WO2009132863A1

WO2009132863A1 - Knock-in mouse for modelling blockade of human tnfalpha

Info

Publication number: WO2009132863A1
Application number: PCT/EP2009/003356
Authority: WO
Inventors: Sergei Nedospasov
Original assignee: Deutsches Rheuma-Forschungszentrum Berlin
Priority date: 2008-04-30
Filing date: 2009-04-29
Publication date: 2009-11-05

Abstract

The invention relates to a transgenic mouse, wherein the complete endogenous murine TNFα-gene including its upstream regulatory sequences is replaced by the human homologous TNFα-gene including its upstream regulatory sequences or a mutant thereof, wherein the mutant is characterized by deletion, addition, substitution, and/or inversion of one or more nucleotides so that the functional properties of the modified human TNFα-gene are conserved and cells, tissues and cell lines derived thereof. The invention further relates to an in-vivo method of screening, investigating and/or preclinical testing of compounds and/or drug candidates for human anti-TNFα therapy using said transgenic mouse and an analogous method using the cells, tissues and cell lines derived thereof.

Description

Knock-in mouse for modelling blockade of human TNFalpha

The present invention relates to a transgenic mouse, in particular a knock-in mouse, tissues, cells and/or cell lines derived from said transgenic mouse and a recombinant vector intended for the generation of such a mouse wherein the endogenous murine TNF-α gene is replaced by the human homologue. Furthermore, the invention relates to stem cells, preferably murine embryonic stem cells comprising said recombinant vector as well as to a method for screening, investigating and/or preclinical testing of compounds and/or drug candidates for human anti-TNF-α therapy using said transgenic mouse and/or tissues, cells and/or cell lines derived thereof, as well as compounds discovered or developed using this method.

TNFα exists as a soluble homotrimer of 17kD protein subunits and as a membrane bound 26kD precursor form . The major source of TNFα are the cells of the monocyte/macrophage lineage, with T lymphocytes, B lymphocytes, neutrophils, mast cells and endothelium all contributing under different circumstances. All potential noxious stimuli, ranging from the physical (UV-light, X-radiation, heat) to the chemical and immunological can rapidly induce TNFα production and release. In-vivo TNFα is the most rapidly produced proinflammatory cytokine, with serum levels detectable in mice within 30 min. TNFα is thought to coordinate the cytokine response to injury and to various infections activating receptors of innate immunity, and it acts as a fire alarm system in-vivo.

By promoting the inflammatory response TNFα can be seen as the master regulator of the activities of other cytokines. Thus, TNFα plays also a key role in autoimmune disorders such as rheumatoid arthritis (RA), ankylosing spondilitis (AS), inflammatory bowel disease (IBD), acute and chronic immune diseases associated with transplantation and others. Hence, the anti-TNF-α therapy has emerged as an efficient treatment of these autoimmune diseases in humans.

At present the only drugs that are in clinical practice or in clinical trials to block TNFα are biologicals, protein-based drugs, either based on antibodies against TNFα or based on TNFα receptor. For example US patent 6,015,557 discloses the use of a chimeric monoclonal antibody binding to TNFα, called Infliximab (trade name: Remicade).

US patent 6,258,562 shows the sequence of a fully human monoclonal antibody binding to TNFα, called Adalimumab (trade name: Humira). A drug based on the TNFα-receptor is for instance Etanercept (trade name Enbrel) which is a fusion protein comprising the TNFα-receptor and the F_c component of human immunoglobulin G1 (IgGI) also disclosed in US patent 6,015,557.

These agents have the major advantage of specificity but also significant disadvantages, including the need for repeated injection and their relative high costs compared to small organic chemical drugs. Thus, the development of novel and better anti-TNFα drugs is highly desirable.

One of the problems of anti-TNFα drug development is the lack of an appropriate animal model for screening, investigating and preclinical testing. At present most of the clinically used anti-TNFα drugs are not active in mice because they recognize human and not murine TNFα. Only Etanercept is able to block murine TNFα as well, so that it can be evaluated in-vivo, by blocking acute liver toxicity induced by LPS-Dgal. However, Infliximab, Adalimumab and some newly developed drugs cannot be compared with Etanercept in-vivo. But these experiments are needed to understand the differences in efficacy and outcome of the different anti-TNFα therapies, for instance why Infliximab is effective in Crohn's disease, and Etanercept is not.

Transgenic mice expressing both human and murine TNFα are known in the state of the art. For instance Hayward M. D. et al. disclose that mice having randomly introduced constructs containing a fragment of the human TNFα gene including the entire coding region and the promoter which is fused to the human β-globin 3 ^'-untranslated region constitutively express low levels of circulating TNFα (Hayward M D. et al., An extensive phenotypic characterization of the hTNFα transgenic mice, BMC PHYSIOLOGY, vol. 7, 10 Dec. 2007). However, these mice show many disadvantages. The human TNFα is expressed deregulated and the murine TNFα gene is intact so that all effects are based on both the human and the murine TNFα. These mice develop quick spontaneous arthritis and dye accordingly. It is possible to use these mice as a model for progressive rheumatoid arthritis, but any other type of chronic disease, such as chronic colitis as well as long-term studies or chronic infection studies cannot be done.

Probert L. et al. suggest backcrossing of a TNFα knock out mouse and a mice overexpressing hTNFα. (Porbert L. et al., TNF-α transgenic and knockout models of CNS inflammation and degeneration, Journal of Neuroimmunology 72 (1997), pp. 137 - 141 ). Thereby, it is possible to get mice which only produce the human gene, but these mice show severe disadvantageous. Said mice often show hidden or obvious gene defects due to random introduction of the new sequences into the genome at several sites. Further the TNFα gene is not expressed in the correct genetic environment. However, chromatin localisation and sequences located far away from the gene of interest are also important for a proper regulation. Mice having randomly introduced human TNFα genes cannot be used for proper systemic disease models.

As a result the anti-TNFα drugs are usually tested in-vitro by neutralization of TNFα activity in-vitro by blocking TNFα-mediated killing of two cell lines L929 and WEHI-231 which both are sensitive to TNFα. But in-vitro testing cannot predict the reaction of a complex biological system.

Thus, it is the object of the present invention to provide an appropriate animal model for screening, investigating and preclinical in-vivo testing of anti-TNFα drugs and drug candidates. It is another object of the invention to provide a method for screening, investigating and preclinical in-vivo testing of anti-TNFα drugs using the appropriate animal model.

It is a further object of the invention to provide cells, cell lines and tissues derived from such an appropriate animal model for a better preclinical in-vitro testing of anti-TNFα drugs and drug candidates and to provide a method for screening, investigating and preclinical in-vitro testing of anti-TNFα drugs and drug candidates using the derived cells, cell lines and tissues of said appropriate animal model.

These problems are solved by the subject matter as characterized in the independent claims and in the description. Further and preferred embodiments of the invention are to be found in the dependent claims and in the description.

In accordance with the invention, the solution to the above-mentioned problem is provided by generation of a transgenic mouse, preferably a knock-in mouse expressing human TNFα protein, wherein the complete endogenous murine TNFα-gene including its upstream regulatory sequences represented by SEQ ID No: 1 is replaced by the human homologous sequence represented by SEQ ID No: 2 or a mutant thereof, wherein the mutant is characterized by deletion, addition, substitution, and/or inversion of one or more nucleotides so that the functional properties of the modified human TNFα-gene are conserved. Thus the invention relates to transgenic mice wherein the human TNFα gene is introduced in the genome instead of the murine gene. Thus, the transgenic mice according to the invention express fully functional human TNFα proteins. The human TNFα gene was introduced together with its complete upstream sequences comprising proximal and distal promoter and enhancer regions and the downstream regulatory elements. Both adjacent regions were needed for functional and physiological regulation of the TNFα gene in human. Thus, the TNFα-gene expression in the transgenic mouse of the invention is regulated by the human regulatory sequences. A "humanized" regulation is very important to produce an adequate murine model for human TNFα related diseases. The sequence of the human TNFα gene is shown in SEQ ID No: 2. The invention further relates to any mutated sequence, wherein the DNA can be translated into a fully functional human TNFα protein. Mutation are for example every single base exchange which does not change the amino acid sequence due to degeneration of the genetic code, but also mutation which results in changes of the amino acid sequence, like deletion, addition or substitution are included, or any mutation in non-coding regions of the gene, which may change its regulation, provided that the TNFα protein retains its biochemical activity.

Usually the mutation are point mutations, but mutations of more nucleotides as well as mutations resulting in more than one mutated amino acid are also included. A substitution is preferably made conservatively. Conservative substitutions or mutations as used herein denotes the replacement of one amino acid residue by another amino acid residue which is biologically similar. That means a cysteine/threonine or serine substitution, an acidic/acidic, a basic/basic or a hydrophobic/hydrophobic amino acid substitution, etc. is preferred. Examples of conservative substitutions include the substitution of one hydrophobic residue such as isoleucine, valine, leucine or methionine for another, or the substitution of one polar residue for another, such as the substitution of arginine for lysine, glutamic acid for aspartic acid, or glutamine for asparagine, and the like. All mutations are allowed according to the invention which nucleotide sequences are translated to a fully functional TNFα protein.

In a preferred embodiment of the invention the mutated sequence show more than 50 % identity with SEQ ID No 2, preferably more than 80 % identity with SEQ ID No 2, more preferred more than 90 % identity with SEQ ID No 2, more preferred more than 95 % identity with SEQ ID No 2 and most preferred more than 98 % identity with SEQ ID No 2.

The recombination arms according to the invention were chosen inside of the flanking murine genes encoding lymphotoxin α and lymphotoxin β, respectively. Advantageously, the regulation of transcription in human and mice is very similar in this part of the genome so that also the replacement of such a big part of the murine genome does not influence the mouse organism negatively.

