and bovine fetal fibroblasts can be used to form feeder layers. These include: 71 Bradley/Baylor mouse STO feeder cells that have been modified to express human LIF (gift from Allan Bradley, Institute for Molecular Genetics, Baylor College of Medicine, Texas Medical Center, Houston, Texas, USA) Robertson/Columbia mouse STO feeder cells that have been modified to express murine LIF (gift from Elizabeth Robertson, Columbia University, New York,

USA)

Several porcine fetal fibroblast lines Several bovine fetal fibroblast lines (the fibroblast lines of and were derived by the present inventors using the procedures described above) 15 The procedure for producing feeder layers is as follows: 1. Rinse one 10 cm tissue culture (tissue cure) dish with gelatin/Milli-Q water solution for 30 min. Aspirate gelatin solution and let dish air-dry.

3. Add 10 ml of STO cell medium to 15 ml centrifuge tube.

4. Remove feeder layer cells frozen in freezing media from liquid N 2 container.

5. Thaw cells by warming vial in hands or in 37 0 C water bath.

6. Transfer STO cells to medium in centrifuge tube.

7. Spin at 1000 rpm for 5 min.

8. Resuspend cells in 10 ml medium and transfer to gelatin-treated tissue culture dish.

9. Incubate at 37 0 C for 3 days.

SPLITTING OF FEEDER LAYER STO CELL/FETAL

FIBROBLASTS:

This procedure is used to expand the number of cells from a single confluent plate/dish; cells are detached from the confluent plate and transferred to fresh plates at sub-confluent densities.

1. Gelatinize five 10 cm tissue culture dishes.

2. Examine incubated STO cells under microscope and check for confluence.

3. If STO feeder monolayer is confluent (cells cover bottom of dish, or nearly so), wash gently with PBS (Ca and Mg free) for 1 min.

4. Aspirate PBS and add 1 ml Trypsin/Versene Working Solution for 1 min (or until cells start to detach).

Check under microscope.

Detach cells by vigorously pipetting, add 1.0 ml STO Cell medium a ratio of 1:1 STO Cell medium:Trypsin/Versene Working Solution) to neutralize trypsin, and transfer to a centrifuge tube containing 10-15 ml STO Cell medium. Wash cells remaining on dish with some of STO cell medium from the tube. Centrifuge at 1000 rpm for 5 min., aspirate supernatant, resuspend pellet in 1 ml STO 15 Cell medium. Resuspend cells to make single cell suspension. Make up to 50 ml with STO Cell medium.

6. Dispense 10 ml into each of the five tissue culture dishes and incubate until confluent 3 days).

INACTIVATION OF FEEDER

LAYERS:

The present inventors use two alternative methods for inactivating feeder layers, which stops the cells from dividing: 'oo*o Mitomycin treatment: 1. Check dishes for confluence of STO cells/fetal fibroblasts.

2. Thaw mitomycin-C solution and use undiluted.

3. Aspirate STO cell medium from feeder cell plate.

4. Add 10 ml aliquot of mitomycin-C to plate and incubate at 370C for 1-3 hours.

5. Aspirate mitomycin-C, wash cells in ix PBS (without Ca++ or for 1 min.

6. Aspirate PBS and add 1 ml trypsin solution for 1 min.

7. Detach cells by vigorously pipetting and transfer to STO cell medium in centrifuge tube.

8. Centrifuge at 1000rpm for 5 min.

9. Resuspend cell pellet in 1 ml ES Cell Medium.

10. Plate out in dishes in preparation for addition of ES cells.

Gamma Irradiation: 1. Check dishes for confluence of STO cells/fetal a fibroblas 2.

3.

4.

5.

6.

15 t.

Trypsinise cells into single cell suspension.

Irradiate cells (3000 rads) in STO cell medium.

Centrifuge at 1000 rpm for 5 min.

Resuspend pellet in 1 ml ES Cell Medium.

Transfer cells to gelatinized tissue culture dishes with ES Cell Medium and place in incubator at 370C until the cells adhere to the dish. NOTE: If cells are not confluent, count using hemocytometer and seed at 5x10 4 cells in 1 ml medium per well of Nunc 4-well plate.

One 10 cm dish of inactivated cells can be split into: Ten 4-well plates (Nunc tissue culture plates), or Eight 3.5 cm tissue culture dishes, or Three 6 cm tissue culture dishes, or Two 20 cm tissue culture dishes.

Demonstration of Totipotency: A. Blastocyst Iniection The ability of embryonic cell lines to form germline chimeric animals is a conclusive test for their totipotency. This can be accomplished by blastocyst injection experiments, using techniques for various mammalian species substantially the same as those established for the mouse. See Example 14, below. See also, Bradley, Production and Analysis of Chimeric 74 Mice, In: Teratocarcinomas and Embryonic Stem Cells: A Practical Approach Robertson, IRL Press, Oxford, pp. 113-52 (1987). However, for porcine manipulations the holding pipette must be somewhat larger as porcine embryos are larger than mouse embryos.

B. Co-Culture of ES Cells/PGC's and Morula Embryos Embryos at the morula stage of development are surgically collected from superovulated animals. For porcine embryos, for example, the zona pellucida is then disrupted using Acid Tyrodes solution and ES cells/PGC' s are cultured in the presence of the zona pellucidadisrupted morulae. ES/PGC cells adhere to the exposed morula cells and, following overnight culture in Whitten's medium, the embryos are transferred to 15 synchronized recipients. Preferably, the zona pellucidadisrupted morula is completely free of the zona pellucida. However, this need not be the case as long as the ES cells/PGC's can gain direct access to at least some of the morula cells.

20 C. Morula Injection ES cells and PGC' s can be injected into a morula embryo prior to formation of the blastocyst cavity. The i technique is similar to blastocyst injection. ES cells or PGC's are drawn into an injection pipette, which is inserted beneath the zona pellucida. Then, the cells are expelled so that they are in contact with the cells of the morula embryo. The injected morula is then cultured overnight in Whitten's medium (porcine) or other appropriate medium to allow blastocyst formation.

D. Nuclear Transfer and Embryo Cloning ES cells and PGC's can be fused to enucleated zygotes that have been derived by in vitro maturation, in vitro culture, in vitro fertilization or collected surgically. Following successful fusion the embryos can be transferred to synchronized recipients. In vitro or in vivo-collected porcine oocytes, for example, are manipulated in Whitten's medium supplemented with BSA Fraction V and 7 Ag/ml cytochalasin B (Sigma). A bevelled micropipette is used to remove the metaphase plate from the oocyte. A single ES cell or PGC (after trypsin treatment to form a single-cell suspension) is inserted through the zona using a bevelled micropipette, such that the cell comes in contact with the oocyte 15 plasma membrane. Fusion is achieved in a 28 V/cm AC field for 5 sec. followed by an 80 V/cm DC pulse of 100 Asec. duration. Subsequent to observed fusion, embryos are incubated at 390 C in 5% C0 2 5% 02, 90% N 2 in microdrops of Whitten's medium supplemented with BSA, until transfer to a synchronized recipient.

EXAMPLE 11 Murine ES Cell Culture *ES cells are able to differentiate spontaneously *into many different cell types, and culture conditions which prevent this differentiation are critical for the continuous passage of these cells in an undifferentiated form, capable of contribution to chimeric mice.

I. CULTURE CONDITIONS ES cells are grown in polystyrene cell culture dishes treated with 0.1% gelatin (made up in PBS or Milli-Q water) for 10 minutes. A feeder layer of mitotically inactivated fibroblasts provides a source of cytokines. The fibroblasts are either primary mouse embryo fibroblasts (PMEFs), or STO fibroblasts, an

I

immortal line. The medium used is DMEM supplemented with glucose, amino acids and nucleosides. Robertson, Embryo- Derived Stem Cell Lines. In: Teratocarcinomas and Embryonic Stem Cells: A Practical Approach (E.J.

Robertson, IRL Press, Oxford (1987). To this medium is added LIF (final concentration of 10 3 U/ml Esgro, AMRAD). FBS is added to 15%. The batch of FBS is chosen on the basis of its ability to support ES cell growth with low levels of differentiation only rare individual cells undergo differentiation. The ES cells are grown in an atmosphere of 5-10% CO2, at 37 0

C

SII. ROUTINE PASSAGE ES cells must be passaged frequently to prevent the colonies from growing too large and differentiating.

15 This is achieved by splitting the cells at a ratio of 1:10 to 1:40, every two to four days.

EXAMPLE 12 Genetic Manipulation of Cells The general procedures set out in this Example 20 provide guidelines that are readily adaptable to individual experimental situations that might employ, for example, different cell lines or equipment supplied by different manufacturers. This Example also provides specific procedures used and results obtained in generating a set of mouse ES cell lines in which the a 1- 3 galactosyltransferase gene was disrupted by homologous recombination. The general procedures provided in this Example are adapted for mouse ES cells. However, the procedures are substantially similar for porcine ES cells.

I. INTRODUCTION OF DNA INTO ES CELLS BY ELECTROPORATION A. Coat required number of plates with 0.1% gelatin (in PBS or Milli-Q water). (Usually 2 X 6 well plates and 8 well plate) B. Thaw 107 embryonic fibroblasts into DMEES (equivalent to ES Cell Medium); inactivate by irradiating at 3000 Rad.

