WO1996017926A1

WO1996017926A1 - Transgenic animal lacking native amyloid precursor protein

Info

Publication number: WO1996017926A1
Application number: PCT/US1995/015672
Authority: WO
Inventors: Hui Zheng; Howard Y. Chen; Myrna E. Trumbauer; Leonardus H. T. Van Der Ploeg
Original assignee: Merck & Co., Inc.
Priority date: 1994-12-05
Filing date: 1995-12-01
Publication date: 1996-06-13
Also published as: JP2000504202A; EP0799305A4; CA2206789A1; EP0799305A1

Abstract

The present invention relates to a transgenic nonhuman animal lacking native amyloid precursor protein. The transgenic mouse of the invention may be used in the study of Alzheimer's Disease and disorders involving the central nervous system.

Description

- 1 -

TITLE OF THE INVENTION

TRANSGENIC ANIMAL LACKING NATIVE AMYLOID

PRECURSOR PROTEIN

CROSS-REFERENCE TO RELATED APPLICATION

This is a continuation-in-part of U.S. Serial No. 08/349,334 filed December 5, 1994, now pending.

FIELD OF THE INVENTION The present invention relates to a transgenic nonhuman animal lacking native amyloid precursor protein.

BACKGROUND OF THE INVENTION

Alzheimer's disease (AD) is a neurological disorder that disproportionately affects the population over 65 years of age.

Incidence of the disease increases from less than 1 % at age 60-65, to 5% at age 75, to as high as 47% at age 85. As a result, 60% to 80% of all cases of dementia in persons over age 65 are caused by AD. Afflicted individuals exhibit impaired cognitive function and memory. Neither a suitable diagnostic procedure nor an effective therapeutic treatment exists for AD. Positive identification of AD requires biopsy or autopsy of the brain.

Although the etiology of AD is unknown, genetic, immunological and environmental factors have been implicated in the development of AD. Distinguishing features Of AD include the presence of senile plaques as well as, neurofibrillary tangles and extensive neuronal loss in the neocortex, hippocampus and associated structures. Senile plaques consist of extracellular deposits containing a β-amyloid core surrounded by a halo of dystrophic neurites, glia and astrocytes. β-amyloid deposits are present in neocortex blood vessel walls. The major component of senile plaques is a 4 kDa peptide referred to as Aβ, that is proteolytically cleaved from a larger 120 kDa amyloid precursor protein (APP). Other components of the plaques include ubiquitin, amyloid P, Apo E, interleukin- 1 , and α-1 -antichymotrypsin. In addition to biochemical evidence supporting Aβ involvement in AD, there are strong genetic data which suggest a link between APP and AD. A clue to the location of a gene involved in AD comes from analysis of Down syndrome patients; in these patients trisomy of chromosome 21 is responsible for the early onset of AD. Karyotype analysis of Down syndrome patients mapped the gene involved to the upper portion of the long arm of chromosome 21. The region involved encodes several genes, including the APP gene. The early onset (~ age 35) of AD in Down syndrome patients suggests that an increase in the gene dosage of the responsible marker(s) on the long arm of chromosome 21 may contribute to the neuropathology noticed in most AD patients.

