Development and characterization of a linear human chromosomal vector to study increased gene dosage in mice

Scientific summaryThe main focus of the human genome lab is the study of X-chromosome Linked Intellectual Disability (XLID). One aspect is to screen XLID-patients for copy-number variations (CNVs) using an in house-developed X-BAC-array, that covers the entire X-chromosome with a resolution of 80 kb, and more recently, an X-oligo-array with a much higher resolution. In addition, our research focuses on the functional characterization of the detected CNVs. For this, three model systems are used within the lab: primary hippocampal neurons, Drosophila melanogaster and the mouse.Whereas early-on it was generally accepted that most genetic disorders were caused by gene deletions or loss-of function mutations, recently it has become clear that duplications can underlie (severe) genetic diseases as well. To study in mice the consequences of gene duplications, transgenic mice overexpressing these genes can be generated. Current overexpression approaches however mostly rely on random integration of the transgene. As a consequence there is no control on where the transgene will insert and on the number of copies to insert. To mimic the subtle overexpression in a controlled way, we focus on the validation and optimization of a new tool for transgenesis, a chromosomal vector (CV), which allows for the controlled insertion of transgenes without any interaction with the host genome. In addition, we report on the generation of transgenic mice that overexpress Huwe1 via BAC pronucleus injection.The human chromosomal vectorWe previously reported on a circular human CV. The CV is maintained as an independent chromosome, is mitotically stable, successfully migrates through the germline, allows for the site specific insertion of genes and displays the physiological expression of human genes. In this thesis, we validated the circular CV as a tool for transgenesis. Furthermore we describe the linearization of the circular CV and the characterization and optimization of the linear CV. Validation of the circular CV for mouse transgenesisAs proof-of-principle for the use of the circular CV for the overexpression of XLID- (candidate-) genes in mice we modified BACs RP11-119A22 and RP11-51C14 encoding MECP2 and SOX3, respectively, and inserted them into the CV via Cre-loxP recombination. MECP2 overexpression in mice had already been described to result in a severe phenotype. As such the mice that overexpress MECP2 from the CV could validate the use of the circular CV for transgenesis. The SOX3 overexpressing mice on the other hand can resolve the phenotypic consequences of SOX3 overexpression for which no animal model is currently available. Hamster clones were identified that contained MECP2- and SOX3-CVs and MECP2-CV-ES-R1 cells were generated. We could demonstrate via RT-PCR the expression of human MECP2 from the CV. FISH analysis however, revealed high numbers of aberrant CVs both for the MECP2 and the SOX3 insertion. The percentage of aberrant CVs was high such that we were unable to proceed to blastocyst injection. Altogether these data demonstrate the successful insertion of human transgenes into the circular CV. In addition, we demonstrated for the first the expression of human transgenes that were inserted into the CV. Transgene integration however, results in gross aberration of the circular CV. Similar aberrations of the circular CV had been observed for the CSN2-CVs and to lesser extent for the wild type circular CV as well. As we believe the instability of the circular CV is inherent to its circular nature we decided no longer to investigate in optimizing the circular CV but focus on its linear counterpart for which a higher stability was predicted. Generation, characterization and optimization of the linear CVThe circular CV was linearized via the Cre-loxP-mediated attachment of 1.6 kb of cloned telomeric repeats. The obtained L-CV was subsequently transferred to a number of cells with different genetic backgrounds via MMCT and via blastocyst injection to mice.Via FISH analysis the presence of the expected 4 telomeric signals at the L-CV could be confirmed for all backgrounds and in mice. FISH analysis moreover demonstrated that the L-CV was maintained as an independent chromosome. TRFA and BAL-31 incubation experiments proved that the L-CV was still linear. As a number of rearrangements were observed at the circular CV in the CHL background we extensively studied the linear CV in the same genetic background by FISH. These data confirmed that 70 94 % of the scored CVs appeared normal, which is much higher than what we observed for the circular CV. Nonetheless aberrations of the CV such as end-to-end fusions to the hamster chromosomes and amplifications of the CV were detected.To calculate mitotic stability for the L-CV, loss-rate percentages were calculated after FISH analysis of L-CV-positive ES-R1 cells that had been cultured for 45 passages without selection. Loss-rate percentages of 0.127% and 0.045% were