PWS Model Development

A Pig Model of Prader-Willi Syndrome

This work is funded by an R21 grant from the National Institutes of Health and includes a subcontract for work with collaborators Randall S. Prather, PhD, and Kevin D. Wells, PhD, at the University of Missouri. The overall goal is to create a pig model of hyperphagia and obesity in Prader-Willi syndrome (PWS) using TALEN- or CRISPR/Cas9-mediated deletion of the PWS-IC or the entire PWS-orthologous imprinted domain. This work has been done in two phases, the first to establish optimal genome editing tools for the pig PWS region using a cell culture model system and the second involving animal work to generate the pig PWS model.

Phase 1: Maintenance of human chromosome epigenetic states within a rodent cell has been demonstrated by using human-rodent somatic cell hybrids (SCH), including for the developmental status of globin gene expression, the active or inactive state of X chromosomes, and genomic imprinting. We have used a pig-rodent SCH panel to assess imprinted gene expression and DNA methylation of candidate imprinted loci. Furthermore, we are using SCH as a cellular model to explore genome editing of pig chromosomes. The SCH lines by cytogenetic analysis contain on average five to seven pig chromosomes for this work including Sus scrofa chromosome 1 (SSC1) or SSC9. Using FISH with a whole SSC1 paint probe and with a specific SNORD116 gene locus probe, we confirmed that five SCH have a single SSC1q18 region – the location of the PWS-orthologous imprinted locus. Gene expression analysis showed that the PWS-orthologous genes SNURF-SNRPN, SNORD116, SNORD107 and MAGEL2, as well as the syntenic PLAGL1 gene, are differentially expressed in the SCH lines, consistent with an imprinted pattern of monoallelic expression. Using the methylation-sensitive restriction enzyme HhaI followed by PCR, the promoter of the bicistronic SNURF-SNRPN locus was unmethylated in three SCH and methylated in two SCH, correlating with gene expression and enabling assignment of paternal or maternal origin of the pig SSC1. Similarly, we are currently analyzing SSC9 candidate imprinted genes. Transfection of SCH lines with vectors encoding TALENs or the CRISPR/Cas9 ribonucleoprotein system designed to target sites in the PWS-orthologous domain resulted in successful induction of double-strand breaks and repair by non-homologous end joining to generate deletions ranging from 2.05 kb to ≥ 1.2 Mb. Levels of genome editing by both designer nucleases were greater on the transcriptionally active, paternal allele than on the silent, maternal allele, with only some CRISPR gRNA pairs functional for the latter. In conclusion, SCH are a valuable resource for imprinting assays of candidate genes in the pig. Further, SCH provide a powerful cell model to test efficacy and mechanisms of genome editing of pig chromosomes by designer nucleases.

Phase 2: Given the greater efficiency and lower cost of CRISPR/Cas9 genome editing we focused on use of this system for translation to development of animal models. We designed five pairs of gRNAs and generated vectors targeted to either side of the PWS-IC as well as two gRNAs for either side of the complete pig PWS-imprinted domain. Of five tested pairs of gRNAs flanking the PWS-IC (in multiple experiments), four pairs efficiently led to 2.05 kb deletions in the two SCH lines with an active, paternal chromosome, with the intensity of the breakpoint PCR bands being two to three times greater than the intensity of the PCR band for the intact allele in each case. Similarly, all four alternatively sized ≥ 1.2 Mb deletions spanning the complete pig PWS-imprinted domain were efficiently generated in a SCH line with an active, paternal chromosome as well as a SCH line with a silent, maternal chromosome, and this was highly reproducible as in 12/12 cases over two experiments the ≥ 1.2 Mb deletion was detected (although with different efficiencies in different experiments). Using two-color fluorescence in situ hybridization (FISH) with two BAC probes both within and outside the 1.2 Mb deletion, we determined the deletion frequency to be ~ 5 percent of cells in the SCH system. DNA sequencing confirmed the presence of locus-specific 2.05 kb PWS-IC deletions or ≥ 1.2 Mb PWS-deletions, respectively, with the deletions clearly generated by repair of two DSBs by the error-prone NHEJ process since the breakpoints have sequence heterogeneity. Interestingly, when we tested three pairs of gRNAs flanking the PWS-IC, which were active in producing deletions on the paternal allele, in two SCH lines having a maternal chromosome, only one of three pairs (PWS-IC gRNA1 + gRNA5) was active and produced deletions of the maternal PWS-IC element in both of these two SCH lines. Therefore, the latter gRNA pair was chosen for use to generate PWS-IC deletions in vivo in pig embryos, while one gRNA pair (PWS-large deletion gRNA2 + gRNA3) was chosen to generate ≥ 1.2 Mb deletions. In Missouri, zygotes are generated by in vitro fertilization using sperm from a Yucatan minipig boar with oocytes from domestic pigs, and these are injected with RNAs encoding each of the two respective gRNAs as well as Cas9 mRNA initially from Sigma but subsequently from TriLink. We had shown that the TriLink Cas9 mRNA modified with pseudouridine and 5-methylcytosine is more stable and resulted in a 10 to 20 times higher genome editing frequency in our SCH system. Currently, the zygote mRNA injections are being repeated and blastocysts generated by in vitro culture for one week with 1) assessment for mutation frequency and 2) embryo transfers performed to a surrogate(s) to generate pregnancies. Following a 114 to 117 day gestation, piglets will be genotyped and phenotyped.

Principal Investigator

Robert D. Nicholls, PhD