Humanized c-Myc Mouse

Background A given tumor is usually dependent on the oncogene that is activated in the respective tumor entity. This phenomenon called oncogene addiction provides the rationale for attempts to target oncogene products in a therapeutic manner, be it by small molecules, by small interfering RNAs (siRNA) or by antigen-specific T cells. As the proto-oncogene product is required also for the function of normal cells, this raises the question whether there is a therapeutic window between the adverse effects of specific inhibitors or T cells to normal tissue that may limit their application, and their beneficial tumor-specific therapeutic action. To address this crucial question, suitable mouse strains need to be developed, that enable expression of the human proto-oncogene not only in tumor but also in normal cells. The aim of this work is to provide such a mouse strain for the human proto-oncogene product c-MYC. Principal Findings We generated C57BL/6-derived embryonic stem cells that are transgenic for a humanized c-Myc gene and established a mouse strain (hc-Myc) that expresses human c-MYC instead of the murine ortholog. These transgenic animals harbor the humanized c-Myc gene integrated into the endogenous murine c-Myc locus. Despite the lack of the endogenous murine c-Myc gene, homozygous mice show a normal phenotype indicating that human c-MYC can replace its murine ortholog. Conclusions The newly established hc-Myc mouse strain provides a model system to study in detail the adverse effects of therapies that target the human c-MYC protein. To mimic the clinical situation, hc-Myc mice may be cross-bred to mice that develop tumors due to overexpression of human c-MYC. With these double transgenic mice it will be possible to study simultaneously the therapeutic efficiency and adverse side effects of MYC-specific therapies in the same mouse.


Introduction
MYC-proteins (c-MYC, MYCN and L-MYC) represent a family of transcription factors which are of particular therapeutic interest because of their strong over-expression in many human cancers [1]. The c-MYC gene codes for two proteins c-MYC1 (p67) and c-MYC2 (p64), that arise from alternative start codons [2]. c-MYC1 is expressed from a CUG start codon in exon 1 and c-MYC2 from the AUG codon in exon 2. In normal cells c-MYC2 is expressed in nearly all proliferating cells, whereas c-MYC1 is expressed when the cells are starved of methionine at high cell density [2][3][4]. Both c-MYC isoforms act as transcriptional activators and also as repressors of several genes. Among the c-MYC target genes are genes which regulate the synthesis of nucleotides, the cell cycle, cell proliferation, cell growth, apoptosis and ribosome biogenesis [5,6]. c-MYC heterodimerizes with a partner protein called MAX to exert transcriptional regulation [6]. MYC-MAX dimers are able to bind to specific DNA-sequences, called E-boxes, for transcriptional regulation of genes. Negative regulators of MYC are MIZ-1 [7] and proteins of the MXD family, e.g. MAD1, MAD3, MAD4 and MXI-1 [6][7][8], which can also interact with MAX instead of MYC and block binding to E-boxes.
In several neoplasms c-MYC2 is overexpressed due to chromosomal translocation, gene amplification or abnormal transcriptional or translational regulation [1,[9][10][11][12]. Interestingly, by using conditional c-MYC-expressing mouse strains [13][14][15][16] or by inhibiting the function of c-MYC in vivo [17] it could be shown that c-MYC overexpression is also important for the maintenance of the cancerous stage of MYC-driven tumors. This MYCdependency emphasizes the potential of MYC-specific cancer therapies.
Several strategies using different kinds of antisense oligonucleotides or peptide nucleic acids have been tested in cell systems and also in mouse strains for their effectiveness in inhibiting the MYCinduced proliferation of cancer cell lines [18][19][20][21][22]. In other studies small molecules were analyzed that interfere with the interaction of MYC and MAX or block the action of MYC in a different fashion [22,23].
In our group we are interested in generating c-MYC-specific T cells for specific elimination of MYC-overexpressing tumor cells. One major question regarding this strategy is, whether there is a therapeutic window in which c-MYC specific T cells eliminate all tumor cells but do not cause fatal damage to healthy tissue. Likewise, for other MYC-targeting strategies it should also be of major concern to analyze in detail the adverse effects on healthy MYC-expressing tissue. However, pre-clinical in vivo model systems are not available to address these issues.
To fill this gap, we report here the generation of a C57BL/6 (B6) mouse strain named hc-Myc, in which the endogenous c-Myc gene is replaced by a humanized c-Myc (hc-Myc) gene. These animals express human c-MYC instead of the endogenous murine protein under physiological conditions. In this mouse strain strategies for targeting the human c-MYC protein can be studied. By cross-breeding humanized c-Myc mice to mouse strains that develop tumors as a consequence of overexpression of human c-MYC in pre B cells, prostate, astroglial, and B cells, like the EnMYC [24], Hi-Myc [25], GFAP/c-Myc [26] or l-MYC mice [27] and targeting the human c-MYC protein by small molecules, siRNAs or T cells, it will become feasible to dissect therapeutic effects from adverse side effects.

