Review of extra-embryonic tissues in the closest arthropod relatives, onychophorans and tardigrades

Abstract

The so-called extra-embryonic tissues are important for embryonic development in many animals, although they are not considered to be part of the germ band or the embryo proper. They can serve a variety of functions, such as nutrient uptake and waste removal, protection of the embryo against mechanical stress, immune response and morphogenesis. In insects, a subgroup of arthropods, extra-embryonic tissues have been studied extensively and there is increasing evidence that they might contribute more to embryonic development than previously thought. In this review, we provide an assessment of the occurrence and possible functions of extra-embryonic tissues in the closest arthropod relatives, onychophorans (velvet worms) and tardigrades (water bears). While there is no evidence for their existence in tardigrades, these tissues show a remarkable diversity across the onychophoran subgroups. A comparison of extra-embryonic tissues of onychophorans to those of arthropods suggests shared functions in embryonic nutrition and morphogenesis. Apparent contribution to the final form of the embryo in onychophorans and at least some arthropods supports the hypothesis that extra-embryonic tissues are involved in organogenesis. In order to account for this role, the commonly used definition of these tissues as ‘extra-embryonic’ should be reconsidered.

This article is part of the theme issue ‘Extraembryonic tissues: exploring concepts, definitions and functions across the animal kingdom’.

1. Introduction

The so-called extra-embryonic tissues are a common feature of embryogenesis in many animal groups and are crucial for embryonic development [1–4]. They serve several functions, such as nutrient uptake and transfer and removal of toxic waste products of metabolism [1–4]. Additionally, such tissues can form a protective barrier around the embryo, providing resistance against mechanical stress or desiccation as well as pathogen resistance by activating the innate immune system [1–3,5,6]. Moreover, they play an active role in morphogenesis and shaping the embryo [1,2,7,8]. Extra-embryonic tissues have been studied well in amniotes (sauropsids and mammals), where they comprise the amnion, the chorion, the yolk sac and the allantois; the latter has been modified considerably in different amniote subgroups (reviewed in [4]). Similarly, extensive research has been undertaken on the extra-embryonic tissues of insects (reviewed in [2,3]). On the other hand, only little is known about corresponding tissues and their functions in other arthropods and, in particular, their closest relatives, Onychophora (velvet worms) and Tardigrada (water bears), which together with Arthropoda comprise Panarthropoda [9,10].

In arthropods, a distinction is made between the ‘embryo proper’ (i.e. the germ bands of the embryo) and the ‘extra-embryonic tissue’, also referred to as ‘extra-embryonic ectoderm’ or ‘extra-embryonic membrane’ [1,5,7,8,11,12]. Typically, these tissues represent thin cell layers extending ventrally (if the germ band is split ventrally) and/or dorsally over the yolk (in most investigated species), thereby physically connecting the dorsal and, if present, ventral edges of the germ band (= ‘embryo proper’) [5]. In the model insect species, the fruit fly Drosophila melanogaster, there is a single-layered tissue that covers the embryo, the amnioserosa [2,8]. In most other insects, however, there are two such tissues, the amnion and the serosa, which have been highly modified in some taxa, likely as a result of functional adaptations to different environmental conditions in which the eggs are deposited [1–3,13]. Based on their relative position with respect to the developing germ band, the extra-embryonic membranes of insects may be homologous with the dorsal tissue of other arthropods; in contrast, the ventral extra-embryonic tissue is only present in a subset of arthropods that undergo ventral splitting of the germ band [5].

Extra-embryonic tissues are commonly defined as tissues that do not contribute to the final form of the body and this indeed holds true for the amnion, the serosa, and the amnioserosa of insects (reviewed in [1]). In other panarthropods, extra-embryonic tissues are identified based on their location of origin outside the embryo proper [5], and the extent to which these tissues contribute to the final form is unclear. We now know, however, after a century of research and the deployment of modern molecular methods including gene expression and knock-out/knock-down experiments, that the extra-embryonic tissue of insects fulfils important and essential functions during development (reviewed in [1–3,13]). The same applies to other arthropods. For example, a recent study in a spider revealed that cells in the dorsal cell layer express typical endodermal and midgut-marker genes, suggesting that these cells do not disappear but rather contribute to the midgut of the adult [14]. Consequently, at least some cells identified as extra-embryonic may become incorporated into the embryo proper.

The following questions still remain with regard to the major role/s and evolutionary history of extra-embryonic tissues across panarthropods: What are the ancestral functions of extra-embryonic tissues in arthropods? Are the tissues that we refer to as ‘extra-embryonic’ in insects and other panarthropods homologous with each other? To what extent do these tissues contribute to the final form of the embryo and adult in different panarthropod groups?

Owing to their key phylogenetic position as outgroups to arthropods, onychophorans and tardigrades might provide insights into the evolution and potential ancestral functions of extra-embryonic tissues in panarthropods. The major aim of this review is therefore to assess the existence of and summarize the current knowledge about the potential homologues of these tissues in onychophorans and tardigrades from the morphological and developmental perspectives. Additionally, we discuss possible functions of these tissues compared to those in insects and other arthropods, including the many highly interesting modifications in different onychophoran lineages. Finally, we summarize the currently sparse data on gene expression in these tissues in onychophorans and discuss their potential implications.

