Distribution changes of non‐self‐test cells and self‐tunic cells surrounding the outer body during Ciona metamorphosis

Background: Ascidians significantly change their body structure through metamorphosis, but the spatio‐temporal cell dynamics in the early metamorphosis stage has not been clarified. A natural Ciona embryo is surrounded by maternally derived non‐self‐test cells before metamorphosis. However, after metamorphosis, the juvenile is surrounded by self‐tunic cells derived from mesenchymal cell lineages. Both test cells and tunic cells are thought to be changed their distributions during metamorphosis, but the precise timing is unknown.


| INTRODUCTION
Depending on the developmental stage, the outer body of a solitary ascidian is surrounded by specific cells: test cells in the larval stage and tunic cells in the juvenile stage (Figure 1). 1 Test cells are maternally derived cells that differentiate from primary follicle cells during oogenesis. 2,35][6][7] Various roles of test cells have been indicated in many studies.Test cells cover the larval tunic with aggregates of secreted granules called "ornaments." 8,9 The cells secrete material that is involved in larval tunic formation at the early tailbud stage. 5Removal of test cells at the early tailbud stage leads to a poorly formed tunic and a lack of a caudal fin. 10 Test cells migrate on the surface of the larval tunic and mostly move slowly in a straight line, but some move quickly in a circular path. 11It has been suggested that test cells may also be involved in cytokine-mediated immune responses.An in situ hybridization assay revealed that the gene for tumor necrosis factor alpha (TNFα), a kind of cytokine, is expressed in test cells. 12n the other hand, tunic cells are distributed within the tunic of juvenile and adult ascidians.The tunic is mainly composed of cellulose outside the body after hatching. 11,13There are several types of tunic cells inside the adult tunic, such as granulocytes, morula-like cells, amoebocytes, pigmentary cells, bladder cells, luminocytes, and phagocytes. 14Although these cells are not common to all ascidians, most of them possess phagocytes.In some species, phagocytic activity has been demonstrated experimentally through the engulfment of labeled microparticles. 15,16In Ciona species, tunic cells differentiate from mesenchymal cells that are derived from A7.6, B8.5, and B7.7 blastomeres. 17est cells and tunic cells have some common features.For example, both are located outside the body and are motile.However, they are distinct cells; test cells are maternally derived non-self-cell, 2,4 and tunic cells are self-cells derived from mesenchymal lineages. 17,18It has been suggested that the distribution of these cells changes during metamorphosis because there are no test cells observed in juveniles, and they are surrounded by tunic cells. 1 However, the exact timing of the changes during development remains unknown.
In our previous study, we developed an artificial metamorphosis induction system by applying mechanical stimulation to the larval palp of a Ciona species. 19We found that two-round Ca 2+ transients occurred in the palp just after the mechanical cue, followed by the wavelike Ca 2+ propagation in the epidermis (epithelial conduction), and backward movement of the trunk epidermis before tail regression.In this study, we investigated that the distribution of the test cells and the tunic cells changes in the outer body by using this system, a transgenic line, and timelapse imaging.

| Cells were released through the epidermis during metamorphosis
To investigate when and how tunic cells emerge in the outer body, we first induced metamorphosis by mechanical stimulation in swimming larvae, from which the test cells had first been removed by dechorionation.Interestingly, we found that single motile cells were released from multiple points in the epidermis of the trunk after 11 min 44 s ± 2 min 50 s (N = 10) after mechanical stimulation (Figure 2A and Movie S1).These cells changed shape with pseudopodia.The release of the cells from the epidermis resembles the extravasation of vertebrate blood cells across the epithelia, so we named them "extravasated cells." We expected the extravasated cells to be tunic-cell precursors.To verify whether extravasated cells were located inside the tunic, the tunic layer was stained with fluorochrome-conjugated wheat germ agglutinin (WGA). 20In WGA-treated swimming larvae, extravasated cells released during metamorphosis were retained inside the WGA-labeled tunic layer (Figure 2B).These results indicate that the extravasated cells are tunic cell precursors.
F I G U R E 2 Timelapse imaging of Ciona trunk region during metamorphosis induction.A, Differential interference contrast (DIC) view of a larva subjected to artificial induction of tail regression by mechanical stimulation.Time after the mechanical stimulation is indicated in each panel.The start of the stimulation was set as time t = 0 min.Each right image is an enlarged view of the white dotted box in the left box.At 11 min, an extravasation of one cell was observed (white arrowhead), and the cell completely extravasated from the epidermal layer at 17 min.Scale bar = 20 μm.B, Merged DIC images and staining of the tunic layer with fluorochromeconjugated wheat germ agglutinin (WGA) in a larva after cell extravasation.Extravasated cells (white arrowheads) located outside of the larval epidermis (dotted lines) and inside the tunic layer (within the WGA-stained region).Scale bar = 50 μm.A, anterior; L, left; P, posterior; R, right.