Generation of the Targeting Vector
The B6 murine c-Myc locus was derived from a bacterial artificial chromosome. The murine c-Myc locus was retrieved as a fragment of about 14 kb by homologous recombination into plasmid pL255 (kindly provided by Ingo Burtscher), which contains a HSV-TK gene. Fragments encompassing about 500 bp each upstream and downstream of the coding region were amplified by PCR and inserted into a plasmid containing the human c-MYC gene, originally derived from the normal allele of a Burkitt lymphoma patient. A Neomycin-Geneticin resistance gene, flanked by loxP-sites, was inserted into the BglII restriction site in the second intron of human c-MYC. The human c-MYC coding region, intron 1, intron 2 and the Neomycin-Geneticin resistance cassette were inserted into the retrieval vector harboring the murine c-Myc locus by homologous recombination in E. coli using kanamycin selection. The coding sequence of the chimeric murine-human construct was thus of human origin from the CUG codon in exon 1 onwards, whereas non-coding regulatory regions in exon 1 and exon 3 were kept murine. In exon 3 the human sequence exceeds the stop codon by 69 bp. The introns 1 and 2 are of human origin. The HSV-TK gene which resided outwards the arms for homologous recombination was used for selection of

Generation of Transgenic Mice
For generation of transgenic mice B6-derived embryonic stem cells (Bruce4) were used [28]. hc-Myc transgenic ES cell clones were generated by standard methods and injected into Balb/c-derived blastocysts [29]. Chimeric mice were cross-bred to B6 mice. To delete the Neomycin-Geneticin resistance gene, mice were crossbred to Cre-transgenic B6 deleter mice [30]. Breeding and maintaining mice was carried out under specific pathogen-free conditions (SPF).

Southern Blot
For Southern blot analysis genomic DNA of ES cells was prepared by incubating ES cells o/n at 56uC in lysis buffer (10 mM Tris pH 7,4, 10 mM EDTA, 10 mM NaCl, 0,5% (w/v) SDS, 10 mg/mL Proteinase K). After cell lysis genomic DNA was precipitated with ethanol and digested o/n with EcoRI. DNAfragments were separated by length on an agarose gel (0.8% agarose) and blotted o/n onto a nylon membrane. DNA was crosslinked to the membrane by baking at 80uC for 1 hour.
With radioactively labeled DNA probes that bind 59 and 39 site of the murine c-Myc locus ( Figure 1) DNA-fragments containing the c-Myc locus were detected. Probes were radioactively labeled with 50 mCi a32-dCTP (Hartmann Analytic) by using the Random Prime Labeling Kit (GE Healthcare) according to manufacturer's instructions.

B Cell Isolation and in vitro Proliferation Assay
For isolation of mature splenic B cells, a-CD43-magnetic beads (Miltenyi Biotec) were used according to manufacturer's instructions.

Flow Cytometry
For flow cytometry, single cell suspensions of various lymphoid organs were prepared and analyzed with a FITC-labeled a-B220antibody (clone RA3-6B2, BD Biosciences) and a PE-labeled a-CD3-antibody (clone 145-2C11, BD Biosciences) with the FACSCalibur TM (BD Biosciences). Data were analyzed using CELLQuest TM software (BD Bioscience). To exclude dead cells TO-PRO-3 staining was performed.

Ethics Statement
All experiments were done complying with the German Animal Welfare act and were approved by the institutional committee on animal experimentation and the Government of Upper-Bavaria.

Statistical Analysis
For statistical analysis a two-tailed student t-test was applied.