C. Count irradiated cells, spin down and resuspend in DMEES to 10 6 cells/ml.

D. Aspirate gelatin from plates and plate cells at: 7 X 105 cells/well (6 well plate) in 2.5ml medium; 7 X 104 cells/well (24 well plate) in 1 ml medium.

Incubate at 37 0 C, 5-10% CO 2 for 3 4 hr.

E. Wash ES cells in 5 ml (250 ml flask) PBS-EGTA 15 and let sit at room temperature for 4 min.

F. Remove PBS, add 5 ml trypsin (CSL) and leave at room temperature for 2 4 min. Wash down cells, add ml DMEES and count. Approximately 5 X 106 to 2 X 107 ES cells are needed for experiments.

G. Centrifuge cells and resuspend in 10 ml PBS.

Centrifuge again and resuspend in 540 pl PBS. Dilute l1 into 10 ml DMEES and culture to determine plating efficiency.

H. Add 5 10 Ag DNA to cells in 10 ul PBS (total 25 volume, 500 p1) and transfer to sterile electroporation cuvette Biorad).

I. Electroporate at 0.22 kV, 500 p/FD (time constant should be This is achieved using a Biorad Gene Pulser unit (Biorad Catalogue No. 1652078) with capacitance extender (Biorad Catalogue No. 1652087), or similar device.

J. Resuspend in 10 ml DMEES with constant pipetting to break up clumps of DNA from lysed cells.

K. Centrifuge cells and resuspend in 5ml DMEES.

L. Take 50 Al, add 50 Al trypan blue solution and count for viability.

M. Culture by dilution plating to determine plating efficiency.

II. SELECTION

CONDITIONS

ES cells that do not express a neomycin resistance gene are selectively killed by treatment with G418 at 200-500 pg per ml of medium. Antibiotic- containing medium is changed daily. A population of cells that has not been electroporated also is treated in order to see how genuinely sensitive cells respond to the G418 treatment. After 6 to 10 days, cells resistant to the antibiotic will be evident as healthy colonies. These cells will have been transformed by the targeting 15 construct and can be screened for homologous recombination screened for gene targeting versus -random integration).

Resistant colonies are picked from the selection dish with a mouth pipette and dispersed into a single 20 cell suspension. Half of these cells are frozen away while the other half is expanded and used to determine whether or not homologous recombination has occurred. If the colonies are small, it is sometimes preferable to a. *e S- expand the whole colony in a 24 well dish, and then to freeze half while further expanding the other half for genetic analysis.

III. PICKING ES CELL COLONIES FOR GENETIC

ANALYSIS

AFTER

SELECTION

A. Method 1: FreezinQ Half Colonies 1. The day before colony picking: a) Coat required number of plates with 0.1% gelatin (in PBS). Two plates per 24 colonies to be picked: one plate is for freezing and one plate is for clone expansion. Start with X 24 well plates.

b) Count irradiated fibroblasts, spin down and resuspend in DMEES.

c) Aspirate gelatin from 10 plates and plate -10 (can use as few as 5 X 104) cells/well in 1ml DMEES. Incubate at 37 0 C, 10% CO 2 overnight (or a minimum of 1 h).

d) Aspirate the gelatin from the other plates.

2. On the day of colony picking: a) Change medium on ES cells before and S* regularly during picking (to remove 15 floating cells).

Sb) Pull plugged pasteur pipettes. Use a fresh pipette after each 24 colonies.

The desired tip is about half a colony in diameter, with the constriction over 1-2cm. The tip should be perpendicular and neat.

Note: after drawing the pipette, rub the glass at the desired break point with freshly drawn glass, then bend.) 25 c) Label multi-tip reservoirs for: 1 PBS-EGTA 2 Trypsin-Versene 3 .DMEES 4 2 X Freezing mix(20% DMSO in

FCS)

d) Using multipipettor, dispense 50 .l PBS-EGTA into 24 wells of 96 well plate.

e) At microscope: Connect finely drawn pasteur pipette to mouth pipette tube. Dislodge colony from plate and transfer (in minimum volume) to one well of a 96 well plate. Expel contents of pipette; the bubbles serve as a location guide. Pick 24 colonies or as many as possible in <10-15 min (preferably a multiple of 6).

f) Back in hood: Add 100 Al trypsin to each well using multipipettor) and leave at RT for 2 min.

g) Pipette up and down 10- 15X to disperse cells, then add 100 pl DMEES. (This should be done within 4-6 min after trypsin addition).

i) Divide cell suspension between freezing and expansion plates using 12 channel pipette with every second tip fitted. Transfer 125 Al to gelatinized 24 well plate (to freeze); the remaining -125 L1 is transferred to a 24 well plate with 15 feeder layer (for DNA). The plates are labelled and carefully aligned to ensure that one clone goes into the same well of each tray.

j) Add 125 L1 2 X freeze mix to each 20 well on freezing plate, mix well by swirling.

k) Seal in ziplock bag or plastic wrap and place in -70 0 C freezer in an equilibrated styrofoam box.

Interleave the plates with styrofoam sheet.

1) Incubate expansion plates until there are sufficient cells for genotype analysis.

30 A. Method 2: Freezing after expansion to 24 wells.

1. The day before colony picking: a) Coat required number of plates with 0.1% gelatin (in PBS).

Start with 10 X 24 well plates.

b) Count irradiated fibroblasts, spin down and resuspend in DMEES.

c) Aspirate gelatin from the plates and plate -10 cells/well in 1ml DMEES.

Incubate at 37oC, 10% CO 2 overnight (or a minimum of 1 h).

2. On the day of colony picking: a) Pick colonies as described for half colonies (method 1, above) but instead of dividing the cell suspension between freezing and expansion plates, the entire cell suspension goes into the expansion plate.

b) After 3-4 days (with daily medium changes) the cells will have grown sufficiently to be frozen. Working one plate at a time (with practice two can be handled), aspirate medium from each well. Flood with PBS/EGTA for 4 minutes. Meanwhile, set up pipette tips to fit alternate channels of a twelve channel multipipettor. Aspirate PBS.

c) Add 100 Al trypsin (using multipipettor and alternate channels) and leave at room temp. for 2 min.

d) Pipette up and down 10- 15X to disperse cells of first row, change tips, then add 100 Al DMEES. Repeat for each row. (This should be done within 6 min of trypsin addition).

*V e) Using 12 channel pipette with every second tip fitted, transfer 125 Al to gelatinized 24 well plate (to 30 freeze). The remaining cells will be expanded for DNA. It is crucial that the plates are labelled and carefully aligned to ensure that the freezing tray matches the expansion tray.

f) Add 125 ~l 2 X freeze mix to each well on freezing plate; mix well by swirling.

g) Seal in ziplock bag or plastic wrap and place in -70 0 C freezer in an equilibrated styrofoam box.

Interleave plates with styrofoam sheets.

h) Add Iml of DMEES to the expansion tray. (There will be sufficient feeder cells to give good plating efficiency). Incubate for 3-4 days until there are sufficient cells for genotype analysis.

IV. THAWING OF ES CELL CLONES FROZEN IN 24-WELL

PLATES

Cells that have been identified to have the desired genetic alteration are recovered from a duplicate plate frozen at -70 0 C. The plate is taken to the laminar flow hood and removed from the plastic bag. Each well is filled with warm medium, and feeder cells are added to the well(s) of interest. The plate is placed in a 37 0

C

incubator for 60 min., then the medium is replaced.

Colonies will appear after two or three days. These colonies are expanded for establishment of new frozen :.*stocks, and tested for 1) karyotype analysis; 2) 15 confirmation of the desired genetic alteration; 3) mycoplasma infection; and 4) ability to form chimeras.

EXAMPLE 13 Production Of Mouse ES Cell Knockouts Using The DNEOaGT10.8B Construct 20 I. TRANSFORMATION A total of 1x10 7 E14 ES cells was electroporated with 5l1 of lg/Al pNeoaGT10.8B DNA (linearized by XhoI digestion) (see Example 9 and Figure 17). Electroporation S- was carried out in 600l in a wide cuvette at 25pF, 350V for 0.5msec. Cells were recovered in 6ml ES complete medium and plated into 6 x 100mm petri dishes, each containing a feeder layer of NeoR STO cells.

Cells were cultured in ES complete medium for 3 days and then medium containing 200-350g/ml G418 was substituted. This medium was changed every second day.

After 9 days, individual NeoR colonies were sufficiently large to be identified and recovered. Colonies were picked in 2 0~l PBS and 20,1 of trypsin solution were added. Forty tl of 60% BRL conditioned medium in ES 83 complete medium were then added. Aliquots of 4 0,l were transferred to single wells of each of two 24-well plates. One plate contained a feeder layer of STO cells in 100il ES complete medium. 140ul of 2x DMSO freezing mix was added to this plate, which was stored at -80 0

C.

Each of the wells of the second 24-well plate contained iml of 60% BRL conditioned medium in ES complete medium.

This plate was incubated at 37 0 C until the colonies were confluent.

II. CONFIRMATION OF HOMOLOGOUS

RECOMBINATION

V Medium was aspirated off confluent colonies and 400pl lysis buffer (10mM Tris pH 7.8, 100mM NaC1, ImM EDTA, 1% SDS, and 500ug/ml Proteinase K) added. The cells were lysed at 37 0 C overnight, extracted with 400al 15 1:1 phenol/chloroform and transferred to Eppendorf tubes containing iml 95% ethanol and 0.2M NaAc. DNA was pelleted by centrifuging at 13,000 rpm in an Eppendorf centrifuge, the pellet washed twice with 80% ethanol and redissolved in 30/il water.