2. No evidence for extra-embryonic tissues in tardigrades

Tardigrades are microscopic invertebrates that are found in marine, limnic and limnoterrestrial habitats around the world [10]. Morphological and molecular evidence suggests that the tardigrade ancestor was secondarily miniaturized, which likely resulted in the reduction of cell number, losses of organs and even the complete loss of a large mid-body region (reviewed in [15]). Tardigrades typically exhibit rapid embryonic development, ranging from around 4 days in Hypsibius exemplaris (formerly identified as H. dujardini [16]) up to 14 days in Halobiotus chrispae [17]. The mature oocytes are deposited either freely into the environment or the moulted exuvium, which forms a protective enclosure [10,18,19]. The eggs are rich with yolk and surrounded by a thin vitelline envelope and a thick chorion (figure 1A–C; (reviewed in [10]). Cleavage is generally regarded as total and isolecithal (i.e. with evenly distributed yolk granules) and embryonic development is similar across the species studied thus far (figure 1A,B; [17,19–22]).

Figure 1.

Selected stages of embryonic development in the tardigrade Hypsibius exemplaris. Light (A–C), scanning electron (D,E) and confocal micrographs (F,G). Nuclei stained with the DNA-selective fluorescent dye 40,6-diamidino-2-phenylindol (DAPI) are depicted in glow scale in (F) and (G). Anterior is up in (D) and left in (D–G); developing legs are numbered. (A) Stage 1 embryo (approx. 1 h after egg deposition). Dividing nuclei are marked with asterisks. (B) Stage 3 embryo (approx. 3 h after egg deposition). (C) Expression of Hox3 in a stage 13 embryo (approx. 20 h after egg deposition); dashed lines indicate segment borders. (D,E) Stage 15 embryo (approx. 30 h after egg deposition) in lateral (D) and dorsolateral views (E). Images in D and E kindly provided by Vladimir Gross and Irene Minich. (F) Stage 13 embryo (approx. 20 h after egg deposition) in ventrolateral view. (G) The same developmental stage as in (D,E) in lateral view. Images in F and G kindly provided by Vladimir Gross. Staging according to [20]. he, head. Scale bars, 25 µm (A–C), 10 µm (D–G).

The initial studies of embryonic development in tardigrades were conducted in the 19^th and early 20^th centuries [19,23–25], while more recent investigations included transmission and scanning electron microscopy, cell lineage tracing using 4D microscopy, in situ hybridization, and fluorescence and confocal microscopy using DNA labelling in conjunction with immunohistochemistry (figure 1A–G; [10,17,20,21,26–31]). Moreover, a comprehensive staging system is available for H. exemplaris [20]. Despite this detailed embryological work, so far, no evidence for the occurrence of extra-embryonic tissues has been found in tardigrades. Owing to the small size of eggs (approx. 50 µm in H. exemplaris; figure 1A; [10,20,28]) and the relatively low number of cells (approx. 1000–3000, depending on the species [17,32]), only little amounts of such tissues, if at all, might be present. Since several tissues, organs and even entire body regions might have been lost as a consequence of miniaturization in tardigrades [15], it seems likely that extra-embryonic tissues might have been reduced as well. Hence, studies explicitly focusing on this topic would be required in order to obtain information about the presence or absence of dorsal and ventral extra-embryonic tissues in tardigrades, including more detailed cell lineage tracings using four-dimensional microscopy to identify cells that do not directly contribute to the final morphology of the animal. State-of-the art single cell mRNA sequencing might reveal cells with genetic fingerprints comparable to those present in extra-embryonic tissues of other panarthropods.

3. Extra-embryonic tissues in onychophorans

In contrast to arthropods, the anatomy of onychophorans has changed little since the early Cambrian, as they resemble fossil lobopodians—putative stem-group representatives of Panarthropoda [33–35]. They are worm-shaped, multi-legged terrestrial animals that can be found in humid habitats, such as decaying logs, soil and leaf litter, of tropical and temperate forests in equatorial regions and on the Southern Hemisphere (figure 2A; [9,37]). Onychophorans are divided in two major subgroups, Peripatidae and Peripatopsidae, and embryonic development has been studied in several representatives of each subgroup [9,37,41–44]. Onychophorans show typical short germ band development, with segments added successively at the posterior of the embryo (figures 2B, 3A–D, 4A–C; [9,38,39,47–54]). While most arthropods form a single, regular germ band with a solid ventral midline, the germ band of onychophorans develops as a paired structure, with both halves connected only at the anterior and posterior ends, thus representing a ventrally split germ band (figures 2B, 3A–F, 4A–C; [5,38,39,45,48–51,55,56]). Consequently, there is a ventral and a dorsal layer of cells that both have been defined as extra-embryonic tissue [5]. Only one instance of long germ band development has been reported from the peripatopsid Opisthopatus cinctipes with extra-embryonic tissues being supposedly absent [57]. This finding, however, has been challenged by a study of the closely related species O. roseus, which does show a typical short germ band development with large, widely expanded dorsal and narrow ventral extra-embryonic tissues that separate the two halves of the split germ band [58].

Figure 2.

