Title:  Method for selecting a transgenic mouse model of alzheimer's disease

United States Patent:  6,717,031

Issued:  April 6, 2004

Inventors:  Games; Kate Dora (411 Central Ave., San Francisco, CA 94117); Schenk; Dale Bernard (605 Sharp Park Rd., Pacifica, CA 94044); McConlogue; Lisa Claire (283 Juanita Way, San Francisco, CA 94127); Seubert; Peter Andrew (222 Northwood Dr., S. San Francisco, CA 94080); Rydel; Russell E. (2211 Village Ct., #6, Belmont, CA 94002)

Appl. No.:  149718

Filed:  September 8, 1998

Abstract

The construction of transgenic animal models of human Alzheimer's disease, and methods of using the models to screen potential Alzheimer's disease therapeutics, are described. The models are characterized by pathologies similar to pathologies observed in Alzheimer's disease, based on expression of all three forms of the .beta.-amyloid precursor protein (APP), APP695, APP751, and APP770, as well as various point mutations based on naturally occurring mutations, such as the London and Indiana familial Alzheimer's disease (FAD) mutations at amino acid 717, predicted mutations in the APP gene, and truncated forms of APP that contain the A.beta. region. Animal cells can be isolated from the transgenic animals or prepared using the same constructs with standard techniques such as lipofection or electroporation. The transgenic animals, or animal cells, are used to screen for compounds altering the pathological course of Alzheimer's disease as measured by their effect on the amount of APP, .beta.-amyloid peptide, and numerous other Alzheimer's disease markers in the animals, the neuropathology of the animals, as well as by behavioral alterations in the animals.

SUMMARY OF THE INVENTION

The construction of transgenic animal models for testing potential treatments for Alzheimer's disease is described. The models are characterized by a greater similarity to the conditions existing in naturally occurring Alzheimer's disease, based on the ability to control expression of one or more of the three major forms of the .beta.-amyloid precursor protein (APP), APP695, APP751, and APP770, or subfragments thereof, as well as various point mutations based on naturally occurring mutations, such as the FAD mutations at amino acid 717, and predicted mutations in the APP gene. The APP gene constructs are prepared using the naturally occurring APP promoter of human, mouse, or rat origin, efficient promoters such as human platelet derived growth factor .beta. chain (PDGF-B) gene promoter, as well as inducible promoters such as the mouse metallothionine promoter, which can be regulated by addition of heavy metals such as zinc to the animal's water or diet. Neuron-specific expression of constructs can be achieved by using the rat neuron specific enolase promoter.

The constructs are introduced into animal embryos using standard techniques such as microinjection or embryonic stem cells. Cell culture based models can also be prepared by two methods. Cells can be isolated from the transgenic animals or prepared from established cell cultures using the same constructs with standard cell transfection techniques.

The constructs disclosed herein generally encode all or a contiguous portion of one of the three forms of APP: APP695, APP751, or APP770, preferably an A.beta.-containing protein, as described herein. Examples of A.beta.-containing proteins are proteins that include all or a contiguous portion of APP770, APP770 bearing a mutation in amino acid 669, 670, 671, 690, 692, and/or 717, APP751, APP751 bearing a mutation in amino acid 669, 670, 671, 690, 692, and/or 717, APP695, and APP695 bearing a mutation in amino acid 669, 670, 671, 690, 692, and/or 717, where each of these A.beta.-containing proteins includes amino acids 672 to 714 of human APP. Some specific constructs that are described employ the following protein coding sequences: the APP770 cDNA; the APP770 cDNA bearing a mutation at amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations; the APP751 cDNA containing the KPI protease inhibitor domain without the OX-2 domain in the construct; the APP751 cDNA bearing a mutation at amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations; the APP695 cDNA; the APP695 cDNA bearing a mutation at amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations; APP695, APP751, or APP770 cDNA truncated at amino acid 671 or 685, the sites of .beta.-secretase or .alpha.-secretase cleavage, respectfully; APP cDNA truncated to encode amino acids 646 to 770 of APP; APP cDNA truncated to encode amino acids 646 to 770 of APP and including at least one intron; the APP leader sequence followed by the A.beta. region (amino acids 672 to 714 of APP) plus the remaining carboxy terminal 56 amino acids of APP; the APP leader sequence followed by the A.beta. region plus the remaining carboxy terminal 56 amino acids with the addition of a mutation at amino acid 717; the APP leader sequence followed by the A.beta. region; the A.beta. region plus the remaining carboxy terminal 56 amino acids of APP; the A.beta. region plus the remaining carboxy terminal 56 amino acids of APP with the addition of a mutation at amino acid 717; a combination cDNA/genomic APP gene construct; a combination cDNA/genomic APP gene construct with the addition of a mutation at amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations; a combination cDNA/genomic APP gene construct truncated at amino acid 671 or 685; and an APP cDNA construct containing at least amino acids 672 to 722 of APP.

These protein coding sequences are operably linked to leader sequences specifying the transport and secretion of the encoded A.beta. related protein. A preferred leader sequence is the APP leader sequence. These combined protein coding sequences are in turn operably linked to a promoter that causes high expression of A.beta. in transgenic animal brain tissue. A preferred promoter is the human platelet derived growth factor .beta. chain (PDGF-B) gene promoter. Additional constructs include a human yeast artificial chromosome construct controlled by the PDGF-B promoter; a human yeast artificial chromosome construct controlled by the PDGF-B promoter with the addition of a mutation at amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations; the endogenous mouse or rat APP gene modified through the process of homologous recombination between the APP gene in a mouse or rat embryonic stem (ES) cell and a vector carrying the human APP cDNA bearing a mutation at amino acid position 669, 670, 671, 690, 692, 717, or a combination of these mutations, such that sequences in the resident rodent chromosomal APP gene beyond the recombination point (the preferred site for recombination is within APP exon 9) are replaced by the analogous human sequences bearing a mutation at amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations. These constructs can be introduced into the transgenic animals and then combined by mating of animals expressing the different constructs.

The transgenic animals, or animal cells, are used to screen for compounds altering the pathological course of Alzheimer's disease as measured by their effect on the amount and/or histopathology of Alzheimer's disease markers in the animals, as well as by behavioral alterations. These markers include APP and APP cleavage products; A.beta.; other plaque related molecules such as apolipoprotein E, laminin, and collagen type IV; cytoskeletal markers, such as spectrin, tau, neurofilaments, and MAP-2; inflammatory markers, such as GFAP, .alpha.2-macroglobulin, IL-1, and IL-6; and neuronal and synaptic neurotransmitter related markers, such as GAP43 and synaptophysin, and those associated with the cholinergic, muscarinic, serotinergic, adrenergic, and adensosine receptor systems.

DETAILED DESCRIPTION OF THE INVENTION

The constructs and transgenic animals and animal cells are prepared using the methods and materials described below.

Sources of Materials.

Restriction endonucleases are obtained from conventional commercial sources such as New England Biolabs (Beverly, Mass.), Promega Biological Research Products (Madison, Wis.), and Stratagene (La Jolla Calif.). Radioactive materials are obtained from conventional commercial sources such as Dupont/NEN or Amersham. Custom-designed oligonucleotides for site-directed mutagenesis are available from any of several commercial providers of such materials such as Bio-Synthesis Inc., Lewisville, Tex. Kits for carrying out site-directed mutagenesis are available from commercial suppliers such as Promega Biological Research Products and Stratagene. Clones of cDNA including the APP695, APP751, and APP770 forms of APP mRNA were obtained directly from Dr. Dmitry Goldgaber, NIH. Libraries of DNA are available from commercial providers such as Stratagene, La Jolla, Calif., or Clontech, Palo Alto, Calif. PC12 and 3T3 cells were obtained from ATCC (#CRL1721 and #CCL92, respectively). An additional PC12 cell line was obtained from Dr. Charles Marotta of Harvard Medical School, Massachusetts General Hospital, and McLean Hospital. Standard cell culture media appropriate to the cell line are obtained from conventional commercial sources such as Gibco/BRL. Murine stem cells, strain D3, were obtained from Dr. Rolf Kemler (Doetschman et al., J. Embryol. Exp. Morphol. 87:27 (1985)). Lipofectin for DNA transfection and the drug G418 for selection of stable transformants are available from Gibco/BRL.

Definition of APP cDNA Clones.

The cDNA clone APP695 is of the form of cDNA described by Kang et al., Nature 325:733-735 (1987), and represents the most predominant form of APP in the brain. The cDNA clone APP751 is of the form described by Ponte et al., Nature 331:525-527 (1988). This form contains an insert of 168 nucleotides relative to the APP695 cDNA. The 168 nucleotide insert encodes the KPI domain. The cDNA clone APP770 is of the form described by Kitaguchi et al. Nature 331:530-532 (1988). This form contains an insert of 225 nucleotides relative to the APP695 cDNA. This insert includes the 168 nucleotides present in the insert of the APP751 cDNA, as well as an addition 57 nucleotide region that does not appear in APP751 cDNA. The 225 nucleotide insert encodes for the KPI domain as well as the OX-2 domain. All three forms arise from the same precursor RNA transcript by alternative splicing. The 168 nucleotide insert is present in both APP751 cDNA and APP770 cDNA.

The sequence encoding APP695 is shown in SEQ ID NO:1. This sequence begins with the first base of the initiation codon AUG and encodes a 695 amino acid protein. The region from nucleotide 1789 to 1917 of SEQ ID NO:1 encodes the A.beta.. The amino acid sequence of APP695 is shown in SEQ ID NO:2. Amino acids 597 to 639 of SEQ ID NO:2 form the A.beta.. The amino-acid composition of the APP695 is A57, C12, D47, E85, F17, G31, H25, I23, K38, L52, M21, N28, P31, Q33, R33, S30, T45, V62, W8, Y17 resulting in a calculated molecular weight of 78,644.45. These sequences are derived from Kang et al. (1988).

The sequence encoding APP751 is shown in SEQ ID NO:3. This sequence begins with the first base of the initiation codon AUG and encodes a 751 amino acid protein. Nucleotides 866 to 1033 of SEQ ID NO:3 do not appear in APP695 cDNA. The region from nucleotide 1957 to 2085 of SEQ ID NO:3 encodes the A.beta.. The amino acid sequence of APP751 is shown in SEQ ID NO:4. Amino acids 289 to 345 of SEQ ID NO:4 do not appear in APP695. This 57 amino acid region includes the KPI domain. Amino acids 653 to 695 of SEQ ID NO:4 form the A.beta.. These sequences are derived from Ponte et al. (1988).

