Structure and functions of stable intercellular bridges formed by incomplete cytokinesis during development

Abstract

Cytokinesis, the final step of cell division, normally proceeds to completion in living organisms, so that daughter cells physically separate by abscission. In certain tissues and developmental stages, on the other hand, the cytokinesis process is incomplete, giving rise to cells interconnected in syncytia by stable intercellular bridges. This evolutionarily conserved physiological process occurs in the female and male germline in species ranging from insects to humans, and has also been observed in some somatic tissues in invertebrates. Stable intercellular bridges have fascinated cell biologists ever since they were first described more than 50 years ago, and even though substantial progress has been made concerning their ultrastructure and molecular composition, much remains to be understood about their biological functions. Another major question is by which mechanisms complete versus incomplete cytokinesis is determined. In this mini-review we will try to give an overview of the current knowledge about the structure, composition and functions of stable intercellular bridges, and discuss recent insights into the molecular control of the incomplete cytokinesis process.

Introduction

Cytokinesis is the final step of cell division that leads to the physical separation of the two daughter cells.1–3 Cytokinesis normally proceeds through sequential steps of cleavage plane specification between the segregated chromosomes, formation and constriction of an actomyosin-based contractile ring, stabilization of the intercellular bridge and finally completion of the cytokinesis process by abscission.1,2,4–7 These processes are governed by a dynamic interplay between the actomyosin cytoskeleton, the mitotic microtubule spindle, membrane trafficking and fusion.1–4,8

During the development of certain tissues, on the other hand, the cytokinesis process is incomplete, manifested by arrest of cleavage furrows and their transformation into stable intercellular bridges that interconnect cells in syncytia.9–12 Stable intercellular bridges are found in the female and male germline in metazoans ranging from insects to humans and have also been described in certain somatic tissues in invertebrates (Table 1).9–15 Plasmodesmata in plants and the septal pore in fungi also represent intercellular bridges formed by incomplete cytokinesis.11,16 The cytoplasmic bridges that form by incomplete cytokinesis are large enough (diameter range of ∼0.2–10 µm, see Table 1) to allow organelles and/or macromolecules to pass.11 This thus represents a very intimate and direct type of intercellular communication and complements the multitude of ways by which cells communicate during the development of multi-cellular organisms, e.g., by signaling via diffusible ligands, cell-cell adhesion, gap junctions or by other types of cytoplasmic connections (such as tunneling nanotubes in animal cells and epithelial bridges between human bronchial epithelial cells that form via projections of filopodia-like extensions between neighboring cells that are remodeled to intercellular tubes).17,18

Since the first descriptions of stable intercellular bridges significant progress has been made concerning their ultrastructure and molecular composition. However, their biological functions and the molecular mechanisms that determine complete versus incomplete cytokinesis remain largely unknown. Here, we will give an overview of the current understanding of these aspects of metazoan stable intercellular bridges and provide insight into recent advances in understanding how incomplete cytokinesis is achieved. We will focus on stable intercellular bridges in Drosophila melanogaster, since they have yet been most thoroughly characterized, but also cover important aspects of mammalian stable intercellular bridges.

Germline Intercellular Bridges

During gametogenesis in both females and males in diverse invertebrate and vertebrate species, gametes undergo incomplete cytokinesis to form clusters of germ cells interconnected by stable intercellular bridges (Table 1). The evolutionary conservation of incomplete cytokinesis in the germline suggests that this germ cell interconnection is important, and indeed stable intercellular bridges have been shown to be required for fertility in insects and mammals (see details below).11,12,19,20

Female germline intercellular bridges.

Stable intercellular bridges that interconnect germ cells in female germline cysts form during oogenesis in species ranging from insects to mammals (Table 1 and Fig. 1).12,15,21 The functions of the syncytial interconnection of germ cells in the mammalian female germline are not well understood, but has been proposed to be important for synchronization of mitotic divisions and entry into meiosis, and possibly coordination of programmed cell death or selective degeneration of oocytes.22–25 Strikingly, however, recent studies in mouse have indicated that the bridges are not required for female fertility.26 In many insects, the germ cell intercellular bridges allow the directional transport of nutrients between cells in a syncytium to promote the growth of one of the cells that will develop into the oocyte, whereas the others, after contributing their cytoplasmic contents to the oocyte, retract and die (Table 1).12,27