The present invention further relates to a transgenic mouse expressing human TNFα protein, wherein the exon sequences of the murine TNFα gene represented by SEQ ID No: 21 , SEQ ID No: 22, SEQ ID No: 23 and SEQ ID No: 24 are replaced by the corresponding human homologous exon sequences represented by SEQ ID No: 25, SEQ ID No: 26, SEQ ID No: 27 and SEQ ID No: 28 or mutants thereof, wherein the mutants are characterized by deletion, addition, substitution, and/or inversion of one or more nucleotides so that the functional properties of the modified human TNFα-gene are conserved and wherein the human exon sequences are introduced in the positions of the murine exon sequences. That means the exon sequences of the human TNFα gene consisting of four exons are grafted exactly into the positions of the murine exon sequences. In detail the first amino acid of the first murine exon (SEQ ID No: 21) is replaced by the first amino acid of the first human exon (SEQ ID No: 25) and so on. That means the second murine exon shown by SEQ ID No: 22 is replaced exactly by the second human exon shown by SEQ ID No: 26, the third murine exon shown by SEQ ID No: 23 is replaced exactly by the third human exon shown by SEQ ID No: 27 and the fourth murine exon shown by SEQ ID No: 24 is replaced exactly by the fourth human exon shown by SEQ ID No: 28. Thereby the murine exon sequences are deleted completely and the murine intron sequences are untouched. In this embodiment of the invention the mice produce only human TNFα, but the regulation of gene expression is still murine. Thus, the transgenic modification in the murine organism is as small as possible.

According to the invention one or more of the human exon sequences can be modified by one or several mutations provided that the TNFα protein retains its biochemical activity. Suitable mutations are shown above. In a preferred embodiment of the invention the mutated sequence show more than 50 % identity with SEQ ID No: 25, SEQ ID No: 26, SEQ ID No: 27 and/or SEQ ID No: 28, preferably more than 80 % identity, more preferred more than 90 % identity, more preferred more than 95 % identity and most preferred more than 98 % identity with SEQ ID No: 25, SEQ ID No: 26, SEQ ID No: 27 and/or SEQ ID No: 28.

The present invention further relates to a transgenic mouse expressing human TNFα protein, wherein the part of the endogenous murine TNFα gene ranging from the ATG-codon to the stop codon represented by SEQ ID No: 29 is replaced by the homologous part of the human TNFα gene ranging from the ATG-codon until the stop codon represented by SEQ ID No: 30 or a mutant thereof, wherein the mutant is characterized by deletion, addition, substitution, and/or inversion of one or more nucleotides so that the functional properties of the modified human TNFα-gene are conserved.

That means the murine gene is replaced completely by its human homologous part, but the regulation of the gene expression is of murine origin. The changes in murine organism are very small so that robust mice are received.

According to the invention the introduced human sequence can be modified by one or several mutations provided that the TNFα protein retains its biochemical activity. Suitable mutations are shown above. In a preferred embodiment of the invention the mutated sequence show more than 50 % identity with SEQ ID No 30, preferably more than 80 % identity with SEQ ID No 30, more preferred more than 90 % identity with SEQ ID No 30, more preferred more than 95 % identity with SEQ ID No 30 and most preferred more than 98 % identity with SEQ ID No 30. The present invention further relates to a transgenic mouse expressing human TNFα protein, wherein the part of the endogenous murine TNFα gene ranging from the TATA-box to the stop codon represented by SEQ ID No: 31 is replaced by the homologous part of the human TNFα gene ranging from the TATA-box until the stop codon represented by SEQ ID No: 32 or a mutant thereof, wherein the mutant is characterized by deletion, addition, substitution, and/or inversion of one or more nucleotides so that the functional properties of the modified human TNFα-gene are conserved.

According to the invention the introduced human sequence can be modified by one or several mutations provided that the TNFα protein retains its biochemical activity. Suitable mutations are shown above. In a preferred embodiment of the invention the mutated sequence show more than 50 % identity with SEQ ID No 32, preferably more than 80 % identity with SEQ ID No 32, more preferred more than 90 % identity with SEQ ID No 32, more preferred more than 95 % identity with SEQ ID No 32 and most preferred more than 98 % identity with SEQ ID No 32.

The mice according to the invention are incapable of expressing endogenous TNFα due to full replacement of the gene by its human homologue. The transgenic mice of the present invention are human TNFα-Knock-in mice. Such true knock-in mouse has many advantages for instance compared to a mouse where the human TNFα gene is put in an arbitrary place somewhere else in the genome and which would be easier to produce. The latter mouse may show obvious or hidden gene defects due to arbitrary integration of the human gene into the mouse genome with potential disruption of another gene or function. Further the endogenous TNFα gene had to be knocked out as well, if drugs to human TNFα induced diseases shall be tested effectively.

The transgenic modification is present in somatic and germ cells of the TNFα- Knock-in mouse according to the invention. According to the invention the terms

"transgenic modification" and "transgenic mutation" means that the murine

TNFα gene is replaced by the human TNFα gene. Transgenic mice at every stage of development, including embryonic, juvenile, adolescent and adult and their offspring are included in this invention. The transgenic mice according to the invention are further viable, fertile and capable of transmitting the transgenic modification, in particular the human TNFα-Knock-in, to its offspring. In a preferred embodiment of the invention the transgenic mice are homozygous for the transgenic modification. That means that the TNFα functions are mediated completely by the genetically introduced human TNFα and no endogenous TNFα is expressed. It is an advantage of the present invention to breed the transgenic mice as a homozygous line, where all mice show the desired genotype.

In a preferred embodiment the transgenic mouse is an established inbred strain of lab mice. Preferably the mouse is a C57BI6 mouse which is preferably used for experimental immunology in mice. These mice are very robust and easy to breed. The transgenic mice according to the invention are grossly normal and possessed apparently normal immune system and normal protective functions. Abnormal behaviour or spontaneous development of diseases were not observed over one year. As the murine TNFα gene is completely knocked out in the homozygous transgenic mice the physiological and the pathogenic TNF-α functions are mediated completely by the genetically introduced human TNF-α. Physiological TNFα functions in homozygous transgenic mice are comparable to wild type animals. Surprisingly, stimulated TNFα response in transgenic mice of the invention is up to 20 times higher than in wild type animals using the endogenous murine TNFα gene. Due to the increased stimulated TNFα response which can be related to pathogenic TNFα functions and TNFα based diseases the transgenic mice of the invention are a very sensitive model organism for TNFα based diseases and can be also used for doses studies.

Thus, the "humanized" TNFα mouse engineered by the invention can be used as in-vivo animal model to compare the efficacy of human anti-TNFα drugs and methods. Biological functions mediated by human TNFα in these mice, for instance using the LPS-Dgal model or a collagen-induced-arthritis model, can be blocked by antibodies and drugs specific for human TNFα, but not by substances specific for murine TNFα. Advantageously the transgenic TNFα Knock-in mice of the invention also develop autoimmune induced hepatitis triggered by Concanavalin A (ConA). This is very surprising because ConA hepatitis is linked to p75 TNF receptor signaling which cannot be initiated by human TNFα effectively. Thus, the mice of the invention can also be used as a model for hepatitis diseases, in particular as a model for autoimmune based hepatitis.

It is a further object of the invention to provide cells, tissues and/or cell lines which are derived from the human TNFα-Knock-in mouse according to the invention. Based on the intended use cells, tissues and/or cell lines are derived from heterozygous or homozygous TNFα Knock-in mice according to the invention. Advantageously, the cellular levels of human TNFα are comparable to that of murine TNFα in cells derived from heterozygous mice of the invention. Thus, standard test systems, routine technical procedures, dosage predictions etc. can be transferred easily to the cells derived from the homozygous mice according to the invention. In particular human TNFα knock-in mice are sacrificed and cells and/or tissues are extracted according to standard protocols. Cells and/or tissues can be derived from somatic and/or from germ cells. Preferred examples for tissues and/or cells which are derived from human TNFα-Knock-in mouse according to the invention are lymphoid organs, tails, blood, mouse embryonic fibroblasts (MEF), bone marrow derived macrophages (BMDM) or dendritic cells, total splenocytes or other sorted cell populations. Cells and tissues are cultivated under appropriate culture conditions. Culture conditions are based on the tissue cells are derived from and the intended use. Several standard protocols are known in the art (Current Protocols in Immunology; Copyright 2007 by John Wiley and Sons, Inc.).

Cell lines can be derived from the cultivated cells and/or tissues by transforming them to unlimited growing. Methods and protocols for cell transformation are known by the skilled person. In particular bone marrow derived macrophages can be transformed using the protocol of Cox GW et al., J Natl Cancer Inst. 1989, 81 :1492-1496. Thus, stable cell lines with macrophage/monocyte properties are obtained which is particularly useful for studying TNFα.

It is another object of the invention to prepare the human TNFα-Knock-in mouse according to the invention. Thus, a method of production of the transgenic mouse is included comprising the steps of a) generating an appropriate recombinant vector comprising the human TNFα-gene b) transfecting said recombinant vector of step a) into murine embryonic stem cells; c) selecting an ES clone with desired mutation in one allele d) microinjecting an embryonic stem cell of step c) into a murine blastocyst and transferring said blastocyst into a receptor mouse; e) selecting the offspring until a human TNFα Knock-In mouse is obtained.

The TNFα-Knock-in mouse according to the invention can be created by methods of gene targeting based on homologous recombination in embryonic stem cells. In a preferred embodiment the recombinant vector is a vector for homologous recombination comprising at least, the target gene, marker genes for positive and/or negative selection and the sequences needed for homologous recombination. The recombinant vector can be transfected into murine ES cells using the standard protocols. After transfection the ES cell are cultured under selection culture conditions to amplify only the ES cell clones which has undergone correct homologous recombination. Amplified cells can be further characterized using molecular biology, such as genomic Southern blot analysis, polymerase chain reaction (PCR) and/or DNA sequencing. ES cells with the desired genetic modification are microinjected into murine blastocysts. Living embryos are transferred into a female pseudo pregnant recipient mouse. The offspring of the female recipient mouse is further analysed and mice which are positive for the transgenic mutation are intercrossed until the transgenic mutation is located in the germ line and homozygous human TNFα Knock-In mice are obtained.

It is another object of the present invention to provide a recombinant vector for homologous recombination in murine ES cells comprising as target sequencethe genomic and/or cDNA sequence encoding the human TNFα-gene including its upstream regulatory sequences represented by SEQ ID No: 2 or a mutant thereof. The genetic construct is based on the genomic clone of human

TNFα which is shown by Nedospasov S.A. et al 1986; Cold Spring Harb Symp Quant Biol., 51 Pt 1 : 611-24.

In addition to the human TNFα sequence represented by SEQ ID No: 2 any mutated human TNFα sequence including alterations in non-coding and regulatory elements can be used which can be translated into a fully functional human TNFα protein which retains its biochemical activity. Mutation can be introduced by the methods of molecular biology known to the skilled person or can appear according to natural mechanisms, for instance mistakes in DNA- replication and/or DNA-repair.