calculated for two independent ES-R1 cell lines, which exceeds 5.7 times the stability that was observed for the circular CV in the same genetic background. After the generation of transchromosomal mice carrying the L-CV, its transmission through the germline was calculated as well. An efficiency of 35% was determined based on the breeding results of 6 males for a time period of 6 months. The mice also demonstrated regulated expression from the L-CV, as the retina-specific expression of the human ABCA4 gene was supported by the L-CV. The mice were subjected to FISH analysis to determine mitotic stability and structural integrity after 6 generations of breeding. Aberrant FISH signals, indicative of amplifications of the CV were observed in 32.5% of the cells. In addition, the L-CV was present in 70% and 30% of the tail fibroblast of 2 littermates. These data highlight that the linear CV is still rather unstable and underline that its stability and integrity require to be monitored carefully. Nonetheless we find that our L-CV is maintained as an independent chromosome both in vitro and in vivo, that it is mitotically stable in vitro and to a lesser extent also in vivo, and that it is successfully transmitted through the germline. Moreover, the L-CV supports the physiological expression of human genes. Given the risk of recircularization the use of the loxP sites for the site-specific insertion of transgenes was no longer favorable. To overcome this limitation we integrated into the CV a cassette for RMCE. The cassette encoded a hygtk fusion gene and 2 heterospecific FRT sites (F3 and FW) to drive recombination. In addition, 2 SAR elements were encoded by the vector that promote RMCE and support transgene expression. The integration of the RMCE cassette into the CV was obtained via HR in the ES-R1 background. For this 2 homologous regions of 3.1 and 2.6 kb respectively, were cloned into the RMCE vector flanking the SAR elements. Upon transfection into the L-CV positive ES-R1 cells the cassette would integrate into the L-CV via HR based on both homologous arms. Following hygromycin selection 197 resistant clones were isolated. LT-PCR across both homologous arms and subsequent sequence-analysis of fragments of the expected length identified 42 cones (21%) with the correctly recombined sequences. Further investigation of these clones was performed by Southern blot analysis. For this, probes were designed that bind to the hygromycin gene of the RMCE cassette (hyg probe) and to L-CV sequences flanking the homologous arms (5 probe). The latter allows for the selection of clones that have inserted the RMCE cassette in the right position while the hyg probe identifies additional randomly integrated vectors. Hybridization of the 42 samples with the hyg probe displayed in addition to random integration events a fragment of the expected length in nearly all clones. According to the 5 probe however, none of them had inserted the cassette in the correct position. There data were apparently inconsistent with the results obtained via LT-PCR and sequence analysis. To resolve whether the cassette had integrated in one or more loci we set up a qPCR experiment. These data were in line with the Southern blot analysis and confirmed that the cassette had not integrated correctly. The presence of randomly integrated RMCE vectors was moreover confirmed by PCR with primers annealing to vector backbone sequences. We cannot explain the inconsistency that we observed (LT-PCR versus Southern blot analysis and qPCR) and up to now we did not succeed in the selection of an L-CV that is equipped with a cassette for RMCE at the correct position only.Alternative applications of the CV: study of the (sub-)telomereChromosomal vectors combine unique properties for transgenic and putative gene therapeutic applications. In addition, they are essential to study the function and architecture of chromosomal structural elements mainly centromeres and telomeres. Moreover, as CVs are propagandized as safe tools for transgenesis their stability and function is a major factor to be investigated. The L-CV enables its transfer to different genetic background via MMCT. As such this supports the study of (sub-) telomeres in an unprecedented way, i.e. at a single chromosome in different genetic backgrounds. A first observation was the variation of telomere length at the L-CV in different genetic backgrounds. The transfer of the L-CV from the BALB/c donor cell lines to the DT40 chicken cell line resulted in a decrease in telomere length. What is more, telomere length decreased to the same extent on the L-CVs from different donor cell lines. Alternatively, the transfer from L-CVs from the CHL cells to the ES-R1 genetic background resulted in an increase in telomere length of about 35 kb. These results indicate that telomere length is dependent on the genetic background rather than it is dictated by the chromosome itself. The observation that in the DT40 cells the length of different L-CVs is altered to the same extent indicates that this process is tightly regulated. Besides