The lateral plate mesoderm

1During gastrulation in vertebrates, the mesoderm forms axial, paraxial and lateral domains that harbor precursor cells for distinct organ systems. The lateral plate mesoderm (LPM) condenses into bilateral sheets of cells at the lateral edge of the developing vertebrate embryo, classically referred to as the lateral plate. While clearly discernible after gastrulation, the dynamic nature of the LPM is challenging to visualize and track during earlier development. Moreover, fate maps derived from various model organisms provide seemingly conflicting data, in part due to differences in lineage-tracing techniques and readouts, as well as uneven nomenclature to describe the LPM as ventral mesoderm, leading-edge mesoderm, visceral mesoderm, ventrolateral mesoderm or lateral mesoderm.

2The LPM also develops into a bewildering array of downstream cell fates. Lineage maps derived from transplantation and cell labeling experiments have linked the LPM to the origin of cardiovascular, hematopoietic, kidney, smooth muscle, craniofacial (head/neck) muscle, mesothelial and limb connective tissue progenitors. While these cell fates contribute essential structures to the adult vertebrate body, their earlier developmental connection is not immediately apparent from the final functional organs.

3Recent breakthroughs in genetic lineage tracing, reporter assays, live imaging and single-cell RNA-sequencing continue to uncover new details of the LPM and its derivatives, detailing its substantial contribution to the evolution of the vertebrate body plan. Here, we outline how the LPM emerges within the embryo and summarize the latest insights into how the LPM generates its diverse interconnected cell fates.

4Although the characteristic stripe architecture of the LPM becomes apparent during the segmentation stages, the LPM originates during gastrulation together with the axial and paraxial mesoderm between ectoderm and endoderm. Prominent signaling cascades influencing early LPM formation include the bone morphogenetic protein (BMP) and Nodal pathways that coordinate the patterning of anterior-posterior (A-P) and dorso-ventral axes. High levels of BMP signaling in the ventral domain of the embryo chiefly specifies the mesoderm territory that forms the LPM in all vertebrates. Nonetheless, coinciding with Nodal activity, LPM-assigned cell fates also emerge along the marginal zone of Xenopus and zebrafish embryos that contribute the cells at the circumference of the forming embryo. Prominently shown in zebrafish embryos, an increasing range or activity of BMP signaling in ventralized mutants results in larger domains of LPM-expressed genes for erythrocyte, pronephros and vascular lineages. Conversely, dorsalizing mutations that affect BMP ligands, BMP regulators or the loss of ventro-posterior transcription factors, such as Cdx4, cause the loss of posterior LPM structures. Nonetheless, BMP and Nodal are not instructive to trigger LPM formation per se, hinting at a more-complex signaling interplay leading to LPM induction. Fibroblast growth factor (FGF), canonical Wnt and retinoic acid (RA) signaling also influence the emerging LPM domains, including the specification of the heart field or of kidney and blood progenitors.

5Following gastrulation, the LPM takes on its characteristic architecture: bilateral stripes (‘plates’) of LPM progenitors form laterally in the embryo and subsequently partition along the A-P and medial-to-lateral axes into dedicated cell fate domains. Convergent extension of the embryo axis involving planar cell polarity (PCP) signaling also affects the final position of the LPM adjacent to the forming somites (paraxial mesoderm) and in relation to the ectoderm and the endoderm; however, LPM-specific mechanisms of cell arrangement have yet to be described in detail. In amniotes, the post-gastrulation LPM splits into a dorsal somatic layer and a ventral splanchnopleuric layer. During body segmentation, the LPM further separates into distinct anterior (ALPM) and posterior (PLPM) domains, while the bilateral cell fields gradually differentiate into descendant cell fates with distinct gene expression patterns.