Although the majority of AD cases appear sporadic, several cases of early onset familial AD (FAD) have been reported. Genetic analysis of FAD families has established that the disorder is inherited as a dominant autosomal gene defect, which maps to the long arm of chromosome 21 and is closely linked to the APP gene. These findings are consistent with genetic data obtained from the analysis of Down syndrome patients. Several FAD families have also been identified in which an early onset of AD is strictly correlated with the presence of a mutation in exon 17 of the APP gene at amino acid 717 (Val-Ile). This mutation within the transmembrane spanning domain of the APP co- segregates with FAD. Since the families afflicted with APP717 FAD are of different ethnic origins (English, Japanese and Canadian), evidence for the involvement of the FAD gene in these cases of AD is strong. The mutation is absent from control individuals, in sporadic AD patients, in Down syndrome patients, in late onset familial AD, and also in most other cases of early onset FAD. Several additional mutations in the APP gene have been identified that can explain the occurrence of AD in other FAD families. The genetic evidence in the five distinct early onset APP717 FAD families strongly supports the hypothesis that the APP717 gene in these FAD families is directly positioned in the pathway of AD progression. The APP gene is approximately 400 kb in length and encodes a glycosylated, transmembrane protein which may be involved in cell-cell interaction. The APP gene has at least 18 exons that create at least 5 distinct APP transcripts by alternative splicing. The predominant transcripts encode proteins of 695, 751 and 770 amino acids (these major forms of APP are referred as APP695, APP751 and APP770, respectively). Transcripts for APP695 are enriched in the brain. Transcripts encoding APP751 and APP770 mRNA species predominate in peripheral tissues. All three isoforms contain the 42 amino acid Aβ domain. APP isoforms 751 and 770 contain an additional 56 amino acid insert encoding the Kunitz type serine protease inhibitor (KPI). APP is proteolytically metabolized by at least two pathways. One pathway involves an α-secretase cleavage site positioned between Lys 16 and Leu 17 of Aβ domain; proteolytic cleavage at this site precludes the formation of an amyloidogenic Aβ entity. The second pathway produces intact, amyloidogenic Aβ (39-42 amino acids) by proteolytic cleavages at the β- and γ-secretase cleavage sites of the full- length APP molecule.

The Aβ laden senile deposits seen in AD patients are also found in aged humans and other aged mammals including non-human primates, polar bears and dogs. However, other aged mammals, such as laboratory rats and mice, do not normally develop Aβ deposits. This could be due to the fact that the three amino acid differences present in the β-amyloid sequence between human and mouse APP prevents mouse Aβ from forming plaques. The lack of a cost-effective, experimental animal model mimicking human pathogenesis hinders the understanding AD neuropathology and developing therapeutics against AD.

Transgenic technology may offer a suitable alternative to this problem. Addition of a gene construct directing high levels of human APP or its components to key regions in the murine central nervous system may cause neuropathological changes resembling the AD phenotype. Attempts to express human amyloid precursor protein segments or the full-length wild type protein in transgenic animals have been successful. Numerous reports exist outlining expression of different wildtype, full-length and truncated APP cDNA isoforms in mouse (Kammesheidt et al., (1992) Proc. Natl. Acad. Sci., 89, 10857- 10861; Sandhu et al., (1991) J. Biol. Chem., _____., 21331-21334; Quon, et al., (1991) Nature, 252, 239-241; Wirak et al, (1991) Science, 253, 323-325; Kawabata et al, (1991 ) Nature 354, 476-478; Patent,

International publication number WO93/02189, Neve, R., Inventor). However, these previous attempts to generate transgenic mouse models for AD have essentially failed. This failure could have been caused by the presence of the endogenous mouse APP gene in the transgenic mouse, which "protects" the human βA4 from depositing in the mouse brain.

Accordingly, it is an object of the present invention to provide a transgenic mouse which does not express mouse amyloid precursor protein. The transgenic mice of the present invention are useful in the determination of the in vivo function of APP and the β- amyloid peptide in the central nervous system and in other tissues. These mice are being bred with transgenic mice expressing the human APP FAD with the aim of producing a strain of mice in which the only APP produced is of human origin. The precise roles of APP in AD is not fully understood at this time. Due to the biological importance of APP in AD and other neurological disorders, the APP gene is an important target for embryonic stem (ES) cell manipulation.

The generation of APP deficient transgenic mice would aid in defining the normal role(s) of APP, and allow an animal model of APP deficiency to be used in the design and assessment of various approaches to modulating APP activity. Such APP modified transgenic mice can also be used as a source of cells for cell culture.

SUMMARY OF THE INVENTION

The present invention relates to a transgenic nonhuman animal lacking native amyloid precursor protein (APP). The transgenic mouse of the invention may be used in the study of Alzheimer's Disease and disorders involving the central nervous system. BRIEF DESCRIPΗON OF THE DRAWINGS

Figure 1 is a genomic map of the mouse APP gene, the location of the cosmid clone isolated from a primary genomic library and different probes and subclones generated from the cosmid clone. ex 1 : exon 1 of the APP gene.