In another embodiment of the invention the target sequence is the murine TNFα gene represented by SEQ ID No: 1 , wherein the exon sequences represented by SEQ ID No: 21 (exon 1), SEQ ID No: 22 (exon 2), SEQ ID No: 23 (exon 3) and SEQ ID No: 24 (exon 4) are replaced by the human exon sequences represented by SEQ ID No: 25 (exon 1), SEQ ID No: 26 (exon 2), SEQ ID No: 27 (exon 3) and SEQ ID No: 28 (exon 4). In addition any mutated human exon sequences can be used which are translated into a fully functional human TNFα protein which retains its biochemical activity. Mutation can be introduced by the methods of molecular biology known to the skilled person or can appear according to natural mechanisms, for instance mistakes in DNA- replication and/or DNA-repair.

In another embodiment of the invention the target sequence is the murine TNFα gene represented by SEQ ID No: 1 , wherein the part ranging from the ATG- codon to the stop codon represented by SEQ ID No: 29 is replaced by the human homologous part of the TNFα gene represented by SEQ ID No: 30. In addition any mutated human exon sequences can be used which are translated into a fully functional human TNFα protein which retains its biochemical activity. Mutation can be introduced by the methods of molecular biology known to the skilled person or can appear according to natural mechanisms, for instance mistakes in DNA-replication and/or DNA-repair.

In another embodiment of the invention the target sequence is the murine TNFα gene represented by SEQ ID No: 1 , wherein the part ranging from the TATA- box to the stop codon represented by SEQ ID No: 31 is replaced by the human homologous part of the TNFα gene represented by SEQ ID No: 32. In addition any mutated human exon sequences can be used which are translated into a fully functional human TNFα protein which retains its biochemical activity. Mutation can be introduced by the methods of molecular biology known to the skilled person or can appear according to natural mechanisms, for instance mistakes in DNA-replication and/or DNA-repair. The recombinant vector according to the invention comprises the following components in functional combination: the target gene, at least one marker gene for positive clone selection, optionally two recognition sequences for a recombinase flanking the marker gene, two sequences for homologous recombination (homologous arms) flanking the marker gene and the target gene and optionally at least one marker gene for negative clone selection flanking the homologous arms.

In a preferred embodiment the recombinant vector according to the invention comprises the target gene comprising the genomic and/or cDNA sequence encoding the human TNFα-gene including its upstream regulatory sequences represented by SEQ ID No: 2 or a mutant thereof.

As marker gene for positive clone selection every gene can be used that, under certain selective conditions, only allow the carrier of this gene to survive. Preferably the selection marker is an antibiotic resistance gene, most preferred the gene is the neomycin resistance gene. Positive selection markers are needed to identify the ES cells which contain the recombinant vector.

As marker gene for negative clone selection every gene can be used that is translated into a toxic protein for normal ES cell. Preferably the selection marker is a gene encoding for a toxin, for instance diphtheria toxin, alpha-toxin, streptozotocin or botulinum toxin, preferably diphtheria toxin. Further genes can be used as negative selection markers which stop the cell cycle under certain selective conditions. For instance by terminating the DNA replication using nucleotide base analogues due to thymidine kinase activity.

Negative selection markers are used to select ES cells with the correct homologues recombination. Since both toxic genes are located outside the homologous arms, upon correct recombination they do not join the TNF-α containing segment of the mouse genome which is being replaced by the human part of the construct. Thus, the fragments containing the negative selection genes do not integrate into the genome and are quickly lost during cell division. To the contrary, when a non-homologous integration event occurs, the linearized vector usually integrates as a whole, together with the negative selection cassettes. In such case, toxic selection genes are transcribed and translated, which creates selective disadvantage for clones with nonhomologous integration. Since non-homologous integration is much more likely than homologous recombination, negative selection is vital for success of a gene targeting experiment.

Usually thymidine kinase is used alone in gene targeting experiments in order to stop proliferation of ES cells with a non-homologous integration of the target gene. Theoretically, these cells would die and the ES cells with a homologous integration of the target gene would proliferate. In practice, many of them survive for various reasons. Thus, according to the invention a second marker for negative selection was used in the recombinant vector. The second marker codes for a protein which is toxic by itself. Thus, the negative selection marker is active without any special culture conditions. The proportion of clones with homologous recombination can be significantly increased using a second negative selection marker.

In a preferred embodiment the recombinant vector according to the invention is a vector, wherein the at least one marker gene for positive clone selection is an antibiotic resistance gene, preferably the gene encoding for neomycin resistance and at least one marker gene for negative clone selection is a gene encoding for a toxin, preferably for diphtheria toxin or a gene encoding for thymidine kinase. Using the second negative marker gene encoding diphtheria toxin a dramatically improved selection of ES cells was achieved.

As recombinant vector any known recombinant vector can be used which is able to integrate the appropriate amount of base pairs. As a preferred recombinant vector a pBluescriptKS(+), also known as pBSKS(+) vector is used for generation of a transgenic mouse according to the invention.

The two recognition sequences for a recombinase flanking the marker gene are preferably loxP sites. The Cre recombinase stems from the E. coli bacteriophage P1 and mediates the site-specific recombination between two identical loxP motives in an intramolecular or intermolecular manner. The recombinase Cre of the E. coli bacteriophage P1 is a site-specific recombinase mediating a DNA reorganization via its DNA target sequence, namely loxP. The loxP sequences consist of an 8 bp spacer region flanked by two 13 bp inverted repeats, which serve as recognition sequences for the DNA binding of Cre. The recombination event is only dependent on these two components and is conducted with absolute reliability. It was appreciated that the Cre-loxP system effectively catalyses recombination events in prokaryotic as well as in eukaryotic cells including those from yeast, plants, insects and mammals. Site- specific recombination systems are used to a large extent as tools for conditional genetic changes in single cells and animals. In the case of an excision the region of a DNA sequence between two loxP recognition sequences is cut out. However after genetic manipulation a single LoxP site still remains in the locus which can affect gene transcription if it hits a control element. According to the invention the remaining loxP site was placed downstream to the 3-UTR of the TNFα gene in order not to destroy the function of the adjacent lymphotoxin-beta gene.

There is a multiplicity of further site-specific recombination systems based on a two-component system that may be used in the present case. All such systems have in common specific repeating DNA sequences. Each of these sequences consists of two recognition sequences that are separated by a spacer and are inversely repetitive to each other. The two components are identical in this case. Further examples are the FIp-FRT system of S. cerevisiae, the zygosaccharomyces rouxii pSR1 , the resolvase-rfsF and the phage Mu Gin recombinase system.

According to the invention the homologous arms for homologous recombination are located on both sides of the target gene and the positive selection marker on the recombinant vector. The sequences chosen for homologues recombination are located on both sides of the target gene in the mouse genome. The particular sequences for the left and right arms are chosen based on homology map between mouse and human TNFα loci to make sure that no insertion or duplication of homologous mouse sequences occurs when merging human TNFα gene into mouse genome. According to the invention the genes located next to the TNFα gene, in particular the sequences of the murine lymphotoxin α and lymphotoxin β genes are used as left and right arms for homologous recombination. The recombination events occur somewhere within the left and right arms. Since recombination occurs with nucleotide precision, no damage is done to either of these two murine genes. Thus, the human TNFα knock-in mouse of the invention does not have any genetic damage due to the genetic modification.

In a preferred embodiment homologous arms correspond to nucleotides 21523738-21528888 (left arm) and 21533435-21538973 (right arm) of Mus musculus chromosome 17 genomic contig, strain C57BL/6J, accession NT_039649.7, GM49268951.

The components of the vector were cloned into the vector backbone using the standard protocols of molecular biology known to the person skilled in the art (T. Maniatis, J. F. Sambrook. Molecular Cloning: A Laboratory Manual. 1989, CSHL Press). Thereby the position of the components and the orientation of the DNA fragments has to be considered. Correct orientation and position of the components were confirmed by digests with restriction enzymes and DNA sequencing. Digests and DNA sequencing can be done according to the known standard protocols (T. Maniatis, J. F. Sambrook. Molecular Cloning: A Laboratory Manual. 1989, CSHL Press).

The recombinant vector can be linearized with an appropriate enzyme. Every enzyme can be used according to the invention which restriction site is located once on the recombinant vector and which leave the targeting insert intact and connected to both negative selection cassettes. Preferably the recombinant vector can be linearized using CIaI. The linearized vector is transfected into the ES cells. Any known method of transfection, for instance calcium phosphate, lipofection, nucleofection, heat shock, electroporation, gene gun, microinjection or magnetofection can be used according to the invention. Preferably electroporation is used for transfection of the recombinant vector. Suitable ES cells according to the invention are any murine ES cells. Preferably murine ES cells are used which are robust and show high transfection and homologous recombination rates. In particular murine ES cells are used which are of C57BI6 origin, most preferred Bruce 4 cells. This has an advantage for immunological research because the transgenic mice produced are immediately on the "clean" genetic background. The poor transfection efficacy of Bruce 4 cells was improved dramatically using the recombinant vector of the invention comprising one positive and two negative selection markers.

The present invention further relates to embryonic stem cells, comprising a recombinant vector according to the invention. In a preferred embodiment the ES cells have been transfected with a vector according to the invention, in particular transfected with a vector comprising the target gene comprising the genomic and/or cDNA sequence encoding the human TNFα-gene including its upstream regulatory sequences represented by SEQ ID No: 2 or a mutant thereof. In another preferred embodiment ES cells of the same genetic background, as the recipient mouse, in particular Bruce 4 cells are used. That means that the ES cells and the recipient mice belong to the same mouse strain. Thus, a genetically modified mouse can be produced on a clean and preferred genetic background, such as C57BI6. Preferably, the invention relates to murine ES cells wherein the target gene of the vector was introduced into the mouse genome due to homologous recombination. ES clones with correct homologous recombination were selected by combining the culture conditions specific for positive and negative selection. Positive ES cell clones were further characterized by genomic Southern analysis, PCR and DNA sequencing. Characterization experiments were carried out due to standard protocols known by the person skilled in the art (T. Maniatis, J. F. Sambrook. Molecular Cloning: A Laboratory Manual. 1989, CSHL Press).

