LAUSR

β-Thalassemia is a common severe genetic disease caused by mutations in HBB and affects approximately 1.5% of the global population (Origa, 2017). In southern China, the carrier rate of β-thalassemia is as high as 6.43%, creating a high socio-economic burden (Xiong et al., 2010). In adult humans, there are three types of hemoglobin: HbA1 (∼97%), HbA2 (∼2%) and HbF (∼1%). HbA1 (α2β2) is composed of two α-globin and two β-globin subunits encoded by HBA and HBB, respectively; HbF (α2γ2) is made up of two α-globin subunits and two γ-globin subunits encoded by HBG. Mutations in the coding region or regulatory region of HBB are involved in β-thalassemia pathogenesis. Except for some rare dominant mutations, most HBB mutations are recessive (Origa, 2017). Depending on the mutation type, the β-globin level will either be reduced or completely depleted, resulting in α-globin accumulation and precipitation. These α-globin precipitates lead to red blood cell death, resulting in anemia and tissue damage, and even death in βthalassemia major patients. Blood transfusions can help slow disease progression but lead to iron overload, ultimately resulting in iron toxicity. Bone marrow transfer is the only cure in the clinic and is available only to a small percentage of patients with human leukocyte antigen-matched donors. Recently, gene therapy and gene editing therapy have shown great promise in curing β-thalassemia (Glaser et al., 2015; Thompson et al., 2018). However, no appropriate animal models are available for evaluating the safety and efficacy of such advanced therapeutic strategies in vivo. β-thalassemia mice are the sole animal model available for research. However, substantial differences have been reported between the types and expression patterns of human and mouse globins (McColl and Vadolas, 2016). Moreover, mice contain no fetal globin gene equivalent, and homozygous mutations of HBB in mouse for early models of β-thalassemia major or Cooley anemia are all embryonic lethal (Huo et al., 2009). Recently, significant phenotype and physiology differences have been reported between SIRT6null mice and the non-human primate model (Zhang et al., 2018). Thus, an appropriate non-human primate model is needed for human β-thalassemia studies and treatments. Macaca fascicularis, a close relative of human, shares many physiological and developmental similarities with human. Genome edited M. fascicularis models are invaluable for mimicking human diseases (Phillips et al., 2014). In M. fascicularis, the composition of the β-globin and β-like globin gene locus is very similar to the human counterparts (Supplementary information, Fig. S1). Moreover, M. fascicularis expresses HbA and HbF at birth, but only HbA in adults (Scott et al., 1986). Whether HBB mutant M. fascicularis recapitulates human β-thalassemia remained unclear. In this study, we designed three guide RNAs (gRNAs, G1, G2 and G3) targeting the M. fascicularis HBB gene locus (Supplementary information, Fig. S2). Three gRNAs and Cas9 mRNAs mixture were injected into the cytoplasm of 97 zygotes of M. fascicularis to generate HBB knockout (KO) monkeys. The injected zygotes were then cultured for 7 days, and 32 zygotes developed to blastocysts. In 10 of the 32 blastocysts, HBB loci were amplified by PCR after wholegenome amplification. A T7E1 assay, combined with Sanger sequencing, showed that 80% (8/10) of embryos were cleaved by Cas9 nuclease (Supplementary information, Fig. S3). TA cloning and sequencing revealed different mutant alleles, including insertions and deletions. Five of the 8 embryos were showed complete KO without the wild-type (WT) allele (Supplementary information, Fig. S4). Another 22 frozen blastocysts were thawed and transplanted into 7 surrogates, resulting in the production of one new-born monkey, marked as the founder monkey (Fig. 1A). Ear and blood samples were used to amplify the HBB locus in this monkey, producing a band approximately 200 base pairs smaller than that in the WT monkey (Supplementary information, Fig. S5A). To test whether there was a PCR bias towards the smaller band in genotyping, the genomic DNA of the founder monkey was diluted with genomic DNA from WT monkey, and we found that we could still detect the WT band even at 160-fold dilution (Supplementary information, Fig. S5B). Next, Sanger sequencing and sequencing alignment revealed pure 217-base pair deletions at the HBB locus in both the ear and blood tissues. Taken together, these data suggested that this founder monkey might be a homozygous KO monkey (Fig. 1A and 1B). Next, we searched for potential off-target sites of these three gRNAs by CasOFFinder and the top 10 sites of each gRNA were PCR