(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Brachyury expression levels predict lineage potential and axis-forming ability of in vitro-derived neuromesodermal progenitors [1]

['Anahí Binagui-Casas', 'Institute For Regeneration', 'Repair', 'Centre For Regenerative Medicine', 'Institute For Stem Cell Research', 'School Of Biological Sciences', 'University Of Edinburgh', 'Edinburgh', 'United Kingdom', 'Anna Granés']

Date: 2026-09 

Neuromesodermal progenitors (NMPs) produce the spinal cord and musculoskeleton in the elongating anterior-posterior axis. In vivo, NMPs possess dual potency, coinciding with regions co-expressing SOX2 and Brachyury (TBXT). In vitro, SOX2/TBXT co-expressing cells can be produced from pluripotent cells and, like their in vivo counterparts, can produce neural tube and somitic mesoderm. However, the functional characteristics of in vitro SOX2/TBXT co-expressing cells remain unclear, confounding comparisons with in vivo data. To address this, we developed a dual Sox2/Tbxt reporter mouse ESC line. SOX2/TBXT reporter-positive cells emerge in vitro from pluripotent populations with dynamics that mirror their appearance in the embryo. Purified SOX2/TBXT co-expressing populations can differentiate towards neurectoderm or mesoderm, including lateral mesoderm upon BMP stimulation. In gastruloids, quantitative live imaging shows that WNT or NOTCH inhibition rapidly leads to downregulation of TBXT expression and diminished axial extension. We show that clonally plated SOX2/TBXT co-expressing cells are bipotent NMPs that can also self-propagate. By combining clonal analysis with mathematical inference, we identify two thresholds of TBXT and/or SOX2 expression, switching clonal output from neural- to mesoderm-biased, and from mesoderm-biased to mesoderm-specified. Image analysis of embryonic NMPs supports a model whereby SOX2 and TBXT independently influence neuromesodermal differentiation. Thus, this Sox2/Tbxt double reporter cell line highlights unsuspected heterogeneity in NMPs, and together with image analysis of embryonic SOX2/TBXT levels, challenges the assumption that neuromesodermal fate choice is primarily governed by mutual antagonism between SOX2/TBXT.

Funding: The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. This work was supported by the Medical Research Council (MR/K011200 and MR/S008799/1 to V.W., and covering salaries for A.B-C., A.G., E.K., Y.H. and F.C.K.W.), Carnegie Trust for the Universities of Scotland (RIG012666 to A.B-C.), the Fondo para la Investigación Científica y Tecnológica de Argentina (PICT-2019-03828 to O.C., and supporting A.S.C.), Biotechnology and Biological Sciences Research Council (BB/X014908/1 to O.C., and covering salary for A.S.C.), UADE (A23T01 and P26T02 to O.C.), Wellcome Trust Senior Fellowship (220298 to S.L., and covering salaries for M.F., R.P., J.A. and A.S.B.), Japan Society for the Promotion of Science KAKENHI grant (JP19K16157 to F.J.W.), EastBio Doctoral Programme (BB/T00875X/1 to R.W.), Erasmus Exchange fund (to A.G. and D.L.R.), Gurdon summer studentship from the British Society for Developmental Biology (to A.G).

Data Availability: All relevant data can be found within the article, supplementary information and repositories as stated below. RNAseq raw data can be found at GEO accession number GSE339040. FCS files and data related to flow cytometry experiments and gene expression analyses can be found in Zenodo DOI: https://doi.org/10.5281/zenodo.21708690 . Code and data for Figs 5E , 5F , S3G – S3I and S4 can be found in Zenodo DOI: https://doi.org/10.5281/zenodo.21650680 . Code and data for Figs 4 , 6 , S5 and S6 can be found in Zenodo DOIs: https://doi.org/10.5281/zenodo.21676159 and https://doi.org/10.5281/zenodo.15802710 .

Interestingly, we find that the Sox2/Tbxt co-expressing population is not functionally homogeneous: the levels of Tbxt-GFP expression predicts clonal lineage bias towards mesoderm in individual progenitors. This intrinsic bias can be tuned by the culture conditions and cell substrates, mirroring the in vivo plasticity of NMC populations [ 8 ]. Investigation of this bias via mathematical inference of in vitro populations and image analysis of early somite stage embryos supports a model whereby TBXT and SOX2 act separately on neuromesodermal fate choice, rather than opposing one another, as suggested previously.