The sequence encoding APP770 is shown in SEQ ID NO:5. This sequence begins with the first base of the initiation codon AUG and encodes a 770 amino acid protein. Nucleotides 866 to 1090 of SEQ ID NO:5 do not appear in APP695 cDNA. Nucleotides 1034 to 1090 of SEQ ID NO:5 do not appear in APP751 cDNA. The region from nucleotide 2014 to 2142 encodes the A.beta.. The amino acid sequence of APP770 is shown in SEQ ID NO:6. Amino acids 289 to 364 of SEQ ID NO:6 do not appear in APP695. This 76 amino acid region includes the KPI and OX-2 domains. Amino acids 345 to 364 of SEQ ID NO:6 do not appear in APP751. This 20 amino acid region includes the OX-2 domain. Amino acids 672 to 714 form the A.beta.. A probable membrane-spanning region of the APP occurs from amino acid 700 to 723. Unless otherwise stated, all references herein to nucleotide positions refer to the numbering of SEQ ID NO:5. This is the numbering derived from the APP770 cDNA. Unless otherwise stated, all references herein to amino acid positions refer to the numbering of SEQ ID NO:6. This is the numbering derived from APP770. According to this numbering convention, for example, amino acid position 717 refers to amino acid 717 of APP770, amino acid 698 of APP751, and amino acid 642 of APP695. The above sequences are derived from Kang et al. (1988) and Kitaguchi et al. (1988).

Unless otherwise noted, all forms of APP and fragments of APP, including all forms of A.beta., referred to herein are based on the human APP amino acid sequence. For example, A.beta. refers to the human A.beta., APP refers to human APP, and APP770 refers to human APP770. As used herein, the term cDNA refers not only to DNA molecules actually prepared by reverse transcription of mRNA, but also any DNA molecule encoding a protein where the coding region is not interrupted, that is, a DNA molecule having a continuous open reading frame encoding a protein. As such, the term cDNA as used herein provides a convenient means of referring to a protein encoding DNA molecule where the protein encoding region is not interrupted by intron sequences (or any other sequences not encoding protein).

Definition of the APP Genomic Locus.

Characterization of phage and cosmid clones of human genomic DNA clones listed in Table 1 below originally established a minimum size of at least 100 kb for the Alzheimer's gene. There are a total of 18 exons in the APP gene (Lemaire et al., Nucl. Acid Res 17:517-522 (1989); Yoshikai et al. (1990); Yoshikai et al., Nucleic Acids Res 102:291-292 (1991)). Yoshikai et al. (1990) describes the sequences of the exon-intron boundaries of the APP gene. These results taken together indicate that the miniimum size of the Alzheimer's gene is 175 kb.

Table 2 indicates where the 17 introns interrupt the APP coding sequence. The numbering refers to the nucleotide positions of APP770 cDNA as shown in SEQ ID NO:5. The starting nucleotide of exon 1 represents the first transcribed nucleotide. It is negative because the +1 nucleotide is the first nucleotide of the AUG initiator codon by convention (Kang et al. (1988)). The ending nucleotide of exon 18 represents the last nucleotide present in the mRNA prior to the poly(A) tail (Yoshikai et al. (1990)). It has been discovered that Yoshikai et al. (1990) and Yoshikai et al. (1991) contain an error in the location of exon 8. FIG. 1 of Yoshikai et al. (1991) includes an EcoRI fragment between EcoRI fragments containing exon 7 and exon 8. In fact, this intervening EcoRI fragment is actually located immediately after exon 8, so that the EcoRI fragment containing exon 7 and the EcoRI fragment containing exon 8 are adjacent to each other.

APP Gene Mutations.

Certain families are genetically predisposed to Alzheimer's disease, a condition referred to as familial Alzheimer's disease (FAD), through mutations resulting in an amino acid replacement at position 717 of the full length protein (Goate et al. (1991); Murrell et al. (1991); Chartier-Harlin et al. (1991)). These mutations co-segregate with the disease within the families. For example, Murrell et al. (1991) described a specific mutation found in exon 17 (which Murrell et al. refers to as exon 15) where the valine of position 717 is replaced by phenylalanine.

Another FAD mutant form contains a change in amino acids at positions 670 and 671 of the full length protein (Mullan et al. (1992)). In one form of this mutation, the lysine at position 670 is replaced by asparagine and the methionine at position 671 is replaced by leucine. The effect of this mutation is to increase the production of A.beta. in cultured cells approximately 7-fold (Citron et al., Nature 360: 672-674 (1992); Lai et al., Science 259:514-516 (1993)). Replacement of the methionine at position 671 with leucine by itself has also been shown to increase production of A.beta.. Additional mutations in APP at amino acids 669, 670, and 671 have been shown to reduce the amount of A.beta. processed from APP (Citron et al., Neuron 14:661-670 (1995)). The APP construct with Val at amino acid 690 produces an increased amount of a truncated form of A.beta..

APP expression clones can be constructed that bear a mutation at amino acid 669, 670, 671, 690, 692, or 717 of the full length protein. The mutations from Lys to Asn and from Met to Leu at amino acids 670 and 671, respectively, are sometimes referred to as the Swedish mutation. Additional mutations can also be introduced at amino acids 669, 670, or 671 which either increase or reduce the amount of A.beta. processed from APP. Mutations at these amino acids in any APP clone or transgene can be created by site-directed mutagenesis (Vincent et al., Genes & Devel. 3:334-347 (1989)), or, once made, can be incorporated into other constructs using standard genetic engineering techniques. Some mutations at amino acid 717 are sometimes referred to as the Hardy mutation. Such mutations can include conversion of the wild-type Val717 codon to a codon for Ile, Phe, Gly, Tyr, Leu, Ala, Pro, Trp, Met, Ser, Thr, Asn, or Gln. A preferred substitution for Val717 is Phe. These mutations predispose individuals expressing the mutant proteins to develop Alzheimer's disease. It is believed that the mutations affect the expression and/or processing of APP, shifting the balance toward Alzheimer's pathology. Mutations at amino acid 669 can include conversion of the wild-type Val669 codon to a codon for Trp, or deletion of the codon. Mutations at amino acid 670 can include conversion of the wild-type Lys670 codon to a codon for Asn or Glu, or deletion of the codon. Mutations at amino acid 671 can include conversion of the wild-type Met671 codon to a codon for Leu, Val, Lys, Tyr, Glu, or Ile, or deletion of the codon. A preferred substitution for Lys670 is Asn, and a preferred substitution for Met671 is Leu. These mutations predispose individuals expressing the mutant proteins to develop Alzheimer's disease. The other listed mutations to amino acids 669, 670, and 671 are known to reduce the amount of A.beta. processed from APP (Citron et al. (1995)). It is believed that these mutations affect processing of APP leading to a change in A.beta. production.

Truncated forms of APP can also be expressed from transgene constructs. For example, APP cDNA truncated to encode amino acids 646 to 770 of APP. The APP cDNA construct truncated to encode amino acids 646 to 770 of APP, and operatively linked to the PDGF-B promoter, is referred to as PDAPPc125.

Nucleic Acid Constructs Encoding A.beta.-containing Proteins.

Constructs for use in transgenic animals include a promoter for expression of the construct in a mammalian cell and a region encoding a protein that includes all or a contiguous portion of one of the three forms of APP: APP695, APP751, or APP770, with or without specific amino acid mutations as described herein. It is preferred that protein encoded is an A.beta.-containing protein. As used herein, an A.beta.-containing protein is a protein that includes all or a contiguous portion of one of the three forms of APP: APP695, APP751, or APP770, with or without specific amino acid mutations as described herein, where the protein includes all or a portion of amino acids 672 to 714 of human APP. Preferred A.beta.-containing proteins include amino acids 672 to 714 of human APP. Preferred forms of such A.beta.-containing proteins include all or a contiguous portion of APP770, APP770 bearing a mutation in amino acid 669, 670, 671, 690, 692, and/or 717, APP751, APP751 bearing a mutation in amino acid 669, 670, 671, 690, 692, and/or 717, APP695, and APP695 bearing a mutation in amino acid 669, 670, 671, 690, 692, and/or 717, where each of these A.beta.-containing proteins includes amino acids 672 to 714 of human APP.

Preferred forms of the above A.beta.-containing proteins are APP770; APP770 bearing a mutation in the codon encoding one or more amino acids selected from the group consisting of amino acid 669, 670, 671, 690, 692, 717; APP751; APP751 bearing a mutation in the codon encoding one or more amino acids selected from the group consisting of amino acid 669, 670, 671, 690, 692, 717; APP695; APP695 bearing a mutation in the codon encoding one or more amino acids selected from the group consisting of amino acid 669, 670, 671, 690, 692, 717; a protein consisting of amino acids 646 to 770 of APP; a protein consisting of amino acids 670 to 770 of APP; a protein consisting of amino acids 672 to 770 of APP; and a protein consisting of amino acids 672 to 714 of APP.

In the constructs disclosed herein, the DNA encoding the A.beta.-containing protein can be cDNA or a cDNA/genomic DNA hybrid, wherein the cDNA/genomic DNA hybrid includes at least one APP intron sequence wherein the intron sequence is sufficient for splicing.

Preferred constructs contain DNA encoding APP770; DNA encoding APP770 bearing a mutation in the codon encoding amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations; a fragment of DNA encoding APP770 which encodes an amino acid sequence comprising amino acids 672 to 714 of APP770; DNA encoding APP751; DNA encoding APP751 bearing a mutation in the codon encoding amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations; a fragment of DNA encoding APP751 which encodes an amino acid sequence comprising amino acids 672 to 714 of APP770; DNA encoding APP695; DNA encoding APP695 bearing a mutation in the codon encoding amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations; a fragment of DNA encoding APP695 which encodes an amino acid sequence comprising amino acids 672 to 714 of APP770; APP cDNA truncated to encode amino acids 646 to 770 of APP; a combination cDNA/genomic DNA hybrid APP gene construct; a combination cDNA/genomic DNA hybrid APP gene construct bearing a mutation in the codon encoding amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations; or a combination cDNA/genomic DNA hybrid APP gene construct truncated at amino acid 671 or 685.

Preferred forms of such constructs are APP770 cDNA; APP770 cDNA bearing a mutation in the codon encoding amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations; a fragment of APP770 cDNA encoding an APP amino acid sequence, the amino acid sequence comprising amino acids 672 to 714 of APP770; APP751 cDNA; APP751 cDNA bearing a mutation in the codon encoding amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations; a fragment of APP751 cDNA encoding an APP amino acid sequence, the amino acid sequence comprising amino acids 672 to 714 of APP770; APP695 cDNA; APP695 cDNA bearing a mutation in the codon encoding amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations; a fragment of APP695 cDNA encoding an APP amino acid sequence, the amino acid sequence comprising amino acids 672 to 714 of APP770; APP cDNA truncated to encode amino acids 646 to 770 of APP; a combination cDNA/genomic DNA hybrid APP gene construct; a combination cDNA/genomic DNA hybrid APP gene construct bearing a mutation in the codon encoding amino acid 669, 670, 671, 690, 692, 717, and a combination of these mutations; and a combination cDNA/genomic DNA hybrid APP gene construct truncated at amino acid 671 or 685.

Construction of Transgenes.