Diversity of developmental modes and extra-embryonic tissue modifications in Onychophora. (A) World map showing geographical distribution of types of reproductive modes. Hatched lines demarcate circumequatorial Peripatidae from southerly distributed Peripatopsidae. Note that lecithotrophic viviparity might be in fact combined matrotrophic/lecithotrophic viviparity, which has only been demonstrated in Euperipatoides rowelli [36]. Modified from [9,37]. (B) Diagrams illustrating embryos of different species displaying distinct reproductive modes and extra-embryonic tissue modifications. Germ band and extra-embryonic tissues are depicted in grey and dark blue, respectively. Note that the embryo of Principapillatus hitoyensis forms a physical connection to the maternal uterine wall and is surrounded by an embryo sac of unresolved origin (light yellow). Note also that the dorsal extra-embryonic tissue is modified into a trophic organ in the matrotrophic species Peripatopsis sedgwicki and Peripatopsis moseleyi, but not in Peripatopsis capensis. Diagrams modified from [38] (Principapillatus hitoyensis), [39] (Peripatopsis sedgwicki, Peripatopsis capensis and Peripatopsis moseleyi), [40] (E. rowelli) and [5] (Peripatoides novaezealandiae).

Figure 3.

Dorsal and ventral extra-embryonic tissues in the combined lecithotrophic/matrotrophic viviparous peripatopsid E. rowelli. Confocal micrographs. Nuclei stained with DAPI are marked in grey, proliferating cells are labelled with the mitosis marker anti-phospho-histone H3 (α-PH3) depicted in cyan, and filamentous actin is labelled with phalloidin-rhodamine marked in red. Anterior is left in (A) and up in all other images. Developing legs are numbered; white arrowheads in (A) and (B) point to the embryonic slit. (A) Stage 0 embryo. The germ disc is visible as an opaque structure. The surrounding blastoderm represents the dorsal extra-embryonic tissue. (B) Early stage I embryo in ventral view. The developing ventral extra-embryonic tissue is separated by the embryonic slit. (C) Late stage I embryo in ventrolateral view. Note that the dorsal extra-embryonic tissue is much larger compared to the ventral extra-embryonic tissue, and extends anteriorly beyond the antennal segment. (D, D′) Stage III embryo in ventrolateral view. (E,F) Details of the border between germ band and ventral tissue of stage III and IV embryos in ventral view; dashed lines demarcate the border between germ band and ventral extra-embryonic tissue. Note that the cells and their nuclei in the ventral extra-embryonic tissue are larger and less dense compared to those of the germ band. Staging according to [45]. as, antennal segment; at, developing antenna; dt, dorsal extra-embryonic tissue; gb, germ band; js, jaw segment; jw, developing jaw; pd, proctodeum; sd, stomodeum; sp, developing slime papilla; ss, slime papilla segment; vt, ventral extra-embryonic tissue. Scale bars: 200 µm (A–D′), 100 µm (E,F).

Figure 4.

Early developmental stages of the placentotrophic peripatids Principapillatus hitoyensis (A,C,C′) and Epiperipatus biolleyi (B,B′). Confocal micrographs. Nuclei stained with DAPI are marked in grey, proliferating cells labelled with α-PH3 antiserum are depicted in cyan. Anterior is up in (B,B′,C,C′). (a) Stalked vesicle stage. Note that the embryo is connected to a placental stalk that develops from dorsal extra-embryonic tissue. (B, B′) Early segmentation stage embryo in ventral (B) and dorsal views (B′). Note embryonic slit (arrowhead) within developing ventral extra-embryonic tissue. (C,C′) Early coiled stage embryo in ventral (C) and dorsal views (C′). Staging according to [46]. as, antennal segment; dt, dorsal extra-embryonic tissue; pd, proctodeum; sd, stomodeum; sk, placental stalk; vt, ventral extra-embryonic tissue. Scale bars: 50 µm (B, B′, C, C′), 25 µm (A).

4. Reproductive strategies and associated modifications of dorsal and ventral extra-embryonic tissues across onychophoran subgroups

In contrast to tardigrades, onychophorans display a remarkable diversity of reproductive strategies ranging from oviparity to different types of viviparity. Viviparous strategies can be further classified into placentotrophic, lecithotrophic, matrotrophic and combined lecithotrophic/matrotrophic viviparity (figure 2A,B; reviewed in [9]). This diversity is reflected by various modifications of dorsal and ventral extra-embryonic tissues.

(a) . Lecithotrophic development

Both oviparous and viviparous lecithotrophic onychophorans produce large, yolky eggs and embryos, which are surrounded by embryonic envelopes that persist until birth (reviewed in [9]). While the deposited eggs of oviparous onychophorans contain embryos of early developmental stages, embryonic development of lecithotrophic viviparous species takes place entirely within the uterus of the mother and hatching and birth occur simultaneously [9,45,59–61]. Notably, it has been shown that the yolky embryo of the peripatopsid species E. rowelli additionally receives nourishment from the mother throughout development, thus displaying a combination of lecithotrophic and matrotrophic viviparity [36]. In both lecithotrophic and combined lecithotrophic/matrotrophic species, the dorsal and ventral extra-embryonic tissues occupy most of the surface of the embryo by covering broad regions that link the widely separated bilateral halves of the germ band (figures 2B, 3B–D, D′; [5,45,51,56,62]). The dorsal extra-embryonic tissue is larger than the ventral one and it typically extends anteriorly beyond the first/antennal segment at early developmental stages (figure 3B,C; [5,45]). In contrast to viviparous onychophorans, studies of embryonic development in oviparous species are scarce owing to the difficulty of obtaining fertilized eggs and relatively long development [60,61,63]. Hence, the dorsal and ventral extra-embryonic tissues of oviparous species have not been described in detail, but their extent and ontogenetic fate are likely similar to those in species with lecithotrophic and combined lecithotrophic/matrotrophic viviparity (cf. fig. 19 in [61]).