Figure 2 is the predicted modification of the mouse chromosomal APP gene by targeted recombination using replacement vector pHZ 038. The targeting vector (pHZ038) contains from left to right:

1.4 kb of 5' homology with the APP locus; a PGKneo expression cassette inserted in the opposite orientation to the APP gene; a fragment of 7.1 kb homologous to the 3' part of the APP promoter region; and an MC1-TK gene. A 3.8 kb sequence containing the 1.0 kb APP promoter, the first exon and part of the first intron was deleted. The homologies of the endogenous APP gene and the targeting vector are represented as shaded rectangles. Targeted recombination between the vector and the wild-type APP locus results in the deletion of the promoter and exon 1 of the APP gene followed by its replacement with a 1.5 kb neo coding sequence. The probes used for Southern blot analysis were the 1.0 kb Xbal-Bglll fragment (5'-probe), the 0.8 kb Bglll-Ncol fragment (3'- probe), both of which are outside the targeting vector, and the neo sequence. EcoRI was used to differentiate the wild-type and the targeted APP alleles, which generates a 9.0 kb and 6.5 kb fragments by the 5'-probe and a 9.5 kb and a 9.0 kb by the 3'-probe, respectively. R: EcoRI, X: Xbal, B: Bglll, N: Ncol. Pr: mouse APP promoter, El : exon 1 of the mouse APP gene. PGK: phosphoglycerate kinase promoter.

Figure 3 is a Southern hybridization analysis of four targeted embryonic stem (ES) clones having an APP knockout.

ES cell DNA (8 μg) from the wild-type AB2.1 cells and four positive clones (76, 123, 174 and 196) were restriction enzyme digested with EcoRI, electrophoresed on a 0.7% agarose gel, transferred onto a Gene Screen Plus nylon membrane (NEN-Dupont) and hybridized with a 5'-probe (labeled a), a 3'-probe (data not shown), and a neo probe (labeled b). A 6.5 kb diagnostic fragment is detected by the 5'-probe in all the targeted clones in addition to the wild-type 9.0 kb fragment. Probing of the same filter with a neo coding sequence (b) shows that the neo gene is present in the 6.5 kb band and is the only integration event in all the clones. The other two bands at high molecular weight corresponds to a nonfunctional neo sequence introduced by a retrovirus in the parental AB2.1 cells.

Figure 4 is a Southern hybridization analysis of tail DNA from transgenic mice having an APP knockout. Southern analysis of genomic DNA from het. x het. crosses yielded the expected number of mice homozygous for the disrupted APP allele.

Genomic DNA isolated from the tails of two week old pups generated from crosses of heterozygous mice was digested with EcoRI, blotted onto filters, and hybridized with the 5'-probe. +/+: wild-type; +/-: heterozygotes; -/-: homozygous APP deficient mice.

Figure 5 is a Northern hybridization analysis for the determination of APP transcripts in the knockout and wild-type control mice. As expected, the brain RNA from the knockout mice did not exhibit any detectable APP expression, whereas wild-type control and heterozygous animals showed a significant amount of APP activity.

Total RNA ( 20 μg) from wild-type (+/+), heterozygous(+/-) and homozygous (-/-) APP mice were isolated (2 mice each) from brain using the RNAzol B method (Biotecx

Laboratories, Inc.) and probed with an APP695 cDNA probe. The lower panel shown in each case is a control hybridization with mouse β- actin cDNA to show that identical amounts of RNA were loaded in each lane. - 7 -

DETAILED DESCRIPTION OF THE INVENTION

A 39 to 43 amino acid β-amyloid peptide is the major component of the neuritic plaques characterizing Alzheimer disease, β- amyloid is derived from a larger amyloid precursor protein (APP).

The APP mRNA undergoes alternative splicing to generate several isoforms, encoding proteins that range from 695 to 770 amino acid residues. Among these, APP695 is expressed predominantly in neurons and APP751 and APP770 can be detected in all the tissues examined. Five different types of point mutations have now been identified in the human APP gene, causative of familial early onset Alzheimer's disease (FAD) in several unrelated families. These affected families provide the strongest evidence yet for the notion that APP processing and the β- amyloid peptide serve a central role in Alzheimer disease progression. APP is one of the most abundant proteins in the brain and the β-amyloid peptide is secreted in cerebrospinal fluid (CSF) of healthy individuals and AD patients. The functions of the APP and β-amyloid in vivo are obscure. The APP has been implicated as a growth factor in vitro in fibroblast cultures. The β-amyloid peptide has been shown to have neuroprotective and neurotoxic actions, dependent on the cell line and protein preparations tested. A Kunitz protease inhibitor domain in the N-terminal portion of the APP may serve a role in regulating protein half-life.