Construction of various APP transgenes can be accomplished using any suitable genetic engineering technique, such as those described in Sambrook et al., Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory, N.Y., 1989). Regions of APP clones that have been engineered or mutated can be interchanged by using convenient restriction enzyme sites present in APP cDNA clones. A NruI site starts at position -5 (relative to the first nucleotide of the AUG initiator codon). A KpnI and an Asp718 site both start at position 57 (these are isoschizomers leaving different sticky ends). A XcmI site starts at position 836 and cuts at position 843. A ScaI site starts at position 1004. A XhoI site starts at position 1135. A BamHI site starts at position 1554. A BglII site starts at position 1994. An EcoRI site starts at position 2020. A SpeI site starts at position 2583. Another EcoRI site starts at position 3076.

The clones bearing various portions of the human APP gene sequence can be constructed in a common manner using standard genetic engineering techniques. For example, these clones can be constructed by first cloning the polyA addition signal from SV40 virus, as a 253 base pair BclI to BamHI fragment (Reddy et al., Science 200:494-502 (1978), into a modified vector from the pUC series. Next, the cDNA coding sequences (APP770, APP751, or APP695) can be inserted. Correct orientation and content of the fragments inserted can be determined through restriction endonuclease mapping and limited sequencing. The clones bearing various carboxy terminal portions of the human APP gene sequence can be constructed through several steps in addition to those indicated above. For example, an APP770 cDNA clone (SEQ ID NO:5) can be digested with Asp718 which cleaves after nucleotide position 57. The resulting 5' extension is filled in using the Klenow enzyme (Sambrook et al. (1989)) and ligated to a hexanucleotide of the following sequence: AGATCT, the recognition site for BglII. After cleavage with BglII, which also cuts after position 1994, and re-ligation, the translational reading frame of the protein is preserved. The truncated protein thus encoded contains the leader sequence, followed by approximately 6 amino acids that precede the A.beta., followed by the A.beta., and the 56 terminal amino acids of APP. The clone is created by converting the nucleotide at position 2138 to a T by site directed mutagenesis in the clone, thus creating a termination codon directly following the last amino acid codon of the A.beta.. APP cDNA clones naturally contain an NruI site that cuts 2 nucleotides upstream from the initiator methionine codon. This site can be used for attachment of the different promoters used to complete each construct.

APP transgenes can also be constructed using PCR cloning techniques. Such techniques allow precise coupling of DNA fragments in the transgenes.

Combination cDNA/Genomic DNA Clones.

Endogenous APP expression results from transcription of precursor mRNA followed by alternative splicing to produce three main forms of APP. It is believed that this alternative splicing may be important in producing the pattern of APP expression involved in Alzheimer's disease. It is also believed that the presence of introns in expression constructs can influence the level and nature of expression by, for example, targeting precursor mRNA to mRNA processing and transport pathways (Huang et al., Nucleic Acids Res. 18:937-947 (1990)). Accordingly, transgenes combining cDNA and genomic DNA, which include intron sequences, are a preferred type of construct.

The RNA splicing mechanism requires only a few specific and well known consensus sequences. Such sequences have been identified in APP genomic DNA by Yoshikai et al. (1990). The disclosed transgenes can be constructed using one or more complete and intact intron sequences. However, it is preferred that the transgenes are constructed using truncated intron sequences that contain an effective amount of intron sequence to allow splicing. In general, truncated intron sequences that retain the splicing donor site, the splicing acceptor site, and the splicing branchpoint sequence will constitute an effective amount of an intron. The sufficiency of any truncated intron sequence can be determined by testing for the presence of correctly spliced mRNA in transgenic cells using methods described below.

Other intron sequences and splicing signals which are not derived from APP gene sequences may also be used in the transgene constructs. Such intron sequences will enhance expression of the transgene construct. A preferred heterologous intron is a hybrid between the adenovirus major late region first exon and intron junction and an IgG variable region splice acceptor. This hybrid intron can be constructed, for example, by joining the 162 bp PvuII to HindIII fragment of the adenovirus major late region, containing 8 bp of the first exon and 145 bp of the first intron, and the 99 bp HindIII to PstI fragment of the IgG variable region splice acceptor clone-6, as described by Bothwell et al., Cell 24:625-637 (1981). A similar splice signal has been shown to enhance expression of a construct to which it was attached, as described by Manley et al., Nucleic Acids Res. 18:937-947 (1990). It is preferred that the heterologous intron be placed between the promoter and the region encoding the APP.

Combination cDNA/genomic expression clones include (1) nucleic acid constructs containing a combination of APP cDNA encoding exons 1-6 and 9-18 and genomic APP sequences encoding introns 6, 7 and 8, and exons 7 and 8, and (2) nucleic acid constructs that do not contain a combination of APP cDNA encoding exons 1-6 and 9-18 and genomic APP sequences encoding introns 6, 7 and 8, and exons 7 and 8.

A preferred APP combination cDNA/genomic expression clone includes an effective amount of introns 6, 7 and 8. Such a transgene can be constructed as follows. A preferred method of construction is described in Example 5. A plasmid containing the cDNA portion of the clone can be constructed by first converting the TaqI site at position 860 in an APP770 cDNA clone to an XhoI site by site-directed mutagenesis. Cleavage of the resulting plasmid with XhoI cuts at the new XhoI site and a pre-existing XhoI site at position 1135, and releases the KPI and OX-2 coding sequence. The plasmid thus generated serves as the acceptor for the KPI and OX-2 alternative splicing cassette.

The alternative splicing cassette can be created through a series of cloning steps involving genomic DNA. First, the TaqI site at position 860 in a genomic clone containing exon 6 and the adjacent downstream intron can be converted to an XhoI site by site-directed mutagenesis. Cleavage of the resulting plasmid with XhoI cuts at the new XhoI site and an XhoI site within either intron 6 or 7. This fragment, containing a part of exon 6 and at least a part of adjacent intron 6, can then be cloned into the XhoI site in a plasmid vector. Second, a genomic clone containing exon 9 and the adjacent upstream genomic sequences is cleaved with XhoI, cleaving the clone at the XhoI site at position 1135 (position 910 using the numbering system of Kang et al. (1987)) and an XhoI site in either intron 7 or 8. This fragment, containing a part of exon 9 and at least a part of adjacent intron 8, can then be cloned into the XhoI site of another plasmid vector. These two exon/intron junction fragments can then be released from their respective plasmid vectors by cleavage with XhoI and either BamHI or BglII, and cloned together into the XhoI site of another plasmid vector. It is preferred that the exon/intron junction fragments be excised with BamHI. It is most preferable that BamHI sites are engineered in the intron portion of the exon/intron junction fragments prior to their excision. This allows the elimination of lengthy extraneous intron sequences from the cDNA/genomic clone.

The XhoI fragment resulting from cloning the two exon/intron junction fragments together can be cleaved with either BamHI or BglII, depending on which enzyme was used for excision step above, and the genomic 6.8 kb BamHI segment, containing the KPI and OX-2 coding region along with their flanking intron sequences, can be inserted. This fragment was identified by Kitaguchi et al. (1988) using Southern blot analysis of BamHI-digested lymphocyte DNA from one normal individual and eight Alzheimer's disease patients using a 212 bp TaqI-AvaI fragment, nucleotides 862 to 1,073, of APP770 cDNA as the hybridization probe. Genomic DNA clones containing the region of the 225 bp insert can be isolated, for example, from a human leukocyte DNA library using the 212 bp TaqI-AvaI fragment as a probe. In the genomic DNA, the 225 bp sequence is located in a 168 bp exon (exon 7) and a 57 bp exon (exon 8), separated by an intron of approximately 2.6 kb (intron 7), with both exons flanked by intron-exon consensus sequences. The exon 7 corresponds to nucleotides 866 to 1,033 of APP770, and the exon 8 to nucleotides 1,034 to 1,090. Exon 7 encodes the highly conserved region of the Kunitz-type protease inhibitor family domain.

After cleavage with XhoI, this alternative splicing cassette, containing both exon and intron sequences, can then be excised by cleavage with XhoI and inserted into the XhoI site of the modified APP770 cDNA plasmid (the acceptor plasmid) constructed above. These cloning steps generate a combination cDNA/genomic expression clone that allows cells in a transgenic animal to regulate the inclusion of the KPI and OX-2 domains by a natural alternative splicing mechanism. An analogous gene bearing a mutation at amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations, can be constructed either directly by in vitro mutagenesis. A mutation to amino acid 717 can also be made by using the mutated form of APP770 cDNA described above to construct an acceptor plasmid.

Promoters.

Different promoter sequences can be used to control expression of nucleotide sequences encoding A.beta.-containing proteins. The ability to regulate expression of the gene encoding an A.beta.-containing protein in transgenic animals is believed to be useful in evaluating the roles of the different APP gene products in AD. The ability to regulate expression of the gene encoding an A.beta.-containing protein in cultured cells is believed to be useful in evaluating expression and processing of the different A.beta.-containing gene products and may provide the basis for cell cultured drug screens. A preferred promoter is the human platelet derived growth factor .beta. (PDGF-B) chain gene promoter (Sasahara et al., Cell 64:217-227 (1991)).

Preferred promoters for the disclosed APP constructs are those that, when operatively linked to the protein coding sequences, mediate expression of one or more of the following expression products to at least a specific level in brain tissue of a two to four month old animal transgenic for one of the disclosed APP constructs. The products and their expression levels are A.beta._tot to a level of at least 30 ng/g (6.8 pmoles/g) brain tissue and preferably at least 40 ng/g (9.12 pmoles/g) brain tissue, A.beta._1-42 to a level of at least 8.5 ng/g (1.82 pmoles/g) brain tissue and preferably at least 11.5 ng/g (2.5 pmoles/g) brain tissue, full length APP (FLAPP) and APP.alpha. combined (FLAPP+APP.alpha.) to a level of at least 150 pmoles/g brain tissue, APP.alpha. to a level of at least 42 pmoles/g brain tissue, and mRNA encoding human A.beta.-containing protein to a level at least twice that of mRNA encoding the endogenous APP of the transgenic animal. A.beta._tot is the total of all forms of A.beta.. A.beta._1-42 is a form of A.beta. having amino acids 1 to 42 of A.beta. (corresponding to amino acids 672 to 714 of APP). FLAPP+APP.alpha. refers to APP forms containing the first 12 amino acids of the A.beta. region (corresponding to amino acids 672 to 684 of APP). Thus, FLAPP+APP.alpha. represents a mixture of full length forms of APP and APP cleaved at the .alpha.-secretase site (Esch et al., Science 248:1122-1124 (1990)). APP.beta. is APP cleaved at the .beta.-secretase site (Seubert et al., Nature 361:260-263 (1993)).

It is intended that the levels of expression described above refer to amounts of expression product present and are not limited to the specific units of measure used above. Thus, an expression level can be measured, for example, in moles per gram of tissue, grams per grams of tissue, moles per volume of tissue, and in grams per volume of tissue. The equivalence of these units of measure to the measures listed above can be determined using known conversion methods.