20 Southern analysis (see, Sambrook et al., sob* supra) was used to identify ES cell clones where homologous recombination had occurred at the 3' end of @o6**f the construct. Aliquots of 15pl of DNA were digested with 20 units of the restriction enzyme BglII according to the manufacturer's recommendations. After incubation at 37 0 C overnight, the DNA was electrophoresed through a 0.8% agarose gel (in a Tris acetate, EDTA buffer) at 1- 2V/cm overnight, using 750ng of HindIII-digested lambda DNA as markers. The DNA was transferred to a Zetaprobe nylon membrane using a Hybaid vacublotter at a vacuum of Hg for 1 hour.

The membrane was prehybridised in a Hybaid hybridization bottle in 10ml of the following hybridization mix for 3 hours at 65 0

C:

0.25M Na 2

HPO

4 pH 7.2 7% SDS 1mM EDTA 100lg/ml salmon sperm DNA 10% PEG Radioactively labeled probe DNA was prepared using a BRESATEC gigaprime oligo labeling kit (Cat. No. GPK-1) according to the manufacturer's recommendations.

Approximately 50ng of a 0.7kb EcoRI/XmnI DNA fragment from beyond the 3' terminus of the construct pNeoaGT10.8B (see Example 9 and Figure 17) were labeled with 32 P-dATP to a specific activity of 5x10 8 cpm/tg. The denatured probe was added to the prehybridising membrane in the 9.

Hybaid bottle and incubated overnight at 65 0

C.

15 The membrane was removed from the Hybaid bottle, rinsed with 0.5xSSC, 0.1% SDS prewarmed to 65 0 C, and then washed 2-3 times with 0.lxSSC, 0.1% SDS at 65 0 C for min each wash. Excess moisture was then blotted from the membrane, the membrane wrapped in plastic wrap and exposed to a phospho-imager screen for 16 hours up to 3 days. The image was visualized on an Imagequant phosphoimager.

Results are shown in Figure 18, which is a Southern blot of DNA from 15 ES cell lines probed with the 25 diagnostic 0.7kb EcoRI/XmnI DNA fragment described above and in Example 9. The 6.4kb band, diagnostic for a homologous recombination event in the a 1-3 galactosyltransferase gene (a 1-3 Gal T) (see Example 9), is seen in 6 of the 15 ES cell lines examined. All of the 6 knockout cell lines appeared to be heterozygous for the inactivated allele since the 8.3kb band, diagnostic for the uninterrupted a-l,3-Gal T gene (see Example 9), was also present in all six lanes.

Two cell lines, designated hereinafter "8D1" and "7C2," were chosen for further analysis. Cell lines 8D1 and 7C2 were identified by Southern analysis to contain an a-1,3-Gal T allele where homologous recombination had occurred at the 3' boundary of the construct.

Long range PCR was then used to determine whether or not homologous recombination had occurred at the boundary of the construct within these cell lines. Two sets of primers were used in separate PCR experiments: 1) Wild-type primers:- MGT-KOex8F and MGT-KOR1 span the intron between exons 8 9, and amplify a 5.5 kb fragment from the wildtype a-1,3-GalT gene (Figure 19)

SEQUENCES:

MGT-KOex8F 5'TGCTGGAAAAGTACTACGCCACACAGAAACTCA-3' (SEQ ID NO: 14) 15 (Nucleotides 1014-1046 in Figure 4) MGT-KOR1 5'AGCCAGAGTAATAGTGTCAAGTTTCCATCACAA-3' (SEQ ID NO: (Nucleotides 1779-1811 in Figure 4) 20 2) Knockout primers:- MGT-KOex8F and MGT-KONeoR span exon 8 to the NeoR gene cassette in the "knock-out" allele and amplify a kb fragment from the knocked out allele (Figure 19)

SEQUENCE:

MGT-KONeoR 5'-GCCACACGCGTCACCTTAATATGCCAAGTGGAC-3 (SEQ ID NO: 16) (Nucleotides 323-355; Figure 16) Each reaction contained -100 ng genomic DNA as template in a reaction volume of 50gl and contained Tris HC1 (pH9.1), 16mM (NH 4 2

SO

4 250 iM dNTPs, 3.5 mM MgCl 2 100 ng each primer, 2 units Taq polymerase and 0.025 units Pfu polymerase. The reactions were heated at 940C for 1 min, then 45 cycles of 94 0 C for 15 sec, 680C for 6 min, followed by a single step of 72 0 C for 10 min.

Genomic DNAs from putative "knock-out" ES cell lines from CBA/C mice (homozygous for the wild-type a-l,3-Gal T allele) were amplified in separate reactions using each set of primers. A 10il aliquot of each PCR was analyzed by Southern blotting (Sambrook et al., 1989).

The results are illustrated in Figure Knockout primers:- A 5.5 kb fragment that hybridized to the 1.3 kb NeoR gene cassette (Figure 16) was generated from 7C2 DNA (Figure 20; lane 4) and 8D1 DNA (not shown). This band was not generated from CBA/cDNA (Figure Slane 3).

Wild-type primers:- 15 A 5.5 kb fragment that hybridized to the a-l,3-Gal T gene probe (isolated by Sal I digestion of paGT-S4.0) was generated from 7C2 and CBA/cDNA's (Figure 20; lanes 1 and 2 respectively) and 8D1 DNA (not shown). This product did not hybridize to the NeoR gene 20 probe.

These results demonstrate that homologous recombination had occurred at the 5' boundary of the construct in cell lines 8D1 7C2.

EXAMPLE 14 Generation of Animals CarryinQ an ES Cell Genome The procedures provided in this Example are adapted for mouse ES cells. However, the general strategy is substantially the same for porcine ES cells and PGC's.

I. PREPARATION OF ES CELLS FOR INJECTION ES cells are split into wells of a 24-well dish at cell densities of 1:2, 1:4, 1:8 and 1:16, relative to the initial density, two and three days before injection.

The most vigorous and least differentiated cultures are chosen on the basis of morphology.

II. EMBRYO INJECTION AND PRODUCTION OF CHIMERIC

MICE

Mouse embryos are collected from either superovulated or naturally mated female mice, approximately days after mating. After overnight culture in M16 medium (Bradley, Production and Analysis of Chimaeras. In Teratocarcinomas and Embryonic Stem Cells a Practical Approach Robertson, ed.) IRL Press, Oxford, pp. 113- 52 (1987)), those that have cavitated to form blastocysts are microinjected with about 12 to 20 ES cells. This microsurgical procedure is performed with instruments drawn from capillary glass, and injection is controlled with micrometer syringe-based hydraulic devices. A differential interference contrast-equipped inverted microscope is used to view the procedure.

After injection, blastocysts are transferred to the uterus of pseudopregnant female mice. Chimeric mice are identified by coat color contribution by the ES cells.

Chimaeric mice show agouti coat colour derived from the host blastocyst, and chinchilla contributed by the ES cells.

Chimeric mice were generated from ES cells carrying the interrupted a-1,3-Gal T allele (including 8D1, 7C2 cells) by injection into C57B1/6J x CBA F2 blastocysts. The ability of individual chimaeric mice to transmit the ES cell characteristics through the germ-line was estimated by glucose phosphate isomerase (Gpi) analysis of sperm (Bradley, supra, (1987)); Mann et al., J. Reprod Fert.

99, 505-512 (1993). Glucose phosphate isomerase catalyses the interconversion of glucose-6-phosphate to fructose-6phosphate. Mice have a single structural Gpi locus with two main alleles Gpi 1A and Gpi lB. Gpi 1A codes for protein which appears as a slow cathodically migrating band during electrophoresis and occurs in strains such as BALB/c and C129. (The ES cells used here were derived from strain 129 mice). Gpi 1B determines an enzyme that moves faster than Gpi 1A and occurs in the wild and in strains such as C57 and CBA (used here to derive host blastocysts).

Heterozygotes have the two parental bands plus an intermediate band which indicates the dimeric structure of the enzyme. Multiple electrophoretic forms occasionally observed are due to oxidation of sulfyhdryl groups and not due to tissue-specific expression. In chimaeric mice, the ratio of Gpi 1A (strain 129-derived) to Gpi 1B (derived from the host blastocyst) indicates the proportion of cells with the ES cell genotype within different tissues. The appearance of Gpi 1A (derived from the ES cells) in the 15 sperm suggests that the mouse is able to transmit the ES cell genotype through the germ-line.

III. GENERATION OF MICE HOMOZYGOUS FOR THE GENETIC CHANGE INTRODUCED INTO THE ES CELLS.

oo Chimaeric mice with sperm derived from ES cells were mated to BALB/c mice. Offspring with the 129/Ola X BALB/c genotype heterozygous for the ES cell genotype) are grey. Half of these grey mice were expected to carry the interrupted allele. Mice heterozygous for the interrupted allele were identified by PCR analysis of genomic DNA obtained from blood.

To generate mice homozygous for the inactivated a-1,3- Gal T gene, the heterozygous mice were mated to each other.

