Collective cell migration is a fundamental process for embryonic development (Szabó and Mayor, 2018) and its dysregulation during morphogenesis is a key contributor to many developmental disorders (Friedl and Gilmour, 2009). Recent studies have additionally placed collective cell migration at the heart of metastasis of certain types of cancer (Clark and Vignjevic, 2015; Wang et al., 2016). These physiological and pathological contexts are linked by the necessity for cells to migrate using unique strategies as they navigate through diverse 3D environments, while maintaining group cohesion and directionality.

In recent years, the Zebrafish Posterior Lateral Line primordium (PLLp) has emerged as a powerful model for studying a wide range of cellular and developmental processes, including cell-cell signaling, tissue patterning, and collective migration (Chitnis et al., 2012; Friedl and Gilmour, 2009; Ghysen and Dambly-Chaudière, 2007). This group of 100–150 cells is initially specified adjacent to the otic vesicle and migrates caudally down the length of the embryo over the course of ~24 hr (Gompel et al., 2001; Nogare et al., 2017). As this primordium migrates, cells in the trailing domain are progressively reorganized into apically constricted epithelial rosettes, each cradling a central sensory hair cell progenitor (Nechiporuk and Raible, 2008). The trailing-most cells in the PLLp eventually lose the ability to sustain collective migration and are deposited by the migrating primordium. Cells that had been incorporated into epithelial rosettes are deposited as neuromasts and will go on to develop into the mature sense organs of the lateral line, a sensory system that detects water flow about the animal (Coombs and Van Netten, 2005). Cells in the PLLp that are not incorporated into these epithelial rosettes are deposited as a continuous stream of so-called ‘interneuromast cells’, which lie between the sense organs (Grant et al., 2005; López-Schier and Hudspeth, 2005; Nogare et al., 2017).

Migration of the PLLp is guided by a stripe of chemokine, Cxcl12a, secreted by muscle pioneer cells lying along the horizontal myoseptum (David et al., 2002; Li et al., 2004). Directional interpretation of this uniform stripe is accomplished by the expression of two chemokine receptors in the PLLp, Cxcr4b, and Ackr3b (previously Cxcr7b). Cxcr4b binds Cxcl12a and engages the G-protein-coupled signaling transduction pathway to determine directed migration in response to a Cxcl12a. Expression of Ackr3b, an atypical chemokine receptor which can bind Cxcl12a with high affinity but cannot activate G-protein-coupled signaling to mediate migration is restricted to trailing cells in the primordium (Boldajipour et al., 2008; Burns et al., 2006; Naumann et al., 2010). It internalizes and degrades bound Cxcl12a generating a local gradient of Cxcl12a with lower levels toward the trailing end and higher levels toward the leading end of the primordium (Dambly-Chaudière et al., 2007; David et al., 2002; Haas and Gilmour, 2006; Lewellis et al., 2013; Wong et al., 2020). Although transcripts for cxcr4b are restricted to the leading two-thirds of the PLLp, the receptor itself is broadly expressed along the entire length of the primordium and it gives primordium cells the potential for directed migration in response to the self-generated Cxcl12a gradient (Donà et al., 2013; Venkiteswaran et al., 2013).

In cell transplantation experiments, basal cryptic lamellipodia are observed extending from PLLp cells in the direction of migration (Haas and Gilmour, 2006; Lecaudey et al., 2008), a common strategy for migrating epithelial cells (Farooqui and Fenteany, 2005). Crucially, these lamellipodia are observed extending from both leading cells, which have a more mesenchymal morphology, and from the basal feet of epithelial cells, which have a more typically epithelial morphology (Haas and Gilmour, 2006), suggesting that cells along the length of the PLLp actively contribute to migration. This is consistent with recent studies showing that chemokine signaling is necessary along the entire Cxcr4b-expressing domain to support effective collective migration (Colak-Champollion et al., 2019). In addition to chemokine signaling, Fibroblast growth factor (Fgf) signaling is also required for migration. The polarization of these basal migratory protrusions appears to be dependent on Fgf signaling in response to Fgfs produced in the leading part of the primordium. Their polarity is lost upon Fgf receptor inhibition, even when chemokine signaling is unperturbed, and this occurs concomitantly with a loss of migratory ability (Lecaudey et al., 2008). Furthermore, experiments with isolated PLLp fragments generated by laser ablation suggest that Fgf could act as a direct migratory cue (Dalle Nogare et al., 2014). These two systems, and potentially others, act together to govern collective migration of the PLLp.

Apart from the fact that underlying muscle pioneer cells are the source of chemokine signals that guide the primordium, the manner in which the PLLp interacts with surrounding tissue as it migrates and what influence surrounding tissue might have on migration and morphogenesis remains poorly understood. Aman et al showed that traversing underlying intersomitic boundaries does not influence the deposition of neuromasts, as the lateral line primordium does not deposit more closely spaced neuromasts in trilobyte mutants, which have more densely packed somites (Aman et al., 2011). Other studies have shown that the directionality of primordium migration does not rely on any extrinsic cues from the surrounding tissue and that its directional migration is an autonomous property of the primordium itself (Haas and Gilmour, 2006). However, the primordium has a dramatic effect on the tissue through which it migrates. The PLLp migrates along the horizontal myoseptum, between the underlying somites and overlying skin. As it migrates, the skin is displaced upwards and is separated from the underlying tissue by the passage of the PLLp, returning rapidly to its original apposition with the underlying somites after the passage of the PLLp.