| Cell extravasation occurred simultaneously with the beginning of the tail regression
We combined an artificial metamorphosis-induction system with Ca 2+ imaging (Movie S2) to measure the timing of the cell extravasation and other metamorphic events observed at the start of mechanical stimulation to the pulp (Figure 3A).Two-round Ca 2+ transients in the adhesive palp were observed after mechanical stimulation, as reported previously. 19After that, a wave-like Ca 2+ propagation (epithelial conduction) was observed on the F I G U R E 3 Imaging of Ca 2+ and cell extravasation of H2B tagged GCaMP6s-expressed larva at the beginning metamorphosis.A, Images of Ca 2+ dynamics at 20 s and 3 min 18 s after metamorphosis induction.Ca 2+ transient observed in palp (A left, 20 s after mechanical stimulation).Ca 2+ propagation in epidermis (epithelial conduction) occurred during 2-6 min (Epi in A right, 3 min 18 s after mechanical stimulation).Scale bar = 50 μm.B, Enlarged images of each white dotted box in (A).Cell extravasation (white arrowhead in upper pictures) and backward movement of posterior trunk epidermis (white dotted line in lower pictures) were shown at 9 min after stimulation.Scale bar = 25 μm.C, Time course of the Ca 2+ transients in palp, epithelial conduction, backward movement of trunk epidermis, and cell extravasation after metamorphosis induction.Relative fluorescence intensity in palp (red) and trunk epidermis (blue).The regions of interest (ROIs) were set as the whole region of GCaMP6s in the palp and the epidermis, respectively.In the palp, two-round Ca 2+ transients, phase I and phase II, were observed after 0 to 1.5 min and 1.5 to 7.5 min, respectively.Epithelial conduction was finished before backward movement (black triangle) and cell extravasation (shaded box).A, anterior; V, ventral; P, posterior; D, dorsal.epidermis as previously reported, 19 and then the posterior trunk epidermis moved backward just before tail regression (Figure 3B).To investigate the precise timing of the cell extravasation, we compared it with the timing of these events (N = 5/5) (Figure 3C and Figure S1).
After mechanical stimulation, two-round Ca 2+ transients in the palp and epithelial conduction occurred within 10 min (Figure 3C).After this ended, the cell extravasation was observed simultaneously with the backward movement of the epidermis in the posterior trunk, as shown in Figure 3B.The figure shows that the posterior trunk shape is elongated in the posterior direction by the backward movement of the epidermis.These observations indicated that the cell extravasation was coupled with this backward movement (Figure 3C arrowhead).

