BIOLOGICAL OVERVIEW

breathless is involved in two distinctly different processes during the development of the fly. First it is used in the development of the trachea, and later it plays a role in migration of specific midline glia and the resultant effect on axonogenesis. These are similar developmental processes to those affected by Drifter, a POU homeodomain transcription factor which could be downstream of breathless.

The ligand is Branchless, and its developmentally regulated distribution is responsible for the spatially restricted activation of Breathless. Function of the extracellular domain of Breathless can be assessed by substituting a Torso sequence for that of the FGF-receptor sequence. Such a hybrid molecule corrects tracheal defects in breathless mutants, suggesting that the Breathless receptor does not receive spatial clues from a spatially restricted FGF ligand (Reichman-Fried, 1994).

Breathless shares a downstream effector pathway with Torso, Sevenless and the EGF-R/Torpedo receptor. Thus the all purpose RAS-RAF pathway is used for each of these receptors. How are specific downstream targets regulated when the signaling pathway is redundant? Tracheal cells specifically express many markers independently of breathless (Krüppel, Caudal and cut for example). These cells form a completely different chemical milieu when compared with retinal cells, follicle cells and terminal cells, all of which have receptors that use the RAS-RAF pathway.

Some cell functions are common to both tracheal cells and midline glia. Cell migration is regulated by Breathless in both types of cells. Use of a heat inducible dominant negative Breathless protein, lacking a functional tyrosine kinase, reveals that breathless is required at the initiation of tracheal cell migration. Distinct subsets of tracheal cells designated as terminal cells extend long processes toward target tissues. Blind-ended tubes called tracheoles are formed connecting to the main tracheal branches. Tracheoles insure the proper supply of air to growing larval tissues. Late induction of the dominant negative form of Breathless blocks tracheole formation. Thus breathless is implicated in a third developmental process, the formation of tracheoles (Reichman-Fried, 1995).

The patterned branching in the Drosophila tracheal system is triggered by Branchless (Bnl), which activates Breathless and the MAP kinase pathway. A single fusion cell at the tip of each fusion branch expresses the zinc-finger gene escargot, leads branch migration in a stereotypical pattern and contacts with another fusion cell to mediate fusion of the branches. A high level of MAP kinase activation is also limited to the tips of the branches. Restriction of such cell specialization events to the tip is essential for tracheal tubulogenesis. Notch signaling plays crucial roles in the singling out process of the fusion cell. Notch is activated in tracheal cells by Branchless signaling through stimulation of Delta (Dl) expression at the tips of tracheal branches and activated Notch represses the fate of the fusion cell. In addition, Notch is required to restrict activation of MAP kinase to the tips of the branches, in part through the negative regulation of Branchless expression. Notch-mediated lateral inhibition in sending and receiving cells is thus essential to restrict the inductive influence of Branchless on the tracheal tubulogenesis (Ikeya, 1999).

Six primary branches form in the tracheal primordia, among which the dorsal branch (DB), anterior and posterior dorsal trunk (DTa, DTp), and anterior and posterior lateral trunk (LTa, LTp) migrate along a stereotyped path to be connected with other branches from adjacent primordia. These fusion branches are capped with fusion cells that express Esg. The remaining visceral branch (VB) migrates to reach the internal organs. Terminal cells expressing Drosophila serum response factor (DSRF) are formed in each primary branch except in DTs and later differentiate multiple tracheoles. High expression of DL mRNA and protein is expressed in the DT of stage-15 embryos. Cells in the tracheal primordium just after invagination expressed Dl uniformly. At early stage 11, Dl expression started to be elevated in 2-3 cells at the tip of the branches in which outgrowth had begun and the number of the Dl-expressing cells was reduced to one at late stage 11. At stage 14, high Dl expression remains only in the DT, which has completed fusion. Ser protein also accumulates at the apical side of the DT cells at the same stage. In the case of the trachea, the level of N protein expression remains uniform, suggesting that the expression level of Dl, or its potentiation, must be crucial for N activation. An esg-lacZ reporter is initially expressed in 2-3 cells at the tip of the fusion branches in mid stage 11 embryos, and is downregulated to be maintained in only a single cell at the tip of each fusion branch at late stage 11. These cells also express a high level of Dl. Therefore, localized elevation of Dl expression in stage 11 correlates well with the selection process of a single fusion competent cell (Ikeya, 1999).