Here we generated a minimally disruptive mouse embryonic stem cell (mESC) line with dual Sox2/Tbxt gene expression reporter activity and use it to examine live cell behaviours, population heterogeneity and functional potency of in vitro NMP-like cells. Clonal plating demonstrates that individual Sox2/Tbxt dual-positive cells are bona fide bipotent progenitors, which can also self-propagate. Additionally, we show that rapid deceleration of axis elongation in gastruloids on WNT or NOTCH inhibition coincides with Tbxt downregulation, while changes in Sox2 expression follow more slowly. We further uncover previously elusive lateral plate mesoderm potency within descendants of NMPs, an attribute shown only in bulk culture [ 21 ].

Addressing whether individual in vitro NMP-like cells co-expressing Sox2 and Tbxt are bi-potent, and how heterogeneity in SOX2 and TBXT levels dictates differentiation outcomes requires the use of cells that report on the expression of both markers. Previous reporters have employed knock-in constructs that disrupt the function of these genes [ 10 , 19 ] or are additive bacterial artificial chromosome transgenics where there is a risk that expression does not accurately mirror endogenous expression [ 20 ].

Recent single-cell RNA-seq studies have found that in vitro cultures harbouring NMC populations are transcriptionally heterogeneous, with varying levels of Sox2 and Tbxt, as well as other NMC-specific transcripts such as Msgn1, Nkx1-2, and Tbx6 [ 12 – 15 ]. One interpretation of this data is that in vitro NMC populations contain contaminating, non-NMP cells. Alternatively, the transcriptional heterogeneity in Sox2 and Tbxt-expressing cells themselves could reflect inherent functional heterogeneity at the single-cell level given that transcription factors such as SOX2 elicit expression level-dependent function [ 16 ], and that, in vivo, levels of TBXT and SOX2 protein appear to correlate with variations in NMC fate [ 8 , 17 , 18 ].

In vitro, FGF and WNT treatment of pluripotent stem cells generates cells that resemble in vivo NMC populations both phenotypically and functionally: a high percentage of cells co-express SOX2/TBXT transcription factors, and these populations can differentiate into neural and mesodermal lineages (for a list of common protocols see [ 1 ]). Indeed, at clonal level a proportion of in vitro-derived NMP-like cells expressing Tbxt (monitored with a fluorescent reporter) have dual potential for neural and mesodermal differentiation [ 9 ]. However, in these experiments, the knock-in reporter created a heterozygous null allele of TBXT [ 10 ], a condition known to affect axial development [ 11 ]. Furthermore, the expression of SOX2 in the starting population was not measured, and therefore the relationship between SOX2 and TBXT co-expression, and bipotency is unknown.

Our current understanding of the literature suggests that NMPs in amniotes (chick, mouse) and anamniotes (fish, frog) are distinct. In amniotes, individual bi-potent NMPs are dual neural/mesodermal fated, while in anamniotes, neuromesodermal potency is maintained in NMC populations in the tail bud, but individual cells adopt either neural or mesodermal fates [ 3 – 8 ]. Thus, NMCs in the neuromesodermal-fated region in different species can consist of bona fide NMPs, or of a mixture of mono-fated cells of unknown potency.

In vertebrates, the head-to-tail (or anteroposterior, AP) body axis forms during early embryonic development. In mice, where this process is well characterised [ 1 ], a transient population of progenitors, collectively called neuromesodermal competent cells (NMCs), produces both paraxial mesoderm and neurectoderm as the AP axis elongates in post-gastrulation stages. NMC populations across vertebrates harbour cells with functional heterogeneity, including bona fide bipotent neuromesodermal progenitors (NMPs) [ 2 ]. Importantly, the hallmark of NMPs is their dual fate towards both neuroectodermal and mesodermal lineages, two separate germ layers, at post-gastrulation stages [ 3 , 4 ]. Fate mapping has shown that cells with neuromesodermal fate are found at the posterior end of various vertebrate embryo species, and this region coincides with the co-expression of the transcription factors SOX2 and Brachyury (TBXT), reviewed in Wymeersch and colleagues [ 1 ].