6The diverse temporal, spatial and combinatorial activities of the signaling pathways involved in LPM specification make their exact influence challenging to dissect. Furthermore, the lack of a concise genetic or molecular definition of the LPM has limited its description. The bulk of our understanding of LPM fates derives from the regionalized, post-gastrulation expression patterns of individual transcription factor genes. Most prominently, the expression of (and transgenes based on) Foxf1, Bmp4, Hoxb6, Hand1, Hand2, Gata4 and Prrx1 have been harnessed in mouse and chick embryos to track various aspects of LPM patterning. However, these gene expression domains are either broader than, or delineate only parts of, the entire LPM. While expression of the T-box factor brachyury is commonly used as reference for mesodermal lineages in mammals, expression and activity of brachyury is dispensable for LPM formation in a variety of animal models, mainly contributing to tail formation and its ancestral role in notochord development.

7Conversely, the +2.0 kb enhancer in the zebrafish draculin (drl) locus (+2.0drl) is specifically active in LPM-primed mesendoderm during zebrafish development by responding to the mesendoderm regulators eomesodermin A, FoxH1 and Mixl1, together with BMP- and Nodal-controlled Smads. Although drl seems to be a zebrafish-specific zinc-finger gene, +2.0drl enhancer-based reporter transgenes also label the emerging LPM in chick, axolotl and lamprey, and in the nonvertebrate chordates Ciona and amphioxus, suggesting that LPM emergence is instructed by a conserved molecular program. It remains to be determined whether LPM formation is universally controlled by eomesodermin A, FoxH1 and Mixl1 orthologs across chordates, which mechanisms induce expression of the genes that pattern the LPM post-gastrulation, and whether any of these conserved genes are LPM specific.

8The seemingly complex developmental relationship of the final organs of the LPM becomes more accessible in the light of their evolutionary connections. In amphioxus, the mesendoderm forms as a continuous dorsal to ventral layer, and a LPM-resembling domain can be recognized in between the dorsal somites and the ventral endoderm. This seemingly simple setup could hint at the original mesendoderm architecture in the last common chordate ancestor, which is set up by the conserved LPM-instructing program. Amphioxus orthologs of genes expressed in the LPM in vertebrates, such as Foxf1, Hand genes, Tbx20 and Nkx2.5, are active throughout the whole ventrolaterally located mesoderm, indicating that the LPM in amphioxus does not segregate into ALPM and PLPM, whereas ALPM and PLPM are clear features in lampreys. These observations suggest that LPM compartmentalization along the embryonic axes is a vertebrate adaptation. Curiously, in Drosophila, visceral mesoderm formation depends on Bap and the Foxf1 ortholog Bin, whereas heart formation relies on GATA factors together with Hand and Tin, the orthologs of the vertebrate heart regulators Hand1, Hand2 and Nkx2.5. This conservation of LPM-associated gene expression hints at a deeply rooted molecular LPM program dating back to early bilaterians.

9The formation of the circulatory system provides an illustrative example for LPM-derived organ development. During early somitogenesis, the heart forming in the ALPM and the endothelial and hematopoietic lineages forming in both ALPM and PLPM become detectable by both shared and specialized gene expression patterns. In the ALPM, endocardial and myocardial progenitor populations become detectable adjacent to cranial endothelial progenitors by expression of Nkx2.5, Etv2, Lmo2 and Scl/Tal1, whereas the PLPM harbors the trunk endothelial and primitive erythrocyte progenitors also expressing Etv2, Lmo2, Scl/Tal1 and Gata1, but not Nkx2.5.

