Spatiotemporal contribution of neuromesodermal progenitor-derived neural cells in the elongation of developing mouse spinal cord

During vertebrate development, the posterior end of the embryo progressively elongates in a head-to-tail direction to form the body plan. Recent lineage tracing experiments revealed that bi-potent progenitors, called neuromesodermal progenitors (NMPs), produce caudal neural and mesodermal tissues during axial elongation. However, their precise location and contribution to spinal cord development remain elusive. Here we used NMP-specific markers (Sox2 and BraT) and a genetic lineage tracing system to localize NMP progeny in vivo. NMPs were initially located at the tail tip, but were later found in the caudal neural tube, which is a unique feature of mouse development. In the neural tube, they produced neural stem cells (NSCs) and contributed to the spinal cord gradually along the AP axis during axial elongation. Interestingly, NMP-derived NSCs preferentially contributed to the ventral side first and later to the dorsal side at the lumbar spinal cord level, which may be associated with atypical junctional neurulation in mice. Our current observations detail the contribution of NMP progeny to spinal cord elongation and provide insights into how different species uniquely execute caudal morphogenesis.


INTRODUCTION
In vertebrates, neuroepithelial cells give rise to the neural tube, which forms through two 50 processes along the anterior-posterior (AP) axis. The first process is primary neurulation, which 51 progresses through convergent extension, elevation, bending, and fusion of the neural plate, 52 forming the rostral neural tube (Pai et al., 2012). By the end of primary neurulation, only the 53 brain and anterior trunk structures of the spinal cord have formed. As the embryo develops, comprises the future caudal domain of the spinal cord, which is in continuity with the neural 58 tube in the trunk derived from the primary neurulation (Schoenwolf, 1984). Defects in this 59 process, termed "secondary neurulation," are often associated with spina bifida, a common 60 congenital malformation in humans (Yang et al., 2014). Secondary neurulation is an embryonic 61 process contributing to axial neural elongation; thus, here we use the term "neural tube 62 elongation" instead of "secondary neurulation." Vertebrates exhibit morphological differences 63 during neural tube elongation. For instance, chick and human tail bud cells undergo cavitation 64 to generate the elongating neural tube; eventually, the elongating neural tube adheres to the 65 primary neural tube at the junctional neurulation zone that is clearly present in the chick and 66 humans (Dady et al., 2014;Saitsu et al., 2004), but not in mouse (Schoenwolf, 1984). In the 67 mouse, elongating neural tube cells are polarized via cell rearrangement, and later migrate 68 toward the lumen of the primary neural tube and fuse, establishing a continuous neural tube 69 without a junctional zone (Colas and Schoenwolf, 2001). In addition, the chick neural tube 70 elongation process shapes the spinal cord and extends to the lumbar region (Colas and 71 Schoenwolf, 2001). A similar anatomical feature is also observed in humans (O'Rahilly and 72 Müller, 2003;Saitsu et al., 2004), which is contrary to that observed in mice, in which only the 73 tail is formed by neural tube elongation (Shum et al., 2010). 74 Another important aspect of caudal tube elongation is that the NSCs at this level are 75 derived from neuromesodermal progenitor cells (NMPs), which are able to produce both neural 76 and mesodermal tissues (Tsakiridis and Wilson, 2015;Tzouanacou et al., 2009). NMPs are 77 bipotent progenitor cells that co-express Sox2 and Brachyury T (BraT), which exist in the tail-78 bud of human, chick (Olivera-Martinez et al., 2012), and mouse embryos (Anderson et al.,79 2013). NMPs persist over extended periods and are eventually depleted toward the end of body 80 axis elongation (Wymeersch et al., 2016). In addition, fate mapping studies at the tail bud stage using TCreER2 transgenic mice further identified BraT-positive cells located in the floor plate 82 of the neural tube (Anderson et al., 2013). Collectively, this indicates that anterior and posterior 83 neural tissues are formed from independent lineages of cells, which challenges the traditional 84 concept of segregation of the three germ layers (ectoderm, endoderm, and mesoderm)