The levels of expression described above need not occur in all brain tissues. Thus, a promoter is considered preferred if at least one of the levels of expression described above occurs in at least one type of brain tissue. Where expression is tissue-specific, it is understood that if the expression level is sufficient in the specific brain tissue, the promoter is considered preferred even though the expression level in brain tissue as a whole may not, and need not, reach a threshold level. It is preferred that this level of expression is observed in hippocampal and/or cortical brain tissue. The promoter can mediate expression of the above expression products to the levels described above either constitutively or by induction. Induction can be accomplished by, for example, administration of an activator molecule, by heat, or by expression of a protein activator of transcription for the promoter operatively linked to the gene encoding an A.beta.-containing protein. Many inducible expression systems which would be suitable for this purpose are known to those of skill in the art.

It is preferred that, in making the above measurements, the brain tissue is prepared by the following method. A brain from a transgenic test animal is dissected and the tissue is kept on ice throughout the homogenization procedure except as noted. The brain tissue is homogenized in 10 volumes (w/v) of 5 M guanidine-HCl, 50 mM Tris-HCl, pH 8.5. The sample is then gently mixed for 2 to 4 hours at room temperature. Homogenates are then diluted 1:10 in cold casein buffer #1 (0.25% casein/phosphate buffered saline (PBS) 0.05% sodium azide, pH 7.4, 1x protease inhibitor cocktail) for a final 0.5 M guanidine concentration and kept on ice. 100x protease inhibitor cocktail is composed of 2 mg/ml aprotinin, 0.5 M EDTA, pH 8.0, 1 mg/ml leupeptin. Diluted homogenates are then spun in an Eppendorf microfuge at 14,000 rpm for 20 minutes at 4^oC. If further dilutions are required, they can be made with cold guanidine buffer #2 (1 part guanidine buffer #1 to 9 parts casein buffer #1).

It is preferred that the following assay be used to identify preferred promoters for their ability to mediate expression of A.beta. to the levels described above. Antibody 266 (Seubert et al., Nature 359:325-327 (1992)) is dissolved at 10 .mu.g/ml in buffer (0.23 g/L NaH₂ PO₄ -H₂ O, 26.2 g/L NaHPO₄ -7H₂ O, 1 g/L sodium azide adjusted to pH 7.4) and 100 .mu.l/well is coated onto 96-well immunoassay plates (Costar) and allowed to bind overnight. The plate is then aspirated and blocked for at least 1 hour with a 0.25% human serum albumin solution in 25 g/L sucrose, 10.8 g/L Na₂ HPO₄ -7H₂ O, 1.0 g/L NaH₂ PO₄ -H₂ O, 0.5 g/L sodium azide adjusted to pH 7.4. The 266 coated plate is then washed 1x with wash buffer (PBS/0.05% Tween 20) using a Skatron plate washer. 100 .mu.l/well of A.beta.1-40 standards and brain tissue samples are added to the plate in triplicate and incubated overnight at 4^oC. A.beta.1-40 standards are made from 0.0156, 0.0312, 0.0625, 0.125, 0.250, 0.500, and 1.000 .mu.g/ml stocks in DMSO stored at -40^oC. as well as a DMSO only control for background determination. A.beta. standards consist of 1:100 dilution of each standard into guanidine buffer #3 (1 part BSA buffer to 9 parts guanidine buffer #1) followed by a 1:10 dilution into casein buffer #1 (Note: the final A.beta. concentration range is 15.6 to 1000 pg/ml and the final guanidine concentration is 0.5 M). BSA buffer consists of 1% bovine serum albumin (BSA, immunoglobulin-free)/PBS/0.05% sodium azide. The plates and casein buffer #2 (0.25% casein/PBS/0.05% Tween 20/pH 7.4) are then brought to room temperature (RT). The plates are then washed 3x with wash buffer. Next, 100 .mu.l/well of 3D6-biotin at 0.5 .mu.g/ml in casein buffer #2 is added to each well and incubated at 1 hour at RT.

Monoclonal antibody 3D6 was raised against the synthetic peptide DAEFRGGC (SEQ ID NO:10) which was conjugated through the cysteine to sheep anti-mouse immunoglobulin. The antibody does not recognize secreted APP but does recognize species that begin at A.beta. position 1 (Asp). For biotinylating 3D6, follow Pierce's NHS-Biotin protocol for labeling IgG (cat. #20217X) except use 100 mM sodium bicarbonate, pH 8.5 and 24 mg NHS-biotin per ml of DMSO.

The plates are then again washed 3x with wash buffer. Then, 100 .mu.l/well of horseradish peroxidase (HRP)-avidin (Vector Labs, cat. # A-2004) diluted 1:4000 in casein buffer #2 is added to each well and incubated for 1 hour at RT. The plates are washed 4x with wash buffer and then 100 .mu.l/well of TMB substrate (Slow TMB-ELISA (Pierce cat. # 34024)) at RT is added to each well and incubated for 15 minutes at RT. Finally, 25 .mu.l/well of 2 N H₂ SO₄ is added to each well to stop the enzymatic reaction, and the plate is read at 450 nm to 650 nm using the Molecular Devices Vmax reader.

It is preferred that the relative levels of mRNA encoding human A.beta.-containing protein mRNA encoding the endogenous APP of the transgenic animal be measured in the manner described by Bordonaro et al., Biotechniques 16:428-430 (1994), and Rockenstein et al., J. Biol. Chem. 270:28257-28267 (1995). Preferred methods for measuring the expression level of A.beta._1-42, FLAPP+APP.alpha., and APP.beta. are described in Example 8.

Yeast Artificial Chromosomes.

Large segments of human genomic DNA, when cloned into certain vectors, can be propagated as autonomously-replicating units in the yeast cell. Such vector-borne segments are referred to as yeast artificial chromosomes (YAC; Burke et al. Science 236:806 (1987)). A human YAC library is commercially available (Clontech, Palo Alto, Calif.) with an average insert size of 250,000 base pairs (range of 180,000 to 500,000 base pairs). A YAC clone of the Alzheimer's gene can be directly isolated by screening the library with the human APP770 cDNA. The inclusion of all of the essential gene regions in the clone can be confirmed by PCR analysis.

The YAC-APP clone can be established in embryonic stem (ES) cells by selecting for neomycin resistance encoded by the YAC vector. ES cells bearing the YAC-APP clone can be used to produce transgenic mice by established methods described below under "Transgenic Mice" and "Embryonic Stem Cell Methods". The YAC-APP gene bearing a mutation at amino acid 717 can be produced through the generation of a YAC library using genomic DNA from a person affected by a mutation at amino acid 717. Such a clone can be identified and established in ES cells as described above.

Genetic Alteration of the Mouse APP Gene.

The nucleotide sequence homology between the human and murine Alzheimer's protein genes is approximately 85%. Within the A.beta.-coding region, there are three amino acid differences between the two sequences. Amino acids Lys 670, Met671, and Val717,which can be mutated to alter APP processing, are conserved between mouse, rat, and man. Wild-type rodents do not develop Alzheimer's disease nor do they develop deposits or plaques in their central nervous system (CNS) analogous to those present in human Alzheimer's patients. Therefore, it is possible that the human but not the rodent form of A.beta. is capable of causing disease. Homologous recombination (Capecchi, Science 244:1288-1292 (1989)) can be used to convert the mouse Alzheimer's gene in situ to a gene encoding the human A.beta. by gene replacement. This recombination is directed to a site downstream from the KPI and OX-2 domains, for example, within exon 9, so that the natural alternative splicing mechanisms appropriate to all cells within the transgenic animal can be employed in expressing the final gene product.

Both wild-type and mutant forms of human cDNA can be used to produce transgenic models expressing either the wild-type or mutant forms of APP. The recombination vector can be constructed from a human APP cDNA (APP695, APP751, or APP770 form), either wild-type, mutant at amino acid 669, 670, 671, 690, 692, 717, or a combination of these mutations. Cleavage of the recombination vector, for example, at the XhoI site within exon 9, promotes homologous recombination within the directly adjacent sequences (Capecchi (1989)). The endogenous APP gene resulting from this event would be normal up to the point of recombination, within exon 9 in this example, and would consist of the human cDNA sequence thereafter.

Preparation of Constructs for Transfections and Microinjections.

DNA clones for microinjection are cleaved with enzymes appropriate for removing the bacterial plasmid sequences, such as SalI and NotI, and the DNA fragments electrophoresed on 1% agarose gels in TBE buffer (Sambrook et al. (1989)). The DNA bands are visualized by staining with ethidium bromide, and the band containing the APP expression sequences is excised. The excised band is then placed in dialysis bags containing 0.3 M sodium acetate, pH 7.0. DNA is electroeluted into the dialysis bags, extracted with phenol-chloroform (1:1), and precipitated by two volumes of ethanol. The DNA is redissolved in 1 ml of low salt buffer (0.2 M NaCl, 20 mM Tris, pH 7.4, and 1 mM EDTA) and purified on an Elutip-D.TM. column. The column is first primed with 3 ml of high salt buffer (1 M NaCl, 20 mM Tris, pH 7.4, and 1 mM EDTA) followed by washing with 5 ml of low salt buffer. The DNA solutions are passed through the column for three times to bind DNA to the column matrix. After one wash with 3 ml of low salt buffer, the DNA is eluted with 0.4 ml of high salt buffer and precipitated by two volumes of ethanol. DNA concentrations are measured by absorption at 260 nm in a UV spectrophotometer. For microinjection, DNA concentrations are adjusted to 3 .mu.g/ml in 5 mM Tris, pH 7.4 and 0.1 mM EDTA. Other methods for purification of DNA for microinjection are also described in Hogan et al., Manipulating the Mouse Embryo (Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1986); in Palmiter et al., Nature 300:611 (1982); in The Qiagenologist, Application Protocols, 3rd edition, published by Qiagen, Inc., Chatsworth, Calif.; and in Sambrook et al., Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1989).

Construction of Transgenic Animals.

A. Animal Sources.

Animals suitable for transgenic experiments can be obtained from standard commercial sources such as Charles River (Wilmington, Mass.), Taconic (Germantown, N.Y.), and Harlan Sprague Dawley (Indianapolis, Ind.). Many strains are suitable, but Swiss Webster (Taconic) female mice are preferred for embryo retrieval and transfer. B6D2F₁ (Taconic) males can be used for mating and vasectomized Swiss Webster studs can be used to stimulate pseudopregnancy. Vasectomized mice and rats can be obtained from the supplier.

B. Microinjection Procedures.

The procedures for manipulation of the rodent embryo and for microinjection of DNA are described in detail in Hogan et al., Manipulating the Mouse Embryo (Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1986), the teachings of which are incorporated herein.