The functions, formation and composition of stable intercellular bridges in the female germline are especially well characterized in Drosophila melanogaster, where the intercellular bridges are called ring canals (Fig. 1). In order to illustrate the processes of incomplete cytokinesis, ring canal formation and ring canal functions in the Drosophila female ovary we will first give an overview of Drosophila oogenesis and then summarize the key steps of incomplete cytokinesis and ring canal formation. More detailed descriptions are found in a number of excellent reviews and original articles (Table 1).12,28

Oogenesis is a continuous process in Drosophila females29 (Fig. 1A). Each ovary consists of 16–20 ovarioles, each of which is a single egg assembly line with egg chambers of increasing age and size from the anterior towards the posterior (Fig. 1A). The egg chamber is the structural and functional unit of the ovary, and consists of 16 germ cells; 1 oocyte and 15 nutritive cells called nurse cells that are surrounded by a single layer of somatically derived epithelial follicle cells (Fig. 1A and B). Egg chambers are classified into 14 developmental stages based on their morphology. Stage 1 represents the 16-cell syncytium immediately after encapsulation by the follicle cells while stage 14 refers to the mature egg where eggshell is completely formed and the nurse cells have degenerated. As described above, the nurse cells provide the developing oocyte with organelles, proteins and RNA through the ring canals, a process required for normal oocyte development and female fertility12 (Fig. 1A–D). The epithelial follicle cells, on the other hand, participate in the oocyte polarity specification and in the secretion of the egg shell during the late stages of oogenesis, a process enhancing survival of the embryo.30

The Drosophila female germ stem cells, located in the germarium at the anterior tip of each ovariole, divide asymmetrically by complete cytokinesis to give rise to another stem cell and a cystoblast (Fig. 1A).12,29 The cystoblasts in turn undergo four synchronized rounds of mitosis via incomplete cytokinesis, giving rise to a cluster of 16 germ cells interconnected by 15 arrested cleavage furrows (Fig. 1A).28,31 The cleavage furrow arrest is accompanied by the appearance of an electron dense outer rim linked to the plasma membrane.28,32–34 Another marker of constriction arrest is the formation of a plug of the fusome inside the ring.28 The fusome is a vesiculated organelle that forms a branching network reaching through each intercellular bridge of dividing germ cell cysts and is involved in oocyte specification.35 After completion of the fourth mitotic division, the fusome starts degrading and the intercellular bridges acquire a less electron dense inner rim consisting of actin filaments.32,36,37 Well after the fourth division, ring canals start expanding in size and grow from a diameter of about 0.5–1.5 µm in the germarium to approximately 10 µm at late stages of oogenesis.12,28 Thus, it is only after completion of cleavage furrow arrest and the fourth mitotic division that ring canal maturation begins.28 Therefore, incomplete cytokinesis and ring canal growth are two separate processes, both temporally and spatially, and the recent detailed mapping of these processes will allow detailed deciphering of the incomplete cytokinesis process in the Drosophila female germline by genetic means.28

Cleavage furrow arrest, ring canal assembly and growth occur by sequential recruitment and removal of a number of identified components, the timing of which can be estimated from the defined regions of the germarium (regions 1, 2a, 2b and 3) (Fig. 1A). Region 1 contains germline stem cells and mitotically active 2-, 4- and 8-cell cysts (Fig. 1A). In this region, the cleavage furrows contain contractile ring actin, the actomyosin-binding protein Anillin, the centralspindlin component Pav-Klp, the glycoprotein mucin-D and the multi-adaptor protein Cindr.12,14,38–40 In late region 1, after the third mitotic division, phosphotyrosine epitopes appear in the cleavage furrows.34 Region 2a contains the 16-cell clusters, and here, upon cleavage furrow arrest, contractile ring actin and Cindr disappear.12,14 At this point Anillin, Pav-Klp, phosphotyrosine epitopes, and possibly further components, build up the outer rim of the nascent ring canal.12 After completion of all four mitotic divisions, the inner rim of the maturing ring canals recruit non-contractile actin filaments, the adducin-like protein Ovhts-RC and the filamin Cheerio.12,36 Phosphotyrosine epitopes, mainly originating from phosphorylation by the Src64 and Tec29 kinases, also become detected in the inner rim at this point.12 In region 2b the cysts flatten, become lens-shaped and surrounded by follicle cells. Region 3 contains spherical stage 1 egg chambers that will leave the germarium and enter the vitellarium for maturation. At this step, Anillin disappears and the actin-bundling protein Kelch begins to localize.12,38 Throughout the continued ring canal growth, Ovhts-RC, Cheerio and Kelch play key roles in promoting bundling of the actin filaments at the inner rim of the ring canal.12 Thus, mature ring canals in the female germline are built of both components of the cleavage furrow (Pav-Klp) and further structural components added after cleavage furrow arrest (e.g., F-actin, Ovhts-RC, Cheerio, Kelch).