(b) . Matrotrophic and placentotrophic development

In contrast to lecithotrophic onychophorans, matrotrophic and placentotrophic species are characterized by small embryos that contain little or no yolk (figure 4A–C; reviewed in [5,9]). Since nutrients are (mostly) not deposited into the oocyte, they have to be supplied by the mother throughout development and the embryos need to be properly equipped to take up these nutrients. As a consequence, the embryos of matrotrophic species are surrounded merely by a thin permeable envelope, whereas those of placentotrophic species do not possess any secreted envelopes whatsoever (reviewed in [9]). These developmental modifications appear to have evolved hand in hand with morphological adaptations of the embryo, possibly owing to functional needs in order to facilitate nutrient uptake (figures 3B, 4A–D; [5,38,39,42,64–69]).

In the matrotrophic species, the ventral extra-embryonic tissue is reduced to a narrow band, whereas the dorsal extra-embryonic tissue is highly enlarged, in extreme cases forming the so-called trophic organ in the neck region of the embryo (figure 2B; [5,39,42,65,66]). While the trophic organ is transient and disappears at early stages in some species, it is retained until late in development in other species [39]. Anderson [5] suggested that the expansion and transformation of the dorsal extra-embryonic tissue into the trophic organ might be owing to a new functional role as an absorptive epithelium. While nutrient uptake is considered to be its main function [5], Willey [65] and Manton [39] suggested that the trophic organ—which typically surrounds the folded developing embryo—additionally serves as a protective barrier, similar to the serosa of insects [2,13].

Representatives of Neopatida (=neotropical Peripatidae [70]) arguably show the most specialized development among onychophorans. The small, yolk-free embryos form a temporary placental relationship to the maternal uterine wall (figures 2B, 4A,B; [5,9,38,46,53,55,69,71]). After implantation into the uterine wall, it becomes surrounded by a cellular embryo sac of uncertain origin and forms a hollow stalk, which is a derivative of the dorsal extra-embryonic tissue in the neck region of the embryo (figures 2B and 4A,B′; [9,38,46,47,53,55,69,71,72]). Hence, the position of the stalk in placentotrophic peripatids resembles that of the trophic organ in matrotrophic peripatopsids (cf. figure 2B). The distal tissue of the stalk proliferates and forms a thick physical connection between the embryo and the maternal uterine wall, the innermost layer of which in turn transforms into a syncytial epithelium (figure 2B; [5,38,46,53,69]). The stalk and the associated tissue were named placenta by von Kennel [38], although this structure only superficially resembles the mammalian placenta [46].

In contrast to the mammalian placenta, nutrients are most likely not transferred via the hollow stalk in placentotrophic onychophorans [9,46,53]. Hence, the question arises how nutrient uptake from the mother is accomplished in these species. It has been suggested that nutrients are absorbed from the maternal haemocoel/haemolymph and stored in the placenta until they are released into the cavity of the embryo sac where they are resorbed by the embryonic surface [53]. This indicates that the placental stalk might serve as an enlarged surface for nutrient uptake, similar to the trophic organ of matrotrophic viviparous peripatopsids. Unlike amniotes, where the placental relationship persists until birth [4], it is only transient in onychophorans [5,38,46,55,71]. After segment formation has been completed, the embryo detaches from the uterine wall and the placental stalk disintegrates by apoptosis, while the remaining dorsal and ventral extra-embryonic tissues connecting the paired germ bands persist as narrow strips until late in development but eventually also disappear [46,55]. The embryo then escapes from the embryo sac and lies freely within the lumen of the uterus for the rest of development [46,71]. Hence, most of the growth might take place after the trophic interaction between the embryo and the uterus has been lost and seems to be succeeded by a matrotrophic phase where nutrients are secreted into the lumen of the uterus/embryo sac and ingested by the embryo via the mouth and absorbed by the gut [38,46,53].

(c) . Evolution of extra-embryonic tissue modifications in onychophorans

The unusual diversity of developmental modes and extra-embryonic tissue modifications in onychophorans had already been discovered by early researchers, which has sparked a debate on how this diversity might have evolved and what the ancestral condition was. Most researchers favour the hypothesis that the last common ancestor of onychophorans showed lecithotrophic viviparity, with yolky embryos and without any maternal contribution [5,39,64,73]. The lack of yolk in placentotrophic and matrotrophic viviparous onychophorans was interpreted as a secondary loss, which may have resulted in secondary modification of extra-embryonic tissues, possibly by increasing the surface of the embryo for nutrient uptake [5,38,46,53,64,69,73]. The occurrence of scattered bits of yolk in some matrotrophic species was interpreted as an ‘intermediate’ state between the ancestral yolky eggs (present in the lecithotrophic, combined lecithotrophic/matrotrophic and oviparous species) and the nearly yolk-free embryos of Neopatida [73]. On the other hand, von Kennel [38] and Willey [65] refuted this hypothesis, stating that the data available at that time were insufficient to answer this question.

Mapping the types of nourishment supply on a recent phylogeny supports the hypothesis that either lecithotrophic or combined lecithotrophic/matrotrophic viviparity might represent the ancestral state [9,44]. Correspondingly, placentotrophic viviparity might be a derived feature of Neopatida and oviparity likely developed independently at least twice within Australasian Peripatopsidae. However, the phylogenetic position of several important subgroups of Onychophora is still unresolved, including some matrotrophic and lecithotrophic species. Hence, a more robust phylogeny is needed to determine the ancestral mode of development in onychophorans.