One quarter of the offspring were expected to be homozygous for the interrupted gene. Homozygotes were identified by PCR analysis of genomic DNA obtained from blood. The PCR strategy was based on the insertion of a NeoR gene in the Sal I site of exon 9 of the a-l,3-Gal T gene (Figure 13).

Wild-type primers:- 89 E9F: 5'TCAGCATGATGCGCATGAAGAC 3' (SEQ ID NO: 17) (homologous to sequence about 40 to 60 bp 5' to the Sal I site of exon 9, corresponding to nucleotides 1257- 1278; Figure 4) E9R2: 5'TGGCCGCGTGGTAGTAAAAA 3' (SEQ ID NO: 18) (homologous to a region about 175 to 195 bp 3' to the Sal I site of exon 9, corresponding to nucleotides 1511- 1492; Figure 4) The expected fragment size generated from the wildtype allele is 255 bp (Figure 21). These primers also can potentially generate a 1596 bp PCR fragment from the interrupted allele. In practice this fragment was not 15 generated when both the wild-type and interrupted alleles were present, probably because the smaller 255 bp product is amplified preferentially.

Knock-out primers:- NeoFl: 5' TCTTGACGAGTTCTTCTGAG 3' 20 (SEQ ID NO: 19) (corresponding to nucleotides 1170-1189; Figure 16) E9R2: (the same primer described above to detect the wild-type allele) The expected fragment size is 364 bp (Figure 21).

Mice were grown to weaning age and bled from the tail. Sodium Heparin was added to about 10 U/ml. PCR amplification was conducted on 1 il of heparinised blood (-104 nucleated cells) in a 50ul reaction volume containing 100 mM Tris-Acetate pH 8.8, 3.5 mM MgC1 2 0.2mM dNTPs, and 2 units Tth DNA polymerase. Each reaction contained both the wild-type and knock-out primers at a concentration of 2ng/Al for each primer. To ensure that Tth polymerase was not inhibited by heparinized blood, each reaction was performed in duplicate.

One of the reactions was spiked with two DNA samples: i) 10 fg (-600 molecules) of linearized KO plasmid pNeoaGT10.8B.

ii) 1 fg (-1000 molecules of a 983 bp RT-PCR product that includes Exon 9.

The other reaction was not spiked. Thus, two separate PCR reactions were set up for each blood sample. In addition, control PCR reactions with no genomic DNA template and with or without spikes were conducted. Each reaction mix was heated at 94 0 C for 3 min., then incubated for 40 cycles at 94 0 C for 40 sec., 530C for 40 sec., and 72 0 C for 40 sec.

Aliquots of 5 pl of each reaction mix were electrophoresed 15 on a 3% agarose gel, and DNA fragments were visualized on a UV light box after staining with ethidium bromide.

HpaII-digested pUC19 plasmid DNA was used for markers.

Results of the PCR analysis for three mice, and a "no DNA" control, are shown in Figure 22. For mouse #42, the 20 KO primers generated a 364 bp band in the spike reaction only. The wild-type primers generated a 255 bp band in the spike and spike reactions. These results demonstrate i that mouse #42 is homozygous for the wild-type allele. For mouse #43, the wild-type primers generated a 255 bp band in the spike reaction only. The KO primers generated a 364 bp band in the spike and spike reactions. These results demonstrate that mouse #43 is homozygous for the interrupted allele. For mouse #44, the KO primers generated a 364 bp band in the spike and spike reactions. The wild-type primers generated a 255 bp band in the spike and spike reactions. These results demonstrate that mouse #44 is heterozygous for the interrupted allele. In the control PCR reactions, no product was evident when template was not included. PCR products of 364 bp and 255 bp were evident when pNeoaGT10.8B and Exon 9 RT-PCR DNA were the only templates included in the control reactions.

EXAMPLE Characterization of Homozyqous Knockout Mice I. ABSENCE OF Gal T mRNA IN Gal T KNOCKOUT

MICE

A. RNA Isolation Total RNA was extracted using the RNAzol"B kit (BIOTECX Laboratories, Inc., 6023 South Loop East, Houston, Texas 77033, USA.), supplied by Bresatec. This extraction S 0 procedure is based on the method described by Chomczynski et al., Anal. Biochem. 162: 156-159 (1987), and involves homogenization in a guanidinium/phenol solution, a chloroform extraction, 2 isopropanol precipitations, and EtOH washes. The RNA was stored as an EtOH precipitate 15 at -20 0 C and quantitated by measuring absorption at wavelenth 260 nm in water. The integrity and quantitation was confirmed by electrophoresis in agarose/formaldehyde gels. Sambrook et al. Molecular Cloning. A Laboratory Manual. Second Edition. (1989) B. RT-PCR First strand cDNA synthesis involved: annealing 2gg of total RNA from kidney, heart or liver with 120ng oligo dT primer (Gibco BRL, M-MLV Reverse Transcriptase Kit) at 65 0 C for 5 minutes in 5~l of 10 mM Tris-HCl,lmM EDTA (pH8).

reverse transcription at 37 0 C for 1-2 hours in a final reaction volume of 20gl utilizing the M-MLV Reverse Transcriptase Kit(Gibco BRL). Each reaction contained 5mM DTT, 0.1~g/gl BSA, 1mM dNTPS, 40 U of human placental RNAse Inhibitor (Bresatec), 200U of M- MLV Reverse Transcriptase and the associated RTase buffer at lX concentration.

C. PCR Analysis of cDNA e-l,3-Gal T cDNA was detected by PCR amplification of oligo dT-primed cDNA template.

Failure to generate this PCR fragment, in conjunction with the control PCR results, indicated that a-l,3-Gal T mRNA was absent from the RNA preparation. To demonstrate that the a-1,3-Gal T primers supported amplification of the a-l,3-Gal T template, each reaction was assembled in duplicate, and one of the reactions was spiked with 0.1 fg (-100 molecules) of a 983 bp mouse a-1,3-Gal T cDNA product (generated by primers 7F and mGT-3UR, spanning exon 7 to the 3' untranslated region). As a second control to demonstrate that cDNA synthesis had occurred, a ferrochelatase PCR fragment was generated from the 15 cDNA template.

1. Primers: Primers to detect a-l,3-Gal T cDNA: 7F: TGGAGATCGCATTGAAGAGC 3' (SEQ ID NO: 20 (corresponding to nucleotides 889-911 within exon 7 (Figure 4) 9R2: TGGCCGCGTGGTAGTAAAAA 3' (SEQ ID NO: 21) (corresponding to nucleotides 1492-1511 within exon 9 (Figure 4) Primers 7F and 9R2 were expected to generate a fragment of -619 bp (Figure 23) from the cDNA template. These primers will not generate a fragment from genomic DNA possibly present in the cDNA preparation, since the primers span two large introns.

mGT-3UR: GGGTTTTGGTTTTGATTGTT 3' (SEQ ID NO: 22) (corresponding to nucleotides 1866-1888 within the 3' untranslated region; Figure 4).

This primer was used with primer 7F to generate the DNA fragment used in the control spike PCRs.

Primers to detect mouse ferrochelatase cDNA (EcoRI linkers, underlined): FC-F: CTGAATTCATGTTAAACATGGGAGGCCCC 3' (SEQ ID NO: 23) (corresponding to nucleotides 215-235, STaketani et al., J. Biol.Chem. 265: 19377-80 (1990)).

gFC-R: CTGAATTCTGCCCACTCCCTGCCGATG 3' (SEQ ID NO: 24) (corresponding to nucleotides 888-908, Taketani et al., J. Biol.Chem. 265: 19377-80 (1990)).

These primers were expected to generate a 709 bp fragment (Figure 23). These primers will not generate a S 20 fragment from genomic DNA possibly present in the cDNA preparation, since the primers span five introns.

Reaction volumes were 50 il, consisting of 4 Al of the first strand cDNA synthesis reaction, 100 ng of each primer, 2 mM MgCl 2 0.3 mM dNTPS, 2U of Taq-Polymerase (Bresatec) and Tag reaction buffer (Bresatec) at IX concentration. Reactions were heated at 94 0 C for 2 min, then 29 cycles of 94°C for 15 sec, 580C for 30 sec and 72°C for 1 min followed by single steps of 720C for 4 min and 4 0 C for 5 min. A 10 il aliquot of each PCR was electrophoresed on a 2% agarose gel and DNA fragments were visualized on a UV light box after staining the gel with ethidium bromide.

Figure 24 shows the PCR fragments generated from RNA isolated from kidney heart and liver of a wild-type mouse, and mice heterozygous or homozygous for the interrupted e-l,3-Gal T allele. Figure 24(i) shows that the 709 bp ferrochelatase fragment was generated from each of the cDNA preparations, indicating that cDNA template was produced from the reverse transcription reaction, and was available for the a-l,3-Gal T gene primers. The 619 bp a-1,3-Gal T fragment was present in each of the reactions spiked with the 983 bp a-l,3-Gal

T

cDNA product (Figure 24(ii)), indicating that amplification of the a-l,3-Gal T cDNA (spike) template had occurred.

In the reactions that were not spiked (Figure 24 4 the 619 bp a-l,3-Gal T fragment was detected in cDNAs synthesized from the wild-type and heterozygous RNAs.

This indicates that a-l,3-Gal T mRNA is present in the 15 kidney, heart and liver of the wild-type and heterozygous mice. The 619 bp fragment was not detected in the unspiked homozygous KO reactions, indicating that a-l,3-Gal T mRNA is not synthesized in the homozygous KO mice.

