Zfp281 is essential for epiblast maturation through a cell-autonomous effect

Dear editor, Pluripotency is recognized as a spectrum of cellular states spanning the naïve pluripotency established in the pre-implantation blastocyst and its transition into primed pluripotency in the post-implantation epiblast cells. The exit from pluripotency is followed by the initiation of gastrulation (Brennan et al., 2001; Nichols and Smith, 2009). Naive to primed pluripotent state transition is essential for cell fate decisions, establishing the correct body plan during mammalian embryo development. Zfp281 is a transcription factor known for its critical role in regulating ESC pluripotency by promoting the naïve to primed pluripotent state transition from our in vitro studies of embryonic stem cells (ESCs) and epiblast stem cells (EpiSCs) (Fidalgo et al., 2011; Fidalgo et al., 2012; Huang et al., 2017). Our previous work using a global null allele to generate the conventional knockout mouse model also confirmed a key role of Zfp281 in the epiblast lineage of the developing embryo, manifested by embryonic lethality due to a failure in epiblast maturation (Huang et al., 2017). Zfp281 is expressed in the pluripotent epiblast and its derivatives and directs the activation of target genes in lineage specification and embryo development. Mechanistically, Zfp281 cooperates with epigenetic regulators at targeted gene promotors and enhancers to transcriptionally regulate the expression of Nodal signaling pathway genes. Zfp281 global loss leads to defects in the anterior-posterior (A-P) specification during epiblast maturation (Huang et al., 2017).

In developmental biology, a cell-autonomous effect is the property conferred to a type of cells by the cells’ own genetic alteration. However, the phenotypic effects in one type of cells (e.g., epiblast) could also be attributed to the same genetic alteration in neighboring cells (e.g., trophectoderm (TE) or primitive endoderm (PrE)) through a non-cell-autonomous effect, i.e., through direct cell-cell interactions or changes in cell microenvironment or signaling events. Since Zfp281 is also expressed in TE and PrE cells (Huang et al., 2017), it is still unclear whether the defects of epiblast maturation in Zfp281 mutant embryo are due to a cell-autonomous effect of Zfp281 ablation in epiblast or a non-cell-autonomous effect of Zfp281 ablation in the extraembryonic tissues. We generated tetraploid (4N) chimeric embryos with wildtype 4N TE and PrE and diploid (2N) Zfp281KO ESCs, and demonstrated that the epiblast-specific loss of Zfp281 produced embryos with a comparable phenotype to that of Zfp281 holo-mutant embryos (Huang et al., 2017). However, it is important to note that, while the 4N experiment has its unique advantages of being straightforward and timesaving in addressing cell-autonomous versus non-cell-autonomous effects, it has the limitations of being indirect evidence in an artificial developmental setting, if not technically demanding, because the functional (2N) TE and PrE are also critical for early development. It is known TE cells are committed to the development of placenta, while PrE cells are capable of forming the outer layers of the yolk sac, contribute to the proper embryo development (Tanaka et al., 2009). To unambiguously address this question, we aimed to generate a Zfp281 conditional knockout (cKO) mouse strain by CRISPR/Cas9 and Cre/loxP systems. To construct the Zfp281 conditional allele, we used the “two-donor floxing method” (Fig. 1A), which uses two sgRNAs and two single-stranded oligonucleotides (ssODNs) containing 34 bp loxP sites to flank a targeted critical exon (Yang et al., 2013; Gurumurthy et al., 2019). Zfp281 gene has a unique 2-exon structure with the Exon2 coding for the full-length protein (Figs. 1A and S1) and the Intron1 containing highly GC-rich sequence (data not shown), presenting a potential challenge for the loxP insertion. Accordingly, we knocked in one loxP site (loxP1) at Intron1 and a second loxP site (loxP2) at the position immediately downstream of the stop codon to minimize the possible disruption of 3’ UTR (Figs. 1A and S1). We designed sgRNAs and the donor ssODNs with the loxP sequence flanked with 60 bp of homologous arms (Fig. S1). To facilitate the detection of correct targeting, each ssODN was engineered to contain a BamHI restriction site (Figs. 1A and S1B). The Cas9 mRNA, sgRNAs, and ssODNs were injected into 150 zygotes, and the in vitro developed blastocysts were then transplanted into the uterus of pseudopregnant females to generate CRISPR founder mice. We obtained 23 (out of 150, or 15.3%) founder pups in total, followed by genotyping of mouse tails through next-generation sequencing (NGS). Four out of 23 (17.4%) mice contained both loxP insertions at the same allele (summarized in Fig. 1B), and the knock-in sequence was confirmed with correctly targeted insertions (Fig. S2A). Our success rate of correct founders (17.4%) is similar to that in a previous report (16%) (Yang et al., 2013). We backcrossed the four founder mice with wildtype (WT) C57BL/6 mice to transmit and segregate the KI allele among the F1 Zfp281^F/+ mice, then intercrossed the Zfp281^F/+ male and female mice to obtain the homozygous Zfp281^F/F mice (Fig. 1C and 1D). The correct integration of loxP sites was further confirmed by Sanger sequencing. The adult Zfp281^F/F mice were normal in body weight and fertility compared to the WT C57BL/6 mice (data not shown). Thus, we successfully generated the first Zfp281 conditional knockout (cKO) mouse model by overcoming the potential loxP insertion hurdle associated with the unique 2-exon structure and the GC-rich Intron 1 of the Zfp281 gene.