We wished to inactivate the APP gene in order to evaluate its role in mouse development and in the central nervous system. In addition, in order to generate a murine model of FAD, APP deficient mice may be useful as acceptors of the human FAD protein. Mouse APP is overall conserved when compared to human APP but differs in three potentially essential amino acid residues of the β-amyloid domain. The murine APP gene is about 400 kb in size and is encoded by at least 18 exons. The murine APP gene was inactivated by deleting its promoter and first exon, which encodes the ATG translation initiation codon. To target the APP gene in murine ES cells, a positive-negative selection strategy was used. The targeting vector pHZ038 (Figure 2) encoded 8.5 kilobases (kb) of DNA derived from the 5' end of the APP gene. A 3.8 kb sequence encoding the APP promoter and the first intron was deleted from this vector and replaced with a positive selectable marker, PGKneo (neomycin-phospho-transferase). A MC1 - TK (thymidine kinase) cassette (labeled HSV-TK in Figure 2) was inserted at the end of the vector for negative selection. Correct homologous recombination between the targeting vector and one of the APP alleles in the ES cells would result in a deletion of the APP promoter and exon 1 , encoding the signal peptide.

The targeting vector was electroporated into AB2.1 ES cells. G418 and FIAU resistant clones were screened by a mini- Southern protocol. A five-fold enrichment was achieved by selecting the cells with FIAU. Six targeted clones were identified and the frequency of targeted recombination versus random integration at the APP locus was 1/160 (Figure 3). Of four clones injected into blastoysts two (no. 76 and 174) transmitted the targeted APP allele to the offspring. Heterozygous matings were set up to produce mice homozygous for the disrupted APP gene.

Homozygous APP knockout mice that resulted from these breedings were generated at expected frequencies (Figure 4). These mice appeared normal and healthy up to 14 weeks of age. Northern blot analysis of RNA isolated from brain using APP695 cDNA as a probe showed that APP mRNA was not produced in mice homozygous for the targeted allele. The APP mRNA level was reduced by approximately 50% in the heterozygous mice as compared to wild-type controls (Figure 5). The present invention utilizes a cloned DNA encoding the

5' portion of the APP gene. Transgenic animals are generated which have an altered APP gene. The alterations to the naturally occurring gene are modifications, deletions and substitutions. Modifications, deletions and substituteions may render the naturally occurring gene nonfunctional, producing a "knockout" animal, or may lead to an APP with altered function. These transgenic animals are critical for drug antagonist or agonist studies, for creation of animal models of human diseases, and for eventual treatment of disorders or diseases associated with APP. Transgenic animals lacking native APP are useful in characterizing the in vivo function of APP. A transgenic animal carrying a "knockout" of APP is useful for the establishment of a nonhuman model for diseases involving APP, and to distinguish between the activities of APP in in vivo and in vitro systems. The term "animal" is used herein to include all vertebrate animals, except humans. It also includes an individual animal in all stages of development, including embryonic and fetal stages. A "transgenic animal" is an animal containing one or more cells bearing genetic information received, directly or indirectly, by deliberate genetic manipulation at a subcellular level, such as by microinjection or infection with recombinant virus. This introduced DNA molecule may be integrated within a chromosome, or it may be extra-chromosomally replicating DNA. The term "germ cell-line transgenic animal" refers to a transgenic animal in which the genetic information was introduced into a germ line cell, thereby conferring the ability to transfer the information to offspring. If such offspring in fact possess some or all of that information, then they, too, are transgenic animals.

The genetic alteration or genetic information may be foreign to the species of animal to which the recipient belongs, or foreign only to the particular individual recipient. In the last case, the altered or introduced gene may be expressed differently than the native gene.