| Extravasated cells originate from mesenchymal cells
Tunic cells are derived from mesenchymal cell lineages. 17o confirm that extravasated cells were derived from the mesenchymal cell lineage, mesenchymal-lineage cells were labeled by AKR::Kaede and tracked by confocal microscopy.Since the aldo-keto reductase (AKR) gene was specifically expressed in mesenchymal cells, 21 we used AKR::Kaede to specifically label mesenchymal cells. 22In the swimming larvae electroporated with AKR::Kaede, mesenchymal cells in both the anterior and posterior trunk were labeled with Kaede.A part of the posterior mesenchymal cells changed color under UV irradiation to red (Figure 4A). 23,24ime-lapse imaging of swimming larvae before the induction of metamorphosis revealed that Kaede-labeled mesenchymal cells migrated inside the trunk (Movie S3 and Figure 4B), which supports the results of a previous study. 17Most mesenchymal cells migrated locally within the anterior or posterior regions of the trunk, but some cells migrated far from anterior to posterior or from posterior to anterior beneath the epidermis (Figure 4B,C).Immediately after UV irradiation, only red Kaede-labeled mesenchymal cells were present in the posterior trunk region, but green cells were observed in the same region 2 h after UV irradiation (Figure 4C).This result indicated that dynamic and long-distance migration of mesenchymal cells occurred within the larval body of Ciona before metamorphosis.
Before induction of metamorphosis, there were no Kaede-labeled cells outside the dechorionated larval body.After 12 min of the mechanical stimulation, Kaedelabeled mesenchymal cells extravasated through the epidermis (Movie S4 and Figure 4D, arrowheads).Extravasated mesenchymal cells were derived from green-labeled cells in the anterior region of the trunk and red-labeled cells in the posterior region.These results indicated that the extravasated cells could be identified as a mesenchymal lineage and these mesenchymal cells acquire an ability to extravasate at the time of metamorphosis.Regardless of their location, they can extravasate at multiple points in the anterior or posterior region of the trunk.

| Extravasated cells divided fast and slow group
Next, we quantitatively investigated the difference of the motility of mesenchymal cells before and after cell extravasation.Three-dimensional timelapse imaging of mesenchymal labeled swimming larvae was performed by light-sheet microscopy to investigate the motility of mesenchymal cells (Movie S5).The velocity of mesenchymal cells migrating inside the trunk and the extravasated cells after metamorphosis were compared.The velocities of most inner migrating cells were 4.0 μm/min or less.The average velocity of migrating cells was 2.51 ± 0.65 μm/min (Figure 5B,C).
Extravasated cells were divided into fast migrating and slow cell groups depending on the velocity (Figure 5A).From the cell number distribution, extravasated cells with a velocity of 4.2 μm/min or higher were grouped as fast cells.The average velocities of fast and slow extravasated cells were 8.79 ± 2.53 μm/min and 1.88 ± 0.78 μm/min, respectively (Figure 5C).Fast cells were about 3.5 times faster than migrating cells, while slow cells were slower.Kaede-labeled extravasated cells did not drop out and migrated around within the tunic.In 2-day-old juveniles, on average 98% of outside cells were Kaede-labeled tunic cells and no test cells were recognized outside the body (N = 5, Table S1, and Figure 6B).The tunic cells in juveniles showed at least two different types: amebocyte-like cells with extended pseudopodia and bladder-cell-like cells with vacuoles (Figure 6E). 14These results indicated that the elimination of test cells started after the tail regression and was completed by at least the juvenile period.