Results

1. A novel reporter line, STR-KI, accurately reports on SOX2 and TBXT expression, and NMP differentiation trajectories in vitro To characterise the dynamics of NMC emergence from mESCs, we constructed a 2A-peptide-linked dual reporter: SOX2::mCherry; TBXT::GFP, hereafter named SOX2-TBXT Reporter Knock-In (STR-KI). CRISPR knock-in transgenesis was used to create heterozygous knock-in alleles that replace the stop codon of the endogenous gene loci with T2A/P2A ‘self-cleaving’ (ribosome-skipping [22]) peptide, retaining the Tbxt and Sox2 open reading frames in their entirety [23]. Since even partial loss of TBXT or SOX2 function affects differentiation [11,16,19,24], we employed this strategy to minimise the chance of adverse phenotypic effects and maximise the faithfulness of the reporters (Fig 1A). Chimeric embryos containing high contribution from STR-KI mESCs showed a normal phenotype, demonstrating that the gene targeting was indeed minimally disruptive to SOX2 and TBXT protein function (Figs 1B and S1C). Furthermore, the fluorescent reporters in pre-streak and somite stage embryos recapitulated the in vivo expression pattern of TBXT and SOX2 in mesodermal and neural lineages, respectively (Figs 1B and S1C) [8]. In pre-streak chimaeras, SOX2::mCherry (Sox2mCh hereafter) was expressed throughout the epiblast, whereas TBXT::GFP (TbxtGFP hereafter) expression was undetectable (S1C Fig). By early somite stages onwards, Sox2mCh was predominantly restricted to neural tissues, and to presumptive primordial germ cells at the base of the allantois, whereas TbxtGFP-expressing cells were observed in the node, notochord and nascent mesoderm at the posterior end (Figs 1B and S1C). Differentiation of STR-KI ESCs into NMP-like cells, following an established protocol [25] produced mCherry and GFP double-positive cells (Sox2mChpos/TbxtGFPpos) (Fig 1C–1F). Immunostaining and quantification of transcription factor (TF) intensity in these in vitro differentiated cells showed that the reporters faithfully co-localised with endogenous SOX2 and TBXT protein (Figs 1C, 1D, and S1B). For both reporters, we also observed a positive correlation between the levels of reporter and endogenous protein in immunofluorescence followed by confocal imaging (SOX2-mCherry R2 = 0.894, TBXT-GFP R2 = 0.838), and intracellular staining followed by flow cytometry (SOX2-mCherry R2 = 0.503, TBXT-GFP R2 = 0.761) (Fig 1C). We therefore adopt the working hypothesis that GFP and mCherry levels reflect TBXT and SOX2 protein production, respectively, although differential posttranscriptional modifications affecting protein levels cannot be excluded. PNG larger image

TIFF original image Download: Fig 1. STR-KI accurately reports on SOX2 and TBXT expression, and NMP differentiation trajectories in vitro. (A) Schematic illustration of the targeting strategy to knock-in GFP and mCherry in the Tbxt and Sox2 loci, respectively. (B) Embryonic day 9 (E9) STR-KI chimaera with high STR-KI contribution shows appropriate localisation of fluorescent reporter expression: high mCherry (magenta) in the prospective brain, neural tube (NT) and primordial germ cells (PGCs). GFP (green) expression marks the notochordal plate (NP) and tail bud (TB). Scale bar = 500 μm. (C) Correlation between immunofluorescent staining for SOX2 or TBXT, and their respective reporter protein in individual cells derived from in vitro NMP differentiation assays. Top panel shows correlations assayed by immunofluorescent confocal imaging and bottom panel by flow cytometry. (D) Immunofluorescence images of differentiated STR-KI colonies following NMP-like differentiation [25]. Scale bar = 50 μm. (E) Representative example from a flow cytometry time-course analysis of STR-KI cells along the differentiation protocol (above), from mESCs (day 0) via NMP-like intermediate (at day 3) to day 5 (n = 4). A trajectory summary highlights population dynamics. (F) Quantification of the reproducibility of STR-KI mESC differentiation to day 3 NMP-like cells shows the percentages of Sox2mCh/TbxtGFP double-positive cells for a cohort of independent experiments (n = 23; Mean ± SD: 67.42 ± 18.28%). Data for Fig 1F: https://doi.org/10.5281/zenodo.21708690. https://doi.org/10.1371/journal.pbio.3003960.g001 We next tracked the reporters’ dynamics over a time course of NMP differentiation (Fig 1E). Briefly, this entails differentiation of ESC to an epiblast-like state, followed by treatment with the WNT agonist CHIR99021 (CHIR) and FGF2 from day (d) 2 onwards [25]. Cells appeared to move along a differentiation trajectory from Sox2mChhigh pluripotent ESC to a Sox2mChlow, putative epiblast state, followed by appearance of Sox2mChpos/TbxtGFPpos (double positive cells) within 3 days. D3 differentiated cells exhibited a spectrum of TbxtGFP levels before exiting the double-positive state towards Sox2mCh-negative (neg)/TbxtGFPhigh mesoderm from d4. Consistent with previous reports, STR-KI ESC differentiation reproducibly resulted in the production of putative NMCs co-expressing mCherry and GFP at a rate of ~67% of the population by d3 (Fig 1F).