Subsequently, the ES cells containing the human TNFα gene at the correct position have to be injected into a blastocyst of a C57BI6 mouse (3 to 4 days after fertilization) in order to obtain chimerae. The blastocyst containing about 100 cells is implanted in pseudo-pregnant females. Blastocyst injection and implantation were done according to standard protocols known by the person skilled in the art. (For reference see for instance Tessarollo L., Manipulating mouse embryonic stem cells. Methods MoI Biol. 2001 ; 158:47-63 or Bonin A., Reid SW, Tessarollo L. Isolation, microinjection, and transfer of mouse blastocysts. Methods MoI Biol. 2001 ; 158: 121 -34.). Usually albino mice are used as recipient mice, if ES cells from black mice are implanted in order to identify the chimeric offspring by mixed colour. In another embodiment the blastocyst can be implanted in normal, i.e. black C57BI6 mice. The offspring is genotyped for the transgene to identify the chimeric animals.

Embryonic stem cells of the mouse are capable of participating in all aspects of the development including the germ track. Therefore, the injected ES cells can also become germ cells, which then transmit the human TNFα gene. The developing mice are chimerae with respect to the changed gene, hence some of the tissues derive themselves from the injected ES cells, others from the normal blastocyst. Thus, the offspring of the mouse has to be PCR typed or analysed by Southern blotting for the expected recombination event, i.e. the human TNFα gene until the germline transmission of the complex genomic modification is achieved. PCR typing and Southern blotting can be done according to standard protocols known by the person skilled in the art (T. Maniatis, J. F. Sambrook. Molecular Cloning: A Laboratory Manual. 1989, CSHL Press). In particular a real clean knock-in, i.e. one gene was taken out, another gene was put on its place should be achieved. Thus, the traces of genetic engineering have to be removed. Neo selection marker which is absolutely needed at the ES stage has to be deleted, either in-vitro or in-vivo. Preferably the Neo selection marker is deleted in the correct ES clone after homologous recombination, for instance by transfecting the cells with another vector comprising the Cre recombinase. However, another manipulation at ES cells stage lowers the change to obtain the desired germline transmission of the mutation. Most preferred the Neo selection marker is deleted after successful germ-line transmission of the mutation by crossing the knock-in mice with a Cre-deleter transgenic mice in which Cre is expressed in many or most of cell types, in particular beta-globin-Cre transgenic mice can be used. After this procedure the Cre-transgene has to be removed itself. Preferably the human TNFα Knock-in mice are intercrossed until the Cre transgene is crossed-out and a clean homozygous human TNFα Knock-In mouse is obtained.

It is another object of the invention to provide an in-vivo method for screening, investigating and/or preclinical testing of compounds and/or drug candidates for human anti-TNFα therapy comprising the following steps: a) providing a transgenic mouse according to the invention b) inducing a pathogenic TNF-α response in said mouse c) treating the mouse of step b) with the compounds and/or drug candidates to be tested d) analysing the human anti-TNFα activity of said compounds and/or drug candidates administered in step c) by evaluating the changes in animal physiology and/or by comparing the changes in animal physiology with control animals e) optionally repeating steps c) and d) with an appropriately modified form of the compounds and/or drug candidates to be tested.

In particular the human TNFα -Knock-in mice according to the invention are used as an in-vivo animal model for screening, investigating and/or preclinical testing of compounds and/or drug candidates which are effective for human anti-TNFα therapy. Recent study has clearly indicated that the consequences of TNFα blockage in patients with regard to organization of lymphoid tissues are the same as in mice, based on earlier knock-out and blockade studies . Thus a humanized TNFα Knock-In mouse is a good preclinical and clinical animal model to investigate the effects of anti-TNFα therapy.

In particular compounds, drugs and/or drug candidates are screened and/or investigated for the ability to effect the properties of human TNFα in said transgenic mice of the invention. According to the invention, properties of human TNFα are any characteristics which belong to the TNFα physiological properties, for instance stability, biological activity, distribution in the organism, degradation or inhibition. In another preferred embodiment compounds which are known to be effective or newly identified compounds in human anti-TNFα therapy can be compared with respect to their properties in order to make human anti-TNFα therapy more specific to distinct diseases and/or the individual needs of various types of patients. According to. the invention, properties of the compounds are any characteristics which belong to the biophysical and/or pharmacological properties, for instance stability, degradation, distribution in the target organism, pharmacological kinetics, efficacy, therapeutic mechanism and/or alleviation or elimination of the symptoms. In a preferred embodiment of the invention therapeutically useful compounds for human anti-TNFα therapy are found to treat human inflammatory diseases in which TNFα plays a key role. In particular human anti-TNFα therapy means the treatment of autoimmune disorders such as rheumatoid arthritis (RA), ankylosing spondilitis (AS), inflammatory bowel disease (IBD), acute and chronic immune diseases associated with transplantation, psoriasis, juvenile arthritis, hepatitis, in particular autoimmune hepatitis, some forms of cancer and others.

In a preferred embodiment of the in-vivo method for screening, investigating and/or preclinical testing of compounds and/or drug candidates for human anti- TNFα therapy LPS-Dgal is injected into said mouse in step b). Thus, acute liver toxicity is induced in said mice. This is a very simple method to induce acute pathological TNFα activity in mice. Alternatively SEB/Dgal can be injected. Untreated mice with acute liver toxicity become moribund within 6 - 10 hours.

In a further preferred embodiment of the in-vivo method for screening, investigating and/or preclinical testing of compounds and/or drug candidates for human anti-TNFα therapy Concanavalin A is injected into said mouse in step b). Thus, autoimmune hepatitis is induced in said mice.

In another preferred embodiment of the in-vivo method for screening, investigating and/or preclinical testing of compounds and/or drug candidates for human anti-TNFα therapy an emulsion of chicken collagen type Il and heat- killed M. tuberculosis H37Ra is injected into said mouse in step b). Thus, collagen-induced arthritis is induced in the human TNFα Knock-in mice of the invention. In addition every other method of inducing acute pathological TNFα activity or a TNFα based disease, in particular autoimmune disorders can be used as well according to the invention. In another preferred embodiment of the in-vivo method for screening, investigating and/or preclinical testing of compounds and/or drug candidates for human anti-TNFα therapy in step c) human TNF-α blocking compounds, preferably Infliximab, Adalimumab, Etanercept, mutant TNFα capable of disrupting existing TNFα trimers, dominant negative mutants and/or small molecules which would interfere with TNF-TNF receptor are injected into a mouse according to the invention. Dominant negative mutants which can be injected as human TNF-α blocking compounds in step c) of said in-vivo method are described in Science 2003 301 (5641): 1895-1898.

In another preferred embodiment of the in-vivo method for screening, investigating and/or preclinical testing of compounds and/or drug candidates for human anti-TNFα therapy the TNF-α plasma levels are analysed, development and/or course of a TNFα-based disease and/or the mortality rate of the mice according to the invention is detected and/or compared to healthy control mice in step d). Further physiological parameters of the transgenic mice of the invention can be investigated to analyse the human anti-TNFα activity of the compounds and/or drug candidates. Suitable parameter of animal physiology are, for instance, occurrence of symptoms of inflammation, such as rubor, tumor, in particular swelling or focal redness of joints, calor, dolor and functio laesa, fever, apathy, loss of appetite, damage to lymphoid organs, changes in gene expression, in particular of the TNFα gene and every other gene belonging to the immune system and/or any other physiological parameter referring to the inflammatory response, such as local cytokine production, acute phase proteins or liver enzymes.

In a preferred embodiment of the in-vivo method LPS-Dgal is injected into the mice according to the invention, human TNF-α blocking compounds, preferably

Infliximab, Adalimumab, Etanercept, mutant TNFα capable of disrupting existing TNFα trimers, dominant negative mutants and/or small molecules which would interfere with TNF-TNF receptor are injected into said mice and the TNFα plasma amounts are analyse, the development and/or course of disease are analysed and/or the mortality rate of said mice is detected and/or compared to healthy control mice.

It is another object of the invention to provide an in-vitro method of screening, investigating and/or preclinical testing of compounds and/or drug candidates for human anti-TNFα therapy wherein the cells, tissues and/or cell lines derived of a mouse according to the invention are used. In a preferred embodiment the in-vitro method comprises the steps of a) deriving cells, cell lines and/or tissues from a transgenic mouse according to the invention and/or using cells, cell lines and/or tissues according to the invention; b) cultivating said cells, cell lines and/or tissues in-vitro; c) inducing a toxic TNF-α response in said cells, cell lines and/or tissues and/or adding human TNF-α to said cells, cell lines and/or tissues; d) adding human TNF-α blocking compounds, in particular Infliximab, Adalimumab, Etanercept, mutant TNFα capable of disrupting existing TNFα trimers, dominant negative mutants and/or small molecules which would interfere with TNF-TNF receptor to said cells, cell lines and/or tissues; e) analysing the TNF-α level and/or marker for cytotoxicity compared to control cells, cell lines and/or tissues; f) optionally repeating steps d) and e) with an appropriately modified form of the test compound.

In a preferred embodiment of the in-vitro method of screening, investigating and/or preclinical testing of compounds and/or drug candidates for human anti- TNFα therapy BMDM are prepared from humanely sacrificed TNFα Knock-in mice. In another preferred embodiment of the in-vitro method the BMDM are stimulated with LPS to induce TNFα production. According to the invention every other method of inducing a toxic TNFα response in cells, in particular in BMDM, cell lines and/or tissues derived from the TNFα Knock-in mice can be used as well. TNFα blocking compounds can be added before or after the stimulation with a toxic TNFα response inducing substance. The level of TNFα, in particular the level of TNFα mRNA and/or TNFα protein, is analysed before and/or after the addition of TNFα blocking compounds. In another preferred embodiment the level of TNFα is compared to non stimulated control cells. Every other marker for cytotoxicity can be analysed in addition and/or instead of the TNFα levels.

In a preferred embodiment of the in-vitro method BMDM are stimulated with LPS to produce TNFα in the cells and the amount of TNFα protein or mRNA is analysed in the presence or absence of a TNFα blocking compound.

The present invention further relates to a compound discovered and/or developed using a method according to the invention. In particular the invention relates to compounds, drug candidates and/or prodrugs identified and/or developed using an in-vivo or in-vitro method based on the human TNFα Knock-In mouse according to the invention and/or cells, tissues and/or cell lines derived thereof.