Male germline intercellular bridges.

Spermatogenesis continues throughout life in the male germline and depends on the continuous production of germline stem cells.41,42 Male germline stem cells divide asymmetrically to give rise to another stem cell and a spermatogonial cell that undergoes mitotic proliferation to form primary spermatocytes. Each spermatocyte in turn undergoes two rounds of meiosis to produce four sperm. During both the mitotic and meiotic cell divisions, cytokinesis is incomplete, giving rise to syncytia of cells interconnected by stable intercellular bridges (Table 1). Interestingly, the mitotic and meiotic cell divisions occur synchronously. In mammals, germ cells form syncytia of hundreds of germ cells interconnected by intercellular bridges and in Drosophila melanogaster each cyst consists of 64 haploid spermatids.42–44

Functionally, it has been proposed that male germline intercellular bridges promote germ cell communication and sharing of cytoplasmic constituents, thereby synchronizing e.g., mitotic cell divisions and entry into meiosis (Table 1).10,45–48 Studies of the mouse male germline have shown that the cytoplasmic bridges may facilitate the sharing of gene products between post-meiotic haploid spermatids so that genetically distinct spermatids remain phenotypically diploid.47,48 Interestingly, transport of mRNA and organelles has been observed between haploid spermatids.48–50 Specifically, the chromatoid body, an organelle controlling RNA processing in haploid spermatids has been shown to move across male germline intercellular bridges in a microtubule-dependent manner.50 Even though a deeper understanding of their physiological roles is still required,46 the importance of male germ cell intercellular bridges is emphasized by the fact that they are required for fertility of insects and mammals.19,20 Female germline intercellular bridges, on the other hand, are required for fertility in Drosophila, but not in mice.12,26

The molecular composition of male germline intercellular bridges has been most thoroughly studied in Drosophila and mouse. Drosophila male germline intercellular bridges contain phosphotyrosine epitopes, Anillin, Pav-Klp and at least three septins (Peanut, Septin 1 and Septin 2).13 Male ring canals do not, on the other hand, contain a thick inner rim of filamentous actin, nor Ovhts-RC or Kelch, major components of female ring canals, nor do they grow in size, but rather remain at a diameter of 1–1.5 µm.13,51 Interestingly, in contrast to the Drosophila female germline, the fusome persists and develops further even after the cease of the mitotic divisions in the male germ cell clusters.13,35

In the mouse male germline, Anillin and several septins (at least 2, 7 and 9) are transiently present in the forming intercellular bridges, but are eventually absent from the mature bridges.52 The mature intercellular bridges contain the testis-expressed gene 14 (TEX14, that shows similarity to Drosophila Src64 and Tec29 in its kinase domain), the two centralspindlin components MKLP1 (Pav-Klp in Drosophila) and MgcRacGAP, the centrosomal proteins CEP55 and pericentrin, as well as e.g., actin, plectin and δ-tubulin.52–54 Thus, Drosophila and mouse male germline intercellular bridges, as well as female and male ring canals in Drosophila are, at least partially, compositionally different, suggesting that there may be a variation in the ways they form.

Somatic Intercellular Bridges

Somatic stable intercellular bridges have been described in Caenorhabditis elegans, hydra and insects (Table 1).14,45,55–59 Several insects have apically localized stable intercellular bridges in the ovarian follicle cell epithelium (Fig. 2A and Table 1).14,56–60 In Drosophila, somatic stable intercellular bridges have also been observed in the larval imaginal disc epithelia (Fig. 2B and Table 1)14,40,58 and in the larval brain.14,40 Cytoplasmic bridges of a more transient nature are found in the cellular blastoderm and extended germ band embryo.11

