Lineage tracing of axial progenitors using Nkx1-2CreERT2 mice defines their trunk and tail contributions

ABSTRACT The vertebrate body forms by continuous generation of new tissue from progenitors at the posterior end of the embryo. The study of these axial progenitors has proved to be challenging in vivo largely because of the lack of unique molecular markers to identify them. Here, we elucidate the expression pattern of the transcription factor Nkx1-2 in the mouse embryo and show that it identifies axial progenitors throughout body axis elongation, including neuromesodermal progenitors and early neural and mesodermal progenitors. We create a tamoxifen-inducible Nkx1-2CreERT2 transgenic mouse and exploit the conditional nature of this line to uncover the lineage contributions of Nkx1-2-expressing cells at specific stages. We show that early Nkx1-2-expressing epiblast cells contribute to all three germ layers, mostly neuroectoderm and mesoderm, excluding notochord. Our data are consistent with the presence of some self-renewing axial progenitors that continue to generate neural and mesoderm tissues from the tail bud. This study identifies Nkx1-2-expressing cells as the source of most trunk and tail tissues in the mouse and provides a useful tool to genetically label and manipulate axial progenitors in vivo.

Timed-pregnant Nkx1-2CreER T2 /YFP mice received tamoxifen at E6.75 and embryos assessed for YFP expression 18 hours later (A) Parasagittal optical section through a late bud stage Nkx1-2CreER T2 /YFP embryo (around E7.5) exposed to tamoxifen at E6.75 and immunolabelled for YFP on whole-mount. YFP + cells were found in the NSB region and CLE which indicates that the expression of the CreER T2 transgene is consistent with the pattern of endogenous Nkx1-2 expression ( Figure  1) (Schubert et al., 1995) (n=6). (Aa) Higher magnification of the region indicated in A. (B) Maximum intensity projection (MIP) of four optical sections (i.e. 16 μm) of the embryo in A. YFP + cells were also found more posteriorly throughout the CLE (arrowheads). (Ba) YFP channel of B. Note that even if early post-implantation embryos are highly autofluorescent (diffuse green signal in A to Ba), cells with clear cytoplasmic YFP stood out above the background signal. (C) To assess the extent of potential spontaneous recombination in Nkx1-2CreER T2 /YFP mice, embryos not exposed to tamoxifen (no TAM) were analysed for YFP + cells on whole-mounts at E8.5. (D) MIP of an Nkx1. 2CreER T2 /YFP embryo at E8.5 in the absence of tamoxifen induction. All embryos analysed (n=5) showed low levels of spontaneous recombination, from 4 to 9 YFP + cells each (6 ± 1 cells/embryo). The close proximity and number of YFP + cells suggest that they originate from a single recombination event at around E7.5 because cells have a cell cycle of ~6-7h (Snow, 1977;Tzouanacou et al., 2009) and it takes ~4h to detect YFP (data not shown). (Ca) Higher magnification of the region indicated in C. Abbreviations are the same as in Fig. 1 Southern blot analyses of genomic DNA from wild type (W), and targeted clones B-E1, B-H9, B-H11 and C-F5. The sizes of the wild type allele (W) and the targeted allele (Targ) are shown in the Southern blot. (A) The genomic DNA was digested with the restriction enzyme SstI and a 5' probe (5ext1) was used to confirm correct homologous recombination at the 5' side. (B) The genomic DNA was digested with the restriction enzyme EcoRV and a 3' probe (3ext2) was used to confirm correct homologous recombination at the 3' side. (C) The genomic DNA was digested with SstI and a probe (cag) that detects a region located within the puro selection cassette used to confirm both correct homologous recombination at the 5' side and single integration of the cassette in all clones. The primer sequences for the PCR amplification of the external probes can be found in Table S2. The black line in the schematics represents the genomic DNA. SHA, short homology arms; LHA, long homology arms.
The Nkx1-1 gene is a paralog of Nkx1-2 and thus both genes could carry out similar functions during embryonic development. To test this, we set out to investigate whether Nkx1-1 is also expressed in Nkx1-2 regions by in situ hybridisation. Simon and Lufkin (Simon and Lufkin, 2003) reported that Nkx1-1 expression "can be detected as early as embryonic stage E10.5". We have tried to revisit Nkx1-1 expression using the following strategies: 1. We used the primers described in the paper (covering the end of intron2/start of exon2) to try and clone the gene, but were unable to amplify the correct product. We think that this could be due to the high GC content of the primers (GC 80%, Tm 70+).
Primer pair 3: mNkx1-1 forward 5'-GGCTACAGCTCGGGACACTA-3', mNkx1-1 reverse 5'-GAGCTGCTCGTAGGTGAAGG-3' 4. We re-checked the genomic sequence of Nkx1-1 and found that the Ensembl version misses out a bit of 5' coding sequence in the first exon. The Kozak sequence is there so it should be translated. Based on Simon and Lufkin 2003, Nkx1-1 mRNA might be more abundant at later stages. We tried again cloning the gene from E13.5 oligodT-primed cDNA, but were still unable to amplify the predicted product.
Given these difficulties we took a different approach and checked whether Nkx1-1 is normally expressed in NMPs by interrogating two published single-cell RNA-sequencing data sets: the Koch (Koch et al., 2017) and

Fig. S8
Bar plot showing the relative expression of Nkx1-1 and Nkx1-2 generated from the bulk mRNAsequencing data of in vitro-derived NMPs (data kindly provided by Robert Blassberg from (Gouti et al., 2014)). The normalised read counts for Nkx1-1 were so low that did not pass filtering steps before downstream analysis.