10In the emerging mouse mesendoderm, Mesp1 expression downstream of Eomes demarcates the earliest cardiac progenitors that upregulate Gata4, Nkx2.5 and Hand2. This initial requirement for Mesp factors holds true for cardiac progenitor formation in Ciona. In contrast, Mesp1 orthologs in Drosophila and in zebrafish seem dispensable for cardiogenesis. These peculiar findings suggest a degree of flexibility in cardiac progenitor initiation, which has yet to be further characterized. Curiously, akin to Eomes, Smarcd3 (BAF60c) expression in the mouse precedes Mesp1 upregulation and is essential for heart formation. Regulatory elements from the mouse Smarcd3 locus actively drive reporter expression in the zebrafish ALPM, while zebrafish Smarcd3 function has been linked to paraxial muscle differentiation through interaction with brachyury/Ntl. These data might indicate that Smarcd3 orthologs act as co-factors to T-box factors such as Eomes or other yet-to-be-determined transcription factors. How universal this interplay is for cardiac progenitor formation or within the LPM in general warrants further investigation.

11The subsequent migration of cardiac progenitors to the midline and the formation of the linear heart tube depends on several factors, including platelet-derived growth factor (PDGF) and Robo-Slit signaling providing extrinsic and intrinsic migration cues. In zebrafish (sox32/cas) and mouse (Sox17) endoderm mutants, multiple heart tubes form within the bilateral ALPM, indicating that cardiac progenitors have an intrinsic propensity to form a rudimentary heart. As a universal trait, the developing heart incorporates cells from two ALPM-associated progenitor fields, the so-called first versus second heart fields (FHF and SHF, respectively). While the FHF descendants set up the initial heart tube with atrium and ventricle for systemic circulation, the addition of SHF progenitors extends the heart on both poles. As a fundamental building block of all vertebrate hearts, the interplay of FHF and SHF influences cardiac conductivity and facilitates sequential contraction; however, why two progenitor pools are required for heart formation remains uncertain. SHF descendants contribute to the increasingly complex compartmentalization in the heart of terrestrial vertebrates, culminating in a right ventricle that is dedicated to pulmonary circulation.

12In vertebrates, blood and endothelium form concomitantly with the heart. While expressing an overlapping set of genes, endothelial and hematopoietic progenitors in ALPM and PLPM develop seemingly disconnected from each other but temporally in sync. Endothelium and blood arise, at least partially, from bipotent hemangioblasts, as well as from fate-restricted angioblasts and hematopoietic progenitors. The zebrafish npas4l/cloche mutant is virtually devoid of blood and endothelium (with exception of few surviving angioblasts), as evident by the broad lack of scl, lmo2 and etv2 expression. While a clear functional Npas4l ortholog is currently unknown outside of fishes, these findings place Npas4l at the top of the developmental hierarchy controlling the formation of endothelial/hematopoietic progenitors. Among the earliest conserved endothelial/hematopoietic transcription factors is the ETS factor Etv2 that, together with Scl/Tal1, governs endothelial/hematopoietic and hemangioblast formation in mouse, chick and zebrafish. With over a dozen family members expressed at different developmental time points downstream of Etv2, ETS factors play a continued role in endothelial differentiation towards a functional vascular network with veins and arteries. For example, the expression of Fli1, Erg and Ets1 provides powerful endothelial markers in various model systems. In addition to transcription factors, vascular-endothelial growth factor (VEGF) signaling is guiding endothelial differentiation. Reflecting this role, VEGF receptor genes such as Vegfr2 and Flk1 and their paralogs are among the earliest genes contributing to hemangioblast formation. Nonetheless, despite a wealth of insights into the mechanisms of vascular system formation, how Etv2 and its related factors (and the even more upstream-acting Npas4l in zebrafish) are selectively activated within the cardiovascular-primed LPM remains unknown.