C. Transgenic Mice.

Female mice six weeks of age are induced to superovulate with a 5 IU injection (0.1 cc, ip) of pregnant mare serum gonadotropin (PMSG; Sigma) followed 48 hours later by a 5 IU injection (0.1 cc, ip) of human chorionic gonadotropin (hCG; Sigma). Females are placed with males immediately after hCG injection. Twenty-one hours after hCG injection, the mated females are sacrificed by CO₂ asphyxiation or cervical dislocation and embryos are recovered from excised oviducts and placed in Dulbecco's phosphate buffered saline with 0.5% bovine serum albumin (BSA; Sigma). Surrounding cumulus cells are removed with hyaluronidase (1 mg/ml). Pronuclear embryos are then washed and placed in Earle's balanced salt solution containing 0.5% BSA (EBSS) in a 37.5^oC. incubator with a humidified atmosphere at 5% CO₂, 95% air until the time of injection. Embryos can be implanted at the two cell stage.

Randomly cycling adult female mice are paired with vasectomized males. Swiss Webster or other comparable strains can be used for this purpose. Recipient females are mated at the same time as donor females. At the time of embryo transfer, the recipient females are anesthetized with an intraperitoneal injection of 0.015 ml of 2.5% avertin per gram of body weight. The oviducts are exposed by a single midline dorsal incision. An incision is then made through the body wall directly over the oviduct. The ovarian bursa is then torn with watchmakers forceps. Embryos to be transferred are placed in DPBS (Dulbecco's phosphate buffered saline) and in the tip of a transfer pipet (about 10 to 12 embryos). The pipet tip is inserted into the infundibulum and the embryos transferred. After the transfer, the incision is closed by two sutures.

D. Transgenic Rats.

The procedure for generating transgenic rats is similar to that of mice (Hammer et al., Cell 63:1099-112 (1990)). Thirty day-old female rats are given a subcutaneous injection of 20 IU of PMSG (0.1 cc) and 48 hours later each female placed with a proven male. At the same time, 40-80 day old females are placed in cages with vasectomized males. These will provide the foster mothers for embryo transfer. The next morning females are checked for vaginal plugs. Females who have mated with vasectomized males are held aside until the time of transfer. Donor females that have mated are sacrificed (CO₂ asphyxiation) and their oviducts removed, placed in DPBS (Dulbecco's phosphate buffered saline) with 0.5% BSA and the embryos collected. Cumulus cells surrounding the embryos are removed with hyaluronidase (1 mg/ml). The embryos are then washed and placed in EBSS (Earle's balanced salt solution) containing 0.5% BSA in a 37.5^oC. incubator until the time of microinjection.

Once the embryos are injected, the live embryos are moved to DPBS for transfer into foster mothers. The foster mothers are anesthetized with ketamine (40 mg/kg, ip) and xylazine (5 mg/kg, ip). A dorsal midline incision is made through the skin and the ovary and oviduct are exposed by an incision through the muscle layer directly over the ovary. The ovarian bursa is torn, the embryos are picked up into the transfer pipet, and the tip of the transfer pipet is inserted into the infundibulum. Approximately 10 to 12 embryos are transferred into each rat oviduct through the infundibulum. The incision is then closed with sutures, and the foster mothers are housed singly.

E. Embryonic Stem (ES) Cell Methods.

1. Introduction of cDNA into ES Cells.

Methods for the culturing of ES cells and the subsequent production of transgenic animals, the introduction of DNA into ES cells by a variety of methods such as electroporation, calcium phosphate/DNA precipitation, and direct injection are described in detail in Teratocarcinomas and Embryonic Stem Cells, A Practical Approach, ed. E. J. Robertson, (IRL Press 1987), the teachings of which are incorporated herein. Selection of the desired clone of transgene-containing ES cells can be accomplished through one of several means. For random gene integration, an APP clone is co-precipitated with a gene encoding neomycin resistance. Transfection is carried out by one of several methods described in detail in Lovell-Badge, in Teratocarcinomas and Embryonic Stem Cells, A Practical Approach, ed. E. J. Robertson, (IRL Press 1987), or in Potter et al., Proc. Natl. Acad. Sci. USA 81:7161 (1984). Lipofection can be performed using reagents such as provided in commercially available kits, for example DOTAP (Boehringer-Mannheim) or lipofectin (BRL). Calcium phosphate/DNA precipitation, lipofection, direct injection, and electroporation are the preferred methods. In these procedures, 0.5x10⁶ ES cells are plated into tissue culture dishes and transfected with a mixture of the linearized APP clone and 1 mg of pSV2neo DNA (Southern and Berg, J. Mol. Appl. Gen. 1:327-341 (1982)) precipitated in the presence of 50 mg lipofectin (BRL) in a final volume of 100 .mu.l. The cells are fed with selection medium containing 10% fetal bovine serum in DMEM supplemented with G418 (between 200 and 500 .mu.g/ml). Colonies of cells resistant to G418 are isolated using cloning rings and expanded. DNA is extracted from drug resistant clones and Southern blots using an APP770 cDNA probe can be used to identify those clones carrying the APP sequences. PCR detection methods may also used to identify the clones of interest.

DNA molecules introduced into ES cells can also be integrated into the chromosome through the process of homologous recombination, described by Capecchi (1989). Direct injection results in a high efficiency of integration. Desired clones can be identified through PCR of DNA prepared from pools of injected ES cells. Positive cells within the pools can be identified by PCR subsequent to cell cloning (Zimmer and Gruss, Nature 338:150-153 (1989). DNA introduction by electroporation is less efficient and requires a selection step. Methods for positive selection of the recombination event (for example, neo resistance) and dual positive-negative selection (for example, neo resistance and gancyclovir resistance) and the subsequent identification of the desired clones by PCR have been described by Joyner et al., Nature 338:153-156 (1989), and Capecchi (1989), the teachings of which are incorporated herein.

2. Embryo Recovery and ES Cell Injection.

Naturally cycling or superovulated female mice mated with males can be used to harvest embryos for the implantation of ES cells. It is desirable to use the C57BL/6 strain for this purpose when using mice. Embryos of the appropriate age are recovered approximately 3.5 days after successful mating. Mated females are sacrificed by CO₂ asphyxiation or cervical dislocation and embryos are flushed from excised uterine horns and placed in Dulbecco's modified essential medium plus 10% calf serum for injection with ES cells. Approximately 10 to 20 ES cells are injected into blastocysts using a glass microneedle with an internal diameter of approximately 20 .mu.m.

3. Transfer of Embryos to Pseudopregnant Females.

Randomly cycling adult female mice are paired with vasectomized males. Mouse strains such as Swiss Webster, ICR or others can be used for this purpose. Recipient females are mated such that they will be at 2.5 to 3.5 days post-mating when required for implantation with blastocysts containing ES cells. At the time of embryo transfer, the recipient females are anesthetized with an intraperitoneal injection of 0.015 ml of 2.5% avertin per gram of body weight. The ovaries are exposed by making an incision in the body wall directly over the oviduct and the ovary and uterus are externalized. A hole is made in the uterine horn with a 25 gauge needle through which the blastocysts are transferred. After the transfer, the ovary and uterus are pushed back into the body and the incision is closed by two sutures. This procedure is repeated on the opposite side if additional transfers are to be made.

Identification, Characterization, and Utilization of Transgenic Mice and Rats.

Transgenic rodents can be identified by analyzing their DNA. For this purpose, tail samples (1 to 2 cm) can be removed from three week old animals. DNA from these or other samples can then be prepared and analyzed by Southern blot, PCR, or slot blot to detect transgenic founder (F₀) animals and their progeny (F₁ and F₂).

A. Pathological Studies.

The various F₀, F₁, and F₂ animals that carry a transgene can be analyzed by immunohistology for evidence of A.beta. deposition, expression of APP or APP cleavage products, neuronal or neuritic abnormalities, and inflammatory responses in the brain. Brains of mice and rats from each transgenic line are fixed and then sectioned. Sections are stained with antibodies reactive with the APP and/or the A.beta.. Secondary antibodies conjugated with fluorescein, rhodamine, horse radish peroxidase, or alkaline phosphatase are used to detect the primary antibody. These methods permit identification of amyloid plaques and other pathological lesions in specific areas of the brain. Plaques ranging in size from 9 to >50 .mu.m characteristically occur in the brains of AD patients in the cerebral cortex, but also may be observed in deeper grey matter including the amygdaloid nucleus, corpus striatum and diencephalon. Sections can also be stained with other antibodies diagnostic of Alzheimer's plaques, recognizing antigens such as APP, Alz-50, tau, A2B5, neurofilaments, synaptophysin, MAP-2, ubiquitin, complement, neuron-specific enolase, and others that are characteristic of Alzheimer's pathology (Wolozin et al., Science 232:648 (1986); Hardy and Allsop, Trends in Pharm. Sci. 12:383-388 (1991); Selkoe, Ann. Rev. Neurosci. 12:463-490 (1989); Arai et al., Proc. Natl. Acad. Sci. USA 87:2249-2253 (1990); Majocha et al., Amer. Assoc. Neuropathology Abs 99:22 (1988); Masters et al., Proc. Natl. Acad. Sci. 82:4245-4249 (1985); Majocha et al., Can J Biochem Cell Biol 63:577-584 (1985)). Staining with thioflavin S and Congo Red can also be carried out to analyze the presence of amyloid and co-localization of A.beta. deposits within neuritic plaques and NFTs.

B. Analysis of APP and A.beta. Expression.

1. mRNA.

Messenger RNA can be isolated by the acid guanidinium thiocyanate-phenol:chloroform extraction method (Chomaczynski and Sacchi, Anal Biochem 162:156-159 (1987)) from cell lines and tissues of transgenic animals to determine expression levels by Northern blots, RNAse and nuclease protection assays.

2. Protein.

APP, A.beta., and other fragments of APP can and have been detected by using polyclonal and monoclonal antibodies that are specific to the APP extra-cytoplasmic domain, A.beta. region, A.beta._1-42, A.beta._1-40, APP.beta., FLAPP+APP.alpha., and C-terminus of APP. A variety of antibodies that are human sequence specific, such as 10D5 and 6C6, are very useful for this purpose (Games et al. (1995)).

3. Western Blot Analysis.

Protein fractions can be isolated from tissue homogenates and cell lysates and subjected to Western blot analysis as described by, for example, Harlow et al., Antibodies: A Laboratory Manual, (Cold Spring Harbor, N.Y., 1988); Brown et al., J. Neurochem. 40:299-308 (1983); and Tate-Ostroff et al., Proc Natl Acad Sci 86:745-749 (1989).

Briefly, the protein fractions are denatured in Laemmli sample buffer and electrophoresed on SDS-Polyacrylamide gels. The proteins are then transferred to nitrocellulose filters by electroblotting. The filters are blocked, incubated with primary antibodies, and finally reacted with enzyme conjugated secondary antibodies. Subsequent incubation with the appropriate chromogenic substrate reveals the position of APP derived proteins.

C. Pathological and Behavioral Studies.

1. Pathological Studies.

Immunohistology and thioflavin S staining are conducted as described elsewhere herein.