0 II. TEST FOR EXPRESSION OF THE GAL EPITOPE IN HOMOZYGOUS KNOCKOUT MICE USING ANTI-GAL ANTIBODIES

WITH

FLUORESCENCE-ACTIVATED CELL SORTING (FACS) A. Solutions Solutions 1 to 5 are 10x isotonic.

1. 1.68M NaC1 9 48.21g/1) Dry salts overnight in hot oven before weighing 2. 1.68M KCl (125g/l) Dry salts overnight in hot oven before weighing 3. 1.12M CaC12 (165g/l CaCl 2 2H 2 0) Dry salts overnight in hot oven before weighing 4. 1.68M MgSO 4 (414g/l MgSO 4 7H20) Do not dry in hot oven Potassium phosphate buffer pH 7.2: a) 1.68M KH 2

PO

4 (229 g/L) b) 1.12M K 2

HPO

4 (226 g/L K 2

HPO

4 3H 2 0 or 195 g/1 K 2

HPO

4 Potassium phosphate buffer is prepared by mixing together equal volumes of solutions a) and To pH the buffer, remove a small sample, dilute 1:50 and read on pH meter.

6. Hepes buffer 1M (CSL, Melbourne Australia) 7. KDS BSS: 0 Add stock solutions in the following order to double-distilled water (DDW): Stock Ratio of Solutions DDW 1210 NaCl 121 KCL 3 CaCI2 3 MgS04 1 Potassium phosphate buffer 2 Hepes 2n

:C

Filter sterlise, store at 4 0

C

8. KDS/BSS/2%HSA/0.02% azide:

KDS/BSS

Human serum albumin (CSL, Melbourne, Australia) 10% Na azide in MT-PBS 244.5ml 9. FITC dilution: Dilute 7.5ul FITC-IgG to 600ul with KDS/BSS Red cell lysis buffer: 0.168M NH 4 C1 in double distilled water 11. 4% paraformaldehyde

(PFA)

Solutions: NaH 2

PO

4 2H 2 0 NaOH paraformaldehyde: 22.6 g/L 25.2 g/L 1) 4 g paraformaldehyde (BDH, Kilsyth, Australia, #29447) dissolved in double distilled water. Heat 70 0 C 2 hours on stirrer in fume hood and a few drops of 2M NaOH are added until the solution becomes clear.

2) 0.54 g glucose is then added.

3) Store RT in light proof bottle.

D. Add together 83 ml of A 17 ml of B.

E. Final 4% PFA fixative solution: 90 ml of D 10 ml of C. pH 7.4 7.6; adjust pH with 1M HC1.

12. Hanks Balanced Salt Solution (Ca and Mg free)(HBBS): KCL 400mg KH2PO 4 NaCl 8g NaHC03 350mg Na 2

HPO

4 2H 2 0 68mg 20 Glucose 1g

H

2 0 to 1 liter adjust to pH 7.0; filter sterilize 13. Sheep antihuman IgG and IgM fluorescein isothiocyanate (FITC) F(ab)2 fragments (Silenus, 25 Hawthorn, Australia): B. Methods 1. Eye bleed mice, collect 300-400ul into prechilled Ependorf tube, store on ice, add EDTA 20mg/ml to o* give final concentration of 2mg/ml.

30 2. Transfer blood (including appropriate human controls) to 10ml plain tube and add 10ml red cell lysis buffer (0.168M NH 4 Cl) pre-warmed to 42 0 C; incubate for several minutes or until cells have lysed.

3. Pellet cells by centrifugation (800 x g, 7 min, 4 0

C).

4. Resuspend cells in 10ml KDS/BSS/2% HSA/0.02% NaN 3 Pellet cells as above; repeat steps 4 6. Resuspend cells in 1000ul KDS/BSS/2% HSA/0.1% NaN 3 transfer aliquots to V bottom FACS tubes.

7 Pellet cells as above.

8. IResuspend cells in iQ0ul KDS/BSS/2% HSA/0.1% NaN 3 9. Add 50u1 of purified anti-GAL antibody (see Example 1, above), normal human serum (NH{S) or HBBS/2% HSA/0.1% NaN 3 and incubate 45 min.

Add 2ml KDS/BSS/2% HSA/0.02% NaN3; centrifuge cells as above.

11. Add 50ul of a 1:80 dilution of sheep antihuman IgG or IgM FITC F(ab)2 fragment (Silenus).

12. Add 2ml KDS/BSS/2% HSA/0.02% NaN3; centrifuge cells as above.

13. Resuspend cells in 300u1 KDS/BSS/2% HSA/0.02% NaN3.

14. Transfer samples to plastic round-bottom

FACS

tubes and add 3 ul of propidium iodide (lO0ug/ml); samples are now ready for analysis; keep on ice.

Analyse on Beckman FACS scan using peripheral blood lymphocyte settings.

C. Results The results of these experiments are given below: median channel peak channel fluorescence fluorescence (log scale) (log scale) MOUSE 129 (Normal) 9 PBL FITC anti- IgG alone (neg. control) MOUSE 19 PBL 197 286 (wild type) GAL IgG MOUSE 21 PBL 22 (Gal KO) GAL IgG MOUSE 129 (Normal) 7 1 PBL FITC anti- 15 IgM alone (neg. control) MOUSE 19 PBL 185 167 (wild type) GAL IgM 20 MOUSE 21 PBL 34 18 (Gal KO) GAL IgM MOUSE 129 PBL (normal) 8 9 PBL FITC IgG alone 25 (neg. control) MOUSE 129 PBL (normal) 120 328 MOUSE 9 PBL 10 9 (Gal KO) GAL IgG The results of human anti-Gal binding to human peripheral blood lymphocytes (negative control) are not shown but were negative. These experiments demonstrate that human anti-Gal (IgG and IgM) antibodies bind to peripheral blood cells of the homozygous al,3 galactosyltransferase knockout mice (mouse 21 and mouse 9) very weakly if at all. This confirms the expected lack of the galactose al,3 galactose (GAL) epitope in 99 such mice. In contrast, peripheral blood cells of normal mice (mouse 129 and mouse 19) of the same strain display clear binding of anti-Gal antibodies.

III. TEST FOR EXPRESSION OF THE GAL EPITOPE IN HOMOZYGOUS KNOCKOUT MICE USING IB 4 LECTIN WITH FACS

IB

4 Lectin has an exclusive affinity for terminal a- D-galactosyl residues, and is demonstrated below to be useful for characterizing the knockout mice.

A. Solutions 1. 4% paraformaldehyde (see above) 2. Mouse Tonicity PBS (MT-PBS) Na 2

HPO

4 2.27g NaH 2

PO

4 2H20 0.62g NaC1 8.7g Make up to 1 liter with DDW 3. Dead Cell Removal Buffer (DCRB): g Sorbitol -7.6 g Glucose monohydrate, (6.93 g if anhydrous)

S.

-12.5 ml KDS/BSS -Make up to 100 ml with DDW -Filter, store at 4 0

C

-Open only under sterile conditions 4. KDS/BSS (Mouse Tonicity, Hepes Buffered Balanced Salt Solution pH 7.2) (see above) Red cell lysis buffer (see above) 6. KDS/BSS/2%HSA/0.02%azide (see above) 7. Hanks Balanced Salt Solution (Ca and Mg above) free) (see B. Methods 1. Remove spleen, hold with curved forceps and collect splenocytes by injecting with a 27 gauge needle bent at 90 0 C, injecting (2.5 ml syringe) 100-200 100 ul buffer into the spleen two or three times. Using the flat surface of the bent needle massage cells out of holes made in spleen. Repeat injections and removal of cells until no cells remain in capsule.

2. Transfer splenocytes to 10ml tube and centrifuge to pellet cells (500xg, 7 min, 4 0

C).

3. Remove supernatant and add 3ml red cell lysis buffer pre-warmed to 42 0 C; incubate for several minutes or until cells have lysed. Underlay with Iml HIFCS (heat inactivated fetal calf serum) and stand on ice 5 minutes. Top to 10ml with KDS BSS/10% HIFCS.

4. Centrifuge as above.

5. Resuspend cells in 3ml dead cell removal buffer; mix well with pipette.

15 6. Pass through a glass pipette plugged with cotton wool and collect cells into a 10ml tube. Don't force cells through, allow to drain under gravity.

7. Underlay cells with 1 ml BSS/10% HIFCS.

8. Centrifuge as above.

20 9. Remove supernatant.

Centrifuge as above.; repeat steps 4 11. Add 0.5 ml cold 4% paraformaldehyde (PFA).

12. Incubate on ice for 5 min with intermittent mixing.

25 13. Add 2 ml ice cold HBBS and centrifuge as 9 above.

14. Repeat washings with 2ml and then 1ml

HBBS.

Resuspend cells in 100ul KDS/BSS/2% HSA/01.% NaN 3 transfer to V bottom FACS tubes.

16. Add FITC IB4 lectin (Sigma, Cat. No. L 2895), 50ul at 20ug/ml, or 50ul HBBS; incubate on ice for min.

17. Add 2ml KDS/BSS/2% HSA/0.1% NaN 3 spin cells as above.

101 18. Resuspend cells in 300ul KDS/BSS/2% HSA/0.1% NaN3.