In arthropods and amniotes, it has been suggested that the presence of specialized extra-embryonic tissues is correlated with terrestrialization and egg deposition in corresponding habitats [3,13,74,75]. Extra-embryonic tissues surrounding the developing embryo might serve as a protective barrier against mechanical stress and desiccation and, thus, allowed arthropods and amniotes to conquer new habitats [3,13,74,75]. On the other hand, while onychophorans are exclusively terrestrial, they are all confined to humid microhabitats irrespective of their developmental mode or specific distribution (figure 2A,B; [9,38,39]). Therefore, the diversity of reproductive modes and extra-embryonic tissue modifications cannot be explained by differences in habitats. Moreover, embryonic development takes place within the maternal uterus in the majority of taxa, except for some oviparous taxa that are found in Australia and New Zealand (reviewed in [9]). In the oviparous species, the extra-embryonic tissues do not seem to exhibit conspicuous modifications that occur in various viviparous species, whereas the secreted chorion surrounding the deposited egg is thickened substantially and serves as a protective barrier [9,61]. Hence, in contrast to many arthropods, the extra-embryonic tissues do not seem to play a role in the protection of the embryo in onychophorans. In summary, there is so far no apparent correlation between terrestrialization and the origin of extra-embryonic tissue modifications in onychophorans, and the reason for the evolution of highly diverse reproductive strategies in onychophorans remains unknown.

5. Development and ontogenetic fate of dorsal and ventral extra-embryonic tissues

Despite considerable variation in reproductive modes and modifications of dorsal and ventral extra-embryonic tissues, their development seems to be similar across the onychophoran species studied thus far (figures 3A–F, 4A–C; [5,39,45,47,55,76]). In oviparous and most viviparous onychophorans, except for placentotrophic species, the dorsal extra-embryonic tissue arises from the blastoderm when the germ disc has begun to form by aggregation of blastomeres and increased cell proliferation (figure 3A; [9,39,47,55,72,76]). By that point, all tissue that is not part of the germ disc is defined as the dorsal extra-embryonic tissue. In placentotrophic species, this tissue arises from the blastula in a similar way, although the embryo undergoes total cleavage and does not form a germ disc (figure 3A; [5,38]). The ventral extra-embryonic tissue emerges only after the formation of the embryonic slit and the onset of segment formation in all species studied (figures 3B and 4B; [38,39,62,76]). As soon as the germ band starts to form and elongate anteriorly, the cells between the embryonic slit and the elongating split germ band become stretched tangentially and their nuclei appear more thinly distributed (figures 3B and 4B; [39,76]). Additionally, this stretching might be facilitated by cell ingression along the lateral lips of the embryonic slit [5,76]. At early developmental stages, the ventral extra-embryonic tissue is a seemingly paired structure with a common posterior end. The two halves of the developing germ band are separated by the embryonic slit and only later fuse anteriorly into a unitary structure (figure 3B–D; [5,9,39,45,49,64,76]). Both the dorsal and ventral extra-embryonic tissues are morphologically distinguishable from the germ band because their cells and nuclei appear larger and less densely packed (figure 3E,F). Furthermore, cell proliferation is considerably higher in the germ band tissue compared to the extra-embryonic tissues (figures 3A–D, 4A–C, B′, C′).

As soon as segmentation has been completed, the two halves of the germ band start to grow towards the dorsal and ventral midlines until they eventually fuse both dorsally and ventrally [38,39,46–49,55,64,77,78]. Growth of the germ band tissue is accompanied by increased cell proliferation (figure 5A,B) and stretching of cells near the borders of the germ band (figure 5E; [39,47,55]). Additionally, an increased number of apoptotic cells are detected in the extra-embryonic tissues, indicating that these tissues at least partially disintegrate during germ band closure (figure 5C,D; [47,55]). In summary, the dorsal and ventral closure in onychophorans is a neatly coordinated process involving morphogenetic cell movements and changes in shape, cell proliferation and apoptosis (figure 5A–E; [55]). Dorsal and ventral closure is also reflected in the development of inner organs, such as the nervous system, where axons grow out from the developing ventral nerve cords and connect at the dorsal and ventral midlines to form the dorsal and ventral commissures [9,79,80]. Interestingly, however, this process occurs before the overlying ectoderm has fused dorsally and ventrally [79].

Figure 5.

Germ band closure in combined lecithotrophic/matrotrophic viviparous E. rowelli. Nuclei stained with DAPI are labelled in grey, proliferating cells labelled with α-PH3 are marked in cyan, apoptotic cells labelled with terminal deoxynucleotidyl transferase-mediated dUTP nick end labelling are illustrated in glow scale, filamentous actin stained with phalloidin-rhodamine is depicted in red. Anterior is up in (A–D and F) and left in (E). During germ band closure, cell proliferation is abundant in the germ bands but absent from the dorsal and ventral extra-embryonic tissues (A,B). Instead, numerous apoptotic cells are detected in the dorsal and ventral extra-embryonic tissues (C,D). Note that the cells at the border between the germ band and dorsal and ventral extra-embryonic tissues become elongated (E), and a prominent net of actin fibres is present in the dorsal and ventral extra-embryonic tissues, which condenses at the midline (F). dt, dorsal extra-embryonic tissue; gb, germ bands; lb, limb bud; vt, ventral extra-embryonic tissue. Scale bars: 100 µm (A–D, F), 50 µm (E).