13Blood emerges in several discrete waves of hematopoiesis at distinct time points and locations within the embryo, and its development is closely intertwined with endothelium formation. In teleosts and amphibians, a specialized primitive wave of myeloid progenitors emerges in the ALPM that might be an ancestral trait. The first primitive wave of PLPM-derived blood consists of transient, embryonic erythrocytes that stem from Scl/Tal1-, Lmo2- and Gata1-expressing progenitors. An intermediate wave of erythro-myeloid progenitors form encased within the developing vessels in zebrafish, and several yolk sack and placenta cell populations have been attributed with intermittent hematopoietic potential in mice. Finally, Runx1-expressing definitive hematopoietic stem cell (HSC) progenitors bud off from the ventral wall of the dorsal aorta (so-called hemogenic endothelium) through an endothelial-to-hematopoietic transition in zebrafish and in mice, while a somite-based contribution of aortic wall and HSCs has also been reported in zebrafish. In addition to hemangioblasts, the repeated interdependence of hematopoietic waves on endothelial cells possibly hints at a joint evolutionary origin. Such scenarios have received further support from the ontogeny of macrophage lineages and from observations made in a variety of invertebrates. Together with the joint expression of key genes in endothelial and hematopoietic progenitors, a common origin of all cardiovascular lineages within the LPM provides the developmental context to tie these interdependent cell fates together.

14Related to endothelium and blood, the origin of lymphatic vessels seems more complex. While trunk and cardiac lymphatics have been shown to originate from LPM-derived lineages, in particular from veins in mouse and zebrafish, recent work in the mouse indicates that Pax3:Cre-expressing paraxial mesoderm is a major source of trunk and cardiac lymphatic vessels. Similarly, Pax3:Cre-expressing cells as paraxial mesoderm contribute to at least parts of the endothelium in the mouse forelimb. Whether Pax3-based lineage tracing in these scenarios is strictly paraxial mesoderm specific, and whether a non-LPM origin for lymphatics and individual endothelial lineages is a universal trait across vertebrates, remain to be elucidated.

15In line with the surprising lineage diversity of the LPM, detailed lineage tracing studies in the mouse have revealed that the ALPM progenitors that form SHF also contribute to neck and craniofacial muscles alongside paraxial mesoderm and neural crest contributions to their connective tissue. Although the nomenclature of mesodermal domains in the developing head suffers from disparities across the literature, at least part of the cardiopharyngeal field (CPF) designates an ALPM-centered progenitor pool that, in addition to forming the heart, also contributes to craniofacial and neck muscles. Comparative anatomical and genetic studies across chordates have demonstrated joint cardiac and branchiomeric muscle formation, together with overlapping expression patterns in the anterior mesoderm of key regulators, including Nkx2.5, Isl1 and Tbx1. The evolutionary timeline of additional cranial muscle groups that are distinct from trunk muscle trajectories coincides with the adaption of multi-chambered hearts. The first traces of a CPF are even detectable in cephalochordates, by the expression of several T-box transcription factors, including Tbx1/10 and Tbx20 in the ventrolaterally located mesoderm of amphioxus. More insights into the CPF promises to reveal new insights into vertebrate head and neck evolution, and how seemingly disparate mesodermal populations in the head, such as the cephalic or cranial paraxial mesoderm, are interconnected.

16In amniotes, the kidney develops in a distinct temporal sequence via three different stages: pronephros, mesonephros and metanephros (adult kidney). In teleosts and amphibians, the mesonephros functions as the adult kidney. The kidney primordia emerge as bilateral fields expressing Wt1, Lhx1, Pax2 and Pax8 in the PLPM during early somitogenesis. During differentiation, besides the rostrally positioned glomerulus, the nephric tubule epithelium is specified along the A-P axis in several segments. Curiously, the mediolateral position of kidney progenitors relative to the paraxial mesoderm is distinct between fish and amniotes: in amniotes, kidney progenitors form directly adjacent to the paraxial somites as the most medially located LPM stripe (also referred to as intermediate mesoderm). Conversely, in fish, kidney progenitors form lateral to the endothelium-hematopoietic progenitors in the LPM, clearly embedding kidney origins in the context of LPM formation.