Targeting Strategy
In order to generate a mouse strain in which the human c-MYC protein is expressed under physiological regulation, we integrated a humanized c-Myc (hc-Myc) gene by homologous recombination into the endogenous murine c-Myc locus in embryonic stem (ES) cells. The hc-Myc gene is composed of human c-MYC-coding sequences and human intron sequences, whereas the non-coding, regulatory regions are of murine origin (see Figure 1). This is expected to lead to normal regulation of the transgenic c-Myc gene by mouse regulatory elements. The human sequence encompasses the region from the CUG start codon in exon 1 up to 69 bp downstream of the stop codon so that both isoforms of the human c-MYC protein can be expressed, c-MYC1 starting from the alternate start codon CUG in exon 1 and c-MYC2 starting from AUG in exon 2. The transgene was flanked by arms for homologous recombination that enabled specific integration into one allele of the endogenous murine c-Myc locus in ES cells.
A Neomycin-Geneticin resistance gene was cloned into the second intron of the transgene for positive selection. The resistance gene was flanked by loxP-sites and could be deleted by crossbreeding the mice to deleter mice that express the Cre-recombinase ubiquitously from early embryonic stages on.

Generation of Recombinant B6 ES Cells, Expressing Human c-MYC
For generation of transgenic hc-Myc mice, ES cells were transfected with the linearized hc-Myc transgene and ES cell clones in which the hc-Myc gene was integrated into one c-Myc allele by homologous recombination were selected by using Geneticin and Ganciclovir. For verification of homologous recombination a Southern blot analysis was performed. Figure 1 shows that the endogenous murine c-Myc locus is flanked by two EcoRI-restriction sites. Since the recombined c-Myc locus carries an additional internal EcoRI-restriction site, it is possible to distinguish between the endogenous and the recombined c-Myc-locus by EcoRIdigestion of genomic DNA and Southern blotting, as indicated in Figure 1. For the endogenous locus one large band (ca. 20 kb) should be detected in the Southern blot, whereas the same digestion of the recombined locus results in two smaller fragments, a 59 fragment and a 39 fragment (each ca. 11 kb). Figure 2A shows an example of a Southern blot analysis of four selected ES cell clones. As a negative control wildtype ES cells were used. All four clones showed a band for the wildtype c-Myc allele and a second band for the 59-(59 probe) as well as the 39-(39 probe) fragment of the recombined locus. Overall 27 ES cell clones could be identified, that had shown this pattern of bands in the Southern blot and thus carried the hc-Myc transgene integrated into one wildtype c-Myc allele.
Additionally we analyzed by Western blotting whether these four ES cell clones expressed the human c-MYC protein. As positive control the EBV-immortalized human lymphoblastoid cell line LCL1.11 was used. As shown in Figure 2B, all four clones expressed the human c-MYC protein (62 kDa), whereas in wildtype ES cells only murine c-MYC of slightly higher molecular weight (64 kDa) was detectable [32]. Furthermore, to verify the human sequence of the transgene and to exclude point mutations, the coding sequence of the hc-Myc gene was sequenced in ES cells.

Generation of the hc-Myc Mouse Strain
The ES cell clones 1, 14, 18 and 19 carrying the hc-Myc gene integrated into one allele of the endogenous c-Myc locus were used for injection into Balb/c blastocysts. From ES cell clones 1, 14 and 19 we could establish chimeric offspring, but only two chimeras deriving from ES cell clone 14 gave rise to transgenic offspring. Figure 3A shows an example of one chimeric mouse derived from clone 14. This chimeric mouse was cross-bred to B6 animals. Brother-sister matings of their offspring gave rise to the mouse line with the homozygous hc-Myc locus. PCR analysis of tail DNA of the offspring revealed that some mice carried the hc-Myc transgene in a heterozygous and others in a homozygous fashion ( Figure 3B). As also proven by PCR-analysis, both copies of the endogenous murine c-Myc gene were deleted in homozygous animals. This indicates that substitution of the murine c-Myc gene, whose deletion is embryonically lethal in mice, with the humanized gene is compatible with life [33].
In hc-Myc mice expression of the human c-MYC protein was additionally confirmed by Western blotting, as shown in Figure 3C. The coding sequence of the hc-Myc gene in transgenic mice was finally verified again by PCR amplification and sequencing of the amplification products.

hc-Myc Mice Display a Normal Phenotype
Progeny were born in the expected mendelian ratio and were apparently normal with respect to fertility, litter size, pre-and postnatal development. In experiments with hypomorphic murine c-Myc alleles it has been shown previously that diminished c-MYC expression in mice resulted in reduced body weight, reduced weight of all organs and reduced proliferation of lymphoid cells in vitro [34]. Thus, if the human c-MYC protein is unable to fully replace its murine homologue, one should expect a similar phenotype. Therefore, heterozygous and homozygous hc-Myc mice were analyzed for these parameters.
As shown in Figure 4, the body weight and the weights of the spleen, kidney and liver was comparable between heterozygous and homozygous hc-Myc mice, as well as wildtype B6 animals. Additionally, the number of total cells, B cells (B220 + ) and T cells (CD3 + ) was determined in the spleen, inguinal lymph nodes and in the bone marrow ( Figure 5). In these experiments again no difference between the three groups of mice was detected. Finally, the in vitro proliferation of splenic B cells was analyzed after stimulation with LPS, or a-CD40 and IL-4 ( Figure 6). Apparently, B cells derived from homozygous hc-Myc mice showed similar proliferation rates as B cells derived from B6 wildtype mice after stimulation with either LPS or a-CD40 and IL-4. Taken together, our results show that the human c-MYC protein substitutes for the functions of its murine ortholog. This indicates that hc-myc mice are ideal candidates for studying the adverse effects of therapies in which the human c-MYC-protein is targeted.