| DISCUSSION
In this study, we used imaging methods to investigate the precise timings of the emergence of tunic cells and elimination of test cells during metamorphosis.Figure 7 shows a summary of this study.Both non-self-test cells and selftunic cells migrated, and their distributions changed during early metamorphosis (Figure 7).The mesenchymal cells of the tunic cell lineage migrate inside the trunk region before metamorphosis (Figure 7A).Although very few mesenchymal cells are observed outside of the body of the late tailbud embryo (data not shown), most mesenchymal cells extravasate at the same time as the backward movement of posterior-trunk epidermal cells after the second Ca 2+ transient (Figure 7B).
Several studies have characterized the molecules and genes that are involved in the reception of mechanical stimulus 25 and in the induction of metamorphic events. 26,27][30][31][32] However, the relationships between these molecules, Ca 2+ transients, and extravasation of mesenchymal cells need to be elucidated in future studies.
The tunic cell precursors are located in the tunic layer after cell extravasation (Figure 1B).Both test cells and mesenchymal cells exist outside of the body for a short time until the elimination of the test cells at the time when the outer tunic layer is removed in early metamorphosis (Figure 6C).Some test cells are known to extend pseudopodia toward the tunic, which they invade in the trunk and tail regions. 11Pseudopodia from a test cell approach the surface of the epidermis.However, none of the test cells remained outside the body in juveniles.
Cell extravasation occurs simultaneously with backward movement of the epidermis after epithelial conduction (Figure 3).Such synchronization of the timing between the mesenchyme and epidermis suggests that cell extravasation may require cooperation of these tissues.There are two possibilities for the mechanism of cell extravasation with penetration into epidermal cells.
One possibility is that extravasating cells enter vacuoles within epidermal cells and are released into the tunic along with the vacuoles (transcellular migration).Electron microscopy imaging suggested that the blood cells enter the base of an epidermal cell, become incorporated within a vacuole of an epidermal cell, and finally enter the tunic after the vacuolar membrane fuses with the apical plasmalemma in Amaroucium constellatum. 33The vacuoles in epidermal cells are reported to disappear before the termination of the tail regression in Ciona. 34,35he role of vacuoles in the epidermis is still unclear, and it may involve promoting cell extravasation.It was observed that the epidermal cells are elongated in shape by the backwards movement.The morphological change of the epidermal cells may facilitate the release of intracellular vacuoles and extravasating cells.
The other possibility is that extravasated cells penetrate between the epidermis (paracellular migration).Some cells extended pseudopodia during cell extravasation.This can help pass through the tunic as well as between the epidermis.The epidermis may move backwards to widen the gap, making cell s easier.It would be interesting to investigate the way of cell extravasation.
In vertebrates, when passing through the endothelial layer in the diapedesis process, leukocytes can either F I G U R E 7 Summary of the order of each metamorphic event occurring after mechanical stimulation.A, Before mechanical stimulation, the outside of the larval body was surrounded by foreign test cells (white cells), and mesenchymal cells (green cells) migrated inside the body.Two-round Ca 2+ transients in the palp and epithelial conduction in the epidermis occurred within 12 min of mechanical stimulation.B, About 12 min after mechanical stimulation, backward movement of the posterior trunk epidermis and extravasation of mesenchymal cells started simultaneously.The extravasated mesenchymal cells differentiate into self-tunic cells in the tunic layer of the outside body.C, After completion of tail regression, test cells are eliminated from the anterior palp region, and tunic cells remain in the tunic.D, All cells outside the body were self-tunic cells in juveniles.
follow a paracellular route or a transcellular one (Mamdouh et al., 2009).In addition, transcellular and paracellular migration is used in cancer cell invasion. 36ome of the genes that are important for these two transmigration processes are shared by Ciona. 37Our findings shed light on the evolution of the mechanisms of the invasion in leukocytes and cancer cells.
The migration velocity of some of the extravasated mesenchymal cells increased after the metamorphosis (Figure 5).In vertebrate, immune cell migration can be classified into two modes: mesenchymal migration and faster amoeboid migration. 38Since immune-related genes are known to be upregulated during metamorphosis, 39 a part of mesenchymal cells are differentiated into immune cell and may be capable of amoeboid migration in response to the two round Ca 2+ transient followed by metamorphosis signal.
In addition, thyroid hormone receptor is highly expressed during metamorphosis, 40 and thyroid hormone initiates metamorphosis and accelerates its different stages. 31,41Thyroid hormone activates immune cells in vertebrates. 42These results support our hypothesis.Differences in migration velocities of extravasated cells may be due to differences in cell types.The tunic cells in juveniles included amebocyte-like cells that extended pseudopodia, as well as bladder-cell-like cells with vacuoles (Figure 6E).The tunic cells of Ciona adults comprise at least four types. 14,17The exact timing of differentiation of tunic cells is unknown, and it is possible that they have already differentiated into several types at the time of cell extravasation.
The body of swimming larvae is covered with test cells, and adults are covered with tunic cells.It has been reported that the gene for TNFα, an inflammatory cytokine, is expressed in test cells. 12TNF is known to be an important factor in innate immunity in vertebrates, 43 so test cells may be responsible for the defense of swimming larvae.On the other hand, some tunic cells are phagocytic and are distributed throughout the tunic. 16These cells are thought to contribute to defense by existing outside the body. 44esenchymal cells extravasated before tail regression, and test cells are removed after tail regression (Figures 1  and 6).So, there are always cells outside the body during tail regression.In addition, test cells are non-self-cells and need to be discarded during development.Cell extravasation occurs before tail regression, allowing replacement of protective cells without loss.Since extravasated mesenchymal cells and test cells are in different layers of outside the body (Figures 2 and 6), extravasated mesenchymal cells are unlikely to directly expel test cells.However, it is possible that mesenchymal cells indirectly eliminate test cells by secreting a kind of repellent substance or digesting the larval tunic.The removal mechanism of test cells needs to be studied in the future.
Our findings of cell extravasation during metamorphosis could help in understanding the metamorphosis of Ciona and also provide a potential model experimental system to elucidate the mechanisms of cell invasion and cell migration, such as those of cancer or immune cells.