99
Presence of neuromesodermal progenitors in the mouse neural tube 100 NMPs emerge near the primitive streak and later remain in restricted regions in the tail 101 bud during axial elongation (Wymeersch et al., 2016). We extended the observation of the 102 spatiotemporal distribution of NMPs during different stages of axial elongation of neural 103 tissues, including the early and late stages of neural folding and neural tube elongation. During 104 an early stage of neural folding (E8), whole-mount immunofluorescence labeling revealed that 105 NMPs, which are double-labeled with Sox2 and BraT, were abundant within caudal 106 neuroepithelial sheets before neural folding (Fig. 1A). Furthermore, a population of NMPs was 107 observed within the neural tissues already forming the neural tube (Fig. 1A), suggesting that at 108 least one population of Sox2 + cells undergoing neurulation maintains BraT expression. At a 109 later stage of neural folding (E10), these double-positive cells were predominantly found within 110 the neural tube, while tail bud cells prominently expressed only BraT (Fig. 1B). At E11, double-111 positive cells were detected in a broad domain of the elongating caudal neural tube, while the expression levels of BraT appeared to be diminished in the tail bud (Fig. 1C), and, by E12, 113 double-positive cells were no more observed within the neural tube (Fig. 1D). Cross-sectioning 114 further demonstrated that the neural tube prominently expresses Sox2 only, and that only a few 115 Sox2-and BraT-expressing cells were restricted to the dorsal caudal tail bud domain (Fig. 1D).

116
Considering that the embryonic stage E10 marks the end of the neural folding process and the 117 beginning of neural tube elongation, our data suggest that NMPs are transiently located at the 118 caudal neural tube, even after depletion of NMPs in the tail bud.

120
Absence of neuromesodermal progenitors in the chick neural tube 121 We described above that double-positive cells, which are potentially NMPs, localize  In this study, we discovered that mouse NMPs localize at the caudal neural tube during 209 the neural tube elongation period, where they contribute to posterior neurulation and elongation.

210
On the other hand, chick NMPs were restricted to tail-bud tissue but not to the neural tube. The   Tissue processing was performed as described in (Lee et al., 2020), and whole-mount IHC 326 was performed as described in (Shaker et al., 2015). In brief, dissected tissues were fixed in 2% 327 PFA in PBS for 20 minutes on ice. Fixed embryos were then washed three times for 10 minutes 328 each on ice. Tissues were then dehydrated in 50% methanol in PBS for 10 minutes and then in 329 100% methanol, two times for 10 minutes on ice. Tissues were either stored at -20 °C or 330 rehydrated in PBST before blocking with 10% BSA in PBST overnight at RT. Primary 331 antibodies (listed above) were diluted in PBST containing 10% BSA and added to blocked 332 tissue for two days at 4 °C. Tissues were then washed three times with PBST for 20 minutes at 333 4 °C before incubation with secondary antibodies for two days at 4 °C. Tissues were washed 334 again with PBST for 20 minutes at 4 °C before mounting and imaging. All steps were 335 performed using a rocker, with gentle rocking for approximately 15 seconds so as to complete 336 one full rocking motion. All images were captured with a Leica TCS SP8 confocal microscope.  Comparisons of lineage-traced cell fates were presented as percent of each sample/tissue 343 section, and multiple samples were used to calculate averages and standard deviations, and for 344 statistical comparisons with one-way ANOVA or unpaired Student's t-test.

346
Statistical analyses were performed using one-way ANOVA or unpaired Student's t-test.

347
All analyses were carried out using the GraphPad Prism 8 software, and the results were 348 presented as mean ± standard deviation. P-values < 0.05 were considered statistically 349 significant.    Sectioned tissues were immunostained for NeuN (Red) and GFP (Green) and counterstained 577 with Hoechst 33342 (Grey). Scale bar = 30 µm.