2. Expression dynamics of Sox2 and Tbxt in vitro matches NMP emergence in vivo A previous report demonstrated that Sox2 expression in the caudal lateral epiblast is activated de novo in mouse via its N1 enhancer activity [26], corresponding approximately to the time at which NMCs are a detectable cell population [8]. Superficially, this would suggest that Sox2 is secondarily activated in the Tbxt-expressing caudal lateral epiblast, contrasting with our observation of the acquisition of Tbxt in Sox2-expressing cells in vitro (Fig 1E). We therefore examined the origin of NMPs in the epiblast. For this, we first analysed SOX2 and TBXT immunofluorescence in embryos staged between mid-streak and late headfold [27] (Fig 2A), including the neural plate (also known as pre-headfold) stage at which TBXT+SOX2+ putative NMPs emerge [8]. Up to early neural plate stage, SOX2 single-positive cells were located anterior to the node, while cells posterior to the node expressed only TBXT. Double-positive cells emerged around the node and just posteriorly to it during neural plate stage. These constituted most of the cells at the posterior node from late neural plate (LNP) stage onwards, as seen by staining for post-nodal TBX6 in this same regions, which delineates mesoderm and the primitive streak (Fig 2B and S1 Movie). Comparison of this staining pattern with focal electroporations at neural plate stage (n = 2 anterior to node, n = 3 posterior to node, Fig 2C and 2D) and the integration of our data with previously published in vivo single cell clonal lineage tracing around the node [28] (Fig 2D) shows that before neural plate stages, cells with dual neuromesodermal fate are located anterior to the node, while those posterior to it are exclusively mesoderm-fated. Thus, NMPs in the embryo emerge from the SOX2+ cell population anteriorly to the node and subsequently acquire TBXT expression (Fig 2E). This indicates that the acquisition of TBXT expression by SOX2+ cells in NMP differentiation conditions in vitro mimics the in vivo ontogeny of NMPs. PNG larger image

TIFF original image Download: Fig 2. NMPs emerge in vivo from a pre-existing SOX2+TBXT- population. (A) Immunofluorescent detection of SOX2 and TBXT in the fate-mapped region (Node-streak border, NSB – dashed line boxed) over developmental time. Upper row: DAPI-stained 3D renderings of mouse embryos (dissected on embryonic days E6.5 (midstreak, MS), E7.0 (late streak early bud, LSEB), E7.5 (late neural plate, LNP) and E8.0 (early and late head fold, EHF, LHF); as staged in [27]). Lower rows: Orthogonal projections of wholemount immunostained mouse embryos for SOX2 (magenta) and TBXT (green), showing mid- and para-sagittal optical sections of the node area (projections represent 8 µm across the Z axis). All scale bars are 50 µm. (B) Left: Immunofluorescent detection of SOX2 (magenta) and TBXT (green) in LSEB-NP and LNP embryos as depicted in A (dashed line boxes), but with the SOX2 field under enhanced exposure (SOX2*). Right: Immunofluorescent detection of TBX6 (cyan) in LSEB-NP and LNP embryos, and counterstained with DAPI (grey). As a reference, the node/notochord is circled by dashed lines. (C) Fate mapping at neural plate (NP) stage via focal electroporation of a ubiquitously expressed CAG-GFP transgene into cells anterior to the node tracks descendants into the neural plate and anterior caudal lateral epiblast (n = 2). Equivalent electroporation in the primitive streak (n = 3) produces descendants in the mesoderm (data summarised in D). (D) Focal electroporations superimposed on data from [28], where cells in or near the node at late streak (LS) to head fold (HF) stages were prospectively mapped to segments of the axis. Data has been re-displayed to indicate clones producing descendants in neurectoderm and/or mesoderm. (E) Summary scheme of data in A–D. Tracing of LSEB-NP and NP stage node region showing expression of SOX2 and TBXT in epiblast and nascent notochord. Expression is non-overlapping at the midline in LSEB-NP transition-stage embryo, while in LNP-stage embryo onwards, SOX2+ TBXT+ cells are evident. The anteroposterior extent of NM fate in the epiblast, extracted from diagrammatic representations in Forlani and colleagues at each stage is indicated, showing N-M fate is initially localised in the SOX2+ dorsal node region, which later becomes SOX2+TBXT+. Px, proximal; Ds, distal; A, anterior; P, posterior; L, left; R, right. All scale bars are 50 µm. https://doi.org/10.1371/journal.pbio.3003960.g002