19. Transfer samples to plastic round-bottom FACS tubes; samples are now ready for analysis; keep on ice.

Analyse on FACS scanner using peripheral blood lymphocyte setting.

C. Samples 1. Mouse 129 splenocytes alone 2. Mouse 129 splenocytes IB4 lectin 3. human PBL alone 4. Human PBL IB4 lectin D. Results Results of these experiments are given below: mean median peak fluorescence fluorescence fluorescence channel channel channel (log scale) (log scale) (log scale) 15 splenocytes alone 1 1 1 (autofluorescence) mouse 19 252 58 16 (wild type) splenocytes mouse 21 3 2 1 (KO mouse) splenocytes The results demonstrate that IB 4 lectin binds spleen cells of the homozygous al,3 galactosyltransferase gene targeted (Gal KO) mouse (mouse 21) very weakly if at all. This confirms the expected lack of the galactose al,3 galactose (GAL) epitope in such mice. In contrast, peripheral blood cells of a normal mouse (mouse 19) of the same strain binds IB 4 lectin strongly.

102 IV. IMMUNOHISTOLOGICAL ASSESSMENT OF MOUSE TISSUES FOR THE PRESENCE OF THE GAL EPITOPE USING ANTI-GAL

ANTIBODIES.

A. Reagents 1. TBS: Tris Buffered Saline NaCI 8g KC1 0.2g Tris base 3g dissolve in 800ml distilled water. Adjust pH to 8.0 with 1 M HC1. Adjust volume to 1L. Sterilise by autoclaving. Store at RT.

Blocking buffer: TBS 2% bovine serum albumin (BSA) rabbit serum: 15 3. Peroxidase conjugates: DAKO (Denmark) peroxidase (POD) conjugated to rabbit anti-human IgG (fragment) and DAKO (Denmark) peroxidase (POD) conjugated to rabbit anti-human IgM (fragment).

Conjugates were both separately pre-absorbed on mouse liver powder at 4 0 C overnight, then centrifuged at 18,000xg for 10 minutes in a Biofuge and then at 30 psi for 30 min in a Beckman airfuge. Conjugated antisera were diluted 1/50 in 2% blocking buffer (TBS 2% BSA 25 2% rabbit serum) with 16% normal mouse serum.

4. Mouse liver powder preparation: As modified from Antibodies, a Laboratory Manual Ed Harber and David Lane, Cold Spring Harbour Laboratories (1988) p663: a) Prepare a fine suspension of mouse liver in mouse tonicity phosphate buffered saline (MT-PBS).

Mash liver through a sieve with a 5 ml plunger. Discard any fibrous tissue. One gram of tissue should be resuspended in approximately 1 ml MT-PBS.

103 b) Transfer the tissue/saline suspension to ice for 5 min.

c) Add 8 ml of acetone (-20 0 C) (Univar 6, Ajax Chemicals) for 10 minutes. Mix vigorously.

Incubate on ice for 30 minutes with occasional vigorous mixing.

d) Collect the precipitate by centrifugation at 10,000g (9,000 rpm in Sorvall RC-5B refrigerated superspeed centrifuge). Spin for 10 minutes.

e) Resuspend the pellet with fresh acetone 0 C) and mix vigorously. Allow to sit on ice for minutes.

f) Centrifuge at 10,000g for 10 minutes.

Transfer the pellet to a clean piece of filter paper.

15 Spread the precipitate and allow to air-dry at room temperature.

g) After the pellet is dry, transfer it to an airtight container. Remove any large pieces that will not break into a fine powder. Dessicate and store at 20 4 0

C.

Yield as approximately 10-20% of the original wet weight.

To use acetone powders, add to a final concentration of Incubate for 30 min at 4 0

C.

Spin at 10,000g for 10 minutes. (13,000 rpm in Biofuge) 25 5. DAB/H 2 0 2 /Imidazole: Peroxidase substrate: 3,3'-Diaminobenzidine tetrahydrochloride (DAB) (Sigma, Missouri) 1 tablet DAB (take out of fridge 10 min before use) 1 tablet urea H 2 0 2 (Sigma, Missouri) add to 15 ml tris HCL, pH 7.6 0.01M imidazole (0.0102g), (Sigma, Missouri) make up immediately before use 6. Tris HCL: 104 1.211g Tris in 200ml double distilled water pH 7.6 7. Animal serum sources: Mouse and rabbit sera were obtained in-house (St. Vincent's Hospital, Dept, of Clinical Immunology).

Sheep serum was obtained from the University of Melbourne Veterinary Clinic and Hospital, Werribee, Australia.

8. Harris Haematoxylin: Haematoxylin C.I. 75290 (BDII, Poole, U.K.

#34037) Absolute ethanol 200ml Potassium alum 200g double distilled water 2000ml 15 glacial acetic acid Preparation: i. Dissolve haematoxylin in absolute ethanol 2. Heat to dissolve alum in double distilled water 20 3. Mix solution 1 and 2 4. Immediately before use add 80 ml 1% sodium iodate and 80 ml glacial acetic acid

S

9. Scott's Tap Water: Sodium hydrogen carbonate 14 g MgSO 4 80 g Tap water 4 litres B. Methods 1. Cut 4 um sections of the relevant tissue on cryostat 2. Tissue should be free of cracks 3. Air dry slides for 30 min 105 4. Apply 10% blocking buffer at room temp in humidified chamber, 60 min Remove blocking antibody with tissue made to fine point 6. Apply 1st antibody, anti-GAL, or 2% blocking buffer as control, 50ul, ensure no air bubbles and incubate at room temp in humidified chamber for 30 min 7. Wash off with Tris buffered saline (TBS) 3 times 2 minutes washes 8. Apply second antibody 1:50 peroxidase

(POD)

conjugated rabbit anti-human IgG and IgM (DAKO, Denmark); incubate 30 min at room temp in :humidified chamber 15 9. Wash off with Tris buffered saline (TBS) 3 times 3 minute washes 10. Wash off with TBS as above 11. Incubate DAB/H 2 0 2 /imidazole for 10 minutes 12. Wash in water 13. Stain with haemotoxylin C 10 seconds 14. Wash in water 15. Place in Scotts tap water for 15 seconds 16. Wash in water 17. Wash in absolute alcohol (x3) (Univar 214, Ajax 25 chemicals) 18. Wash in absolute xylene (x3) (Univar 577, Ajax chemicals) 19. Coverslip using automatic coverslip machine (Tissue Tek) Controls: 1. Buffer only POD conjugated rabbit anti-human IgM (negative)

I

106 2. Buffer only POD conjugated rabbit anti-human IgG (negative) 3. Human kidney (negative) 4. Pig renal cortex (positive) Samples: Mouse 129 SV (control) mouse 9 (Gal Knockout) mouse 21 (Gal Knockout) kidney kidney kidney 0O S

S.

S.

S r

S..

S S

S.

*SOS

5@ 5

S

9e e 055455 C. Results

SKIDNEY

GLOMERULI ENDOTHELIUM comments MOUSE 129 POSITIVE

POSITIVE

anti-IgM MOUSE 9 NEGATIVE NEGATIVE weak adventitial anti-IgM staining MOUSE 21 NEGATIVE NEGATIVE weak adventitial anti-IgM staining MOUSE 129 POSITIVE

POSITIVE

anti-IgG MOUSE 9 NEGATIVE

NEGATIVE

anti-IgG MOUSE 21 NEGATIVE

NEGATIVE

anti-IgG POD conjugated ALL

ALL

antibody alone NEGATIVE

NEGATIVE

These results indicate that human anti-Gal IgG and IgM antibodies do not bind kidney tissue of the al,3 galactosyltransferase gene targeted (Gal KO) mice (mouse 21 and mouse This confirms that lack of the galactose al,3 galactose (GAL) epitope in the gene targeted (KO) mice. In contrast, these antibodies react strongly with the endothelium of blood vessels and the glomeruli of a normal mouse of the same strain (129).

N

107 V. IMMUNOHISTOLOGICAL EXAMINATION OF MOUSE TISSUES USING IB 4

LECTIN

A. Reagents 1. Blocking buffer: TBS 2% BSA 10% sheep serum 2. FITC IB 4 (Sigma, Missouri, USA #L-2895) 1 mg diluted in 1 ml HBBS to give stock solution, then dilute to final volume of 20 ug/ml in TBS 2% BSA 2% sheep serum 3. Peroxidase anti-FITC Boehringer anti-fluorescein POD Fab fragments; dilute 1/300 in 2% blocking buffer 4. DAB/H 2 0 2 /Imidazole see above 5. Tris HCL see above 6. Animal serum sources see above 15 7. Harris Haematoxylin see above 8. Scott's Tap Water see above B. Methods 1. Preparation of Sections; same as Section 4B, steps 1-7 above.

2. Apply 50 pi FITC conjugated IB4 (Sigma 1-2894) 20 pg/ml, incubate in a humidified chamber for 30 minutes.

S. 3. Wash with TBS, 3 minutes (x3).

4. Apply 50 pl per oxidase conjugated anti FITC Fab fragments (Boehringer Mannheim), diluted with TBS 2% BSA 2% sheep serum. Incubate for 30 minutes in humidified chamber.

Wash with TBS, 3 minutes (x3).

6. Processing for microscopy same as Section

IVB

steps 14-22.