To date, the functions of somatic intercellular bridges remain unknown. It is believed, by analogy to the functions of germline intercellular bridges, that they promote exchange of cytoplasm and maybe even organelles, thereby facilitating intercellular communication, synchronization of cell division or differentiation, and coordination of cell behavior.11,14,40,56–58 Studies of the Drosophila follicle cell epithelium by injections of a fluorescent dye into a single follicle cell indeed revealed the presence of follicle cell clusters, in line with the cytoplasmic continuity observed by electron microscopy.57 Many such independent clusters build up the follicle cell epithelium. Interestingly, these clusters do not show a uniform morphology, sometimes occurring in linear arrays or with asymmetric branching, suggesting that cell divisions by incomplete cytokinesis in the follicle cell epithelium are not as tightly synchronized as in the germline.57,60 It will be essential to elucidate the roles of somatic stable intercellular bridges during development and the identity of the signals that may be shared within somatic cell clusters, as well as whether stable somatic intercellular bridges exist in mammals.

Only a handful of proteins have yet been identified as constituents of somatic stable intercellular bridges in Drosophila. These include Anillin, Pav-Klp, Cindr and mucin-D, all of which are found in intercellular bridges in the egg chamber follicle cell and larval imaginal disc epithelia, as well as in the larval brain (Table 1).14,35,39,40

Ultrastructure of Intercellular Bridges

Female and male germline.

Detailed analyses of the ultrastructure of the ring canals have been performed only in the female germline in Drosophila melanogaster.32,61 The ring canals consist of two rims: An outer rim that appears electron dense in electron micrographs and an inner thicker rim that is less electron dense61 (similar in the olive fruit fly Dacus oleae; Fig. 1E–G). Horizontal thin sections through the ring canal and decoration with subfragment 1 of myosin reveal that the inner rim consists of a layer of circumferentially oriented actin filaments attached to the plasma membrane at the periphery of each canal. The actin filaments in the ring are of mixed polarities, similar to those found in the contractile ring formed during cytokinesis. In vertical sections through the canal the actin filaments appear as dense dots.61 As the ring canal diameter increases during oogenesis (from about 0.5 µm at stage 2 to about 10 µm at late stages), the thickness of the ring canal wall increases (from 0.06 µm at stage 2 to 0.37 µm at stage 9) and the ring canal length, too (from 0.22 µm at stage 2 to 1.87 µm at stage 9). At developmental stage 2 there are 82 actin filaments in the ring, by stage 6 there are 717 and by stage 10 there are 726. Taking into account the diameter, this indicates that there are 170 µm of actin filaments/canal at stage 2, 14,000 µm at stage 9 and approximately 23,000 µm at stage 11 or one inch of actin filament.61 The density of actin filaments remains unchanged throughout development. Studies in mutant flies that lack the ring canal component Kelch revealed that actin filaments form normally but individual bundles that comprise the fibers of the net are not bound tightly together. This results in bundles to enter into the ring canal lumen, partially blocking the canal.61

In vertebrates, the ultrastructure of ring canals in the female germline has been described in frog, chicken, rabbit, mouse, rat, hamster and human ovaries (Table 1).21–23,62–64 Mammalian female ring canals consist mainly of one rim that appears electron dense in electron micrographs. A second less electron dense inner rim, as evident in Drosophila, is not clearly seen in electron micrographs of mammalian female germline ring canals. Their diameter is on average 0.5 to 1 µm.21,23

The ultrastructure of stable intercellular bridges in the male germline has not yet been described in as much detail as for the bridges found in female germline. However, Drosophila male germline intercellular bridges consist mainly of one rim that appears electron dense in electron micrographs and they have a diameter of 1–1.5 µm (Table 1).51 In mammals, the ultrastructural morphology of male ring canals has been described in mouse, rat, guinea pig, Chinese hamster, squirrel, rabbit, cat, goat, monkey and human testis.9,19,44,45 Interestingly, they also consist mainly of one electron dense rim and have a diameter of µ1–1.5 m.9,19,44,45

Somatic intercellular bridges.