3. In vitro derived Sox2mCh/TbxtGFP co-expressing cells are NMCs To characterise the cell populations emerging along the ESC-to-NMP-like cell trajectory, we probed the gene expression profiles of various subpopulations for diagnostic marker genes from d2 to d4 via qRT-PCR (Figs 3A, 3B, S2A, and S2B). Subpopulations were delineated based on the intensities of mCherry and GFP reporter detection in FACS. Comparison of mRNA expression levels of endogenous Sox2 and Tbxt with their respective reporters mCherry and GFP showed that in all sorted populations, mRNA expression reflected endogenous gene expression levels. The levels of Sox2 and Tbxt mRNAs exceeded those of the corresponding reporters, as expected, since the reporter allele, which transcribes both endogenous mRNA and reporter, is heterozygous (Fig 3B). Interestingly, the putative NMC population (Sox2mChpos/TbxtGFPpos) displayed relatively constant Sox2mCh fluorescence but a spectrum of TbxtGFP expression. Since the levels of TbxtGFP in the double-positive population increased before the appearance of TbxtGFP single-positive mesoderm cells, we hypothesised that low and high TbxtGFP populations might represent different states in the wider NMC population. Specifically, high TbxtGFP-positive cells might be further along a mesoderm differentiation trajectory compared to TbxtGFP-low cells [8,17,18]. PNG larger image