When tracheal cells became unresponsive to Bnl due to btl mutation, no sign of primary branching and Dl upregulation is observed. On the contrary, when Btl is hyperactivated by overexpression of Bnl in all the tracheal cells, primary branching is severely inhibited and Dl expression is elevated. These results suggest that elevation of Dl expression is triggered by the external signal Bnl. The results also suggest that the N/Dl pathway may mediate the Bnl signal to control cell migration and cell fate decisions (Ikeya, 1999).

In N^ts1 embryos grown under non-permissive condiditons a misrouting defect was observed in DB, which normally elongates to the dorsal midline where it meets its counterpart from the other side of the metamere. DBs are often curved in the anteroposterior direction and make contact with the tip of DB from the same side. The misrouted DBs accumulate a luminal component detectable by 2A12 antibody at the ectopic contact sites, but do not appear to fuse properly. Cell migration defect is also observed in DB. DB consists of a total of 5-7 cells in the case of Tr5, of which two specialized cells are located at the tip. One is the terminal cell, from which a thin terminal branch sprouts: the other is the fusion cell. The remaining stalk cells are located between the tip and DT at regular intervals. In N^ts1 mutants, the number of cells at the tip is increased with a corresponding decrease in the number of stalk cells, the latter having become unusually elongated. The total number of cell nuclei do not change compared to the controls, so no additional mitosis occurs. This cell migration phenotype suggests that stalk cells and terminal cells acquire a property of fusion cells to become localized at the tip of DB. Since Esg represses DSRF expression and terminal branching, the loss of terminal branching in N^ts1 embryos may be the consequence of ectopic Esg expression. Similar defects in Esg and DSRF expression are also observed in null N mutants. These results suggest that in N minus embryos, several terminal and stalk cells are recruited to the fate of fusion cells (Ikeya, 1999).

A three-step model is presented for tubule formation in fusion branches. During induction, exogenously supplied Bnl activates its receptor Btl in equivalent tracheal cells where N is inactive. The signal is transduced by activation of MAPK to stimulate Dl expression. The expression of Bnl is also regulated negatively by N signaling. During lateral inhibition, induced Dl activates N in neighboring cells, which in turn, represses esg transcription. N may also repress Dl expression. However, a high level of Dl inhibits N signaling in a cell autonomous manner, allowing activation of esg and MAPK. A small difference in the response to Bnl within equivalent tracheal cells is amplified to select out a single fusion cell with a high level of Dl and esg expression. During tubule formation, the fusion cell becomes the only cell that responds to Bnl and becomes motile. Maintenance of Btl activity by Bnl would limit the migration toward the source of Bnl. Other tracheal cells follow fusion cells to become stalk cells (Ikeya, 1999).

Mitotic cell rounding accelerates epithelial invagination

Mitotic cells assume a spherical shape by increasing their surface tension and osmotic pressure by extensively reorganizing their interphase actin cytoskeleton into a cortical meshwork and their microtubules into the mitotic spindle. Mitotic entry is known to interfere with tissue morphogenetic events that require cell-shape changes controlled by the interphase cytoskeleton, such as apical constriction. However, this study shows that mitosis plays an active role in the epithelial invagination of the Drosophila tracheal placode. Invagination begins with a slow phase under the control of epidermal growth factor receptor (EGFR) signalling; in this process, the central apically constricted cells, which are surrounded by intercalating cells, form a shallow pit. This slow phase is followed by a fast phase, in which the pit is rapidly depressed, accompanied by mitotic entry, which leads to the internalization of all the cells in the placode. It was found that mitotic cell rounding, but not cell division, of the central cells in the placode is required to accelerate invagination, in conjunction with EGFR-induced myosin II contractility in the surrounding cells. It is proposed that mitotic cell rounding causes the epithelium to buckle under pressure and acts as a switch for morphogenetic transition at the appropriate time (Kondo, 2013).