17A powerful example of the evolutionary adaptability of the LPM is the connective tissue and skeleton of the paired appendages, which has been repeatedly reviewed. Fore- and hindlimb buds emerge from the somatic LPM at specific positions along the A-P axis. RA signaling and Hox genes are involved in properly positioning the progenitor fields. In close interplay with the ectoderm secreting FGF ligands (i.e. Fgf8), LPM-expressed Tbx5 and Tbx4, among other factors, contribute to initiating limb formation in the ALPM and PLPM, respectively. The limb skeleton and connective tissue have been predominantly fate-mapped to an LPM origin, while limb musculature is contributed by paraxial/somitic mesoderm. Curiously, Tbx5 is required for both forelimb and heart development, and its expression encompasses the forelimb and heart field progenitors that emerge adjacently within the LPM. Whether Tbx5 expression is driven by a joint program in both progenitors or if each progenitor field responds to separate inputs remains uncertain.

18The epithelial lining of the coelomic cavities, the so-called mesothelium, is also LPM derived. The mesothelium covers the body cavities (parietal layers) and the organs within (visceral layers), which have been linked to somatic and splanchnic LPM origins, respectively. In mice, Wt1 and Gata4, as well as Msln expression, provide the earliest (E9.0) genetic and lineage markers for the visceral coelomic epithelium, which is detectable lateral to the urogenital progenitors. The coelomic epithelium is increasingly recognized as the source of a wide range of cell types. In particular, smooth muscles surrounding the gastrointestinal and respiratory tract, and the vascular system, as well as additional fibroblast-like lineages, have been tracked in mice and chick. Genetic lineage tracing in zebrafish further supports a conserved LPM origin for the visceral intestinal smooth muscles. Nonetheless, which smooth muscle lineages (or if all) feature a LPM origin remains unresolved.

19The spleen develops from condensations of the Wt1-positive coelomic epithelium in mice and humans, and visceral white adipose tissue (WAT) depots in mice have been suggested to derive from LPM, as indicated by genetic lineage tracing with Wt1:creERT2 and HoxB6:creERT2. Finally, gonadal structures have been assigned to have a common origin in the adrenogonadal primordium, which in mouse and rat arises from a thickening of the coelomic epithelium. In these models, Gata4, Tbx18, Tcf21 and Wt1 are among the earliest activated genes expressed throughout the genital ridge, recapitulating a gene expression signature observed in other regions of the coelomic epithelium. Lineage tracing in mouse with Cre driver lines including Wt1:creERT2 and Tbx18:creERT2 have shown that coelomic epithelium-derived precursor cells give rise to several cell types within the primitive gonads, such as the Sertoli cells in the testis and the follicular cells in the ovary. During gonadal differentiation, the primordial germ cells (PGCs) arrive at the genital ridge and are retained within the forming gonads. Despite these fascinating links to LPM origins, how the initial mesothelium-associated lineages arise within the LPM remains to be charted.

20The LPM connects diverse organ systems in the vertebrate body plan. How individual LPM-derived lineages emerge during development is of major interest for controlled stem cell reprogramming. Two principal approaches promise the generation of therapeutically relevant, human LPM derivatives. First, direct reprogramming into specified LPM fates has been achieved with defined combinations of developmental transcription factors in iPSCs, ESCs and somatic cells, as exemplified with the programming of fibroblasts into beating cardiomyocytes using GATA4, HAND2, MEF2C and TBX5. Second, the stepwise recapitulation of developmental signaling towards a LPM gene expression signature can be achieved in ES cells, as demonstrated with timed exposure to TGFβ and BMP. Nonetheless, characterizing and categorizing LPM-derived cells through reprogramming still relies on limited marker signatures. Nonetheless, LPM-based organoid models promise to provide a potent source of clinically relevant cell types and new platforms to probe the basic mechanisms of LPM patterning. Together, these new and exciting models enable elucidation of the mechanisms driving LPM emergence from embryonic mesendoderm. Easily rivaling the cell fate potential of the neural crest, we are slowly unraveling the cellular properties that render the LPM capable of forming its diverse descendant cell fates and organ systems.