In situ Hybridizations:

Radioactive or enzymatically labeled nucleic acid probes can be used to detect mRNA in situ. The probes are degraded or prepared to be approximately 100 nucleotides in length for better penetration of cells. The hybridization procedure of Chou et al., J. Psych. Res. 24:27-50 (1990), for fixed and paraffin embedded samples is briefly described below although similar procedures can be employed with samples sectioned as frozen material. Paraffin slides for in situ hybridization are dewaxed in xylene and rehydrated in a graded series of ethanols and finally rinsed in phosphate buffered saline (PBS). The sections are post-fixed in fresh 4% paraformaldehyde. The slides are washed with PBS twice for 5 minutes to remove paraformaldehyde. Then the sections are permeabilized by treatment with a 20 .mu.g/ml proteinase K solution. The sections are re-fixed in 4% paraformaldehyde, and basic molecules that could give rise to background probe binding are acetylated in a 0.1 M triethanolamine, 0.3 M acetic anhydride solution for 10 minutes. The slides are washed in PBS, then dehydrated in a graded series of ethanols and air dried. Sections are hybridized with antisense probe, using sense probe as a control. After appropriate washing, bound radioactive probes are detected by autoradiography or enzymatically labeled probes are detected through reaction with the appropriate chromogenic substrates.

2. Behavioral Studies.

Behavioral tests designed to assess learning and memory deficits are employed. An example of such as test is the Morris water maze (Morris, Learn Motivat. 12:239-260 (1981)). In this procedure, the animal is placed in a circular pool filled with water, with an escape platform submerged just below the surface of the water. A visible marker is placed on the platform so that the animal can find it by navigating toward a proximal visual cue. Alternatively, a more complex form of the test in which there are no local cues to mark the platform's location will be given to the animals. In this form, the animal must learn the platform's location relative to distal visual cues, and can be used to assess both reference and working memory. A learning deficit in the water maze has been demonstrated with PDAPP transgenic mice. An example of behavioral analysis for assessing the effect of transgenic expression of A.beta.-containing proteins is described in Example 9.

Operant Behavior Studies of Memory Function:

Memory function of the disclosed transgenic animals can be assessed by testing memory-related feeding behavior (Dunnett, "Operant delayed matching and non-matching position in rats" in Behavioral Neuroscience, Volume I: A Practical Approach (Sagal, ed., IRL Press, N.Y., 1993) pages 123-136; Zornetzen, Behav. Neur. Biol. 36:49-60 (1982)). Transgenic and non-transgenic mice, are trained to earn food rewards in a two component operant procedure. One component features a delayed spatial alternation schedule. Under this schedule, the mouse must remember over a variable time delay which lever it has pressed in the previous trial so that it can earn a reward by pressing the alternate lever on the current trial. This provides a measure of the animal's recent or "working" memory. The second component features a discrimination spatial alternation schedule. Under this schedule, the mouse earns a reward by pressing whatever lever is illuminated. This discrimination behavior is an example of reference memory. These two groups of mice, transgenic and non-transgenic, can be chronically studied over time, for example, from 3 months of age until the end of their useful life span, in order to assess the development of sensitivity to cholinergic antagonists and behavioral impairment on these memory tasks. It is expected that the disclosed transgenic mice will model the cognitive deficits of Alzheimer's disease with enhanced sensitivity to the memory-disrupting effects of cholinergic antagonists and impairment on "working" and reference memory tasks. Dose-response challenges with the cholinergic antagonist can be conducted at various ages. These memory behavioral tests can also be used to compare the effect of compounds on the behavioral impairment of the disclosed transgenic animals. In this case, the two groups of mice are transgenic mice to which a test compound is administered and transgenic mice to which the compound is not administered.

Emotional Reactivity and Object Recognition:

Various functions of the disclosed transgenic animals can be assessed by testing locomotor activity, emotional reactivity to a novel environment or to novel objects, and object recognition. A first set of assessments are performed in the same animals at different ages (each animal is its own control) in order to test their performance in terms of locomotor activity, emotional reactivity to a novel environment or to novel objects, and object recognition, a form of memory which is severely impaired in AD patients. On the first day, transgenic and non-transgenic control mice are individually placed in a square open field with a central platform. For 30 minutes, horizontal and vertical activity, and crossings of the platform, are recorded by blocks of 5 minutes for each animal. On the second day, each animal is submitted to two trials with an intertrial of 1 hour. On the first trial, two identical objects are placed in the open field and the animal is allowed 3 minutes of exploration. On the second trial, one of the objects is replaced by a new object and the time spent by the animal in exploring the familiar and novel object is recorded during the next 3 minutes (Ennaceur and Delacour, Behav. Brain Res. 31:47-59 (1988)). Animals are then tested for neophobic behavior, which is considered as an index of anxiety, in a free exploration situation, in which animals are given the opportunity to move freely between a familiar and a novel environment.

Thereafter, the same animals are submitted to various learning tasks to investigate their learning and memory capacities. They are first tested for spatial recognition memory in a T-maze delayed alternation task at 6 hour and 24 hour delays. This form of memory has been shown to be very sensitive to hippocampal damage. One half of the animals of each group is then trained in a positively reinforced lever-press task as described above. This can be used to measure post training improvement in performance of the animals, which has been shown to involve hippocampal activation. The other half of the animals is trained in spatial discrimination in an 8 arm radial maze (Oltons and Samuelson, J. Exp. Psychol. [Animal Behav.] 2:97-116 (1976)) in order to evaluate working and reference memory and to analyze their strategies (angle preference), which give a better index of memory capacities. The animals trained and tested in the bar-lever press task at 2 to 3 months old can be trained and tested in radial maze at 9 to 10 months old, and vice-versa. A working memory deficit has been demonstrated in PDAPP transgenic mice in the radial arm maze.

Two additional groups can also be submitted to the same behavioral tests as above at 9 to 10 months old in order to determine whether behavioral screening performed at 2 to 3 months old influenced further learning and memory capacities.

These memory behavioral tests can also be used to compare the effect of compounds on the behavioral impairment of the disclosed transgenic animals. In this case, the two groups of mice are transgenic mice to which a test compound is administered, and transgenic mice to which the compound is not administered.

The procedures applied to test transgenic mice are similar for transgenic rats.

D. Preferred Characteristics.

The above phenotypic characteristics of the disclosed transgenic animals can be used to identify those forms of the disclosed transgenic animals that are preferred as animal models. Additional phenotypic characteristics, and assays for measuring these characteristics, that can also be used to identify those forms of the disclosed transgenic animals that are preferred as animal models, are described in Example 6. These characteristics are preferably those that are similar to phenotypic characteristics observed in Alzheimer's disease. APP and A.beta. markers which are also useful for identifying those forms of the disclosed transgenic animals that are preferred as animal models are described below. Any or all of the these markers or phenotypic characteristics can be used either alone or in combination to identify preferred forms of the disclosed transgenic animals. For example, the presence of plaques in brain tissue that can be stained with Congo red is a phenotypic characteristic which can identify a disclosed transgenic animal as preferred. It is intended that the levels of expression of certain APP-related proteins present in preferred transgenic animals (discussed above) is an independent characteristic for identifying preferred transgenic animals. Thus, the most preferred transgenic animals will exhibit both a disclosed expression level for one or more of the APP-related proteins and one or more of the phenotypic characteristics discussed above. Especially preferred phenotypic characteristics (the presence of which identifies the animal as a preferred transgenic animal) are the presence of amyloid plaques that can be stained with Congo Red (Kelly (1984)), the presence of extracellular amyloid fibrils as identified by electron microscopy by 12 months of age, and the presence of type I dystrophic neurites as identified by electron microscopy by 12 months of age (composed of spherical neurites that contain synaptic proteins and APP; Dickson et al., Am J Pathol 132:86-101 (1988); Dickson et al., Acta Neuropath. 79:486-493 (1990); Masliah et al., J Neuropathol Exp Neurol 52:135-142 (1993); Masliah et al., Acta Neuropathol 87:135-142 (1994); Wang and Munoz, J Neuropathol Exp Neurol 54:548-556 (1995)). Examples of the detection of these characteristics is provided in Example 6. It is most preferred that the transgenic animals have amyloid plaques that can be stained with Congo Red as of 14 months of age.

Screening of Compounds for Treatment of Alzheimer's Disease.

The transgenic animals, or animal cells derived from transgenic animals, can be used to screen compounds for a potential effect in the treatment of Alzheimer's disease using standard methodology. In such AD screening assays, the compound is administered to the animals, or introduced into the culture media of cells derived from these animals, over a period of time and in various dosages, then the animals or animal cells are examined for alterations in APP expression or processing, expression levels or localization of other AD markers, histopathology, and/or, in the case of animals, behavior using the procedures described above and in the examples below. In general, any improvement in behavioral tests, alteration in AD-associated markers, reduction in the severity of AD-related histopathology, reduction in the expression of A.beta. or APP cleavage products, and/or changes in the presence, absence or levels of other compounds that are correlated with AD which are observed in treated animals, relative to untreated animals, is indicative of a compound useful for treating Alzheimer's disease. The specific proteins, and the encoding transcripts, the enzymatic or biochemical activity, and/or histopathology of those proteins, that are associated with and characteristic of AD are referred to herein as markers. Expression or localization of these markers characteristic of AD has either been detected, or is expected to be present, in the disclosed transgenic animals. These markers can be measured or detected, and those measurements compared between treated and untreated transgenic animals to determine the effect of a tested compound.

Markers useful for AD screening assays are selected based on detectable changes in these markers that are associated with AD. Many such markers have been identified in AD and have either been detected in the disclosed transgenic animals or are expected to be present in these animals. These markers fall into several categories based on their nature, location, or function. Preferred examples of markers useful in AD screening assays are described below, group as A.beta.-related markers, plaque-related markers, cytoskeletal and neuritic markers, inflammatory markers, and neuronal and neurotransmitter-related markers.

A. A.beta.-related Markers.

Expression of the various forms of APP and A.beta. can be directly measured and compared in treated and untreated transgenic animals both by immunohistochemistry and by quantitative ELISA measurements as described above and in the examples. Currently, it is known that two forms of APP products are found, APP and A.beta. (Haass and Selkoe, Cell 75:1039-1042 (1993)). They have been shown to be intrinsically associated with the pathology of AD in a time dependent manner. Therefore, preferred assays compare age-related changes in APP and A.beta. expression in the transgenic mice. As described in Example 6, increases in A.beta. have been demonstrated during aging of the PDAPP mouse.