The invagination of epithelial placodes converts flat sheets into the three-dimensional structures that form complex organs, and it is a key morphogenetic process in animal development. A major mechanism of invagination is apical constriction, which is driven by actomyosin contraction. However, not all constricted cells invaginate, and some cell internalization occurs without apical constriction, suggesting that additional mechanisms of inward cell movement contribute to invagination (Kondo, 2013).

To obtain three-dimensional information about cell behaviour during invagination, live imaging was performed of the Drosophila tracheal placode. Ten pairs of tracheal placodes, each of which is composed of about 40 cells, are formed in the ectoderm at mid-embryogenesis, and each placode initiates invagination simultaneously. Using an adherens junction marker, DE-cadherin-green fluorescent protein (E-cad-GFP), it was found that the adherens junctions of the central placode cells slowly created a depression by apical constriction, which became the tracheal pit. After 30 to 60 min of slow movement (slow phase), the tracheal pit was suddenly enlarged, and the tracheal cells were rapidly internalized (fast phase) and eventually formed L-shaped tube structures (Kondo, 2013).

After the fast transition, all the tracheal cells and surrounding epidermal cells entered mitosis 16, the final round of embryonic mitosis. It was noticed that the fast invagination was always associated with the mitotic entry of central cells that were frequently the first to enter mitosis 16. Intriguingly, mitotic rounding of the central constricted cells occurred simultaneously with the rapid depression of their apices, followed by chromosome condensation 10 min later. In this study, this atypical mitotic rounding associated with apical depression in an internalized cell is called 'rounding', to distinguish it from canonical surface mitosis (surface cell rounding) (Kondo, 2013).

To determine whether cell rounding is required for invagination, zygotic mutants were examined of the cell-cycle gene Cyclin A (CycA), which fail to enter mitosis 16, and double parked^a3 (dup^a3), which show a prolonged S phase 16 and delayed entry into mitosis 16. Tracheal invagination was initiated normally in the CycA and dup^a3 mutants, but proceeded more slowly than in controls, indicating that entry into mitosis 16 is required for proper timing of the fast phase (Kondo, 2013).

Although delayed, the accelerated invagination in the CycA or dup^a3 mutants eventually occurred, allowing the formation of tube structures and suggesting that additional mechanisms are involved. After invagination, fibroblast growth factor (FGF) signalling is activated in the tracheal cells to induce branching morphogenesis through chemotaxis. To examine the contribution of FGF signalling to invagination, mutants of the FGF ligand branchless (bnl) or the FGF receptor breathless (btl) were analyzed. These mutants invaginated normally, indicating that chemoattraction to FGF is dispensable for invagination (Kondo, 2013).

Next, to assess FGF's role in the mitosis-defective condition, double mutants were analyzed for CycA and bnl or CycA and btl, and it was found that they showed slower invagination than CycA single mutants. Furthermore, the invagination in these double mutants was incomplete, in that the cells failed to form L-shaped tubular structures. Therefore, FGF signalling is critical for invagination when mitosis is blocked, serving a back-up role. Tracheal-specific CycA expression rescued the defects in invagination speed and tube structure in the CycA btl mutants. In addition, mitosis of cells outside the pit was occasionally observed that occurred before the mitosis of the central apically constricted cells and was not correlated with the fast invagination phase. Thus, mitosis of the surrounding epidermal cells is dispensable for tracheal invagination. Taken together, it is concluded that mitotic entry of central cells is a major mechanism for accelerating tracheal invagination (Kondo, 2013).

To distinguish the role of cell rounding from that of cell division in the fast phase, the microtubule inhibitor colchicine was used to arrest the cell cycle after cell rounding. Colchicine treatment after mitosis 15 induced M-phase arrest at mitosis 16, but the fast invagination movement accompanied by cell rounding was not affected. This result indicates that cell rounding, but not cell division, is responsible for the acceleration phase of the tracheal invagination (Kondo, 2013).

Mitosis of cells in the columnar epithelium normally occurs at the apical surface after surface rounding. It was next asked how the apical surface of the central cells becomes depressed during internalized cell rounding. One possible model explains internalized cell rounding as cell-autonomously controlled by the association of the cells with the basement membrane or underlying mesodermal cells. However, genetic removal of basement-membrane adhesion by the maternal and zygotic mutation of βPS-integrin (also known as mys) did not compromise the speed of invagination, and snail-twist double-mutant embryos, which lack mesodermal cells, still showed tracheal invagination with internalized cell rounding. These results suggest that anchoring to the basal side is probably not required (Kondo, 2013).