A similar process is evident in arthropods, indicating that germ band closure is a common feature of onychophorans and arthropods. In the fruit fly D. melanogaster, active cell movements and elongation of the epidermal cells lead to dorsal closure and subsequent internalization of the amnioserosa [8,81–83]. These morphogenetic movements are facilitated by a thick actin cable in the leading edge cells at the border between germ band tissue and amnioserosa, which pulls the germ band together in a zipper-like fashion [8,82]. Correspondingly, ablation of the amnion/amnioserosa in insects leads to severe defects in dorsal closure, indicating that they are involved in this process [83,84]. In D. melanogaster, this includes regulation of the volume and stiffness of the amnioserosa, followed by apical constriction and internalization. The internalized cells of the amnioserosa undergo apoptosis after the dorsal closure has been completed [8].

Although there is currently no evidence for the occurrence of an actin cable at the edges of the germ band in onychophorans, a prominent network of actin fibres is present in the dorsal and ventral extra-embryonic tissues (figures 3F and 5F). These fibres seem to constrict along the midline in an anterior-to-posterior progression during the germ band closure (figure 5F), indicating that the germ bands also close in a zipper-like fashion and that the dorsal and ventral extra-embryonic tissues might be actively involved in this process. Since the function of these actin fibres is currently unknown, future studies should focus on identifying the processes involved in the germ band closure of onychophorans by analysing and comparing the morphological and molecular changes in the germ band and the dorsal and ventral extra-embryonic tissues during development.

While dorsal closure has been described from most arthropods studied thus far, ventral closure is only seen in a few taxa that undergo germ band splitting [5,7]. However, this process differs substantially from that in onychophorans. In spiders, for example, the germ band initially forms as a unitary structure on the ventral surface of the yolky egg and the dorsal edges of the germ band are separated by an extensive dorsal extra-embryonic tissue (reviewed in [7]). Later in development, ectodermal cells along the ventral midline undergo apoptosis, leading to the splitting of the germ band into two halves that remain connected anteriorly and posteriorly [85]. Only a single layer of cells (i.e. the sulcus) persists between the two halves of the germ band, which then move dorsally around the yolk and reconnect along the dorsal midline (reviewed in [7]). During this so-called inversion, the sulcus widens more or less extensively, depending on the species. The germ band reconnects ventrally only after both halves have fused dorsally [7,85]. Hence, the dorsal and ventral closures do not occur simultaneously, like in onychophorans, but rather consecutively in spiders. Interestingly, numerous apoptotic cells occur in the dorsal extra-embryonic tissue during germ band inversion in spiders, which is on the other hand similar to onychophorans [85]. It is, however, unclear whether or not all cells of the dorsal and/or ventral extra-embryonic tissue undergo apoptosis in both onychophorans and spiders. To clarify this, statistically relevant quantitative data about the ratios of cell proliferation and cell death would be required from those tissues over the entire course of embryonic development.

6. Gene expression associated with dorsal and ventral extra-embryonic tissues

Some of the current research on the evolution and development of onychophorans and tardigrades increasingly focuses on the molecular aspects. This includes the study of gene expression patterns via whole-mount in-situ hybridization (WISH), which allows the spatial and temporal investigation of gene transcription. Most of these studies represent classic candidate gene approaches, in which known orthologues of important developmental genes, such as the Hox genes or the orthologues of the D. melanogaster segmentation gene cascade, have been studied (e.g. [30,86–90]). Two closely related representatives of Onychophora, E. rowelli and Euperipatoides kangangrensis, have been established as the major species for this kind of research (reviewed in [9]).

Gene expression analysis is inter alia used for the homologization of embryonic regions, such as tissues, organs, segments, tagmata and other structures. Such expression data, however, can also provide insight into the likely function of a given tissue, cell type or even a single cell. With respect to the dorsal and ventral extra-embryonic tissues of panarthropods, in particular outside insects, however, our knowledge about the gene expression is scarce. In D. melanogaster and other insects, the derived Hox gene zen (Hox3) is prominently expressed in the dorsal extra-embryonic tissues, including serosa or amnioserosa (e.g. [91,92]; reviewed in [1]). In other arthropods and onychophorans, however, Hox3 is not expressed in the dorsal extra-embryonic tissue (reviewed in [93,94]). Other examples of genes that are expressed in the amnioserosa of D. melanogaster include genes that are needed for providing immune response to pathogens such as the GATA factor serpent (e.g. [95,96]). This gene, however, is also an important factor for fat body [96] and midgut development [97], and the fat body indeed is also involved in providing immune response [98], possibly linking these processes and the function/s of the tissues involved.

As mentioned before, it is still unclear whether the amnioserosa, amnion and/or serosa of insects are homologous with the dorsal extra-embryonic tissue of other arthropods, and it is also unclear whether the dorsal extra-embryonic tissues of onychophorans and arthropods are homologous. If those tissues are indeed homologous, one would expect a certain degree of similarity and, thus, similar gene expression profiles. Recent research in the cobweb spider Parasteatoda tepidariorum revealed that its dorsal extra-embryonic tissue harbours a high number of cells that express endoderm- and midgut-specific genes, such as hepatocyte nuclear factor-4 and the aforementioned serpent [14]. Although the fate of these cells is not fully resolved, it is likely that they later contribute to the midgut. If so, these cells in fact cannot be regarded as extra-embryonic. serpent is one of the genes that is also expressed in the amnioserosa of D. melanogaster and its expression might be linked to the metabolism in this species [14], representing an example of a shared, and possibly conserved, ancestral function.