Preferred targets for assay measurement are A.beta. markers known to increase in individuals with Alzheimer's disease are total A.beta. (A.beta._tot) A.beta. 1-42 (A.beta._1-42 ; A.beta. with amino acids 1-42), A.beta._1-40 (A.beta. with amino acids 1-40), A.beta. N3(pE) (A.beta._N3 (pE)); A.beta._X-42 (A.beta._X-42 ; A.beta. forms ending at amino acid 42); A.beta. X-40 (A.beta._X-40 ; A.beta. forms ending at amino acid 40); insoluble A.beta. (A.beta._Isoluble); and soluble A.beta. (A.beta._Soluble ; Kuo et al., J. Biol. Chem. 271(8):4077-4081 (1996)). A.beta._N3 (pE) has pyroglutamic acid at position 3 (Saido, Neuron 14:457-466 (1995)). A.beta._X-42 refers to any of the C-terminal forms of A.beta. such as A.beta._13-42. A.beta._Insoluble refers to forms of A.beta. that are recovered as described in Gravina, J. Biol. Chem. 270:7013-7016 (1995). APP.beta. can also be specifically measured to assess the amount of .beta.-secretase activity (Seubert et al., Nature 361:260-263 (1993)). Several of these A.beta. forms and their association with Alzheimer's disease are described by Haass and Selkoe (1993). Detection and measurement of A.beta._tot, A.beta._1-42, and A.beta._X-42 are described in Example 6. Generally, specific forms of A.beta. can be assayed, either quantitatively or qualitatively using specific antibodies, as described below. When referring to amino acid positions in forms of A.beta., the positions correspond to the A.beta. region of APP. Amino acid 1 of A.beta. corresponds to amino acid 672 of APP, and amino acid 42 of A.beta. corresponds to amino acid 714 of APP.

Also preferred as targets for assay measurement are APP markers. For example, different forms of secreted APP (termed APP.alpha. and APP.beta.) can also be measured (Seubert et al., Nature 361:260-263 (1993)). Other APP forms can also serve as targets for assays to assess the potential for compounds to affect Alzheimer's disease. These include FLAPP+APP.alpha., full length APP, C-terminal fragments of APP, especially C100 (the last 100 amino acids of APP) and C57 to C60 (the last 57 to 60 amino acids of APP), and any forms of APP that include the region corresponding to A.beta._1-40.

APP forms are also preferred targets for assays to assess the potential for compounds to affect Alzheimer's disease. The absolute level of APP and APP transcripts, the relative levels of the different APP forms and their cleavage products, and localization of APP expression or processing are all markers associated with Alzheimer's disease that can be used to measure the effect of treatment with potential therapeutic compounds. The localization of APP to plaques and neuritic tissue is an especially preferred target for these assays.

Quantitative measurement can be accomplished using many standard assays. For example, transcript levels can be measured using RT-PCR and hybridization methods including RNase protection, Northern analysis, and R-dot analysis. APP and A.beta. levels can be assayed by ELISA, Western analysis, and by comparison of immunohistochemically stained tissue sections. Immunohistochemical staining can also be used to assay localization of APP and A.beta. to particular tissues and cell types. Such assays were described above and specific examples are provided below.

B. Plaque-related Markers.

A variety of other molecules are also present in plaques of individuals with AD and in the disclosed transgenic animals, and their presence in plaques and neuritic tissue can be detected. The amount of these markers present in plaques or neuritic tissue is expected to increase with the age of untreated transgenic animals. Preferred plaque-related markers are apolipoprotein E, glycosylation end products, amyloid P component, advanced glycosylation end products (Smith et al., Proc. Natl. Acad. Sci. USA 91:5710 (1994)), growth inhibitory factor, laminin, collagen type IV (Kalaria and Perry (1993); Ueda et al. (1993)), receptor for advanced glycosylation products (RAGE), and ubiquitin.

While the above markers can be used to detect specific components of plaques and neuritic tissue, the location and extent of plaques can also be determined by using well known histochemical stains, such as Congo Red and thioflavin S, as described above and in some examples below.

C. Cytoskeletal and Neuritic Markers.

Many changes in cytoskeletal markers associated with AD have also been detected in transgenic PDAPP mice. These markers can be used in AD screening assays to determine the effect of compounds on AD. Many of the changes in cytoskeletal markers occur either in the neurofibrillary tangles or dystrophic neurites associated with plaques (Kosik et al. (1992); Lovestone and Anderton (1992); Brandan and Inestrosa (1993); Trojanowski et al. (1993); Masliah et al. (1993)).

The following are preferred cytoskeletal and neuritic markers that exhibit changes in and/or an association with AD. These markers can be detected, and changes can be determined, to measure the effect of compounds on the disclosed transgenic animals. Spectrin exhibits increased breakdown in AD. Tau and neurofilaments display an increase in hyperphosphorylation in AD, and levels of ubiquitin increase in AD. Tau, ubiquitin, MAP-2, neurofilaments, heparin sulfate, and chrondroitin sulphate are localized to plaques and neuritic tissue in AD and in general change from the normal localization. GAP43 levels are decreased in the hippocampus and abnormally phosphorylated tau and neurofilaments are present in PDAPP transgenic mice.

D. Inflammatory Markers.

Alzheimer's disease is also known to stimulate an immunoinflammatory response, with a corresponding increase in inflammatory markers (Frederickson and Brunden (1994); McGeer et al. (1991); Wood et al. (1993)). The following are preferred inflammatory markers that exhibit changes in and/or an association with AD. Detection of changes in these markers are useful in AD screening assays. Acute phase proteins and glial markers, such as .alpha.1-antitrypsin, C-reactive protein, .alpha.2-macroglobulin (Tooyama et al., Molecular & Chemical Neuropathology 18:153-60 (1993)), glial fibrillary acidic protein (GFAP), Mac-1, F4/80, and cytokines, such as IL-1.alpha. and .beta., TNF.alpha., IL-8, MIP-1.alpha. (Kim et al., J. Neuroimmunology 56:127-134 (1995)), MCP-1 (Kim et al., J. Neurological Sciences 128:28-35 (1995); Kim et al., J. Neuroimmunology 56:127-134 (1995); Wang et al., Stroke 26:661-665 (1995)), and IL-6, all increase in AD and are expected to increase in the disclosed transgenic animals. Complement markers, such as C3d, C1q, C5, C4d, C4bp, and C5a-C9, are localized in plaques and neuritic tissue. Major histocompatibility complex (MHC) glycoproteins, such as HLA-DR and HLA-A, D,C increase in AD. Microglial markers, such as CR3 receptor, MHC I, MHC II, CD 31, CD11a, CD11b, CD11c, CD68, CD45RO, CD45RD, CD18, CD59, CR4, CD45, CD64, and CD44 (Akiyama et al., Brain Research 632:249-259 (1993)) increase in AD. Additional inflammatory markers useful in AD screening assays include .alpha.2 macroglobulin receptor, Fibroblast growth factor (Tooyama et al., Neuroscience Letters 121:155-158 (1991)), ICAM-1 (Akiyama et al., Acta Neuropathologica 85:628-634 (1993)), Lactotransferrin (Kawamata et al., American Journal of Pathology 142:1574-85 (1993)), C1q, C3d, C4d, C5b-9, Fc gamma RI, Fc gamma RII, CD8 (McGeer et al., Can J Neurol Sci 16:516-527 (1989)), LCA (CD45) (McGeer et al. (1989); Akiyama et al., Journal of Neuroimmunology 50:195-201 (1994)), CD18 (beta-2 integrin) (Akiyama and McGeer, Journal of Neuroimmunology 30:81-93 (1990)), CD59 (McGeer et al., Brain Research 544:315-319 (1991)), Vitronectic (McGeer et al., Canadian Journal of Neurological Sciences 18:376-379 (1991); Akiyama et al., Journal of Neuroimmunology 32:19-28 (1991)), Vitronectin receptor, Beta-3 integrin (Akiyama et al. (1991)), Apo J, clusterin (McGeer et al., Brain Research 579:337-341 (1992)), type 2 plasminogen activator inhibitor (Akiyama et al., Neuroscience Letters 164:233-235 (1993)), CD44 (Akiyama et al., Brain Research 632:249-259 (1993)), Midkine (Yasuhara et al., Biochemical & Biophysical Research Communications 192:246-251 (1993)), Macrophage colony stimulating factor receptor (Akiyama et al., Brain Research 639:171-174 (1994)), MRP14, 27E10, and interferon-alpha (Akiyama et al., Journal of Neuroimmunology 50:195-201 (1994)). Additional markers which are associated with inflammation or oxidative stress include 4-hydroxynonenal-protein conjugates (Uchida et al., Biochem. Biophys. Res. Comm. 212:1068-1073 (1995); Uchida and Stadtman, Methods in Enzymology 233:371-380 (1994); Yoritaka et al., Proc. Natl. Acad. Sci. USA 93:2696-2701 (1996)), I.kappa.B, NF.kappa.B (Kaltschmidt et al., Molecular Aspects of Medicine 14:171-190 (1993)), cPLA₂ (Stephenson et al., Neurobiology Dis. 3:51-63 (1996)), COX-2 (Chen et al., Neuroreport 6:245-248 (1995)), Matrix metalloproteinases (Backstrom et al., J. Neurochemistry 58:983-992 (1992); Bignami et al., Acta Neuropathologica 87:308-312 (1994); Deb and Gottschall, J. Neurochemistry 66:1641-1647 (1995); Peress et al., J. Neuropathology & Experimental Neurology 54:16-22 (1995)), Membrane lipid peroxidation, Protein oxidation (Hensley et al., J. Neurochemistry 65:2146-2156 (1995); Smith et al., Proc. Natl. Acad. Sci. USA 88:10540-10543 (1991)), and diminished ATPase activity (Mark et al., J. Neuroscience 15:6239 (1995)). These markers can be detected, and changes can be determined, to measure the effect of compounds on the disclosed transgenic animals.

E. Neuronal and Neurotransmitter-related Markers.

Changes in neuronal and neurotransmitter biochemistry have been associated with AD and in the disclosed PDAPP animals. In AD there is a profound reduction in cortical and hippocampal cholinergic innervation. This is evidenced by the dramatic loss of the synthetic enzyme choline acetyltransferase and decreased acetylcholinersterase, synaptosomal choline uptake (as measured by hemicholinium binding) and synthesis and release of acetylcholine (Rylett et al. (1983); Sims et al. (1980); Coyle et al., Science 219:1184-1190 (1983); Davies and Maloney, Lancet 2:1403 (1976); Perry et al., Lancet 1:189 (1977); Sims et al., J. Neurochem. 40: 503-509 (1983)) all of which are useful markers. These markers can be used in AD screening assays to determine the effect of compounds on AD. There is also a loss of basal forebrain neurons and the galanin system becomes hypertrophic in AD.

In addition to changes in the markers described above in AD, there is also atrophy and loss of basal forebrain cholinergic neurons that project to the cortex and hippocampus (Whitehouse et al., Science 215:1237-1239 (1982)), as well as alterations of entorhinal cortex neurons (Van Hoesan et al., Hippocampus 1:1-8 (1991). Based upon these observations measurement of these enzyme activities, neuronal size, and neuronal count numbers are expected to decrease in the disclosed transgenic animals and are therefore useful targets for detection in AD screening assays. Basal forebrain neurons are dependent on nerve growth factor (NGF). Brain-derived neurotrophic factor (BDNF) may also decrease in the hippocampus in the disclosed transgenic animals and is therefore a useful target for detection in AD screening assays.