The compound, drug candidate and/or prodrug is not particularly limited, but examples thereof include peptides, proteins, in particular antibodies and/or protein receptors, non-peptide compounds, synthetic compounds, fermented products, cell extracts, inorganic compounds and/or organic compounds.

With an in-vivo human TNFα animal model of the invention it is possible to compare the efficacy of various known and/or novel TNFα related drugs to each other, to understand the mechanisms of action of various TNFα related drugs which are in clinical use, for instance why Infliximab is active in Crohn's disease patients and Etanercept is not, minimizing the side effects of anti-TNFα therapy by sparing beneficial physiological functions of TNFα, such as maintenance of bactericidal granulomas in protection against Mycobacteria. It is further possible to understand the mechanisms of action of TNFα in TNFα related diseases. In summary, it is an advantage of the invention to model TNFα-mediated diseases in mice and investigate and/or treat them with the human therapies in order to test the various novel therapies. Without "humanized" mice such experiment would not be possible.

Short description of the figures Fig. 1 shows the pBSKS(+) recombinant vector used for homologous recombination.

Fig. 2 shows the different stages of the gene Knock-In procedure

Fig. 3 shows the analysis strategy of the transgenic mice using Southern Blotting. Fig. 4 shows the results of genomic Southern Blot analysis after Asp718l and BamHI digest, respectively.

Fig. 5 shows the results of genomic PCR analysis using primers specific for the right junction between human TNFα gene and mouse locus.

Fig. 6 shows the efficacy of Infliximab in the human TNFα-Knock-ln mice after induction of acute liver toxicity.

Fig. 7 shows the production of murine and human TNFα in BMDM derived from heterozygous human TNFα-Knock-ln mice compared to WT mice and TNFα KO mice.

Fig. 8 shows the efficacy of Infliximab in human TNFα-Knock-ln mice after induction of autoimmune hepatitis.

Fig. 9 shows the induction of collagen-induced arthritis in human TNFα-Knock- In mice compared to WT mice.

Fig. 10 shows the humoral response of the immune system in TNFα-Knock-ln mice after induction of collagen-induced arthritis. Fig. 11 shows the cellular response of the immune system in TNFα-Knock-ln mice after induction of collagen-induced arthritis. Fig. 12 shows the results of treatment of collagen-induced arthritis in human TNFα Knock-In mice using anti-human TNFα antibodies.

Fig. 13 shows the TNFα levels in homozygous human TNFα-Knock-ln mice induced by lipopolysaccharide (LPS).

The present invention is explained in greater detail by the following examples. These examples are illustrative of the present invention and are not to be construed as limiting thereof.

Example 1: Tissue culture of feeder cells

Tissue culture dishes for ES cell culture has to be prepared with feeder cells which provide a LIF enriched environment for the ES cells. The ES cells are grown on top of these feeder cells. Primary mouse embryo fibroblasts (PMEFs) are prepared directly before use from healthy C57BI6 mice (Tessarollo L. Manipulating mouse embryonic stem cells. Methods MoI Biol. 2001 ; 158:47-63); PMEFs that are treated with mitomycin C (MMC, Sigma # M0503) are called feeders or feeder layer. For MMC solution 5ml DPBS (Invitrogen # 14190-169) were added to the 2mg vial of MMC using a syringe. The 5ml were added to 195ml of DMEM [1000 ml bottle of DMEM (4.5g/L glucose, with L-Glutamine, without Sodium Pyruvate, Chemicon # SLM-121-A), 1ml 0.1 M Betamercaptoethanol (BME; 70μl of 14.3 M BME, Sigma # M7522, in 1OmIs of DPBS, filtered through a 0.2μM filter and stored at 4°C), 5ml 10,000units Penicillin; 10mg/ml Streptomycin (Invitrogen # 15140-122)] + 10% FBS (Invitrogen # 10439-024) (D10%) or DMEM + 15 % FBS (D15%). The medium was filtered through 0.2μM filter and transferred at 12 ml each to 15ml tubes and frozen at -2O⁰C. To prepare a 100mm feeder dish a confluent 150mm PMEF dish was obtained. Medium was aspirated and cells were treated with 5-7 ml MMC solution for 2-4 hours at 37°C. After incubation the cells were washed with DPBS three times to remove all MMC. 5 ml trypsin/EDTA (Trypsin/EDTA (Invitrogen # 25200-056) + 1% sterile filtered chicken serum (Invitrogen # 16110-082) were added to the cells and the cells were incubated at room temperature until the PMEFs detach from bottom of the dish. 2OmIs of DMEM + 10 % FBS (D10%) were added to each dish and cells were collected into a 50ml tube. Cells were counted with a hemocytometer and calculated in cells/ml. 3x10⁶ feeders were seeded into one 100mm dish in a total of 10ml D10%. After PMEFS are MMC treated they are incubated for 1-1.5 weeks. MTK-Neomycin 2 (Neo2) PMEFS are prepared equally to PMEFs from any neomycin-resistant source, for instance any classical KO mouse. Usually MTK-Neomycin 2 (Neo2) PMEFS are made from classical LT-alpha knockout mouse (De Togni et al, Science 1994, 264(5159):703-707). Neo2 PMEFs are treated equally to PMEFs and can be cultured up to passage number 5. D10% is used for PMEFs alone and D15% is used if ES cells are cultured on feeders.

Example 2: Tissue Culture of ES cells

Bruce 4 cells (directly received from their author Colin Stewart, Singapore) are cultured in D15% on a feeder layer. One vial (about 3 million cells) are cultured in the 100mm feeder dish, in a total volume of 10ml D15%. They will be confluent in 3-4 days. ES cells need to be trypsinized within 4-6 days after seeding or when confluent, preferably within 1-2 days after confluency or they will begin to differentiate. For trypsinyzing warm DPBS, trypsin/EDTA are used. Media in should be warmed to 37°C in water bath before use.

Example 3: Preparation of the recombinant vector

The recombinant vector is shown in Fig 1 and is generated using standard cloning protocols (T. Maniatis, J. F. Sambrook. Molecular Cloning: A Laboratory Manual. 1989, CSHL Press). The recombinant vector is based on a pBSKS(+) backbone (Stratagene). The negative selection markers diphtheria toxin alpha (DTalpha) and Herpes Simplex thymidine kinase (HS-TK) are located at both ends of the insert and outside the homologous recombination arms (left arm and right arm). Downstream to the DTalpha cassette (SEQ ID NO 12; 1.2 kb) the left arm (SEQ ID NO 3; 5.1kb) is localised, followed by the positive selection marker, the neo cassette encoding for a neomycin resistance (SEQ ID NO, 4; Floxed Neo, 2.2kb). After the neo cassette the human TNFα gene (SEQ ID NO 2; hTNF locus, 4.1kb) is cloned, followed by the right arm (SEQ ID NO 5; 5.3 kb). The cloning of the insert is completed by adding the second negative marker the HS-TK cassette (SEQ ID NO 6; 2.9 kb). The vector can be linearized upstream of the DTalpha gene with the restriction enzyme CIaI. The position of further restriction sites are marked together with the name of the appropriate digesting enzymes. These sites can be used for control digests of the recombinant vector. The final vector was fully sequenced using standard sequencing service.

Example 4: Transfection of the recombinant vector A 100mm dish of ES cells is grown until confluency as shown in example 2. 1- 3 days before electroporation, six 60mm feeder dishes are prepared by MMC treatment, washed, trypsinized, counted and seeded with 8 x 10⁵ feeders/dish in a total of 5ml D15% in six 60mm dishes as shown in example 1.

For electroporation the recombinant vector of example 3 is linearized with CIaI (New England Biolabs # R0197). The linearized DNA is purified with phenol- chloroform (Fluka # 77617), chloroform (Sigma # C2432), precipitated with NaCI (Carl Roth # 3957.2) and 100% ethanol (Carl Roth # 9065.2) and washed with 70% ethanol. Cleaned DNA is resuspended in TE (Part of Qiagen # 12362) or dH₂O at a concentration of 1 μg/μl. Each electroporation uses 30μg DNA. For electroporation the media from the six 60mm feeder dishes is aspirated and 4ml of D15% are added to each dish. Cells are re-incubated at 37°C. The media from the ES cells in the confluent 100mm dish is aspirated as well and cells are rinsed with DPBS. Afterwards 5ml of trypsin/EDTA are added and ES cells are incubated at room temperature until cells are detached from the bottom of the dish and individual ES cells can be detected. 25ml of D15% are added, cells are collected in a 50ml tube, counted with a hemocytometer and calculated as cells/ml. 1-3 x 10⁷ ES cells are needed for one electroporation. The feeders do not have to be excluded in the cell counts because there are not enough feeder cells to make a difference. 1-3 x10⁷ cells are added to one 15ml tube, centrifuged at IOOOrpm for 5 minutes, rinsed with DPBS once and spinned again. The procedure is repeated for a total of three PBS washes. 6 x 10⁶ cells are treated equally and resuspended in 1 ml of D15%. These cells are added to one of the 60mm feeder dishes and are used as control cells. They will be G418 sensitive.

To the 1-3 x 10⁷ cells for electroporation 0.8ml of ice cold DPBS are added and cells are transferred to an electroporation cuvette w/ 0.4cm gap. 30 μg (30μl) of recombinant vector are added to the electroporation tube and mixed well with a

1ml pipette without making bubbles. The electroporation machine is set to

250μF, 320 volts. The electroporation cuvette is put into the chamber with notch on the outside and electroporated. The electroporated cells are added immediately to 4.5ml of D15% and given at 1ml each to the remaining five feeder dishes.

Example 5: Tissue culture under selection conditions

Media are aspirated the one day after electroporation and all six dishes of Example 4 are treated with D15% containing 350μg/ml G418 (Invitrogen # 10131-019; 20 ml of 50 mg/ml solution). Only cells carrying the positive selection marker, in particular the neomycin resistance, will survive in the G418 media. The non-electroporated control cells shall die. Further the media contains 2 μM ganciclovir (Sigma # G2536-100MG) for negative selection using the thymidine kinase. In cells with non-homologous recombination the DNA replication will be terminated due to incorporation of ganciclovir. As the diphtheria toxin alpha is toxic by itself no special culture conditions are needed for functional negative selection. Cells are re-feed in 48 to 72 hours later or when cells begin to die. Cells can be re-feed every day if desired. Positive targeted clones should be picked after 6-8 days.