A second model proposes that the apical depression of the rounding cells is driven by local planar interactions among the tracheal cells. Before and during tracheal invagination, myosin II is enriched at the cell boundaries tangential to the centre of the placode and regulates cell intercalation. It was noted that the myosin II level in the central cells was lower than in the surrounding, intercalating cells. Nevertheless, the apices of the central cells were constricted during the slow phase, strongly suggesting that the surrounding cells exerted centripetal pressure on the central cells through myosin II cables. Myosin II cables fail to form in EGFR signalling mutants (such as rho, the rhomboid endopeptidase required for EGF ligand maturation, and Egfr), and apical constriction is impaired in these mutants. The first few cells undergoing mitosis 16 in the tracheal placode of rho or Egfr mutants showed surface cell rounding with expanded apices, indicating that EGFR signalling is required to couple the mitotic cell rounding with fast apical depression. It is speculated that the columnar shape of the central cells resists centripetal movements, resulting in the accumulation of inward pressure during the slow phase. The existence of such resistance was supported by the results of a physical perturbation experiment using a pulsed ultraviolet lase. The cell rounding associated with mitotic entry would release the stored inward pressure by means of cytoskeletal remodelling that causes rapid depression of apical surface together with the active shortening of cell height, leading to rapid buckling of the apical surface and the fast phase of invagination (Kondo, 2013).

Even with the loss of both EGFR and FGF signalling, the tracheal placodes form moderately invaginated structures, compared to the flat tracheal placode observed in the rho-bnl-CycA triple mutant at the same stage, indicating that cells needed to undergo mitosis 16 to induce invagination, independent of EGFR and FGF signalling. In rho bnl double mutants, although the cells undergoing the earliest mitoses showed surface cell rounding, some of the subsequent mitotic events were coupled to apical depression and internalized cell rounding. Unlike the earlier mitotic events on the surface, the internalized rounding cells in the rho bnl embryos showed constricted apices and were surrounded by apically rounded cells before mitosis. Internalized rounding with a constricted apical surface were shared properties of cells in mitoses leading to invagination, in both control and rho bnl embryos. It is suggested that the first few cells undergoing surface cell rounding compress the adjacent interphase cells and restrict their apical area, so that they are forced to move internally after rounding, causing the epithelial layer to buckle and invaginate (Kondo, 2013).

Although invagination was largely blocked in the rho-bnl-CycA triple mutants, any double mutant combination permitted invagination to some degree, indicating that three qualitatively distinct mechanisms, mitotic cell rounding, myosin II contractility (EGFR) and active cell motility (FGFR), can independently trigger invagination. In the normal context of wild-type development the combination of cell rounding and EGFR signalling may optimize the timing and speed of invagination, and then invaginated tracheal sacs subsequently respond to FGF emanating from several target tissues guiding branching morphogenesis (Kondo, 2013).

These observations demonstrates a new role for mitosis in tissue morphogenesis to generate mechanical force through cell rounding, independent of cell division. This is distinct from previously described invagination mechanisms involving cell-autonomous constriction by the apical activation of actomyosin contractility, which is incompatible with mitosis. Mitosis 16 outside the tracheal placode occurs in clusters on the ectoderm surface, but does not lead to invagination, suggesting that the tracheal placode is sensitized to invaginate upon mitosis, independent of EGFR and FGFR signalling. Future research to uncover the properties of the tracheal placode that enables it to respond to clustered mitosis will explain not only this new mode of morphogenesis, but also the homeostasis mechanisms of epithelial architecture (Kondo, 2013).