Most information about the detailed morphology of somatic stable intercellular bridges comes from studies in insects and especially from Drosophila.14,56–60 Somatic stable intercellular bridges found in the ovarian follicle cell epithelia and the epithelial cells of the larval imaginal discs all have a stable diameter of about 0.25 to 1 µm and they do not grow during development (Fig. 2).14,57,58 In the follicular epithelium of Drosophila ovarian follicles, the bridges have a cylinder shape and an open lumen filled with cytoplasm that is continuous with both cells and they have an average luminal diameter of 0.25–0.3 µm and a length of 0.40 µm, which is smaller than that of ring canals in the female germline (Fig. 2A). Their wall consists of an outer and an inner rim. The outer rim is more electrondense compared to the inner rim that appears filamentous.57 By decoration with the S1 fragment of myosin, these filaments have been confirmed as actin filaments.57 Stable intercellular bridges observed in the epithelial cells of the Drosophila imaginal discs exhibit a very similar ultrastructural morphology (Fig. 2B).14,58

Molecular Mechanisms of Incomplete Cytokinesis

By which molecular mechanisms does the cytokinesis process halt during incomplete cytokinesis as opposed to conventional cytokinesis that proceeds to abscission? In the next paragraphs we will give an overview of specific features of stable intercellular bridges and some recent advances in the field that have shed light on the molecular control of the incomplete cytokinesis process.

Specific morphological features of stable intercellular bridges.

As evident from the ultrastructural studies of both germline and somatic intercellular bridges, the formation of a stable intercellular bridge is accompanied by the appearance of an electron dense rim just beneath the plasma membrane.14,19,32,44,57,61 One possibility is that this rim protects the plasma membrane of the intercellular bridge from membrane remodeling and vesicle fusion events that normally promote cytokinesis completion and abscission, thereby stabilizing the bridge.2,8 Interestingly, stable intercellular bridges indeed have smooth bridge walls, as opposed to the transient intercellular bridges during conventional cytokinesis which display membrane foldings and protrusions.14,57 The protrusions indicate active membrane remodeling prior to abscission of these transient intercellular bridges.14,65

Another characteristic of mature stable intercellular bridges is the absence of mitotic microtubule spindle bundles in their lumen.14,44,52,57,66 The mitotic microtubule spindle is required for cytokinesis progression during conventional cytokinesis.2 Electron microscopy studies of mammalian male germ cell intercellular bridges show that germ cell cytokinesis initially appears identical to conventional cytokinesis, but eventually the microtubule spindle and midbody break down.44,66 Similarly, stable somatic intercellular bridges in the Drosophila follicle cell and larval imaginal disc epithelia, as well as in the larval brain, do not contain visible microtubule bundles in their bridge lumen.14,56–58 Interestingly, actively dividing cells in the same tissues, on the other hand, such as larval neuroblasts that will eventually undergo abscission, display very prominent microtubule spindles.2,14 Moreover, ovarian follicle cells that undergo complete cytokinesis until developmental stage 6 of oogenesis display mitotic spindles in their lumen.57 In contrast, after stage 6, when stable intercellular bridges have formed, and until the end of oogenesis, microtubule bundles are not any more detected within their lumen.14,57 Of note is that microtubules may still be involved in intercellular transport, as evident in the female and male germline in insects and mammals.12,50 A deeper understanding of the formation, composition and functions of the electron dense bridge rim, the mechanism by which the mitotic spindle and midbody break down and their possible cross-talk may give important clues to the control of the incomplete cytokinesis process.

Halting constriction of the contractile ring.

Evidence for arrest of the actomyosin ring constriction as a mechanism of incomplete cytokinesis has recently been described in the Drosophila female germline. An essential motor for contractile ring constriction is non-muscle myosin II (hereafter referred to as myosin II). Phosphorylation of the regulatory myosin II light chains activates myosin II and thereby contractile ring constriction, whereas their dephosphorylation by myosin light chain phosphatase (MLCP) negatively regulates myosin II-mediated contraction.67 Recent studies of the incomplete cytokinesis process in the germarium of the Drosophila female germline have shown that inactivation of MLCP phosphatase activity by mutation of its myosin-binding subunit MYPT leads to contractile ring over-constriction during incomplete cytokinesis in the germarium and thereby formation of small ring canals.68,69 These data suggest that dephosphorylation of myosin II is a prerequisite for contractile ring arrest and thereby incomplete cytokinesis in the Drosophila female germline. Whether similar mechanisms apply during the formation of other stable intercellular bridge types remains to be determined.

Stabilization of the intercellular bridge.