| Electroporation
Epi1::H2B-GCaMP6s and AKR::Kaede were introduced into fertilized eggs by electroporation; 40 μL of each plasmid construct at 500 ng/μL or more and was combined with 360 μL of 0.77 M mannitol in 10% seawater.At 30 min after fertilization, eggs in 400 μL of this solution were placed in a cuvette for electroporation.After electroporation, eggs were washed with seawater and incubated until observation.

| Fixation of swimming larvae
A PLL hydrobromide solution (final concentration = 0.02 mg/mL) was placed in a 12-mm glass base dish until it covered the bottom of the glass.The petri dish was incubated overnight or longer.Immediately before observation, the PLL solution was removed from the glass base dish.The dish was washed twice with seawater.A single larva was oriented with water flow by a glass pipette and placed on a petri dish.The lateral side of the trunk adhered to the bottom, but the palp did not contact the bottom surface.

| Artificial mechanical stimulation
The palp of swimming larvae fixed the glass base dish were mechanically stimulated by the artificial mechanical stimulation system using a manipulator (NARISHIGE MM-89) and an injection holder set (IM-H1).A glass needle was prepared by pulling a NARISHIGE GC-1 glass tube with a Sutter P-97/IVF puller.The needle tip was heated with a lighter until it became spherical.This needle was used as a mechanical stimulator.

| Microscopy
Swimming larvae were observed with confocal fluorescence microscopy, fluorescence microscopy (FM) with a 3CCD camera and light-sheet microscopy (LSM).For GCaMP6s, jRGECO1α and Kaede imaging, we followed previous methods. 50The time interval was set as 3 to 5 s per frame.An Olympus Â20, Â40, or Â60 oil immersion lens was used.For FM and LSM, we followed previous methods. 19For LSM, the time interval was set as 30 s per frame.Each image was downscaled to 800 Â 1024 pixels for registration.The diSPIMFusion plugin for ImageJ was used for registration and fusion of multiple views of the same sample. 51

| Cell lineage tracking
The photoconversion of Kaede was performed by bleaching a SIM scanner by manual operation.The Kaede signal of mesenchymal cells of swimming larvae (21 hpf) was photoconverted by irradiation with a 405-nm laser for 5 s.Z-stack images of the juveniles of the Tg [MiCiAKRK]1 transgenic line 22 were acquired using an Olympus FV 1000 microscope.

| Cell tracking analysis
Swimming larvae injected with mRNA EKAREV-NLS 52 were observed by LSM.EKAREV-NLS labeled all types of cells but strongly labeled mesenchymal cells in particular.The migration trajectories of mesenchymal cells were analyzed using IMARIS software.In the fluorescence image, cells were distinguished inside and outside the body based on cell density.The following parameter values were used for the analysis: XY diameter = 2.0 μm, maximum distance = 3.35 μm, and maximum gap size = 1.In cells inside the body, those with track lengths of 10.5 μm or more were defined as migrating cells.The migration velocity was determined as the average migration velocity over time.Tracking data yielded the position of each cell over time.The directionality ratio was obtained by dividing the distance between two points for each time width by the track length.