Controls 108 1. Buffer only POD anti-FITC 2. Human kidney 3. Pig renal cortex Samples 1st Experiment (negative) (negative) (positive) 1.

2.

3.

4.

Mouse 129 SV normal mouse mouse 6 wild type mouse 7 heterozygote KO mouse 9 homozygous KO heart liver kidney lung heart liver kidney lung heart liver kidney lung heart liver kidney lung Samples 2nd Experiment 1. mouse 19 wild type heart liver kidney lung 2. mouse 20 heterozygote KO heart liver kidney lung 3. mouse 21 homozygous KO heart liver kidney lung a.

109 C. Results Kidney a a a.

*aa.

a.

a 9* a a a a. a a a a.

Liver ENDOTHELIUM BILE DUCT 1l29 MOUSE POSITIVE

POSITIVE

1MOUSE POSITIVE

PSTV

MOUSE 7 POSITIVE

PSTV

MOUSE 9 NEGATIVE

NGTV

MOUSE 19 POSITIVE

PSTV

MOUSE 20 POSITIVE

PSTV

MOUSE 21 NEGATIVE

NEGATIVE

anti-FITC alone ALL

ALL

NEGATIVE

110 Heart *5 ENDOTHELIUM PERINUCLEAR

ENDO-

MYOCARDIUM

129 MOUSE POSITIVE POSITIVE

POSITIVE

MOUSE 6 POSITIVE POSITIVE

POSITIVE

MOUSE 7 POSITIVE POSITIVE

POSITIVE

MOUSE 9 NEGATIVE NEGATIVE

NEGATIVE

MOUSE 19 POSITIVE POSITIVE

POSITIVE

MOUSE 20 POSITIVE POSITIVE

POSITIVE

MOUSE 21 NEGATIVE NEGATIVE

NEGATIVE

anti-FITC alone ALL NEGATIVE ALL NEGATIVE ALL NEGATIVE Lung ENDOTHELIUM BRONCHI

PARENCHYMA

129 MOUSE POSITIVE POSITIVE

POSITIVE

MOUSE 6 POSITIVE POSITIVE

POSITIVE

MOUSE 7 POSITIVE POSITIVE

POSITIVE

MOUSE 9 NEGATIVE NEGATIVE

NEGATIVE

MOUSE 19 POSITIVE POSITIVE

POSITIVE

MOUSE 20 POSITIVE POSITIVE

POSITIVE

MOUSE 21 NEGATIVE NEGATIVE

NEGATIVE

I

anti-FITC alone ALL NEGATIVE ALL NEGATIVE ALL NEGATIVE These results indicate that IB 4 lectin does not bind kidney, heart, liver or lung tissue of the al,3 galactosyltransferase gene targeted (Gal KO) homozygous mice (mouse 21 and mouse This confirms the lack of the galactose al,3 galactose (GAL) epitope in the gene targeted (KO) mice. In contrast these antibodies react strongly with the tissues of a normal mouse and heterozygous KO mice (mouse 129, mouse 6, mouse 7, mouse 19, mouse 20) of the same strain.

111 VI. RESISTANCE OF SPLEEN CELLS FROM KNOCKOUT MICE TO LYSIS BY HUMAN SERUM Lysis of spleen cells by human serum was tested through use of a 51 chromium release assay. See in general Example 4, above.

A. Preparation of Mouse Splenocytes Shortman, K.J. et al, Immunological Methods. 1:273-287 (1972).: -Dissect out spleen, avoid damaging outer membranes and carefully remove mesentery tissue and fat.

-Place in petri dish, with 1 ml RPMI 1640 (Gibco BRL) Heat-inactivated foetal calf serum (HI-FCS). (Heatinactivation 40 Min at 56 0

C).

-Gently tease out cells into petri dish, collect and Scentrifuge 500xg, 5 min, 4oC 15 -Remove RPMI/10% HIFCS, gently resuspend cells in 3 ml 0.9% NH4C 1 (0.168M), using a Pasteur pipette. (Use Pasteur pipettes or wide-bore pipettes for all re suspension and transfer procedures) -Transfer to 10 ml tube, underlay with 1 ml HIFCS, stand 20 on ice, 5 min.

-Transfer supernatant to clean tube, centrifuge 500xg, 7 min, 4 0

C

-Discard supernatant, re-suspend cells in 3 ml dead cell removal buffer, mix well with pipette.

i 25 -Pass through cotton wool plug in glass pipette (under gravity, do not force through), collect cells into 10 ml tube.

-Underlay cells with 1 ml HI-FCS.

-Centrifuge 500xg, 7 min, 4 0

C

-Remove supernatant, re-suspend cells in 50 p1 RPMI, HI-FCS. Store cells on ice.

B. Preparation of Serum: Human Collect whole blood from a pool of normal donors; allow to stand at room temp. for 2 hours.

112 Wring the clot with an 'Orange stick'; spin Remove and pool serum. Store half at -70 0 C in 3 ml aliquot's (normal human serum); heatinactivate the other half, see below.

Fetal calf serum purchased from Gibco BRL, and stored at -20 0

C.

C. Cell Counting: 1. Add 5 ~l cells to 95.0 il RPMI, 10% HI-FCS 2. Remove 10 Al cells, add 10 il Acridine 10 Orange/Et Br solution, (Lee, S.K. et al. Eur J. Immunol.

1975. 5: 259-262) 3. Count cells, (viable green, deads orange).

4. Cell viability should be approx. 90-100 5. Calculate cell number.

15 D. 51 Chromium Labelling: Cell Type Incubation conditions Time Amount 51 Cr/107 cells Freshly prepared cells: -2 hours -150-300 Cl splenocytes or 20 lymphocytes) Cultured Cells: o o a°oo..

f e -1 hour -100 AC1 Labelling: Combine: cells (2 X 107) 51 Cr) Sodium Chromate in 0.9% NaCl solution (the volume added depends on cell type as indicated above and on the specific activity of the 51 Cr) Sodium Chromate).

RPMI/2% HIFCS up to a total of 200 1l Incubate at 370C for time shown above with gentle agitation every 15 min.

113 E. Washing -Place 4ml HI-FCS into 10ml tube and carefully layer labelling reaction on top with a swirling motion; centrifuge 5 min, 500xg, 4 0

C.

-Remove top two layers with a careful circular motion using a glass pipette.

-Resuspend cells in 1ml RPMI/2% HI FCS -Pellet cell suspension through another 4 ml HI FCS -Resuspend cell pellet in iml RPMI/2% HI FCS, store on ice.

F. Release Assay: -Perform assay in 96-well microtire plate (ICN-

FLOW).

I

-Assay should be set up in quadruplicate.

15 -Assay is performed in a total volume of 180 il.

Assay: NHS *HI-NHS 16% SDS CELLS RPMI/2% HIFCS MAX Release 90p1 22.5pl 251l 42.5gl SSpont.Release 90 25 5% NHS 9j1 81 25 NHS 18 72 25 20% NHS 36 54 25 30% NHS 54 36 25 NHS 72 18 25 50% NHS 90 25 *HI heat inactivated -All volumes indicated are in il -Reaction components are added to the plate in the order: RPMI, Serum and 51 Cr-labelled cells.

-Cover plate with plate-sealer -Incubate, 4 hours, 37 0

C.

-Spin plate, 1500 rpm, 5 min.

-Remove plate-scaler, remove 80 il from each wall, count released chromium on gamma counter.

-Calculate specific lysis for each well according to the formula: Specific Lysis (Test cpm Spontaneous release cpm) X 100 (Maximal release cpm Spontaneous release cpm) 114 Calculate mean and standard deviation for each experimental point. Graph Human serum (X axis) against Specific lysis (Y axis) for each type of cell (wild type, heterozygote KO and homozygous KO) The results of these experiments are depicted in Figure 25. The results indicate that spleen cells from a homozygous knockout mouse are relatively resistant to lysis by human serum, in comparison to spleen cells derived from mice heterozygous for the interrupted allele or from wild-type mice.

EXAMPLE 16 Generation of Knockout Animals Through Microinjection of Eggs Transgenic animals are generated routinely 15 by microinjection of DNA into the pronuclei of fertilised eggs. Generally this technology results in the random integration of the transgene in the genome. However, conventional transgenic technology has resulted in homologous recombination between the injected transgene 20 and the endogenous gene. See, for example, Brinster et al., Proc. Nat. Acad. Sci. USA 86: 1087-91 (1989).

Described below are procedures for inactivating the a- 1,3-Gal T gene in pigs'through microinjection of eggs with gene targeting constructs.

25 I. GENE TARGETING CONSTRUCTS The frequency of homologous recombination in embryos is improved if the gene targeting constructs are prepared with isogenic DNA. Therefore the "knock out" constructs are prepared from DNA isolated from the boar used to fertilize the oocytes used for microinjection. DNA is isolated from the tail or ear tissue, and genomic fragments from both a-l,3-Gal T alleles of the boar, encompassing exons 8 9 are cloned using long range PCR or conventional genomic library technologies. Clones for each of the a-l,3-Gal T alleles 115 are identified using restriction fragment length polymorphism identification and DNA sequencing.

Constructs to target both alleles are made by interrupting the coding sequence of exon 9, either by deletion or by inserting a heterologous DNA fragment.

The constructs contain at least 8 kb of homologous DNA to promote efficient homologous recombination.