Clearly, stabilization of the intercellular bridge contributes to the maintenance of incomplete cytokinesis. In the Drosophila female germline, one molecular explanation of the cleavage furrow arrest was described above.69 Subsequently, during ring canal assembly and growth, their structural burden is carried largely by the thick actin bundles, Ov-htsRC and Kelch at the inner rim of the bridge.12,61 In the Drosophila male germline ring canals, on the other hand, thick actin bundles are not evident.13,57 Here, members of the septin family may contribute to stabilizing the intercellular bridge.13 Septins are components of the stable bud neck in yeast and form filamentous oligomeric ring-like arrays beneath the plasma mem- brane.70,71 Septins are missing from ring canals in the Drosophila female germline, but have been proposed to be components of somatic intercellular bridges.11,72 Moreover, somatic intercellular bridges have a monolayer of actin filaments at their inner rim that by analogy to female germline intercellular bridges may contribute to their structural stability.14,57 Interestingly, during conventional cytokinesis, actin disassembly is a prerequisite for abscission,4 and in human cells with chromosome bridges, actin has been shown to accumulate to stabilize intercellular bridges in order to delay abscission until chromosome segregation has been completed.73

Other components that may play a role in intercellular bridge stabilization include Anillin, Cindr and MKLP1/Pav-Klp. Anillin, a stable component of both Drosophila male germline and somatic stable intercellular bridges,13,14,35,38,74 acts at several steps of the cytokinesis process,75 and has been proposed to be involved in intercellular bridge stabilization during conventional cell divisions and incomplete cytokinesis.14,76,77 Drosophila Anillin was recently found to interact with Cindr which is a component of cleavage furrows during female germline cell divisions and of stable somatic intercellular bridges.14 Together, Anillin and Cindr may play a role in stabilizing actin filaments at the inner rim of somatic intercellular bridges.14 Moreover, the fact that MKLP1/Pav-Klp represents a conserved component of all stable intercellular bridge types with known molecular composition (Table 1)14,26,39,52,78 may suggest that it plays a role in stabilizing them. Interestingly, MKLP1/Pav-Klp is required for the formation of the electron dense midbody matrix during cytokinesis, and electron microscopy studies implicate that there may be a relationship between the midbody matrix and the electron dense rim of stable intercellular bridges.65,79 It will be exciting to learn more about the mode of action of these proteins, other intercellular bridge components, and their effectors during incomplete cytokinesis.

Inhibition of abscission.

A further mechanism that may contribute to promoting incomplete cytokinesis is controlled inhibition of the abscission process. Even though much still needs to be understood about the detailed molecular requirements for cytokinetic abscission, some important aspects have been discovered.1,2,4,6 Targeted fusion of vesicles derived from the Golgi, endosomes or recycling endosomes to the intercellular bridge, and membrane docking factors, are required.4,8,80–82 Centrosomal proteins, such as centriolin, CEP55, FYVE-CENT and TTC19, also participate in cytokinesis completion.80,83–85 Centriolin and CEP55 recruit factors required for abscission, including SNAREs and components of the exocyst complex (recruited by centriolin),80 the endosomal sorting complex required for transport (ESCRT) (recruited by CEP55) and ALIX (recruited by CEP55).83,84,86 Specifically, CEP55 has been shown to recruit the ESCRT-I component TSG101, and ALIX, to the midbody via interactions between its hinge region and GPPX3Y motifs present in TSG101 and ALIX, to promote recruitment of ESCRT-III components and thereby abscission.83,84 It can be speculated that compartmentalized absence of active abscission factors, e.g., by developmentally regulated expression, their inactivation or by inhibition of their recruitment, may contribute to incomplete cytokinesis. Interestingly, such a case was recently described in the mouse male germline. These recent studies have shown that the intercellular bridge component TEX14 also contains a GPPX3Y motif, with which CEP55 strongly interacts, thereby inhibiting the interactions of CEP55 with TSG101 and ALIX.53 This suggests that TEX14 prevents the completion of cytokinesis, including the localization of ALIX, by altering the fate of CEP55, from a midbody organizer that recruits proteins for abscission, to a major component of the stable intercellular bridge.53 Thus, when TEX14 is bound to CEP55, TSG101 and ALIX are not recruited to the midbody to complete cytokinesis.53 In this way, TEX14 plays an important role in promoting incomplete cytokinesis in the mouse male germline.

Hence, incomplete cytokinesis may be accomplished by mechanisms acting at multiple steps, including arrest of the actomyosin ring, structural stabilization of the intercellular bridge, disassembly of the microtubule spindle, inhibition of membrane fusion events, as well as of abscission. Given some morphological and compositional differences between female germline, male germline and somatic stable intercellular bridges, it will be interesting to learn what may be the conserved and specific features of the incomplete cytokinesis process when comparing tissue types and species.