Figure 1. Generation of an epiblast-specific Zfp281 conditional knockout (cKO) mouse strain and confirmation of Zfp281 KO lethality through a cell-autonomous effect.

A. Schematic depiction of the generation of Zfp281 cKO allele. The Zfp281 exons, intron, 3’ UTR, the start (ATG) and stop (TAA) codons, the knock-in BamHI restriction sites and two loxP sites, and the primers used for genotyping are labeled.

B. Statistics of loxP insertion at two sites of the Zfp281 locus using CRISPR/Cas9 targeting in in vitro developed blastocysts and the targeted live pups (founder mice).

C. Schematic depiction of breeding strategy to generate desired genotypes in mice. (1) The Zfp281 founder mice (usually mosaic) were backcrossed with WT C57BL/6 mice to transmit and segregate the Zfp281 loxP allele, then the heterozygous (Zfp281^F/+) mice were intercrossed to obtain the homozygous (Zfp281^F/F) mice. (2) The Zfp281^F/F and Mox2^Cre/+ mice were bred to transmit the loxP allele with the Cre recombinase to obtain the Zfp281^+/−Mox2^Cre/+ mouse strain. (3) Zfp281^+/−Mox2^Cre/+ and Zfp281^F/F mice were bred to generate four theoretical genotypes of the pups.

D. Zfp281^F/F adult mice used in this study. The identity of these mice was confirmed by genotyping and sequencing.

E. Statistics of the genotyping result of live pups by mating Zfp281^F/F and Zfp281^+/−Mox2^Cre/+ mice.

F. Examples of genotyping PCR results showing different alleles. “*” indicates the potential hybrid DNA of amplified Flox and wildtype bands.

G. Statistics of the genotyping result of dissected embryos by mating Zfp281^F/F and Zfp281^+/−Mox2^Cre/+ mice. “N/A” means genotyping is not successful because of very limited materials.

H. Epiblast-specific Zfp281 cKO embryo (Zfp281^−/−Mox2^Cre/+) exhibits abnormal morphology at E7.0, compared to the other (Zfp281^F/+, Zfp281^F/−, Zfp281^+/−Mox2^Cre/+) littermates. A=Anterior, P=Posterior, Pr=Proximal, D=Distal.