Example 6: Picking of G418 resistant clones

Using a dissecting microscope undifferentiated clones are picked with a pipet tip outfitted on a mouth pipette by scraping the clone up with the tip while drawing it into the tip using the mouth pipet. After picking a clone is expelled into one well of a 48 well feeders plate prepared as shown in example 1. Approximately 20 G418 resistant ES clones from each of the five 60mm dishes of example 4 cultured as shown in example 5 are picked. Optionally the clone can be trypsinized before adding to the 48 well by adding each clone to a well of a 96 well plate that has 100μl of trypsin/EDTA added and incubating at room temperature for a few minutes. After incubation the clones are broken up by triturating with a P-200 pipet. The dissociated clones are added to a well of the 48 well feeder plate. Single clones are grown to confluency as shown in example 2, used for DNA preparation or trypsinized and frozen by -80⁰C in 350μl freezing media (D15% with a final concentration of 7% DMSO (Sigma # D4540)).

Example 7: ES Cell lysis and DNA preparation

600μl of lysis buffer (5OmM Tris-HCI (Carl Roth # 4855.3 and # 9277.1 respectively); 10OmM EDTA (Sigma # E5134), pH8; 10OmM NaCI; 1% SDS (Sigma # L4390) with 420 μg/ml of proteinase K (Biodeal # CH1005,0200) are directly added to each well of a 48 well plate. Cells are incubated at 37°C overnight, shaking is not necessary. Lysed cells can be stored at this point by - 20⁰C.

For DNA preparation the lysed cells are transferred to 1.5 ml eppendorf tubes and mixed on an eppendorf shaker for 5 minutes to get to room temperature. All the following steps are done at room temperature. 200μl of saturated NaCI (approximately 6M (Carl Roth # 3957.2) are added to each tube and mixed again for 5 minutes on the eppendorf shaker. Then the tubes are spun for 10 minutes in an eppendorf centrifuge at 14k rpm. The 600 μl supernatant are collected, transferred to a new tube and 500μl of isopropanol (Sigma # 19516) are added to precipitate the DNA. The precipitated DNA is spun at 14K rpm for 5 minutes, washed twice with 500μl of 70% EtOH and dried. Pellets are resuspended in 100μl TE buffer.

Example 8: DNA preparation from the transgenic mice

For genotyping of the transgenic mice a short piece of an ear or the tail are cut and transferred into an eppendorf vial containing 600μl of lysis buffer (example 7) with 420 μg/ml of proteinase K. The tissue is incubated at 37°C overnight under continuous shaking. DNA preparation was done as shown in Example 7.

Example 9: Stages of gene replacement

Fig. 2 shows the stages of gene Knock-in before and after homologous recombination and deletion of the Neo cassette. Only one allele is shown. Before the experiment Bruce 4 cells contain two alleles with the murine TNFα gene as shown in Fig. 2b. After transfection the recombinant vector shown in Fig. 2a is localized in the cytosol in addition to the murine TNFα gene (Fig. 2b). After the recombination event and the deletion of the Neo cassette the human TNFα gene replaces the murine homologous. The human TNFα Kl is shown in Fig. 2c. The only trace of the gene modification is the loxP site located at the end of the human TNFα gene. A mouse having one allele shown in Fig. 2b and one allele shown in Fig. 2c would be heterozygous for the gene modification (Fig. 2d). Such a mouse produces both the human and the murine TNFα. A homozygous mouse (Fig. 2e) having both alleles modified as shown in Fig. 2c. This mouse is a real humanized TNFα mouse with no endogenous murine TNFα.

Example 10: Genotyping of the ES cells and the transgenic mice by Southern Blotting The extracted DNA of examples 7 and 8 was digested with different restriction enzymes for Southern blotting. Fig. 3 shows the enzymes used and the expected DNA fragments using appropriate hybridization probes for identifying the different configurations of the TNFα-locus. Using Asp718l (Roche) a 15 kb fragment is expected for the wild type mouse TNF/LT locus. After successful recombination the size of the detected fragment is 11.8 kb and the size of the fragment after deletion of the neo cassette is 9.7 kb. Using Bam HI (New England Biolabs) a 7 kb fragment is expected for the human TNFα-KI after recombination and a 5 kb fragment is expected after deletion of the neo marker.

After digest with either Asp718l or Bam HI the DNA fragments were electrophoresed on an agarose gel to separate them by size. After separation the DNA was denatured and transferred to a nitrocellulose membrane Hybond- N (Amersham Biosciences # RPN303N) using the upward capillary transfer method. Digest, electrophoresis and Southern blotting were done according to standard protocols known to the skilled person (Brown T. Analysis of DNA sequences by blotting and hybridization. Current Protocols in Molecular Biology (1999) 2.9.1-2.9.20, John Wiley & Sons, Inc). The membrane was hybridized with either the hTNFα-KI-26/9 DNA probe (SEQ ID NO 13) for Asp718l digest or with hTNFα-KI-BamHI probe (SEQ ID NO 14) for BamHI digest.

Example 11: Results of Genotyping DNA from ES cells was extracted as shown in example 7 and genotyped according to example 10. Digest was performed with Asp718l. Fig. 4 A shows the results after hybridisation with human TNFα-KI-26/9 DNA probe. Clone 1 and 5 show a single band at 15 kb. This band is specific for the mouse TNFα wild type allele. In these clones the recombination event has not happened successfully. Clones 2, 3 and 4 show one band at 15 kb and one band at 11.8 kb. The shorter band is specific for the human TNFα Kl after recombination. Thus, clones 2, 3 and 4 are heterozygous for the required mutation. They contain one allele with the mouse TNFα and one allele with the human TNFα, These clones are selected for blastocyst injection.

DNA from four mice #14, #15, #19 and #20 was extracted as shown in example 8 and genotyped according to example 10. Digest was performed with BamHI. Fig. 4 B shows the results after hybridisation with hTNFα-KI-BamHI probe. Mouse #14 and #15 show a single band at 7.0 kb. This band is specific for mice that carry the human TNFα gene and still the neo-selection marker (neo+). After the Cre mediated recombination event the neo gene is cut out. Thus, the 7.0 kb band (neo+) is absent in mouse #19 and #20 and a shorter band of 5.0 kb (neo-) appeared. The mouse #19 and #20 show the genotype desired according to the invention. Example 12: Genotyping of the ES cells and the transgenic mice by PCR

The genetic modification was confirmed by identifying the two junctions between mouse and human sequences that did not exist before the genetic modification. As shown in Example 9 the loxP site is still located at the end of the human TNFα gene after the effective gene knock-in. The sequence of the loxP site is shown in SEQ ID NO 7 and can be used as marker for effective gene knock-in using PCR.

Upstream of the introduced human TNFα gene a human/mouse right junction fragment occurs. This fragment is combined of human and mouse parts and several nucleotides of the cloning junction between the mouse and human parts The sequence is shown in SEQ ID NO 8. This fragment can be amplified by PCR as a marker for effective gene knock-in. As the fragment does not occur in nature the PCR does not work on either wild type mouse or wild type human DNA. PCR was done under following conditions. Initial degeneration for 3 min at 94°C, followed by 35 cycles of degeneration at 94°C for 30 sec, annealing at 60⁰C for 30 sec and amplification at 72°C for 30 sec; and final amplification for 5 min at 72°C.

The following three primers are being used in one PCR reaction: hTNF_kn_in_1 : δ'-ATGTACCGCAGTCAAGATATG (SEQ ID NO 9) hTNFkn_in_2: δ'-TTGAGTCCTGAGGCCTGTGTT (SEQ ID NO 10) hTNFkn_in_3: δ'-TGTTGTATAGGACCCTGAGAA (SEQ ID NO 11 )

Primers 1 and 3 amplify a control mouse fragment of 472 bp which is absent in a mouse with a homozygous Knock-In mutation, while primers 1 and 2 amplify the unique human/mouse right junction fragment of 335 bp. In heterozygous mice and/or cells both fragments are amplified. In homozygous human TNFα Kl mice and/or cells only the 335 bp fragment can be amplified. Figure 5 shows the results. In the first and second lane both fragments appear. These mice are heterozygous for the mutation. Lane three shows only a single band by 335 bp identifying a homozygous genotype. Lane four shows a molecular weight marker (Fermentas SM0322). Example 13: Test of the TNFα Knock-In mice as animal in-vivo model for human anti-TNFα therapy

For testing the efficacy of human anti-TNFα drugs, in particular Infliximab, acute liver toxicity was induced in mice by injecting intraperitoneal (i.p.) lipopolysaccharide (LPS) (Sigma # L3024, 10mcg/mouse) and D-galactosamine (D-GaI) (Sigma # G1639, 20mg/mouse). TNFα blockers (2mg/mouse) against human TNFα (Infliximab) or against mouse TNFα (anti-mTNFα antibody) were injected i.p. 30 min later. Survival rate was monitored during 48 hours after LPS/D-Gal injection as a marker for efficacy of the injected TNFα blocker. Moribund mice were humanely sacrificed. For control wild type (WT) mice and TNFα knock-out (TNF KO) mice were treated accordingly. Tests were performed with three mice per group.

Figure 6 shows the result of these tests. WT mice are injected with LPS/D-Gal alone, with LPS/D-Gal + anti-mTNFα antibody and with LPS/D-Gal + Infliximab. All animals died, if LPS/D-Gal is injected alone due to acute liver toxicity. Animals which were injected with the anti-mTNFα antibody survived because the TNFα activity was neutralized by the mouse antibody. Infliximab does not work in WT mice, so that all animals died. TNF KO mice were injected with LPS/D-Gal alone, but acute liver toxicity could not be induced due to the complete knock out of TNFα. Thus, all animals survived. Humanized knock-in mice according to the invention are injected as the WT mice. Mice treated with Infliximab survived and mice treated with the anti-mTNFα antibody died. It can be seen that Infliximab only works in "humanized" knock-in mice according to the invention and the anti-mTNFα antibody only works in wild-type mice and have no effect in humanized knockin mice. Thus, the humanized TNFα Knock- in mouse according to the invention can be used as an excellent animal model for testing of human anti TNFα therapy. Example 14: Production of bone marrow derived macrophages (BMDM) from TNFα Knock-In mice

L929 cells are grown in DMEM medium (Invitrogen: 31966) containing 10%

Fetal Calf Serum, non-essential amino acids, L-Glutamine, penicillin and streptomycin (PAA laboratories GmbH: P11-010). Cells are collected during logarithmic growth and plated at concentration 10,000 cells/ml (30 ml) into T75 culture flasks. Cells are grown for 7 days in CO₂ incubator at 37°C and 5% CO₂.