Concluding Remarks and Perspectives

Since the first ultrastructural studies of stable intercellular bridges, we now have a greater understanding of their molecular composition and some insight into their functions and the molecular control of complete versus incomplete cytokinesis. However, questions still remain as to the physiological and developmental roles of stable intercellular bridges, to the identity of factors and signals that are exchanged, to how they can coordinate cell behavior in a syncytium, as well as to the temporal, spatial and molecular control of the incomplete cytokinesis process. Future challenges in the field include answering these questions. The available knowledge and scientific interest in the topic as well as the rapid development of advanced molecular genetics and imaging techniques ensure that the understanding of these issues will grow at a fast pace. A deeper understanding of the incomplete cytokinesis process may also give further insights into the control of abscission initiation during conventional cytokinesis.

Abbreviations

ESCRT	=	endosomal sorting complex required for transport
TEX14	=	testis-expressed gene 14
Drosophila	=	Drosophila melanogaster

Figures and Tables

Figure 1 Stable intercellular bridges in the Drosophila melanogaster female germline. (A) Schematic overview of Drosophila oogenesis, the germarium and incomplete cytokinesis of germ cells in the germarium. Cleavage furrows and ring canals are schematically shown in black. Details are described in the text. (B) Confocal micrograph of an early stage 9 wild-type Drosophila melanogaster egg chamber stained with phalloidin-FITC (fluorescein isothiocyanate) (green) and propidium iodide (red) for visualizing filamentous actin and nuclei, respectively. Scale bar: 35 µm. (C) Confocal micrograph of a late stage 9 wild-type Drosophila melanogaster egg chamber stained with phalloidin-FITC for visualizing filamentous actin. The image is a projection of a z-stack of 18 serial sections of 2 µm thickness each. Scale bar: 15 µm. (D) Light micrograph of a 1 µm thick resin section stained with toluidine blue. It illustrates the anterior pole of a wild-type stage 11 Drosophila melanogaster egg chamber. Bipolar arrows: Ring canals, (F) follicle cell, (N) nurse cell. Scale bar: 10 µm. (E–G) Transmission electron micrographs of a stage 6 wild-type olive fruit fly Dacus oleae egg chamber. (E) illustrates a ring canal in longitudinal section. Arrows point to the ring canal edges. (F and G) are higher magnifications of the ring canal edges shown in (E). Arrows in (F and G) point to ring canal rims. Scale bars: (E) 1 µm; (F and G) 0.5 µm.

Figure 2 Stable intercellular bridges in somatic Drosophila tissues. (A) Stable intercellular bridges in the Drosophila melanogaster follicle cell epithelium. Left and middle: Cross section through a stage 8 egg chamber (left) and a superficial section through the follicle cell epithelium of the same egg chamber (middle) showing stable intercellular bridges (arrowheads) between cells detected with antibodies against the bridge component Cindr (green). Rhodamine-phalloidin was used to visualize F-actin (red). Scale bars: 10 µm (left) and 5 µm (middle). Right: Electron micrograph showing a longitudinal section through a stable intercellular bridge of a stage 8 egg chamber. The electron dense outer rim is evident (arrowhead) and a less electron dense inner rim (arrow) is also seen. Scale bar: 250 nm. (B) Stable intercellular bridges in the Drosophila melanogaster wing disc epithelium. Left: Stable intercellular bridges (arrowheads) in a wing disc epithelium are labeled with antibodies against the bridge component Cindr. Scale bar: 20 µm. Middle: Stable intercellular bridges (arrowheads) containing Cindr (green) are present at cell borders labeled by Dlg (red). Nuclear DNA was labeled with Hoechst (pseudo-colored blue). Scale bar: 5 µm. Right: Electron micrograph showing a longitudinal section through an intercellular bridge in the wing disc epithelium. The electron dense outer rim (arrowhead) and the less electron dense inner rim (arrow) are evident. Scale bar: 250 nm.