telomere length also CV-subtelomere methylation differed between the different backgrounds. In the ES-R1 cells a de novo methylation of the tk-neo cassette was observed that was never detected on the linear CV in other genetic backgrounds nor on the circular CV in the same genetic background. Taken together, these data indicate that subtelomeric methylation is dictated by the genetic background. As the L-CV telomere lengths are highest in ES-R1 cells a link between telomere length and de novo subtelomeric methylation is suggested.Pedram et al suggested a correlation between subtelomeric DNA methylation and telomere silencing in ES-R1 cells. Therefore, we quantified expression levels of the neomycin selection marker that is encoded by the subtelomeric region of the linear CV. In the ES cells in which subtelomeric de novo methylation occurs on the L-CV, we observed a drop in neomycin expression compared to the circular CV in the same genetic background. Moreover, the expression levels of the L-CV in the other genetic backgrounds, without significant subtelomeric methylation, seem to correlate with that of the circular CV. In addition to telomere length and subtelomeric methylation we thus find that subtelomere silencing also depends on the genetic background. We confirm the data observed by Pedram et al and suggest a link between telomere length, subtelomeric methylation and telomere silencing. Generation of Huwe1 transgenic mice via BAC pronucleus injectionThe technique that is used most extensively for the introduction of transgenes into mice is pronucleus injection. The latter enables the generation of multiple founders with unique insertion sites and variable copy numbers of the transgene following a single round of injections. The comparison of founders with variable copy numbers permits to investigate the link between transgene expression levels and the severity of the phenotype. The simultaneous analysis of multiple founders on the other hand is indispensable to link the phenotype to the inserted transgene as random transgene integrations might mutate the host genome and as such induce a phenotype that is unrelated to the inserted transgene. Recently, we identified in the lab a new gene dosage-sensitive ID-candidate gene, HUWE1. The latter was present in the minimal overlapping region of nonrecurrent duplications at Xp11.22 that were detected in 13 unrelated families with a mild to moderate non-syndromic ID. HUWE1 moreover was found to contain missense mutations in 3 other families with ID. HUWE1 encodes an E3 ubiquitin ligase and is highly expressed in brain, mostly hippocampus. Mouse Huwe1 has been shown to play an important role in normal brain development as it regulates neuronal development. To unravel the impact of Huwe1 overexpression in vivo we generated via BAC pronucleus injection transgenic Huwe1 mice. For this, we selected BAC RP23-360D23 which contains the entire Huwe1 mouse gene and 133 kb of flanking genomic sequences. No other genes were included. After preparation of the BAC it was linearized and purified via sepharose column chromatography and analyzed via PFGE to ensure that no fragmentation of the BAC had occurred. BAC pronucleus injection was outsourced to the Transgene Animal Model Core at the University of Michigan. In total 3 rounds of injections were performed resulting in 12 potential founders with variable copy numbers for the BAC that each carry at least 1 extra copy of the Huwe1 coding region (CDS). Two of the founders were lost in quarantine leaving 10 positive founders. So far 3 out of 10 founders transmitted the Huwe1 transgene to their offspring. The remaining founders either did not reproduce (4/10) or failed to transmit the transgene (3/10). Of the 3 transgenic lines that we established only 2 support the expression of the transgene. The single positive pup of founder 5 was lost and no additional pups were obtained. Therefore, we are currently restricted to a single transmitting positive line. Transgene expression levels were calculated. A female with 2 extra copies was expected to result in 3-fold increased expression levels of Huwe1 whereas only a 2-fold increase was calculated via qPCR. The female however, lacked most of the genomic sequences flanking the 3 end of the CDS. The other pups displayed variable Huwe1 expression levels (1 - 3.5 fold compared to wild type littermates). Surprisingly, the transgenic male founder of this line transmitted the Huwe1 transgenes to all female but never to male pups. Based on this transmission pattern the founder likely has the Huwe1 transgene inserted on its X-chromosome thereby fully mimicking the situation in our patients. As such, random X-inactivation can explain the variation in expression that is observed in the female pups. Currently, breeding of the founders is ongoing in order to set up additional transgenic lines with the aim to study the impact of Huwe1 overexpression on brain development and cognition. These further studies will be part of another PhD project.

Jaar van publicatie:2012

Toegankelijkheid:Closed