In myriapods, many segmentation genes are expressed in the dorsal extra-embryonic tissue in a similar pattern to the germ band (e.g. [12,99,100]), but again, it is unclear what their function may be here. It may be that they pattern the dorsal extra-embryonic tissue and later become part of the embryo. In any case, it appears clear that the dorsal extra-embryonic tissue of most (if not all) arthropods expresses developmental genes. Therefore, they are likely to actively contribute to the development of the embryo. The type of genes expressed in the dorsal extra-embryonic tissues may differ between different arthropod groups, as exemplified by the midgut genes, because in contrast to the spider they are not expressed in the dorsal extra-embryonic tissue of myriapods ([14]; R. Janssen 2015, unpublished data on gut markers in the diplopod Glomeris marginata). This may be owing to differences in midgut development between these two arthropod groups (e.g. [7,12]).

Currently, there is no compelling evidence for gene expression in the dorsal extra-embryonic tissue of onychophorans (but see Treffkorn & Mayer [40] for an unconfirmed report of decapentaplegic expression). Here, we present the expression of the Iroquois complex (Iro-C) gene iroquois, which is weakly expressed in this region in the lecithotrophic viviparous peripatopsid Euperipatoides kanangrensis (figures 6A, A′, 7A). The potential function of this gene in the dorsal extra-embryonic tissue of onychophorans, however, is unclear. In arthropods, its orthologue is inter alia expressed along the dorsal edge of the germ band, where it may function as a dorsal selector (e.g. [102]). Interestingly, in the red flour beetle Tribolium castaneum, iroquois has been suggested as a marker of the amnion [103,104]. It is thus tempting to speculate that the dorsal extra-embryonic tissue of onychophorans and the dorsal amnion of insects may represent homologous tissues that might also share the same or similar function/s.

Figure 6.

Overview of gene expression in the dorsal and ventral extra-embryonic tissues of lecithotrophic viviparous Euperipatoides kanangrensis. Whole-mount in-situ hybridization performed as per [101]. Light (A–F) and fluorescent (A′–D′) micrographs. Nuclei stained with the nuclear marker SYBR^® Green are labelled in green. Anterior is left and dorsal is up in all images; ventral view in (E, E′) and lateral view in all other images. Arrows point to expression in the split ventral midline, arrowheads in A and A′ point to expression at the dorsal edge of the germ band. Developing legs are numbered. Note expression of iro (A, A') in the dorsal tissue, and expression of net (B, B′), sog (C, C′), sim (D, D′) and vnd (E, E′, F) in the ventral extra-embryonic tissue. as, antennal segment; at, developing antenna; dt, dorsal extra-embryonic tissue; es, embryonic slit; iro, iroquois; jw, developing jaw; js, jaw segment; net, netrin; sim, single minded; sog, short gastrulation; sp, developing slime papilla; ss, slime papilla segment; vnd, ventral nervous system defective; vt, ventral extra-embryonic tissue. Scale bars: 200 µm (A–F).

Figure 7.

Details of gene expression in the dorsal and ventral extra-embryonic tissues of lecithotrophic viviparous E. kanangrensis. Light micrographs. Anterior is left and dorsal is up in all images; ventral view in (E) and lateral view in all other images. (+) and (–) point to expression or lack thereof in the dorsal and ventral extra-embryonic tissues, respectively. Developing legs are numbered. Note expression of iro (A) in the dorsal extra-embryonic tissue, and expression of net (B), sog (C), sim (D) and vnd (E) in the ventral extra-embryonic tissue. dt, dorsal extra-embryonic tissue; net, netrin; sim, single minded; sog, short gastrulation; vnd, ventral nervous system defective; vt, ventral extra-embryonic tissue. Scale bars: 200 µm (B), 100 µm (A,C,D), 50 µm (E).

The ventral extra-embryonic tissue is missing in most arthropods, but even in species in which it is present, this tissue develops differently from that in onychophorans. Again, relatively little is known about the transcriptome of the corresponding cells. In spiders, the ventral extra-embryonic tissue expresses inter alia the axonal guidance cue encoding gene netrin [105], suggesting that it is more than just a tissue covering the yolk. Another example is the expression of a Wnt gene (Wnt6) in the ventral sulcus of the tarantula Acanthoscurria geniculata [106]. Whether or not these cells become part of the ventral nervous system or any other tissue, however, is again unclear. Similar to its spider orthologue, netrin is expressed in the ventral extra-embryonic tissue of the onychophoran E. kanangrensis (figures 6B, 7B; see also fig. 1 in [107]). Possibly, this guidance cue is required for the correct neurogenic wiring in, across and around the ventral midline in panarthropod species with a split ventral germ band. In any case, expression of netrin in the ventral extra-embryonic tissue appears to be important and suggests that the corresponding cells have a specific function in neural development, but again it is unclear whether or not they become part of the later embryo and, finally, the adult.