A single-cell survey of Drosophila blood

Drosophila blood cells, called hemocytes, are classified into plasmatocytes, crystal cells, and lamellocytes based on the expression of a few marker genes and cell morphologies, which are inadequate to classify the complete hemocyte repertoire. This study used single-cell RNA sequencing (scRNA-seq) to map hemocytes across different inflammatory conditions in larvae. Plasmatocytes were resolved into different states based on the expression of genes involved in cell cycle, antimicrobial response, and metabolism, together with the identification of intermediate states. Further, rare subsets within crystal cells and lamellocytes were discovered that express fibroblast growth factor (FGF) ligand branchless and receptor breathless, respectively. These FGF components were identified as required for mediating effective immune responses against parasitoid wasp eggs, highlighting a novel role for FGF signaling in inter-hemocyte crosstalk. This scRNA-seq analysis reveals the diversity of hemocytes and provides a rich resource of gene expression profiles for a systems-level understanding of their functions (Tattikota, 2020).

Previous studies have identified three major Drosophila blood cell types essential for combating infections in this species. This study used scRNA-seq of larval fly blood to gain deeper insights into the different cell types and their transition states in circulation during normal and inflammatory conditions. Comprehensive scRNA-seq data provide information on subpopulations of plasmatocytes and their immune-activated states. Importantly, scRNA-seq could precisely distinguish mature crystal cells and lamellocytes from their respective intermediate states, which are less well understood and for which marker genes were not previously available. Thus, new marker genes identified in this study should facilitate further study of these states. Moreover, it was possible to identify the gene signature of self-renewing plasmatocytes and suggest their role as extra-lymph gland oligopotent precursors. In addition to the identification of various states of mature cell types, this study also suggests novel roles for a number of genes and pathways in blood cell biology. In particular, a putative new Mtk-like anti-microbial peptide (AMP) was identified, and a role was proposed for the FGF signaling pathway in mediating key events leading to the melanization of wasp eggs. Finally, a user-friendly searchable online data mining resource was developed that allows users to query, visualize, and compare genes within the diverse hemocyte populations across conditions (Tattikota, 2020).

Blood cell types are dynamic in nature and several transient intermediate states exist in a continuum during the course of their maturation in several species. This scRNA-seq analysis provides a framework to distinguish cell types from their various states including oligopotent, transient intermediate and activated states (Tattikota, 2020).

Oligopotent state: ScRNA-seq analysis identified PM2 as the oligopotent state of plasmatocytes based on the enrichment of several cell cycle genes including polo and stg. This signature suggests that PM2 corresponds to self-renewing plasmatocytes located in the circulatory and sessile compartments of the Drosophila hematopoietic system where plasmatocytes are the only dividing cells identified. Further, previous studies suggested that lamellocytes derived from embryonic-lineage hemocytes are readily detectable in circulation prior to their release from the lymph gland, and that terminally differentiated crystal cells can also derive from preexisting plasmatocytes in the sessile hub. Hence, it is proposed that PM2 corresponds to the oligopotent state that not only drives expansion of plasmatocytes, but importantly can also give rise to crystal cells and lamellocytes. Monocle3 analysis indicates that cell cycle genes decrease over pseudotime and there is ample evidence in support of the notion that cell cycle arrest may be required for terminal differentiation of various cell types in flies and vertebrates. Our in vivo data also indicates that cell cycle arrest can lead to the generation of terminally differentiated lamellocytes. Interestingly, recent evidence in hemocytes suggests that perturbing cell cycle by knocking down jumu, which is upstream of polo, can also lead to the generation of lamellocytes by activating Toll. In contrast, forced expression of certain oncogenes such as activated Ras and Hopscotch/JAK in hemocytes can also lead to overproduction of plasmatocytes and lamellocytes. It is, however, speculated that the proliferation and differentiation of hemocytes in these contexts may be linked to cell cycle. Thus, it is important to address this paradoxical role of cell cycle in the maintenance of oligopotency and transdifferentiation of plasmatocytes. Studies using lineage tracing methods such as G-TRACE or CRISPR-based in vivo cellular barcoding techniques may help further characterize the contribution of proliferating oligopotent plasmatocytes to blood cell lineages (Tattikota, 2020).