Various approaches can be used to detect gene targeting events, depending on the strategies used in designing the knockout constructs. Several such approaches, and the corresponding strategies for construction of constructs, are provided below: a) PCR of Genomic DNA: Homologous DNA on one side of the interrupting

DNA

15 fragment is constructed to be less than 1 kb, allowing PCR amplification of a short diagnostic fragment.

(Amplification of small fragments generally is relatively efficient).

b) Reverse Transcription/PCR: 20 A deletion of about 100 bp within exon 9 is made, allowing synthesis of a shortened a-1,3-Gal T mRNA in correctly targeted cells. The shortened mRNA is detected by RT/PCR, using primers that amplify a fragment extending from exon 8 and encompassing the deletion site.

25 c) Green Fluorescent Protein (GFP) gene expression: GFP is a protein from the bioluminescent jelly fish Aequorea victoria. It absorbs blue light (395 nm) and fluoresces to emit green light (509 nm). GFP is a useful marker for gene expression. Chafie et al., Green Fluorescent Protein as a Marker for Gene Expression.

Science 263: 802-5 (1994). The a-1,3-Gal T gene is interrupted within exon 9 by in-frame insertion of the GFP coding region. Expression of the GFP gene (with 116 resulting fluorescence at 509 nm) is driven by the a-1,3- Gal T gene promoter in correctly targeted cells.

II. GENERATING EMBRYOS FOR MICROINJECTION Fertilized embryos are generated as described by Nottle et al., (1993). Proc Aust Soc for Reproductive Biol 26, 33. The protocol involves: a)Sperm from the boar providing DNA for the targeting construct is collected and stored frozen in liquid

N

2 b) Superovulation of donor gilts: Gilts are mated at the second oestrus, and aborted between days 25-40 days of gestation to synchronise the subsequent oestrus cycles. Abortion is achieved by intramuscular injection of 1 mg cloprostenol 15 (a prostaglandin F2a analogue), followed by a second mg injection 24 hours later. Gilts are superovulated by injection of 1000 i.u. equine chorionic gonadotrophin S(eCG) or pregnant mare serum gonadotrophin at the time of the second cloprostenol injection, and a subsequent injection 72 hours later of 500 i.u. human chorionic :e gonadotrophin (hCG).

c) Fertilization: Superovulated gilts are artificially inseminated 20-30 hours after the hCG injection, followed by a second 25 insemination 2-4 hours later, with semen from the boar that provided DNA for the targeting construct.

d) Embryo collection: Embryos are collected surgically 50-56 hours after hCG injection prior to fusion of the pronuclei. oviducts are flushed with 15-20 ml phosphate saline buffer containing 1% fetal calf serum. One-cell embryos are recovered by searching oviductal flushings using low magnification microscopy.

117 III. MICROINJECTION OF EMBRYOS Embryos are centrifuged at 12000 x g for 8 min to stratify the cytoplasm and allow the pronuclei to be visualised, and held in Dulbecco's Minimal Essential Medium with 25 mM Hepes and 5 mg/ml bovine serum albumin.

Pronuclei are injected, using differential interference contrast optics, with 4-10 picolitres of DNA (10 ng/il) in PBS. Gene targeting with isogenic DNA is maximized by coinjecting both allelic constructs derived from the boar into the male pronucleus.

oa IV. TRANSFER OF INJECTED EMBRYOS TO RECIPIENT

GILTS

The oestrus cycles of recipient gilts are synchronized with those of donors. The recipients are mated and aborted using the protocol described above, and injected with 500 i.u. eCG. Injected embryos are transferred surgically (20-40 per oviduct) to recipients on the same day that they are collected from donor gilts.

V. SCREENING FOR HOMOLOGOUS RECOMBINATION Homologous recombinants can be detected by analysis of tissue from the born piglets. Screening procedures involve PCR technology, the precise strategy depending on the design of the gene targeting construct.

Because many a-l,3-Gal T mRNA molecules are synthesized 15 from a single a-l,3-Gal T gene in expressing cells, the RT/PCR approach can be more sensitive than PCR amplification of genomic DNA. The RT/PCR screening strategy relies on successful transcription of the interrupted gene and relative stability of the shortened 20 mRNA.

Alternatively, constructs that promote expression of heterologous genes (eg: GFP) in correctly targeted cells allow embryos to be screened at the blastocyst stage for marker gene expression

GFP

expression can be detected by measuring fluorescence within blastocysts at 509 nm). The microinjected embryos are cultured in vitro until blastocyst development, screened for fluorescence, and fluorescing embryos transferred into recipients.

Claims (17)

1. A DNA construct when used for inactivating an a-1, 3 galactosyltransferase gene, comprising a DNA sequence for porcine a-1, 3 galactosyltransferase, wherein said DNA sequence carries a disruption of said DNA sequence such that functional a-1, 3 galactosyltransferase is not encoded by said DNA sequence.

2. The DNA construct of claim 1, wherein said disruption is within exon 4, exon 7, exon 8 or exon 9 of the porcine a-1, 3 galactosyltransferase gene.

3. The DNA construct of claim 1, wherein said DNA sequence comprises a selectable marker selected from the group consisting of the neo R gene, and the hygR gene.

4. The DNA construct of claim 3 wherein said selectable marker is bordered at the 5' and the 3' ends by FRT DNA elements, and wherein stop codons for each of the three reading frames have been inserted 3' to the selectable marker. A DNA construct when used for inactivating an a-1, 3 galactosyltransferase gene, comprising a DNA sequence for murine a-1, 3 galactosyltransferase, wherein said DNA sequence carries a disruption of said DNA sequence such that functional a-1, 3 galactosyltransferase is not encoded by said DNA sequence.

6. The DNA construct of claim 5, wherein said disruption is within exon 4, exon 7, exon 8 or exon 9 of the murine a-1, 3 galactosyltransferase gene. 30/7/99 120

7. The DNA construct of claim 5, wherein said DNA sequence comprises a selectable marker selected from the group consisting of the neo gene, and the hyg gene.

8. The DNA construct of claim 7, wherein said selectable marker is bordered at the 5' and 3' ends by FRT DNA elements, and wherein stop codons for each of the three reading frames have been inserted 3' to the selectable marker. A method for generating a mammalian totipotent cell having at least one inactivated a-1, 3 galactosyltransferase allele, said totipotent cell derived from a mammalian species having a functional a-1, 3 galactosyltransferase gene, comprising: providing a plurality of cells characterized as totipotent cells of said mammalian species; introducing into said totipotent cells a DNA construct effective for inactivating said a-1, 3 galactosyltransferase gene, said DNA construct comprising a DNA sequence for a-1, 3 galactosyltransferase, wherein said DNA sequence carries a disruption of said DNA sequence such that functional a-1, 3 galactosyltransferase is not encoded by said DNA sequence; and identifying a totipotent cell or mitotic descendants of a totipotent cell having at least one inactivated a-1, 3 galactosyltransferase allele. The method of claim 9 in which said totipotent cell is a murine ES cell.

11. The method of claim 9 in which said totipotent cell is a murine egg. 121

12. The method of claim 9 in which said totipotent cell is a porcine ES cell.

13. The method of claim 9 in which said totipotent cell is a porcine PGC.

14. The method of claim 9 in which said totipotent cell is a porcine fertilized or unfertilized egg. A method for generating a non-human mammal lacking a functional a-1, 3 galactosyltransferase gene, said mammal belonging to a species having a functional a-1, 3 galactosyltransferase gene, comprising: providing a mammalian totipotent cell having at least one inactivated a-1, 3 galactosyltransferase allele, said totipotent cell derived from a mammalian species having a functional a-1, 3 galactosyltransferase gene; manipulating said totipotent cell such that mitotic descendants of said cell constitute all or part of a developing embryo; recovering a neonate derived from said embryo; and raising and breeding said neonate to obtain a non-human mammal homozygous for said inactivated a-1, 3 galactosyltransferase and lacking the GAL epitope as determined by failure of somatic cells of said mammal to bind anti-GAL antibodies and IB4 lectin and resistance of said cells to lysis by human serum.

16. The method of claim 15, wherein said totipotent cell is a murine ES cell and said manipulating comprises injecting said ES cell into the blastocyst cavity of a murine blastocyst and implanting said injected blastocyst into a murine recipient female. 30/7/99 122

17. The method of claim 15, wherein said totipotent cell is a murine egg, and said manipulating comprises implanting said egg into a murine recipient female.

18. The method of claim 15, wherein said totipotent cell is a porcine ES cell and said manipulating comprises injecting said ES cell into the blastocyst cavity of a porcine blastocyst and implanting said injected blastocyst into a porcine recipient female.

19. The method of claim 15, wherein said totipotent cell is a porcine ES cell and said manipulating comprises injecting said ES cell into a porcine morula. 00

20. The method of claim 15, wherein said totipotent cell is a porcine ES 00cell and said manipulating comprises co-culture ofs said ES cell with a zonenucleated Sporellucine zygoda-disrupted porcine morula.

22. The method of claim 15, wherein said totipotent cell is a porcine ellg and said manipulating comprises implanting said egg into a porcine recipient female. Dated this 16th day of July 1998 BRESAGEN LIMITED AND ST VINCENT'S HOSPITAL (MELBOURNE) LIMITED Patent Attorneys for the Applicants: PETER MAXWELL ASSOCIATES PETER MAXWELL ASSOCIATES