To create epiblast-specific KO of Zfp281, we bred the Zfp281^F/F mice with Mox2-Cre (heterozygous, Mox2^Cre/+) mice (Tallquist and Soriano, 2000) (kind gift from Dr. Soriano in Mt. Sinai), which expresses the Cre recombinase from the endogenous Mox2 locus. This Mox2-Cre line drives the expression of Cre throughout the epiblast and its derivatives following implantation as early as E5.0 (Tallquist and Soriano, 2000). By crossing the Zfp281^F/F and Mox2^Cre/+ mice (Fig. 1C), we expected to achieve epiblast-specific Zfp281 deletion in early mouse embryos. We obtained two genotypes in the pups following this mating strategy, Zfp281^F/+(Mox2^+/+) and Zfp281^+/−Mox2^Cre/+, confirmed by genotyping PCR (Fig. 1F, PCR primers shown in Fig. 1A, see details in Table S1). This result confirms that the Mox2-Cre line works well with the Zfp281 floxed allele, resulting in the Mox2-Cre excised Zfp281 null allele. We further mated Zfp281^+/−Mox2^Cre/+ mice with the Zfp281^F/F mice (Fig. 1C), which would theoretically generate pups of four genotypes, Zfp281^F/+(Mox2^+/+), Zfp281^F/−(Mox2^+/+), Zfp281^+/−Mox2^Cre/+, and Zfp28^−/−Mox2^Cre/+, with an expected ratio of 1:1:1:1. Because Mox2 is only expressed in the epiblast and its derivatives, the Zfp281^−/−Mox2^Cre/+ embryos should have an epiblast-specific deletion of Zfp281, while embryos with the other three genotypes should have at least one allele with normal Zfp281 protein expression. Of note, since Mox2 is not expressed in extraembryonic tissues, the extraembryonic lineages should be either Zfp281^F/+ or Zfp281^F/− genotype, with at least one allele of normal Zfp281 protein expression. With this mating strategy, we obtained a total of 47 pups from 9 litters (an average of 5.2 pups per litter). By PCR genotyping, we obtained 20 (42.6%), 13 (27.6%), and 14 (29.8%) pups with Zfp281^F/+, Zfp281^F/−, Zfp281^+/−Mox2^Cre/+ genotypes, respectively (summarized in Fig. 1E, and part of the genotyping result is shown in Fig. S2B), but zero Zfp28^−/−Mox2^Cre/+ pups. Further staged embryo analysis failed to recover any normal E8.5 embryos with the Zfp281^−/−Mox2^Cre/+ genotype (data not shown). Our results demonstrate that the epiblast-specific deletion of Zfp281 in the Zfp281^−/−Mox2^Cre/+ embryos is embryonic lethal, which is consistent with the Zfp281 conventional KO results in our previous study (Huang et al., 2017). This study thus confirms that Zfp281 is essential for mouse embryo development, and importantly, conclusively demonstrates a cell-autonomous effect of Zfp281 in epiblast development.

Vertebrates have three principal body axes: the anterior-posterior (A-P), the dorsal-ventral (D-V), and the left-right (L-R) axes. The A-P axis is the first established and morphologically discernible axis of the body during mouse embryo development. During early post-implantation embryo development (E4.5 to E6.5), the distal visceral endoderm (DVE), a subpopulation of PrE, forms at the distal tip of the embryo, migrates to the future anterior side, and derives signals to induce the head structures on the nearby epiblast (Tam and Loebel, 2007). Nodal, as well as several of its pathway components, such as Lefty1, Cripto, Dkk1, and Foxh1, are essential for A-P axis establishment (Ding et al., 1998; Brennan et al., 2001; Yamamoto et al., 2001; Kimura-Yoshida et al., 2005). We previously showed that Zfp281 mutant embryos display indistinguishable A-P axis, compared to the WT littermates as early as E6.0–6.5 (Huang et al., 2017). To determine whether the embryos with the epiblast-specific Zfp281 deletion phenocopies the defects of conventional Zfp281 mutant embryos in early development, we dissected the embryos at early post-implantation stages (E6.5–7.5) after crossing the Zfp281^+/−Mox2^Cre/+ and Zfp281^F/F mice (Fig. 1C). The Zfp281^−/−Mox2^Cre/+ embryos were identified by abnormal morphology, including A-P symmetry and shrink epiblast, which was further confirmed by PCR genotyping (summarized in Fig. 1G, typical embryos of each genotype are shown in Fig. 1H). In contrast, all the embryos with the other three genotypes (Zfp281^F/+, Zfp281^F/−, Zfp281^+/−Mox2^Cre/+) were normally developed. These results demonstrate that epiblast-specific deletion of Zfp281 phenocopies the failure of Zfp281 conventional KO mutant in epiblast maturation during early embryo development.

In conclusion, we established a conditional Zfp281 KO mouse model and, combined with epiblast-specific Cre excision, demonstrated conclusively a cell-autonomous effect of Zfp281 functions in mouse embryo development. This novel Zfp281 conditional KO mouse strain will also be a valuable resource to investigate other tissue-specific functions of Zfp281 in development and disease in the future.