Supernatant is collected and cells are grown with fresh medium for additional 7 days at +37°C and 5% CO₂. Resultant supematants are mixed, filtered and being used for preparing medium for bone marrow derived macrophages

(BMDM) generation.

To produce BMDM heterozygous human TNFα Knock-In mice are humanely sacrificed. Tibias and femurs are removed and bone marrow is flushed out with sterile ice-cold PBS. Erythrocytes are lysed by RBC lysis buffer (0,15 M NH₄CI (Merck: 168320), 10 mM KHCO₃ (Merck: 104854), 0,1 mM Na₂EDTA (ROTH: 8043.2), pH 7.2-7.4) during 5 minutes on ice. Cells are centrifuged at 1 ,200 rpm at +4°C and are resuspended in concentration 1 ,000,000 cells/ml in BMDM medium, which contains 50% DMEM (Invitrogen: 31966), 20% horse serum (Invitrogen: 26050), 30% supernatant from L929 cells, non-essential amino acids, L-Glutamine, penicillin and streptomycin. Cells are plated in 75 cm nonadherent Petri dishes in 6 ml of total volume. Cells are grown in CO₂ incubator at 37°C and 5% CO₂ during 7 days. At day 7, supernatant is decanted and 5 ml of sterile cold PBS without Ca²⁺ and Mg²⁺ is added to the cells. Then, cells are kept 20 min at +4°C and detached cells are collected from plate surface, centrifuged at 1,200 rpm at +4°C, resuspended and plated in 75 cm nonadherent Petri dishes in 6 ml of total volume. At day 10, BMDMs are collected as described above and 1 mln of cells is used for stimulation.

Example 15: Production of human TNFα by BMDM

To investigate the TNFα response BMDM are prepared as shown in Example 14 from WT mice, TNFα KO mice and heterozygous TNFα Knock-In mice of the invention. Cells are stimulated with lipopoylsaccharide (100ng/ml) during 90 minutes to induce the TNFα production. To analyse the amount of TNFα protein the supernatant of the cells is collected and human and murine TNFα is measured by ELISA (eBiosceince: ELISA Ready-SET-Go! kit: 88-7346-88 (human TNFα), 88-7324-88 (mouse TNFa)).

To investigate the TNFα mRNA cells are collected and mRNA is prepared using TRI reagent (Sigma Co., T9424) according to manufacturer^'s protocol. RNA is treated with RQ 1 RNase-free Dnase 1 (Promega; M6101) and cDNA is synthesized using ImProm-ll™ Reverse Transcriptase (Promega; A3802). Realtime PCR is performed using LitghtCycler (Roche Deutschland Holding GmbH) using following PCR conditions: initial degeneration at 95°C 10 sec, then 45 cycles consisting of degeneration at 95°C for 10 sec, annealing at 60⁰C for 10 sec, and amplification at 72°C for 20 sec. The mRNA levels were normalized to β-actin as housekeeping gene.

The following primers are being used in PCR reaction:

TNFHF: 5'-AGGAGGACGAACATCCAACC (SEQ ID NO. 15)

Human TNFα

TNFHR 5'-TAAGGTCCACTTGTGTCAATTT (SEQ ID NO. 16)

TNFMF: 5'-GGGTGTTCATCCATTCTCTAC (SEQ ID NO. 17)

Murine TNFα

TNFMR: 5'-TGAGATCTTATCCAGCCTCAT (SEQ ID NO. 18) β-actin FW: δ'-CCAAGGTGTGATGGTGGGAATG (SEQ ID No. 19) β-actin β-actin RV: 5'-CCAGAGGCATACAGGGACAGC (SEQ ID No. 20)

Figure 7 shows the result of these tests. BMDMs generated from WT mice produce around 2 ng/ml murine TNFα, whereas human TNFα Knock-In mice bearing one allele of murine TNFα (hTNF KI/WT) produce around 1 ng/ml murine TNFα protein (figure 7A). At the same time, hTNF KI/WT mice produce around 0,8 ng/ml human TNFα (figure 7B). Thus, the amount of human and mouse TNFα protein produced by BMDMs stimulated with LPS is similar in cells bearing both a human and a mouse TNFα allele. The TNF KO cells neither produce murine nor human TNFα protein . Compared to the WT mice mRNA levels of murine TNFα were reduced to one third in the heterozygous TNFα Knock-In mouse of the invention (figure 7C). Human TNFα mRNA can be detected only in the TNFα Knock-In mice according to the invention (figure 7D).

Example 16: Induction of Concanavalin A - induced hepatitis in human TNFα Knock-In mice

For testing the efficacy of human anti-TNFα drugs, for instance Infliximab, in model autoimmune diseases autoimmune hepatitis triggered by Concanavalin A was induced by injecting Concanavalin A (20 mg/kg; C5275, Sigma-Aldrich) into homozygous TNFα Knock-In mice according to the invention. Infliximab (200μg/mouse) was injected intraperitoneal 1 hour before. Mice were bled 2 hours after Concanavalin A injection and human and murine levels of TNFα protein in blood serum were measured by ELISA (eBiosceince: ELISA Ready- SET-Go! kit: 88-7346-88 (human TNFα), 88-7324-88 (mouse TNFa)). Mice were bled 8 hours post-injection for the second time and serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were measured by ELISA.

For histological analysis mice were humanely sacrificed 8 hours after Concanavalin A injection, livers were excised and were embedded in Tissue

Tek OCT compound (Sakura: 4583). Liver sections were cut at 8 μm thickness and were fixed in ice-cold acetone (ROTH: 9372.4). Then, hematoxilin (0,2%

Hematoxilin (ROTH: 3816.1), 0,025% Sodium iodate (Merck: 106525), 6,25%

Aliminiumkalisulfate (Merck: 101047), 6,25% chloralhydrate (ROTH: K3182), 0,125% crystal citric acid (Fluka:27485)) and eosin (1 ,25% eosin (Merck:

115935), 70% absolute ethanol (Fluka: BCR656)) staining was performed according to standard protocols to estimate level of hepatitis (T. Maniatis, J. F.

Sambrook. Molecular Cloning: A Laboratory Manual. 1989, CSHL Press). For control wild type (WT) mice and TNFα knock-out are treated accordingly. Tests are preformed with three mice per group. Figure 8 shows results of these tests. Homozygous human TNFα Knock-In mice of the invention produce human TNFα protein upon injection of Concanavalin A (figure 8A) and the levels of human TNFα protein are slightly higher than levels of murine TNFα protein in WT mice (figure 8B). Autoimmune hepatitis can be detected by elevated levels of AST and ALT in blood sera and by histological analysis. WT mice and human TNFα Knock-In mice injected with Concanavalin A show elevated levels of both proteins (figures 8C and 8D), i.e. they develop autoimmune hepatitis. Histological analysis (figure 8E - 8H) confirmed the AST/ALT results (figure 8C, 8D). Human TNFα Knock-In mice (figure 8E) and WT mice (figure 8F) show necrotic areas with extensive lymphocyte infiltration, whereas human TNFα Knock-In mice, injected with Infliximab (figure 8H), and TNF KO mice (figure. 8G) do not exhibit signs of necrosis and inflammation TNF KO mice do not develop autoimmune hepatitis. Development of hepatitis in human TNFα Knock-In mice can be inhibited by anti-human TNFα antibody, Infliximab as shown by nearly normalized AST and ALT levels and histology without pathological findings.

Example 17: Induction of collagen-induced arthritis (CIA) in human TNFα Knock-In mice.

Induction of collagen-induced arthritis (CIA) is performed in human TNFα Knock-In mice of the invention. Chicken collagen type Il (Sigma Co., C9301) is dissolved in 1OmM acetic acid (ROTH: TT179.2) overnight at +4° C with continuous stirring at final concentration 2 mg/ml. Further, collagen solution (200 μg of collagen) is emulsified with equal volume of Complete Freund Adjuvant (Sigma Co., F5881) enriched by heat-killed M. Tuberculosis H37Ra (DIFCO; 231141), sonicated at 7Ox cycles, 70% power (sonicator: Bandelin Sonopuls HD2200) till antigen became hard consistency. Mice are injected subcutaneously at the base of tail with 100 μl emulsion total. Injections are performed twice: at day -21 and at day 0 of the experiment. For control C57BI/6 (WT) mice and mice with TNF deletion (TNF KO) are treated accordingly. Development of arthritis is evaluated every day according to clinical grades: no swelling (grade 0); swelling or focal redness of finger joints (grade 1); mild swelling of wrist or ankle joints (grade 2); severe swelling of the entire paw (grade 3). Incidence of arthritis is calculated by dividing the number of mice showing swelling of any paws with number of total mice tested. Tests are performed with 10 mice per group. Figure 9 shows the result of these tests. Human TNFα Knock-In mice develop collagen-induced arthritis with an incidence around 50%, similar to WT mice. In contrast, TNF KO mice do not develop arthritis. This indicates that collagen- induced arthritis can be triggered by human TNFα in the mice according to the invention as well as by murine TNFα in WT mice.

Example 18: Measurement of humoral response of the immune system of human TNFα Knock-In mice during collagen-induced arthritis (CIA)

To induce CIA mice were treated as shown in Example 17. Untreated mice according to the invention are used as control. 16 days after 2^nd injection blood was taken from CIA-TNFa Knock-In mice and control TNFα Knock-In mice via tail vein. Collected blood samples were incubated at 4° C overnight. Blood was centrifuged at 1 ,200 rpm at +4°C to remove the blood cells. Resulting sera are stored at -80⁰C before use. Chicken collagen type II, dissolved at 5 μg/ml in PBS, is added to each well of a 96-well plate and incubated overnight at +4°C. Next day, collagen solution is removed and blocking solution (0.1% (v/v) Tween 20, 2% BSA (PAA laboratories GmbH: K51-001), 1.9mM NaH₂PO₄ (Sigma: S 3139), 8.4 mM Na₂HPO₄ (Merck: 106586), 500 mM NaCI (Merck, 106404)) is added to the plate and incubated 30 min at +4°C. Then, the plate is washed 3 times with washing buffer (0.15 M NaCI/0.05% Tween 20) to remove non- adsorbed collagen. Serial sera dilutions of CIA-TNFa Knock-In and control mice starting from 1 :100 are applied to the plate and incubated overnight at +4°C, then the plate is washed three times with washing buffer and incubated with goat anti-mouse antibodies conjugated with alkaline phosphatase (Southern Biotech., goat anti mouse IgM-AP: 1020-04; goat anti-mouse IgG-AP: 1030- 04;goat anti-mouse IgGI-AP: 1070-04; goat anti-mouse lgG2a-AP: 1080-04) for 2 hours at +4°C. Finally, the plate is washed three times with washing buffer and substrate is added to each well (Sigma Co., Alkaline Phosphatase Yellow (pNPP) Liquid Substrate System: P7998). Reaction is stopped after 30 minutes by adding 3M NaOH and optical density of each well is measured at 490 nm in an ELISA plate reader.