To gain further insights into gene expression in the ventral extra-embryonic tissue of onychophorans, we investigated genes in E. kanangrensis that are typically expressed along the ventral midline in arthropods. We found that these genes, including netrin (net), short gastrulation (sog), single minded (sim) and ventral nervous system defective (vnd) are indeed expressed along the ventral edges of the split germ band (the split ventral midline), but some of them are also expressed in the space in between, i.e. in the ventral extra-embryonic tissue (figures 6B–F, 7B–E). It is thus possible that this tissue is part of the ventral midline in onychophorans.

The expression of midgut and endoderm patterning genes in E. kanangrensis has been analysed previously [108]. This study showed that at least the genes Hepatocyte nuclear factor 4 (Hnf4), GATA456/serpent and forkhead (fkh) are expressed in the ventral extra-embryonic tissue early in development, linking this tissue to gut development and possibly also nutrition uptake and metabolism. In contrast to spiders, however, in onychophorans these genes are expressed merely in the ventral extra-embryonic tissue and only early in development, although metabolization of yolk surely proceeds until later in development. It is thus still an open question whether or not the thin-layered extra-embryonic ‘membranes’ or tissues that cover the yolk in insects, spiders and onychophorans play a role in nutrition uptake. Various modifications of the dorsal extra-embryonic tissue in different onychophoran lineages support this assumption at least for onychophorans (figure 2B; [5,38,39]).

In summary, assessment of the available data reveals that gene expression analyses in the dorsal and ventral extra-embryonic tissues of arthropods and onychophorans are clearly underrepresented and that the functions of expressed genes are mostly unknown. It would therefore be important to firmly investigate these tissues by applying tissue-specific transcriptomics, single-cell sequencing and whole-mount in situ hybridization.

7. Conclusion and future directions

Extra-embryonic tissues are crucial for embryonic development in many animals and they serve a variety of functions, including morphogenesis, nutrient uptake, waste removal and resistance against mechanical stress, desiccation and pathogens. The evolution of extensive and complex extra-embryonic tissues is thought to have played a major role in the diversification and success of animal groups, such as amniotes and insects. Historically, the term 'extra-embryonic tissue' is based on the assumption that it is segregated from the embryo and does not contribute any cells to the adult. While this might hold true for the amnion and serosa of insects, in which these tissues serve important functions and embryonic development is severely affected when these functions are disrupted, it is possible that the extra-embryonic tissues of other arthropods do contribute cells to the adult.

In order to identify putative ancestral roles of extra-embryonic tissues in arthropods, it is also important to focus on their closest relatives, tardigrades and onychophorans. Detailed studies of morphogenesis in onychophorans and tardigrades were conducted as early as in the ninteenth century. Despite the remarkable detail in these studies, still virtually nothing is known about the occurrence of extra-embryonic tissues in tardigrades. On the other hand, studies on embryonic development in different onychophoran taxa revealed an extraordinary diversity of modes of development and nourishment supply.

The available data from onychophorans summarized in the present review support the hypothesis that extra-embryonic tissues are actively involved in embryonic development and are not just tissues covering the yolk and not contributing to the adult body. Modifications of the dorsal extra-embryonic tissue in yolk-free embryos and the expression of gut marker genes in the ventral extra-embryonic tissue of the lecithotrophic species suggest a role in nutrient uptake and gut development. Moreover, cytological and gene expression data indicate an involvement of extra-embryonic tissues in other developmental processes, such as germ band closure and neural development [1,7,8,106,107]. Similarities in gene expression and morphology may hint at a common origin of these tissues in onychophorans and arthropods.

Taken together, the available data from onychophorans and arthropods suggest that the characterization of extra-embryonic tissues needs to be revised and expanded, taking morphological, molecular and functional data into account. The potential incorporation of ‘extra-embryonic’ cells into adult tissues in onychophorans and spiders further suggests that the term ‘extra-embryonic tissue’ might be misleading and should be reevaluated at least for these taxa. Our review further highlights the importance of paying more attention to outgroups, such as onychophorans and tardigrades, in order to distinguish between ancestral versus derived character states. Future studies should focus, in particular, on the following topics and open questions:

• — Search for (rudiments of) extra-embryonic tissues in tardigrades to clarify their existence or potential loss;
• — Tracing of individual cell lineages in extra-embryonic tissues of arthropods and onychophorans to clarify their ontogenetic fate;
• — Identification of genes specific to extra-embryonic tissues in onychophorans (and other panarthropods) using differential transcriptomics and single-cell sequencing;
• — Comparative gene expression study of dorsal extra-embryonic tissues across onychophoran species to clarify the extent of evolutionary modifications.

Data accessibility

All data and material are publicly available.

Authors' contributions

S.T.: conceptualization, writing—original draft, writing—review and editing; G.M.: conceptualization, writing—original draft, writing—review and editing; R.J.: conceptualization, writing—original draft, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

The authors declare that they have no competing interests.

Funding

S.T. received support from the Doktorandenförderplatz of the University of Leipzig (grant no. G00021) and the Abschlussstipendium of the University of Kassel. G.M. gratefully acknowledges support from the German Research Foundation (Emmy Noether Program of the DFG: grant no. MA 4147/3-1). R.J. was funded by the Swedish Natural Science Council (VR) (grant no. 2015-04726) and the Marie Skłodowska-Curie Action (MSCA), Innovative Training Network (ITN), H20202-MSCA-ITN-2017 ‘EvoCELL’ (grant no. 766053).