Discussion
We describe the generation of a mouse strain that contains a humanized c-Myc (hc-Myc) gene integrated into the murine c-Myc locus and expresses the human c-MYC protein. Thus, in the homozygous situation the endogenous murine c-Myc gene is deleted. Homozygous hc-Myc mice have a normal phenotype with respect to body weight, organ weights and cell numbers in lymphoid organs, and fertility indicating that the human c-Myc protein can functionally replace its murine ortholog.
The hc-Myc gene contains human coding sequences and murine regulatory sequences upstream of the start codon in exon 1 and downstream of the stop codon in exon 3. To redirect homologous recombination to the sequences upstream and Figure 5. Normal cell numbers in lymphoid organs of hc-Myc mice as compared to wildtype mice. Cell numbers in the spleen (A), both inguinal lymph nodes (B) and bone marrow of one tibia (C) was analyzed of 10 -11 week old heterozygous (B6 hc-Myc +/2 , n = 5), homozygous (B6 hc-Myc +/+ , n = 5) hc-Myc mice, and of wildtype B6 mice (B6 wt, n = 5). Depicted are total cell numbers of the lymphoid organs as well as B cell numbers (B220 + ) and T cell numbers (CD3 + ). Total cell numbers were counted using a Neubauer chamber. The percentage of B and T cells was determined by flow cytometric analysis. Mean values are indicated in the graph with black lines and shown at the top of the symbols. Where ever small differences were observed, they were not significant. doi:10.1371/journal.pone.0042021.g005 downstream of the coding region and to prevent additional recombination events within the introns, the humanized transgene also harbors human intron sequences. As regulatory sequences for the less frequently used promoter P3 may reside in the introns of the Myc gene [35], promoter P3 may not be regulated in hc-Myc mice in exactly the same fashion as in the endogenous murine c-Myc locus. Yet, these small differences may apparently not be important for two reasons. Firstly, the differences in the regulation of the P3 promoter between the humanized and endogenous c-Myc gene on Myc transcription and translation, if they exist, are probably minimal because human and murine intron sequences are relatively similar. Secondly and most importantly, 95% of transcripts are normally generated from promoter P1 and P2 which reside upstream of the coding region in exon 1 [35]. Thus, it may be anticipated that in hc-Myc mice P1 and P2 are under the control of the regulatory elements of the endogenous murine locus.
The hc-Myc mouse strain was generated for the analysis of therapeutic prospects of c-MYC targeting strategies. To this end, it will be useful to cross-breed hc-Myc mice with animals already available that develop tumors due to an overexpression of human c-MYC in pre B, B cells, prostate cells and astroglial cells [24][25][26][27]. In this situation the animals shall develop tumors that overexpress human c-MYC and shall express human c-MYC also in healthy cells with high proliferative and regenerative capacity, like hematopoietic and intestinal epithelial cells. Thus it will be possible to figure out in this pre-clinical model whether a therapeutic window exists for various strategies targeting human c-MYC in malignant cells. For strategies using DNA or RNA oligonucleotides as well as for strategies using peptide nucleic acids it will be important to verify their location of binding. If they bind to coding sequences in the exon or to intron sequences of the human c-MYC gene they should have specificity for the hc-Myc transgene and should be applicable in this mouse strain. Especially for the evaluation of strategies that target the human c-MYC gene product by antigen-specific T cells, the hc-Myc mouse will be useful and necessary. We could show that it is possible to induce a T cell response against the human c-MYC protein in wildtype B6 mice by vaccination with c-MYC peptides covering non homologous regions between murine and human c-MYC (Helm et al., unpublished). Furthermore one immunogenic CD8 + T cell epitope of the human c-MYC protein could be identified in C57BL/6 mice that can be used for exploring the feasibility of c-MYC-specific T cell therapies in hc-Myc mice.