The altered APP gene should not fully encode the same APP as native to the host animal, and its expression product should be altered to a minor or great degree, or absent altogether. However, it is conceivable that a more modestly modified APP gene fall within the scope of the present invention.

A type of target cell for transgene introduction is the embryonal stem cell (ES). ES cells may be obtained from pre- implantation embryos cultured in vitro and fused with embryos (M. J. Evans ai-, Nature 292: 154-156 (1981); Bradley et ah, Nature 309: 255-258 (1984); Gossler et aL Proc. Natl. Acad. Sci. USA 83: 9065-9069 (1986); and Robertson et aj., Nature 322, 445-448 (1986)). Transgenes can be efficiently introduced into the ES cells by a variety of standard techniques such as DNA transfection, microinjection, or by retrovirus-mediated transduction. The resultant transformed ES cells can thereafter be combined with blastocysts from a non-human animal. The introduced ES cells thereafter colonize the embryo and contribute to the germ line of the resulting chimeric animal (R. Jaenisch, Science 240: 1468-1474 (1988)).

Since the proposed APP functions are complex, they must be examined in a variety of ways. One approach to the problem of determining the contributions of individual genes and their expression products is to use isolated genes to selectively inactivate the native wild-type gene in totipotent ES cells (such as those described herein) and then generate transgenic mice. The use of gene-targeted ES cells in the generation of gene-targeted transgenic mice was described 1987 (Thomas et a]_.., Cell 51 :503-512, (1987)) and is reviewed elsewhere (Frohman et al., Cell 56:145-147 (1989); Capecchi, Trends in Genet. 5:70-76 (1989); Baribault et aL, Mol. Biol. Med. 6:481-492, (1989); Wagner, EMBO J. 9: 3025-3032 (1990); Bradley et aj., Bio/Technology 10: 534-539 (1992)). Techniques are available to inactivate or alter any genetic region to any mutation desired by using targeted homologous recombination to insert specific changes into chromosomal alleles. However, in comparison with homologous extrachromosomal recombination, which occurs at frequencies approaching 100% homologous plasmid-chromosome recombination was originally reported to only be detected at frequencies between 10-6 and 10-3 (Lin et al., Proc. Natl. Acad. Sci. USA 82:1391-1395 (1985); Smithies et al, Nature 317: 230-234 (1985); Thomas et ah, Cell 44:419-428, (1986); Song et ah, Proc. Natl. Acad. Sci. USA 84:6820-6824 (1987)). Nonhomologous plasmid-chromosome interactions are more frequent, occurring at levels 105-fold (Lin et ah, Proc. Natl. Acad. Sci. USA 82:1391-1395 (1985)) to 102-fold (Thomas et al-, Cell 44:419-428 (1986); Song et aL, Proc. Natl. Acad. Sci. USA 84:6820-6824 ( 1987)) greater than comparable homologous insertion.

To overcome this low proportion of targeted recombination in murine ES cells, various strategies have been developed to detect or select rare homologous recombinants. One approach for detecting homologous alteration events uses the polymerase chain reaction (PCR) to screen pools of transformant cells for homologous insertion, followed by screening individual clones (Kim el -, Nucleic Acids Res. 16:8887-8903 (1988); Kim et al-, Gene 103:227-233 (1991)). Alternatively, a positive genetic selection approach has been developed in which a marker gene is constructed which will only be active if homologous insertion occurs, allowing these recombinants to be selected directly (Sedivy et al-, Proc. Natl. Acad. Sci. USA 86:227-231 (1989)). One of the most general approaches developed for selecting homologous recombinants is the positive-negative selection (PNS) method developed for genes (such as APP) for which no direct selection of the alteration exists (Mansour et ah, Nature 336:348-352: (1988); Capecchi, Science 244:1288-1292, (1989); Capecchi, Trends in Genet. 5:70-76 (1989)). The PNS method is more efficient for targeting genes which are not expressed at high levels because the marker gene has its own promoter. Nonhomologous recombinants are selected against by using the Herpes Simplex virus thymidine kinase (HSV-TK) gene and selecting against its nonhomologous insertion with the herpes drugs such as gancyclovir (GANC) or FIAU (l-(2-deoxy 2-fluoro-B-D-arabinofluranosyl)-5-iodouracil). By this counter-selection, the number of homologous recombinants in the surviving transformants can be enriched.