TIFF original image Download: Fig 3. Gene expression and lineage potential analysis of Sox2mCh/TbxtGFP fluorescent subpopulations during differentiation. (A) Differentiation scheme and representative gating strategy for FACS and further qRT-PCR analyses performed at day 3 of the differentiation protocol. (B) qRT-PCR gene expression analysis of the populations in A. n = 5 independent differentiation experiments. (C) Top: Post-d3 differentiation scheme in which populations in A were isolated by FACS, replated and further differentiated for an additional 3 days. Bottom: Representative flow cytometry graphs of one time-course experiment (n = 3, see S2C Fig), following the Sox2mCh and TbxtGFP expression dynamics over three more days. A summary of the lineage outcomes in three experiments over time is provided. The thickness of arrows represents the relative contribution to the indicated trajectory, with dashed arrows indicating a presumed trajectory based on the proportions of cells at the end of the experiment. (D) Experimental schematic for RNA-seq analysis of NMP-derived populations. Top: schematics of the sorting strategy: day 3 NMPs are sorted and cultured for two days and resorted and cultured for seven additional days under FGF/CHIR conditions, generating TbxtGFPpos and Sox2mChpos subpopulations at day +2, and Sox2mChpos and double-negative (DN) subpopulations at day +7. Bottom: summary of the differentiation routes, culture conditions and populations collected for RNA-seq. For NMP n = 3, for +2d Sox2+ and Tbxt+ n = 2, for Tbxt+ derived DN n = 3, for Sox2+ derived DN n = 1 per culture condition (FGF or FGF/CHIR), for Sox2+ derived Sox2+ n = 2 in FGF and 1 in FGF/CHIR. (E) PCA of RNA-seq transcriptomes from all sorted populations. Each dot represents an independent biological sample, coloured by population identity. PC1 (21.7%) and PC2 (12%) separate samples by differentiation state. (F) Heatmap of lineage-associated gene expression across NMPs and 48h sorted populations (TbxtGFPpos and Sox2mChpos). Genes are grouped by lineage identity: NMP, NMP+Mesoderm, NMP+Neuro, Mesoderm, and Neuroectoderm. (G) Heatmap of mesodermal and neuroectodermal marker gene expression across 48 h and day 7 sorted populations, showing transcriptional divergence between TbxtGFPpos -and Sox2mChpos-derived cells and the effect of FGF/CHIR vs. FGF treatment on day 7 Sox2mChpos and DN populations. Expression values are row-normalised (row min to max). Data for Fig 3B–3G: https://doi.org/10.5281/zenodo.21708690. https://doi.org/10.1371/journal.pbio.3003960.g003 To test this, we divided the NMC-like population into low-TBXT (Sox2mChpos/TbxtGFPlow) and high-TBXT (Sox2mChpos/TbxtGFPhigh) expressing subpopulations, where SOX2 levels were relatively constant. We analysed subpopulations based on marker gene expression: pluripotency-associated Nanog and Oct4; axial progenitor Fgf8, Evx1, Cdx2 and Wnt3a; paraxial mesoderm Tbx6; lateral plate mesoderm (LPM) Flk-1; and neural-associated Otx2, Sox1 and Pax6. Although both low-TBXT and high-TBXT populations expressed similar levels of NMC markers, high-TBXT cells were enriched for mesodermal markers such as Wnt3a and Tbx6, while expression of neural genes like Sox1 were low. The absence of Sox2, coupled with high expression of Tbx6 and Flk-1 in both Sox2mChneg/TbxtGFPpos and Sox2mChneg/TbxtGFPneg (double negative) populations, together with the downregulation of NMC/caudal markers like Fgf8 and Wnt3a from d3 to d4, suggests that these two populations represent nascent and differentiated mesoderm, respectively (S2A and S2B Fig). Finally, the expression of endoderm markers Sox17 and Foxa2 was universally low, suggesting negligible fractions of contaminant non-NMC cells. Thus, gene expression analysis in populations undergoing differentiation confirms that our double reporter cell line faithfully recapitulates endogenous gene expression and can be used to purify neural, mesodermal and NMC-like cell populations based on Sox2mCh and TbxtGFP expression levels (Figs 3B, S2A, and S2B). Furthermore, this analysis supports the idea that high levels of TbxtGFP mark a more mature mesodermal state than low TbxtGFP levels. To test whether the apparently dynamic changes in Sox2mCh and TbxtGFP cell populations during NMP differentiation (Fig 1E) represent one or more lineage trajectories, the lineage potentials of sorted subpopulations at d3 of NMP differentiation were assayed by FACS over 3 further days in continuous FGF and CHIR, which are permissive for both neural and mesodermal potential [9,25] (Figs 3C and S2C). Most (>80%) Sox2mCh single-positive cells remained single-positive. Together with the enrichment of Sox1 and Nanog in the starting population (Fig 3B), this suggests a neural and/or epiblast identity of Sox2mCh single-positive cells [29]. A small proportion of the progeny lost expression of both reporters, likely representing further maturation, since SOX2 is downregulated during neural differentiation. Sox2mChpos/TbxtGFPlow cells produced Sox2mCh single-positive cells by d6, but also a recapitulation of the pattern shown in Fig 1E: emergence of Sox2mChpos/TbxtGFPhigh cells, followed by a high proportion of Sox2mChneg/TbxtGFPpos and later double-negative cells. This suggests that Sox2mChpos/TbxtGFPlow NMCs produce both neural and mesodermal cells. Interestingly, Sox2mChpos/TbxtGFPhigh cells produced few Sox2mCh single-positive cells and followed a predominantly mesoderm trajectory, producing Sox2mChneg/TbxtGFPpos and then double-negative cells. Finally, Sox2mChneg/TbxtGFPpos populations only produced double-negative cells. Together with its enrichment in expression of mesoderm markers, it suggests this is a mesoderm-committed population. Thus, using our dual reporter, we describe a lineage trajectory in which a largely self-sustaining SOX2-positive population generates a SOX2-positive;TBXT-low dual-potency population, which progresses through a mesoderm-biased SOX2-positive;TBXT-high state towards a nascent mesodermal TBXT-high and, finally, a double-negative state (Fig 3D, see summary). This indicates that the TBXT-high NMC population is predisposed towards mesoderm differentiation. Therefore, within the Sox2mChpos/TbxtGFPpos double-positive NMC population, there are cells with apparent differing competence to differentiate to neural or mesodermal outcomes. To investigate the identity of the later populations, we next characterised intermediate progenitors and their main derivative populations by bulk RNA sequencing (Fig 3D and 3E). We sorted d3 NMC-like Sox2mChpos/TbxtGFPpos cells and cultured for two further days to capture their descendant Sox2mChpos/TbxtGFPneg and Sox2mChneg/ TbxtGFPpos populations (Fig 3F). We then cultured these cells for a further 7 days, sorting these final populations to their constituent Sox2mCh/TbxtGFP phenotypes. We found that culturing Sox2mChpos cells over seven days in FGF/CHIR did not enrich for a neural phenotype and therefore included an FGF-only condition for this population (Fig 3G). A principal component analysis (PCA) plot showed that replicate samples clustered together, and that Sox2mChpos and TbxtGFPpos cells cluster apart from each other and from the starting NMC-like population. The TbxtGFPpos-derived cells plus 7 days clustered far from the undifferentiated samples, as did the FGF-treated Sox2mChpos-derived samples. In contrast, FGF/CHIR-treated Sox2mChpos-derived samples did not diverge far along either PC axis, suggesting limited differentiation (Fig 3E). We next examined a set of mesoderm- and neural-specific markers for E8–8.5 recently differentiated mesoderm and neurectoderm. These were drawn from the top 50 differentially expressed genes (DEGs) between a derivative cell and its progenitor. We validated the mesoderm or neural character of DEGs between 2d samples and NMPs via comparison with a mouse single-cell RNA-seq atlas [30]. To assess the neural or mesodermal character of 7d samples, we manually curated the top 50 DEGs between 7d populations and their progenitors through searching E9.5–10.75 images and their annotations in the Mouse Genome Informatics database [31]. During the 2 days after NMP cell plating, cells in the TbxtGFPpos compartment had lost NMP-specific expression at the expense of early mesoderm differentiation (Fig 3F). A further 7 days of culture led to most of the cells becoming double negative (DN) and expressing relatively mature mesoderm markers (Fig 3G). In contrast, day 2 Sox2mChpos cells retained some NMP markers while enriching some neural markers (Fig 3F). Further culture in FGF enriched for neural differentiation markers, whereas FGF/CHIR did not result in neural marker upregulation (Fig 3G). These results support our starting hypothesis that TBXTGFPpos cells are mesoderm-fated, eventually downregulating TBXT as they pass from nascent to mature mesoderm, confirming that the predominant DN population in TBXTGFPpos-derivatives is mesoderm. We additionally suggest that WNT pathway activation in Sox2mChpos cells is incompatible with progression to a more differentiated neural phenotype, while FGF treatment alone permits neural differentiation. Previous reports have shown that treatment of pluripotent cells with FGF2 and CHIR leads to the expression of both paraxial and lateral plate mesoderm (LPM) markers [14,15]. In vivo, while homotopic grafts of NMCs overwhelmingly give rise to paraxial mesoderm [8,32], heterotopic grafts to the posterior primitive streak lead to LPM differentiation [21]. This suggests that NMPs may have LPM potency, although it cannot be excluded that this LPM arose from non-NMCs within the graft. To directly test the LPM-potency of NMCs, we examined the potential of different FACS-enriched populations at d3 of the NMP protocol upon BMP treatment, which induces LPM differentiation in bulk NMP cultures (S2D Fig) [21]. In Sox2mChpos/TbxtGFPneg and Sox2mChpos/TbxtGFPlow populations, we observed not only FLK-1 protein (a lateral plate/endothelial cell marker) labelling in colonies but also the formation of endothelial sheets and primitive tubes similar to those seen in classical endothelial assays [33] (S2E Fig). These populations were markedly different in their potential to generate and maintain SOX2+ and TBXT+ populations, since Sox2mChpos/TbxtGFPlow cells are able to generate and maintain both while only SOX2+ cells seem to remain in culture from the Sox2mChpos/TbxtGFPneg population. Surprisingly, NMP populations characterised by high TbxtGFP expression produced virtually no FLK-1 positive cells by the end of the culture and showed no TBXT and SOX2 staining, suggesting that high Tbxt expression at the NMP stage is incompatible with subsequent endothelial fate acquisition under these conditions. These findings suggest that the most immature epiblast-like or NMC populations (marked by no or low-Tbxt expression, respectively) can respond to BMP4 to produce LPM whereas Tbxt-high NMC populations might be already committed to a paraxial fate, a finding supported by the enrichment of paraxial mesoderm markers (e.g., Dll1, Dll3, Tbx6, Meox1, Tcf15, Lef1, Cited1, Ald1a2), in TbxtGFPpos populations. Nevertheless, our in vitro results support the idea that in vivo Tbxtlow NMCs themselves, and not contaminating populations, can contribute to both paraxial and lateral mesoderm if provided with permissive environmental cues.