Table 1 Stable intercellular bridges in the germline and in somatic tissues in diverse metazoans

Organism	Bridge type and tissue	Known intercellular bridge components	Diameter and morphology	Proposed intercellular bridge functions	References
Female germline
Mammals (Human, rat, hamster, rabbit, see mouse below)	Ring canals interconnecting germ cells in cysts	N.D.	0.5–1 µm, one electron dense rim	Synchronization of mitotic germ cell divisions and entry into meiosis, germ cell degeneration and intercellular transport of organelles	22–25, 87, 88
Mouse (Mus musculus)	Ring canals interconnecting germ cells in cysts	TEX14, MKLP1	0.5–1 µm, one electron dense rim	Synchronization of mitotic germ cell divisions and entry into meiosis, germ cell degeneration and intercellular transport of organelles, not required for fertility	21, 26, 89
Chicken	Ring canals interconnecting germ cells in cysts	N.D.	0.5–1 µm, one electron dense rim	Synchronization of cell cycles and intercellular communication	63
Froglet (Xenopus laevis)	Ring canals interconnecting germ cells in cysts	Actin, Ov-htsRC, Kelch	0.5–1 µm, one electron dense rim	Likely directional transport of nutrients to the developing oocyte	64
Fruit fly (Drosophila melanogaster)	Ring canals interconnecting germ cells in cysts	Mature RC: Actin, Ovhts-RC, Kelch, Cheerio, Pav-Klp, Tec29, mucin-D. Transient: Cindr, Anillin.	0.5–10 µm, two rims, an outer electron dense and an inner less electron dense	Directional transport of cytoplasm, nutrients, organelles, proteins and mRNA from the nurse cells to the developing oocyte through the ring canals, required for fertility	12, 14, 34, 38–40, 61, 90–92
Other insects (Honey bee, mosquito, aphid)	Ring canals interconnecting germ cells in cysts	Probably actin	∼2–4 µm, one electron dense rim	Directional transport of nutrients to the developing oocyte, oocyte determination	59, 60, 93, 94
Male germline
Mammals (Human, goat, cat, rabbit, rat, hamster, squirrel, see mouse below)	Ring canals interconnecting germ cells in cysts	Actin (shown in rat and squirrel)	1–1.5 µm, one electron dense rim	Synchronization of development and entry into meiosis, sharing of haploid gene products	9, 10, 44, 45, 50, 66, 95–98
Mouse (Mus musculus)	Ring canals interconnecting germ cells in cysts	Actin, TEX14, MKLP1, CEP55, Pericentrin, plectin, δ-tubulin	1–1.5 µm, one electron dense rim	Synchronization of development and entry into meiosis, sharing of haploid gene products, required for fertility	19, 47, 48, 52–54
Fruit fly (Drosophila melanogaster)	Ring canals interconnecting germ cells in cysts	Anillin, septins, Pav-Klp, pTyr, mucin-D	1–1.5 µm, one electron dense rim	Synchronization of development and entry into meiosis, required for fertility	13, 20, 40, 51, 74, 78
Somatic intercellular bridges
Nematode (Caenorhabditis elegans)	Bridges called cell division remnants	Actin	∼1 µm, N.D.	Regulation of embryonic cell divisions	55
Fresh water polyp (Hydra oligactis)	Bridges between clusters of cnidioblasts	N.D.	∼0.5 µm, one electron dense rim	Synchronization of cnidoblast differentiation into nematocysts	45
Fruit fly (Drosophila melanogaster)	Bridges between epithelial follicle cells	Anillin, Pav-Klp, Cindr, mucin-D, Actin	0.2–0.5 µm, two rims, an outer electron dense and an inner less electron dense	Intercellular communication to coordinate cell behavior	14, 35, 39, 40, 56, 57
Fruit fly (Drosophila melanogaster)	Bridges between cells in larval imaginal disc epithelia	Anillin, Pav-Klp, Cindr, mucin-D	0.2–0.5 µm, two rims, an outer electron dense and an inner less electron dense	Intercellular communication to coordinate cell behavior	14, 40, 58
Fruit fly (Drosophila melanogaster)	Bridges between cells in the larval brain	Anillin, Pav-Klp, Cindr, mucin-D	0.2–0.5 µm, two rims, an outer electron dense and an inner less electron dense	None	14, 40
Other insects (Honey bee, mosquito, stable fly)	Bridges between epithelial follicle cells	N.D.	0.2–0.5 µm, two rims, an outer electron dense and an inner less electron dense	Intercellular communication to coordinate cell behavior	59, 60, 99

Known stable intercellular bridge components, ultrastructural morphology and proposed functions are listed.

Acknowledgements

K.H. acknowledges support from the International Human Frontier Science Program Organization and The Research Council of Norway. I.P.N. acknowledges support from the European Research Council and H.S. from the Norwegian Cancer Society.