As used herein, a "targeted gene" or "Knock-out" (KO) is a DNA sequence introduced into the germline of a non-human animal by way of human intervention, including but not limited to, the above described methods. The targeted genes of the invention include DNA sequences which are designed to specifically alter cognate endogenouos alleles. The methods for evaluating the targeted recombination events as well as the resulting knockout mice are readily available and known in the art. Such methods include, but are not limited to DNA (Southern) hybridization to detect the targeted allele, polymerase chain reaction (PCR), polyacrylamide gel electrophoresis (PAGE) and Western blots to detect DNA, RNA and protein.

The following is presented by the way of examples and is not to be construed as a limitation on the scope of the invention:

EXAMPLE 1

Preparation of a murine embryonic stem cell genomic library

Evidence suggested that homologous genomic targeting in embryonic stem (ES) cells is strongly inhibited (>100x) by subtle base-pair differences in their genomic DNAs (Riele et al., Proc. Natl. Acad. Sci. USA 89:5128-5132 (1992); Deng et al., Mol. Cell. Biol. 12:3365-3371 (1992)). To circumvent this potential problem, genomic libraries were constructed from ES cells grown in the absence of feeder cell layers for the isolation of genes to be used for subsequent ES cell targeting. Genomic libraries were prepared from AB2.1 cells according to the in situ procedure described in (Mudgett et ah, Genomics 8:623-633, (1990)). The cosmid vector sCos-1 was chosen, as it allows both the vector and the insert to be dephosphorylated. This prevents concantamer formation and generally results in genomic libraries of better quality and quantity (up to 5x106 clones per package) than is achieved with other vectors (Evans et ah, Gene 79:9-20, (1989)).

The genomic library was constructed with bacterial hosts SURE. The SURE host line (Stratagene) was used to stably maintain indirect repeats and allow isolation of methylated DNA. EXAMPLE 2

Isolation and characterization of mouse APP cosmid clone

The primary mouse cosmid library of Example 1 was screened using a 1.0 kb Hindlll-PvuTJ fragment located in the promoter of the APP gene as a probe (Izumi ______ Gene 112: 189-195 (1992)).

The conditions for library plating and screening are described by Sambrook et al. in "Molecular cloning: A Laboratory Manual". Cold Spring Harbor Press, (1989). About 6 x 10^ clones were plated onto a 150mm NZYDT/amp (100 μg/ml) plate. A total of 1.2 x 10⁶ independent clones were screened. To reduce the chance of false positives, duplicate lifting was used. The hybridization and washing conditions were maintained at high stringency to reduce the chance of isolating different but homologous genes and related pseudogenes. Only one positive clone was identified which was further purified through secondary and tertiary screenings. The cosmid DNA was prepared using standard conditions (Sambrook et al.. supra).

The cosmid DNA was digested with different restriction endonucleases, electrophoresed on agarose gels, and hybridized with different probes isolated from the plasmid HH5.0 (Izumi et al.. Gene 112: 189-195 (1992)) to map the boundary as well as the promoter and exon 1 of the mouse APP gene (Fig.l).

EXAMPLE 3

Construction of APP gene targeting vector

From the knowledge of the genomic organization of mouse APP gene with regard to restriction sites, promoter and the first exon (Example 2), a gene targeting vector for inactivating the APP gene was prepared using standard cloning techniques (Sambrook et al.. supra): a). A three-way ligation was performed using the 1.4 kb Bglll-Xhol fragment in the 5' portion of the APP cosmid clone upstream of exon 1 (ex 1), the 7 kb Xhol-Bglll fragment in the 3' portion of the cosmid clone, and BamHI digested pKS vector. The resulting plasmid was named pHZ036. b). Plasmid pHZ036 was partially-digested with Xhol and ligated with the 1.5 kb Xhol-Sall fragment of PGKneo. A ligation product with neo inserted in between the two APP fragments was selected and referred to as pHZ037. c). A 2 kb Xhol fragment from pKS-TK was inserted into the Sail digested pHZ037 vector. The resulting plasmid, pHZ038, is the complete construct for targeting of the mouse APP gene.