It has also been shown that APP and A.beta. release are affected by stimulation of muscarinic receptors both in vitro in tissue culture as well as in brain slices. Similar findings have also been obtained with application of other agonists linked to phosphoinosital turnover (Nitsch et al. (1992); Hung et al., J. Biol. Chem. 268:22959-22962 (1993); Nitsch et al., Proceedings of the Eighth Meeting of the International Study Group on the Pharmacology of Memory Disorders Associated with Aging 497-503 (1995); Masliah and Terry (1993); Greenamyre and Maragos (1993); McDonald and Nemeroff (1991); Mohr et al. (1994); Perry, British Medical Bulletin 42:63-69 (1986); Masliah et al., Brian Research 574:312-316 (1992); Schwagerl et al., Journal of Neurochemistry 64:443-446 (1995)). Based upon these observations, it is possible that neurotransmitter agonists will reduce the production of A.beta. in the disclosed transgenic animals. Based on this reasoning, screening assays that measure the effect of compounds on neurotransmitter receptors can possibly be used to identifying compounds useful in treating AD.

In addition to the well-documented changes in the cholinergic system, dysfunction in other receptor systems such as the serotinergic, adrenergic, adenosine, and nicotine receptor systems, has also been documented. Markers characteristic of these changes, as well as other neuronal markers that exhibit both metabolic and structural changes in AD are listed below. Changes in the level and/or localization of these markers can be measured using similar techniques as those described for measuring and detecting the earlier markers.

The following are preferred cytoskeletal and neuritic markers that exhibit changes in and/or an association with AD. Levels of cathepsin (cat) D,B and Neuronal Thread Protein, and phosphorylation of elongation factor-2, increase in AD. Cat D,B, protein kinase C, and NADPH are localized in plaque and neuritic tissue in AD. Activity and/or levels of nicotine receptors, 5-HT₂ receptor, NMDA receptor, .alpha.2-adrenergic receptor, synaptophysin, p65, glutamine synthetase, glucose transporter, PPI kinase, drebrin, GAP43, cytochrome oxidase, heme oxygenase, calbindin, adenosine A1 receptors, mono amine metabolites, choline acetyltransferase, acetylcholinesterase, and symptosomal choline uptake are all reduced in AD.

Additional markers that are associated with AD or after treatment of cells with A.beta. include (1) cPLA₂ (Stephenson et al., Neurobiology of Diseases 3:51-63 (1996)), which is upregulated in AD, (2) Heme oxygenase-1 (Premkumar et al., J. Neurochemistry 65:1399-1402 (1995); Schipper et al., Annals of Neurology 37:758-768 (1995); Smith et al., American Journal of Pathology 145:42-47 (1994); Smith et al., Molecular & Chemical Neuropathology 24:227-230 (1995)), c-jun (Anderson et al., Experimental Neurology 125:286-295 (1994); Anderson et al., J. Neurochemistry 65:1487-1498 (1995)), c-fos (Anderson et al. (1994); Zhang et al., Neuroscience 46:9-21 (1992)), HSP27 (Renkawek et al., Acta Neuropathologica 87:511-519 (1994); Renkawek et al., Neuroreport 5:14-16 (1993)), HSP70 (Cisse et al., Acta Neuropathologica 85:233-240 (1993)), and MAP5 (Geddes et al., J. Neuroscience Research 30:183-191 (1991); Takahashi et al., Acta Neuropathologica 81:626-631 (1991)), which are induced in AD and in cortical cells after A.beta. treatment, and (3) junb, jund, fosB, fra1 (Estus et al., J. Cell Biology 127:1717-1727 (1994)), cyclin D1 (Freeman et al., Neuron 12:343-355 (1994); Kranenburg et al., EMBO Journal 15:46-54 (1996)), p53 (Chopp, Current Opinion in Neurology & Neurosurgery 6:6-10 (1993); Sakhi et al., Proc. Natl. Acad. Sci. USA 91:7525-7529 (1994); Wood and Youle, J. Neuroscience 15:5851-5857 (1995)), NGFI-A (Vaccarino et al., Molecular Brain Research 12:233-241 (1992)), and NGFI-B, which are induced in cortical cells after A.beta. treatment.

F. Measuring the Amounts and Localization of AD Markers.

Quantitative measurement can be accomplished using many standard assays. For example, transcript levels can be measured using RT-PCR and hybridization methods including RNase protection, Northern analysis, and R-dot analysis. Protein marker levels can be assayed by ELISA, Western analysis, and by comparison of immunohistochemically stained tissue sections. Immunohistochemical staining can also be used to assay localization of protein markers to particular tissues and cell types. The localization and the histopathological association of AD markers can be determined by histochemical detection methods such as antibody staining, laser scanning confocal imaging, and immunoelectron micrography. Examples of such techniques are described in Masliah et al. (1993) and in Example 6 below.

In the case of receptors and enzymatic markers, activity of the receptors or enzymes can be measured. For example, the activity of neurotransmitter metabolizing enzymes such as choline acetyltransferase and acetylcholine esterase can be measured using standard radiometric enzyme activity assays.

The activity of certain neurotransmitter receptors can be determined by measuring phosphoinositol (PI) turnover. This involves measuring the accumulation of inositol after stimulation of the receptor with an agonist. Useful agonists include carbachol for cholinergic receptors and norepinephrine for glutaminergic receptors. The number of receptors present in brain tissue can be assessed by quantitatively measuring ligand binding to the receptors.

The levels and turnover of receptor ligands and neurotransmitters can be determined by quantitative assays taken at various time points. Dopamine turnover can be measured using DOPAC and HVA. MOPEG sulfate can be used to measure norepinephrine turnover and 5-HIAA can be used to measure serotonin turnover. For example, norepinephrine levels have been shown to be reduced 20% in the hippocampus of 12 to 13 month old PDAPP transgenic mice relative to controls. Generally, the above assays can be performed as described in the literature, for example, in Rylett et al. (1983); Sims et al. (1980); Coyle et al., Science 219:1184-1190 (1983); Davies and Maloney, Lancet 2:1403 (1976); Perry et al., Lancet 1:189 (1977); Sims et al., J. Neurochem. 40: 503-509 (1983). These markers are also described by Bymaster et al., J. Pharm. Exp. Ther. 269:282-289 (1994).

G. Screening Assays Using Cultured Cells.

Screening assays for determining the therapeutic potential of compounds can also be performed using cells derived from animals transgenic for the disclosed APP constructs and cell cultures stably transfected with the disclosed constructs. For example, such assays can be performed on cultured cells in the following manner. Cell cultures can be transfected generally in the manner described in International Patent Application No. 94/10569 and Citron et al. (1995). Derived transgenic cells or transfected cell cultures can then be plated in Corning 96-well plates at 1.5 to 2.5x10⁴ cells per well in Dulbecco's minimal essential media plus 10% fetal bovine serum.

Following overnight incubation at 37^oC. in an incubator equilibrated with 10% carbon dioxide, media are removed and replaced with media containing a compound to be tested for a two hour pretreatment period and cells were incubated as above. Stocks containing the compound to be tested are first prepared in 100% dimethylsulfoxide such that at the final concentration of compound used in the treatment, the concentration of dimethylsulfoxide does not exceed 0.5%, preferably about 0.1%.

At the end of the pretreatment period, the media are again removed and replaced with fresh media containing the compound to be tested as above and cells are incubated for an additional 2 to 16 hours. After treatment, plates are centrifuged in a Beckman GPR at 1200 rpm for five minutes at room temperature to pellet cellular debris from the conditioned media. From each well, 100 .mu.L of conditioned media or appropriate dilutions thereof are transferred into an ELISA plate precoated with antibody 266 (an antibody directed against amino acids 13 to 28 of A.beta.) as described in International Patent Application No. 94/10569 and stored at 4^oC. overnight. An ELISA assay employing labelled antibody 6C6 (against amino acids 1 to 16 of A.beta.) can be run to measure the amount of A.beta. produced. Different capture and detection antibodies can also be used.

Cytotoxic effects of the compounds are measured by a modification of the method of Hansen et al., J. Immun. Method. 119:203-210 (1989). To the cells remaining in the tissue culture plate, 25 .mu.L of a 3,(4,5-dimethylthiazol-2-yl)2,5-diphenyltetrazolium bromide (MTT) stock solution (5 mg/mL) is added to a final concentration of 1 mg/mL. Cells are incubated at 37^oC. for one hour, and cellular activity is stopped by the addition of an equal volume of MTT lysis buffer (20% w/v sodium dodecylsulfate in 50% dimethylformamide, pH 4.7). Complete extraction is achieved by overnight shaking at room temperature. The difference in the OD_562nm and the OD_650nm is measured in a Molecular Device's UV_max microplate reader, or equivalent, as an indicator of the cellular viability.

The results of the A.beta. ELISA are fit to a standard curve and expressed as ng/mL A.beta.. In order to normalize for cytotoxicity, these results are divided by the MTT results and expressed as a percentage of the results from a control assay run without the compound.

Claim 1 of 56 Claims

We claim:

1. A method of selecting a transgenic mouse as a model of Alzheimer's disease, comprising

providing a plurality of transgenic mice, each comprising a nucleic acid construct stably incorporated into the genome, wherein the construct comprises a promoter for expression of the construct in a mammalian cell and a region encoding an Au-containing protein, wherein the promoter is operatively linked to the region,

wherein the region comprises DNA encoding the A.beta.-containing protein, wherein the A.beta.-containing protein consists of all or a contiguous portion of a protein selected from the group consisting of

APP770 bearing a mutation in one or more of the amino acids selected from the a group consisting of amino acid 669, 670, 671, 690, 692, and 717, APP751 bearing a mutation in one or more of the amino acids selected from the group consisting of amino acid 669, 670, 671, 690, 692, and 717, and APP695 bearing a mutation in one or more of the amino acids selected from the group consisting of amino acid 669, 670, 671, 690, 692, and 717; a protein consisting of amino acids 672 to 770 of APP; and a protein consisting of amino acids 672 to 714 of APP;

determining expression levels of APP, APP.beta., APP.alpha. and/or A.beta. in each of the transgenic mice;

identifying a transgenic mouse wherein A.beta.tot is expressed at a level of at least 30 nanograms per gram of brain tissue of the mouse when it is two to four months old, A.beta.1-42 is expressed at a level of at least 8.5 nanograms per gram of brain tissue of the mouse when it is two to four months old, APP and APP.alpha. combined are expressed at a level of a least 150 picomoles per gram of brain tissue of the mouse when it is two to four months old, APP.beta. is expressed at a level of at least 40 picomoles per gram of brain tissue of the mouse when it is two to four months old, and/or mRNA encoding the A.beta.-containing protein is expressed to a level at least twice that of mRNA encoding the cndogenous APP of the transgenic mouse in brain tissue of the mouse when it is two to four months old;

using an offspring of the identified transgenic mouse as a model of Alzheimer's disease.

____________________________________________
If you want to learn more about this patent, please go directly to the U.S. Patent and Trademark Office Web site to access the full patent.