4. Live imaging of STR-KI cells uncovers a role for TBXT in WNT and NOTCH-driven axis elongation In vivo, wildtype TBXT levels are not only critical for NMP differentiation, but also to maintain axial elongation [1]. However, the dynamic relationship between TBXT, SOX2 and axial elongation remains untested. To determine in real time the relationship between reporter expression, axial elongation and factors known to affect both aspects, we next used live-tracking of STR-KI reporter expression. For this, we tracked Sox2mCh and TbxtGFP expression in vitro in a standard gastruloid model of axis elongation, in which aggregates of mESCs grown in N2B27 remain SOX2 positive/TBXT negative until 48 hours post-aggregation (hpa), then exit pluripotency [34,35]. At 48 hours, a 24-hour pulse of CHIR is used to activate WNT globally and ensure robust symmetry breaking. WNT activation transiently induces TBXT expression in all cells [35], before it resolves into polarised expression at the posterior pole by 72 hpa and is then followed by elongation. We first compared the dynamics of mesoderm differentiation in STR-KI-gastruloids with previous reports (S2F Fig). After treatment with CHIR, most cells became TbxtGFPpos by 72 hpa, and down-regulated Sox2mCh. In the absence of CHIR, STR-KI cells remained Sox2mChhigh, but a small proportion of cells began to express TbxtGFP. These trends were maintained at the population level by 96hpa (S2F Fig). We also observed downregulation of Sox2mCh, symmetry breaking and polarisation of TbxtGFP to the posterior growing axis only upon the CHIR pulse (S2G Fig). We conclude that our STR-KI reporter can generate gastruloids with similar timings to cell lines in other studies (including a TBXT-GFP reporter line, [35]) and thus it is ideally suited for studying TF dynamics within them. We next assessed how the manipulation of WNT and NOTCH signalling pathways, which regulate mesoderm differentiation from NMPs [8,36], impact gastruloid axial elongation. We treated elongating gastruloids (96 hpa) with either WNT or NOTCH inhibitors (IWP-2 [37] and LY411575 (LY) [38], respectively) and live imaged them every hour for 24 hours to monitor major axis length and reporter expression (Fig 4A and 4B). Both WNT and NOTCH inhibition decelerated major axis growth (elongation) starting from as early as 2–3 h post-treatment (Fig 4C). PNG larger image