EXAMPLE 4

Targeted disruption of the APP gene in murine ES cells

The targeting vector used in the APP gene disruption experiments was the pHZ038 vector of Example 3. When this vector recombined with the wild-type APP allele to generate the APP knockout (APP KO), the promoter and exon 1 of the APP gene encoding the ATG translation iniation codon and the signal peptide were deleted (Fig. 2 ). The mouse embryonic stem cell line AB2.1 was electroporated with NotI digested pHZ038 to linearize the plasmid, leaving the pks sequence attached to the end of the TK cassette. All AB2.1 ES cells were cultured on SNL feeder cells as described (Robertson, in Teratocarcinomas and embryonic stem cells, E L Press, pp. 71-112 (1987)). Electroporations were performed with 1x107 ES cells and 25 μg linearized vector in 0.8 ml PBS buffer at 230v, 500 μF using a Bio-Rad Gene Pulser. ES cell transformants were selected with the antibiotic geneticin (Gibco G418: 200 μg/ml active G418) 24 hr post electroporation, and some transformants were counter-selected with FIAU (Bristol Myers Squibb; 0.4 μM) 48 hours later for enhancement of homologous recombinants. Murine leukemia inhibitory factor (LIF; ESGRO, Gibco BRL, Inc.) was used at 200 U/ml. Selection with FIAU resulted in about five-fold fewer colonies as compared to G418 selection alone, thereby enhancing the isolation of targeted transformants. G418- and FIAU -resistant ES clones were isolated, grown up and analyzed by a mini-Southern protocol (Ramirez-Solis, R. et al. Anal. Biochem. 201:331-335, 1992). A total of six targeted clones were identified from 200 double resistant colonies analyzed. Therefore, the frequency of targeted recombination vs. random integration at the APP locus is 1/160. Detailed Southern blot analysis of the targeted clones using 5'-, 3'- and neo probes showed the expected integration pattern both within the APP gene and in the APP flanking region. Integration events other than the targeted recombination could not be detected.

EXAMPLE 5

Injection of APP KO clones into donor blastocvsts All APP-targeted AB2.1 cell lines were characterized by

Southern hybridization analysis to confirm that APP was disrupted. Targeted cell lines which grew normally and did not contain an abnormal proportion of differentiated cells (Robertson, in supra) were then separated from their feeder cells by treating the cell culture with trypsin, allowing the feeder cells to attach for about 30 min, and removing the unattached ES cells. The ES cells were injected into recipient blastocysts. Four APP targeted ES clones (76, 123, 174 & 196) were injected into C57BL/6J recipient blastocysts in separate experiments using techniques described previously (Bradley, A. "Production and analysis of chimeric mice-. In Teratocarcinomas and Embryonic Stem Cells: A Practical Approach", E.J. Robertson, ed.Oxford:IRL Press, (1987), ppl 13-151). The injected C57B1/6J recipient blastocysts were reimplanted into the uteri of day 3 pseudopregnant Tac:SW(fBR) mice and allowed to develop to term. Progeny were screened initially by coat color chimerism, the agouti color (which is the ES cell background strain) being an indicator of ES cell chimerism. APP targeting was confirmed by Southern hybridization analysis performed on genomic DNA isolated from tail samples obtained from these mice. Injection of the APP targeted lines yielded 12 male chimeras and one female chimera, with the chimerism ranging from 20% to 100%. As the ES cell line AB2.1 is homozygous for the agouti (A) coat color gene, penetrance of ES cells into the injected (black coat color) C57B1/6 blastocyst gives rise to chimeric coat color mice.

EXAMPLE 6

Breeding chimeric mice

The chimeric coat color mice were bred to wild-type C57BL/6 (black coated) and 129/J (agouti coated) female mice. Some of the progeny from the chimera X C57BL/6 cross were expected to be agouti if the chimeric male had ES cell genetic material incorporated into its germline (agouti is dominant to black coat color). The chimera X 129/J cross would yield only agouti mice. These crosses were performed to transfer ES cell genetic information, including the disrupted APP allele, to its offspring. Breeding of three male chimeras from both clone 76 and 174 resulted in agouti pups when crossed with C57B1/6J females.