F
I G U R E 1 Threedimensional reconstructed images and stereomicroscopic images of Ciona larvae and juveniles.Embryos were stained by alexa phalloidin 488 for the 3D images.A, Test cells (white arrow) surrounded swimming larvae at 21 h postfertilization (hpf).In the left image, the trunk region of the swimming larvae were sucked and fixed with a glass needle.B, Tunic cells (white arrowhead) surrounding juveniles at 45 hpf.Those on the left are fixed embryos, and those on the right are living (white box; enlarged view of each white dotted box, inset).Scale bar = 50 μm.
Timelapse imaging of a swimming larva electropolated with AKR::Kaede.Mesenchymal cells were labeled with Kaede.A, 405-nm laser was directed at the regions of interest (ROI) (white circle), and only the posterior mesenchymal cells changed from fluorescent green (before) to red (after).B, C, Timelapse images of mesenchymal cell migration in 21-hpf swimming larva.White box is an enlarged view of each white dotted box.The positions of one mesenchymal cell at 0 min are marked as white dotted circles showing cell migration beneath the epidermis.D, Larva electroporated with AKR::Kaede before (above) and after (below) metamorphosis induction (MI).Kaede-labeled mesenchymal cells extravasated 20 min after MI (arrowheads).White dotted boxes in the left images are enlarged in the right.White dotted line on the right indicated the outer epithelial surface of the larva.Scale bar = 50 μm.A, anterior; V, ventral; P, posterior; D, dorsal.
Next, we investigated the timing of the elimination of test cells.Timelapse imaging of AKR::Kaede transgenic swimming larvae (without dechorionation) was performed.Before metamorphosis, unlabeled test cells surrounded the outer body of the swimming larvae (Figure6A, arrows).Kaede-labeled mesenchymal cells were in the anterior and posterior regions within the trunk (Figure6A, green color).During tail regression, Kaede-F I G U R E 5 Quantitative analysis of the cell motility in extravasating cells and inner migrating cells.Histogram for extravasating cells migration velocity (A) and inner migrating cells (B).C, Cell migration velocity.n = 92 cells (fast extravasating cells), 90 cells (slow extravasating cells), and 42 cells (migrating cells).Among 92 fast cells and 90 slow cells, some cells that moved out and re-entered to the field of view might have double counted during the 1-h recording.The actual average number of extravasated cells per one flame was 34 ± 6.On the other hand, juvenile with fully extravasated cells (2 days post fertilization) had an average of 185 ± 29 cells even after cell proliferation of tunic cells (N = 5; data not shown).Therefore, at least 18.6% of extravasated cells outside the larval body were analyzed.Mann-Whitney U test; **P < 0.0001.
labeled mesenchymal cells extravasated (data not shown).In addition, unlabeled test cells and Kaedelabeled extravasated cells coexisted outside the body (Figure6C,D; arrows and arrowheads).After the tail regression, the test cells began to drop out from the vicinity of the palp or wiped out posteriorly, and finally, most of the test cells removed within several hours (Movie S6).

F I G U R E 6
Changes in localization of test cells and tunic cells during metamorphosis.A, B, Images of a swimming larva at 23 hpf (A) and juvenile at 44 hpf (B) from Tg[MiCiAKRK]1 transgenic line.Mesenchymal cells are labeled with Kaede.The white dotted box in each upper image is enlarged in the lower image.White dotted line in the lower image indicates the outline of the body epitheria.A, Unlabeled tunic cells (white arrows) surrounded the tunic on the body.B, Kaede-labeled mesenchymal cells were shown in the tunic on the outside of the body after tail regression (white arrowheads).C, Time-lapse images of larvae about 1 h after tail regression is complete.The elimination of the test cells started from the anterior trunk (arrows).D, Enlarged images of the white dotted boxes in (C).At 0 min (above), test cells (white arrows) and Kaede-labeled tunic cells (white arrowhead) coexisted outside the body.At 4 h 40 min (below), most of the test cells are eliminated, and only Kaede-labeled tunic cells remain inside the tunic.Scale bar = 50 μm.A, anterior; V, ventral; P, posterior; D, dorsal.E, Images of 2 types of tunic cells in the yellow dotted box in (B) (left: like amebocyte, right: like bladder cell).Scale bar = 5 μm.

4. 10 |
Three-dimensional reconstructed images Ciona intestinalis type A were obtained from Misaki Marine Biological Station (University of Tokyo) and Maizuru Fisheries Research Station (Kyoto University) through the National Marine Bioresource Project (NBRP) in Japan.Eggs and sperm were collected by dissecting gonoducts and were mixed in seawater in petri dishes to obtain fertilized eggs.Fertilized eggs were incubated at 20 C until fixation and observation.As Ciona larvae acquire competence of tail regression after stage 31, 1 Stage 30 larvae were used after 26 h postfertilization (hpf).