TIFF original image Download: Fig 4. STR-KI cell analysis indicates that Tbxt-expressing cells are responsible for WNT and NOTCH-driven gastruloid elongation. (A) Experimental setup for the generation of STR-KI gastruloids in 10% matrigel at 96 hours with DMSO, WNT signalling inhibitor IWP2 (2.5 µM) or Notch signalling inhibitor LY411575 (150 nM). Gastruloids were imaged on a confocal microscope every hour for 24 h after embedding. Examples of timelapses can be seen in S2–S6 Movies. (B) Quantification pipeline of 3D gastruloid confocal image stacks. Blue circle = high TbxtGFP region (top 90% in the gastruloid, see methods). All scale bars indicate 100 µm. (C) Axial length after treatment of gastruloids with IWP-2 or LY411575. (D) Heatmaps showing the average Tbxt-GFP or Sox2-mCherry signal per binned distance from the pole along time post Matrigel embedding. (E) Average difference in fluorescence signal of the Sox2/Tbxt region between control and treated conditions. N = 3 independent experiments. Total gastruloids analysed: DMSO = 35, LY = 38, IWP-2 = 41. Shaded areas indicate the mean ±0.95 confidence intervals, calculated by x̄ ± z*(σ/sqrt(n)). Data for Fig 4B–4E: Data file 1–3, https://doi.org/10.5281/zenodo.21676159. https://doi.org/10.1371/journal.pbio.3003960.g004 To determine how reporter expression responded to WNT or NOTCH inhibition during this period, whole gastruloid tissues were segmented and traces extracted along the extending long axis marked at the presumptive posterior end by TbxtGFP expression. Treatment with IWP2 or LY rapidly decreased TbxtGFP signal in the NMP region (within ~2–3 h). In contrast, Sox2mCh signal in this region increased after ~10–15 hours following an initial minor decrease compared to the DMSO control (Fig 4D), which could be the result of a delayed overproduction of neural cells or an overall Sox2 level increase on specification to neural identity. These changes in gene expression coincide in time with growth deceleration (Fig 4D and 4E). Thus, the rapid TbxtGFP signal downregulation coupled with slowdown of axis elongation in response to WNT or NOTCH inhibition suggests AP axis elongation is driven predominantly by Tbxt-expressing cells.–.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3003960

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/