To deteimine the APP genotypes, genomic DNA was purified from about 1 cm of tail taken from each mouse at about two weeks of age. The genomic DNA was isolated as described (Laird et al., supra), followed by phenol hloroform extractions and ethanol precipitation. Southern hybridization analysis (as described in Example 5) were used to identify offspring which contained the disrupted APP allele. These transgenic offspring were heterozygous for the APP disruption. Both transgenic heterozygous and nontransgenic mouse (tail) genomic DNAs were digested with EcoRI, and were hybridized with 5' flanking DNA probe to confirm the transgenic APP structure. Southern hybridization analysis confirmed that the structure of the altered APP allele was identical to that predicted, and previously characterized in the APP targeted ES clones. EXAMPLE 7

Breeding heterozygous mice and generation of homozygous APP deficient mice. Male and female transgenic mice, each of which contained one copy of the altered APP allele (heterozygous mice), were mated with each other to generate mice in which both copies of the APP gene encoded the targeted, altered APP allele. It was predicted that one fourth of the mouse embryos would be homozygous for the altered APP gene. Surviving offspring were genotyped by Southern hybridization as described above (Fig. 4). It was determined that 21 ( 24%) of the 87 offspring mice were homozygous APP-/-, 30 ( 34%) were wild-type APP+/+, and 36 (41 %) were heterozygous APP+/-. These numbers indicate that there was no significant decrease in the number of APP deficient transgenic mice which survived at two weeks of age.

EXAMPLE 8

Characterization of homozygous APP deficient mice

Surviving homozygous APP deficient mice of Example 7 were bred with wild-type or heterozygous mates to determine if they were fertile. All homozygous APP-/- males and females tested were fertile. Significant differences in gross morphology or histology between the APP deficient mice and the wild-type or heterozygous mice were not observed.

EXAMPLE 9

Confirmation of APP inactivation by Northern analysis

Total brain RNA was prepared from the wild-type, heterozygous and homozygous APP KO mice. Northern hybridizations were carried out by standard procedures using APP695 cDNA as a probe (Fig. 5). No RNA was detected in the two homozygous APP knockout mice. EXAMPLE 10

Cell Culture

The transgenic animals of the invention may be used as a source of cells for cell culture. Cells of brain tissues lacking the APP gene may be cultured using standard culture techniques.

Claims

WHAT IS CLAIMED IS:

1. A transgenic animal whose somatic and germ cells lack a gene coding for an mouse amyloid precursor protein (APP).

2. The transgenic animal of Claim 1 wherein the animal is a mouse.

3. The mouse of Claim 2, wherein the mouse is fertile and capable of transmitting the altered amyloid precursor protein gene to its offspring.

4. The mouse of Claim 2, wherein the altered amyloid precursor protein gene has been removed from an ancestor of the mouse at an embryonic stage by microinjection of altered embryonic stem cells into mouse blastocysts.

5. The mouse of Claim 2, wherein the altered amyloid precursor protein gene has been introduced into the mouse at an embryonic stage either by microinjection of altered embryonic stem cells into mouse blastocyts or coincubation of altered embryonic stem cells with fertilized eggs or morulae.

6. A method of producing a mouse whose somatic and germ cells lack a gene coding for amyloid precursor protein, which comprises:

(a) providing a gene encoding an altered form of APP designed to target an APP allele of mouse embryonic stem cells; (b) introducing the altered gene into mouse embryonic stem cells; (c) selecting embryonic stem cells which contain the altered gene; (d) injecting the embryonic stem cells containing the altered APP gene into mouse blastocysts;

(e) transplanting the injected blastocysts into a pseudopregnant mouse, and

(f) allowing the embryo to develop to term; to produce a founder transgenic mouse.

7. The method of Claim 6 wherein the introduction of step (d) is by microinjection.

8. The method of Claim 7 which further comprises the steps:

(g) breeding the chimeric transgenic mice to wild- type mice to obtain heterozygous (Fl) mice; and

(h) breeding the heterozygous (Fl) mice to generate homozygous (F2) APP transgenic mice.

9. A cell line derived from a transgenic animal according to Claim 1.

10. A genomic mouse cosmid clone spanning the APP promoter region.