Figure 10 shows the results of antibody production. Human TNFα Knock-In mice with CIA produce antibodies against chicken collagen type Il (figure 10A). Antibodies were further analysed and two subspecies could be detected: IgGI (figure 10B) and lgG2a (figure 10C). The appearance of lgG2a shows that a severe CIA has been developed in treated mice. In contrast, naϊve human TNFα Knock-In control mice do not produce any antibodies.

Example 19: Measurement of cellular response of the immune system of human TNFα Knock-In mice during collagen-induced arthritis (CIA)

Human TNFα Knock-In mice treated as shown in Example 17 and non treated mice according to the invention are humanely sacrificed. Spleens are collected and cell suspensions are prepared by meshing organs through cell strainer (Becton Dickinson: 352350). Erythrocytes are lysed using RBC lysis buffer. Cells are washed twice with ice-cold RPMI-1640 medium (Invitrogen: 61870) and resuspended at final concentration 8,000,000 cells/ml. To activate the T- lymphocytes cells (800,000 cells/ml) are stimulated by anti-mouse CD3 (eBioscience: 16-0031-85) and anti-mouse CD28 antibodies (eBioscience: 16- 0281-85) with a presence of GolgiStop (Becton Dickinson: 554715) at 37°C and 5% CO₂ during 6 hours. Finally, cells are washed with PBA buffer (PBS containing 1% BSA), incubated with anti-CD4 antibody coupled with PE-Cy7 fluorochrome (Becton Dickinson: 552775), washed three times with PBA buffer, permeabilized and stained intracellular^ against IFNy (clone: AN 18.17.24) and IL-17(Becton Dickinson: 559502). Cells are acquired using BD FACScalibur (Becton Dickinson). Data are analyzed using FlowJo (Tree Star, Inc.).

Figure 11 shows the result. CD4 T cells isolated from human TNFα Knock-In mice with CIA and restimulated in-vitro have significantly higher numbers of IFNγ (figure 11A) and IL-17 (figure 11 B) producing cells than cells from naϊve human TNFα Knock-In mice which were used as control. Example 20: Treatment of collagen-induced arthritis in human TNFα Knock-In mice using anti-human TNFα antibodies (Infliximab).

Human TNFα Knock-In mice are treated as shown in Example 17. 10 days after 2^nd injection of collagen, human TNFα Knock-In mice which develop arthritis (CIA-human TNFα Knock-In mice) are blindly divided into two groups. One group is injected intravenously with Infliximab (20 mg/kg), the second group is injected with control human IgGI (20mg/kg). Clinical score is evaluated as described in Example 17.

Figure 12 shows the result. CIA - human TNFα Knock-In mice treated with Infliximab show significant reduction of clinical score compared to CIA - human TNFα Knock-In mice treated with control antibody. These results show that specific human TNFα blockade in CIA - human TNFα Knock-In mice results in arthritis inhibition.

Example 21: Expression of human TNFα in homozygous human TNFα-Knock- In mice induced by lipopolysaccharide (LPS).

The expression of TNFα after induction with LPS was studied in homozygous (TNF KI/KI) and heterozygous (TNF KI/WT) TNFα-Knock-ln mice. Expression of TNFα was induced in mice by injecting intraperitoneal (i.p.) LPS (Sigma # L3024, 100mcg/mouse). For control wild type (WT) mice and TNFα knock-out (TNF KO) mice were treated accordingly. Tests were performed with three mice per group. 90 min after stimulation with LPS mice were bled and the amount of human and murine TNFα protein was measured by ELISA (eBiosceince: ELISA Ready-SET-Go! kit: 88-7346-88 (human TNFα), 88-7324-88 (mouse TNFa)).

Figure 13 shows the results of these tests. The levels of human TNFα are shown in Figure 13A the levels of murine TNFα are shown in Figure 13B.

TNF KO mice do not express neither human nor murine TNFα protein. The concentration of murine TNFα in WT is 6 ng/ml, the concentration in heterozygous Kl mice is slightly reduced. That means the replacement of one murine TNFα allele by the human allele does not change the amount of expressed murine TNFα significantly. In contrast the levels of human TNFα are significant higher. In the heterozygous mice the amount of human TNFα is 6 times higher than the amount of murine TNFα. In the homozygous mice which only express the human TNFα protein 20 times more human TNFα protein was detected than murine protein in the corresponding WT mice. This increase in TNFα expression is surprising, but very advantageous. Due to the high expression levels, the TNFα-KI mice represent a sensitive TNFα animal model in which dosage studies for TNFα blocking substances can easily be done.

Claims

1. A transgenic mouse expressing human TNFα protein, wherein the complete endogenous murine TNFα-gene including its upstream regulatory sequences represented by SEQ ID No: 1 is replaced by the human homologous TNFα-gene including its upstream regulatory sequences represented by SEQ ID No: 2 or a mutant thereof, wherein the mutant is characterized by deletion, addition, substitution, and/or inversion of one or more nucleotides so that the functional properties of the modified human TNFα-gene are conserved.

2. A transgenic mouse expressing human TNFα protein, wherein the exon sequences of the murine TNFα gene represented by SEQ ID No: 21 , SEQ ID No: 22, SEQ ID No: 23 and SEQ ID No: 24 are replaced by the corresponding human homologous exon sequences represented by SEQ ID No: 25, SEQ ID No: 26, SEQ ID No: 27 and SEQ ID No: 28 or mutants thereof, wherein the mutants are characterized by deletion, addition, substitution, and/or inversion of one or more nucleotides so that the functional properties of the modified human TNFα-gene are conserved and wherein the human exon sequences are introduced in the positions of the murine exon sequences.

3. A transgenic mouse expressing human TNFα protein, wherein the part of the endogenous murine TNFα gene ranging from the ATG-codon to the stop codon represented by SEQ ID No: 29 is replaced by the homologous part of the human TNFα gene ranging from the ATG-codon until the stop codon represented by SEQ ID No: 30 or a mutant thereof, wherein the mutant is characterized by deletion, addition, substitution, and/or inversion of one or more nucleotides so that the functional properties of the modified human TNFα-gene are conserved.

4. The transgenic mouse of any one of claims 1 to 3, wherein the transgenic modification is present in somatic and germ cells.

5. The transgenic mouse of any one of claims 1 to 4, wherein said transgenic mouse is vital, fertile and capable of transmitting the transgenic modification to its offspring.

6. The transgenic mouse of anyone of claims 1 to 5, wherein the mouse is homozygous for the transgenic modification and/or wherein the TNF-α functions are mediated completely by the genetically introduced human TNF-α.

7. The transgenic mouse of anyone of claims 1 to 6, wherein the mouse is a C57BI6 mouse.

8. Cells, tissues and/or cell lines, preferably bone marrow derived macrophages (BMDM) derived from the transgenic mouse of anyone of claims 1 to 7.

9. A recombinant vector for homologous recombination in murine embryonic stem cells comprising the genomic and/or cDNA sequence encoding the human TNFα-gene including its upstream regulatory sequences represented by SEQ ID No: 2 or a mutant thereof, at least one marker gene for positive clone selection, optionally two recognition sequences for a recombinase flanking the marker gene, two sequences for homologous recombination (homologous arms) flanking the marker and the target gene and optionally at least one marker gene for negative clone selection flanking the homologous arms.

10. The recombinant vector of claim 9, wherein the at least one marker gene for positive clone selection is an antibiotic resistance gene, preferably the gene encoding for neomycin resistance and/or the at least one marker gene for negative clone selection is a gene encoding for a toxin, preferably for diphtheria toxin or a gene encoding for thymidine kinase.

11. Murine embryonic stem cells transfected with the recombinant vector of claims 9 or 10.

12. An in-vivo method of screening, investigating and/or preclinical testing of compounds and/or drug candidates for human anti-TNFα therapy comprising the following steps: a) providing a transgenic mouse of anyone of claims 1 to 7 b) inducing the production of a pathogenic TNF-α amount in said mouse c) treating said mouse of step b) with the compounds and/or drug candidates to be tested d) analysing the human anti-TNFα activity of said compounds and/or drug candidates administered in step c) by evaluating the changes in animal physiology and/or by comparing the changes in animal physiology with control animals e) optionally repeating steps c) and d) with an appropriately modified form of the compounds and/or drug candidates to be tested.

13. The method of claim 12, wherein in step b) LPS-Dgal is injecting into said mouse and/or in step c) human TNF-α blocking compounds, preferably Infliximab, Adalimumab, Etanercept, mutant TNFα capable of disrupting existing TNFα trimers, dominant negative mutants and/or small molecules which would interfere with TNFα-TNF receptor are injected into said mouse and/or in step d) the TNF-α plasma amounts are analysed, development and/or course of a TNFα-based disease are analysed and/or the mortality rate of said mice is detected and/or compared to healthy control mice.

14. An in-vitro method of screening, investigating and/or preclinical testing of compounds and/or drug candidates for human anti-TNFα -therapy wherein the cells, tissues and/or cell lines of claim 8 are used.

15. The method of claim 14, wherein BMDM are stimulated with LPS to produce TNFα in the cells and the amount of TNFα protein or mRNA is analysed in the presence or absence of a TNFα blocking compound.

6. A method of production of the transgenic mouse of anyone of claims 1 to 7 comprising the steps of a) generating an appropriate recombinant vector comprising the human TNFα-gene b) transfecting said recombinant vector of step a) into murine embryonic stem cells; c) selecting an ES clone with desired mutation in one allele d) microinjecting an embryonic stem cell of step c) into a murine blastocyst and transferring said blastocyst into a receptor mouse; e) selecting the offspring until a human TNFα Knock-In mouse is obtained.