Immune-activated states: PM5 from the scRNA-seq data is enriched in several genes that encode glutathione S-transferase family of metabolic enzymes, which are known to catalyze the conjugation of reduced glutathione (GSH) to xenobiotics for their ultimate degradation. It has been demonstrated that a subset of hemocytes accumulate high GSH levels in Drosophila, in support of these data. Further, the two AMP clusters PM6-7 (PM^AMP) were identified as part of the immune-activated states of plasmatocytes. A recent study has demonstrated that AMPs are highly specific and act in synergy against various pathogens. The scRNA-seq analysis reveals the remarkable difference in the expression of a set of AMPs in the two clusters. Future studies with PM^AMP-specific perturbation of various AMPs identified within plasmatocytes should clarify their contribution in killing specific pathogens. Moreover, the role of Mtkl against pathogens needs further characterization. Pseudotime analysis showed that PM^AMP ends in the same lineage as lamellocytes suggesting a common mode of activation for these cell types and states. Interestingly, induction in hemocytes of Toll, which is upstream of Drs, can lead to the production of lamellocytes, suggesting that LM^int cells may act as the common branch point between immune-activated states and lamellocytes (Tattikota, 2020).

Transient intermediate states: In addition to the oligopotent and immune-activated states, plasmatocytes showed several subpopulations, which most likely are transient intermediate states. Although it remains to be seen whether they exist throughout the larval development, it is possible that these transient states exist along the continuum of cell maturation process. On the other hand, the transcriptomic composition of CC1 and LM1 clusters suggested the presence of intermediates for crystal cells and lamellocytes, respectively. Further analysis by Monocle3, which placed these clusters prior to their terminally differentiated cell types, confirmed the hypothesis that CC1 and LM1 correspond to CC^int and LM^int states, respectively. In the context of the CC lineage within the lymph gland, ultrastructural studies have revealed the presence of immature crystal cells, called procrystal cells, alongside mature crystal cells. This study furthered this observation by demonstrating in vivo that crystal cells exist in a continuum (PPO1^low to PPO1^high), validating the Monocle3 and scRNA-seq data. Moreover, clear gene signatures between the CC^int and LM^int states and their mature counterparts revealed that these intermediates most likely emerge from preexisting Hml⁺ plasmatocytes. With regards to the LM lineage, several groups have speculated that intermediates, called podocytes, or also lamelloblasts, may exist based on cell morphology and size. The scRNA-seq and Monocle3-based data clearly demarcate mature lamellocytes from LM^int at the transcriptomic level. In addition, sub-clustering analysis revealed that LM^int possessed a PM signature demonstrating that these intermediates are presumably derived from PM2 (Tattikota, 2020).

A novel role for the FGF signaling pathway in hemocyte crosstalk: In addition to the known hemocyte - tissue crosstalk, Drosophila hemocytes must act in a coordinated fashion to combat harmful pathogens and foreign entities such as wasp eggs. However, the signaling pathways that mediate the interactions among hemocytes and wound sites or wasp eggs have been unclear. The scRNA-seq uncovered a novel role for the FGF signaling pathway in controlling hemocyte differentiation and subsequent effects on the melanization of wasp eggs. The FGF ligand bnl and its receptor btl were among the genes identified in rare subsets of crystal cells and lamellocytes, respectively, highlighting the power of scRNA-seq in capturing and detecting these small populations of cells. Based on the in vivo data, it is proposed that Bnl⁺crystal cells interact with Btl⁺ lamellocytes to coordinate lamellocyte differentiation and possible migration towards parasitoid wasp eggs. Furthermore, because lamellocytes are also enriched in additional core components of the FGF signaling pathway, future studies involving a comprehensive analysis of this pathway will advance understanding of blood cell communication, differentiation, and migration in the context of immune response (Tattikota, 2020).

In summary, these scRNA-seq data provides a resource for a comprehensive systems-level understanding of Drosophila hemocytes across various inflammatory conditions (Tattikota, 2020).

GENE STRUCTURE

Bases in 5' UTR - 60

Exons - two

PROTEIN STRUCTURE

Amino Acids - 1061

Structural Domains

The extracellular region encodes five immunoglobulin-like domains, and an N-terminal signal sequence. The cytoplasmic kinase domain exhibits a high degree of similarity to the vertebrate FGF-Rs with the typical split kinase and comparably sized juxtamembrane and carboxy-terminal regions (Glazer, 1991 and Klambt, 1992).

date revised: 2 December 2018 

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.