In this study, we focus on flat superficial PLLp cells that lie above the deeper epithelialized cells that form protoneuromasts. We show that these cells extend directional migratory processes apposed to the overlying skin and that the directionality of these processes, like that of the basal cryptic lamellipodia, is dependent on Fgf signaling. Furthermore, we show that mechanically removing the skin prevents PLLp migration, which is subsequently recovered when the skin heals back over the PLLp. Loss of the overlying skin disrupts polarization of both superficial and basal migratory processes. Embedding skin-removed embryos in Matrigel partially restores both superficial and basal protrusions, but only the polarization of the basal protrusions are recovered, and the primordium remains unable to migrate. Taken together, these data suggest that the PLLp coordinates collective migration by extending lamellipodia both basally, against the underlying tissue, and superficially, against the skin, and that the presence of the overlying skin is essential for PLLp migration.

We imaged the morphology of individual cells in the PLLp by transplanting cells labeled with a membrane localized EGFP driven by the cldnb promoter (Tg(cldnb:lyn-egfp)) into Tg(cldnb:lyn-mscarlet) transgenic embryos. The resulting chimeric embryos contained mosaically labeled primordia, with a small percentage of donor cells expressing membrane-localized EGFP while the remaining host-derived cells expressed membrane-localized mScarlet (Figure 1A).

Figure 1 with 1 supplement see all

Download asset Open asset

In most cases, cells in the trailing domain had the characteristic shape of cells incorporated into an epithelial protoneuromast -- a basally positioned cell body and a tightly constricted apical domain that connects with the apical domains of other protoneuromast cells to form a microlumen (Durdu et al., 2014). In a smaller number of cases, cells in the leading mesenchymal-like domain were labeled. These cells had a flat morphology without any clear apical-basal polarity and extended numerous membranous protrusions against the underlying tissue. However, we also observed cells whose cell bodies lay above the level of the protoneuromast apical constrictions. In some cases, these cells were clearly connected to the apical constriction of a nearby protoneuromast, and they extended across the top of the PLLp and wrapped around the lateral sides (Figure 1B,C). In other cases, these superficial cells had no apparent direct connection to the protoneuromast apical constriction, often lying between protoneuromasts. During the course of a timelapse video, we observed these cells extending multiple broad, flat protrusions reminiscent of lamellipodia (Figure 1D–F, Figure 1—video 1). These protrusions were extended closely apposed to the basal surface of the skin, and their similarity to basal cryptic lamellipodia suggested that they might also contribute to the migration of the PLLp.

To further characterize this superficial population, we performed high-resolution imaging of the PLLp in Tg(cldnb:lyn-egfp);TgBAC(cxcr4b:h2a-mcherry) double transgenic fish, in which both the membranes and nuclei of the lateral line primordium are labeled, using both DiSPIM to generate image volumes with isotropic resolution (Figure 2A–C), and Airyscan super-resolution confocal microscopy (Figure 2D). While the cells that make up the core of the protoneuromast have a basally positioned nucleus, the nuclei of these overlying cells were above the level of the apical constrictions (arrowheads in Figure 2B,C), and directly underneath the skin (Figure 2D). Slicing this image along the long axis of the PLLp (red dashed line in Figure 2A) shows this population of cells occupying a position above the apical constrictions of the protoneuromast cells (Figure 2B). Likewise, slicing this image across the cross-section of the PLLp (blue dashed line in A) revealed a series of nuclei occupying a superficial and circumferential position around the apical side of the protoneuromast (Figure 2C).

Figure 2 with 4 supplements see all

Download asset Open asset

To determine the distribution of these cells along the anterior-posterior extent of the PLLp, we mapped the position of all cells whose nuclei were directly apposed to the overlying skin (red in Figure 2E,F), and which resided above the level of the closest protoneuromast apical constriction (blue in Figure 2E,F). We then compared their distribution with the distribution of all the cells of the PLLp (gray in Figure 2E,F) in both single primordia (for example see Figure 2E), or when PLLp were aligned based on the position of the most leading protoneuromast and the positions aggregated (Figure 2F). These data show that these cells lie above the fully formed protoneuromasts in the PLLp, but their frequency decreases toward the back of the PLLp, where the trailing-most protoneuromast is preparing to deposit. It should be noted that the apparent lack of these cells in the leading domain is a consequence of our selection criteria, where leading cells, typically flatter and lying closer to the underlying migratory substrate, are excluded. Segmentation and reconstruction of these cells showed that they formed a broad ‘sheath’ that covered the apical side of the PLLp protoneuromasts, forming a layer between the apical constrictions of the protoneuromast and the overlying skin cells. For illustration, Figure 2G,H and Figure 2—video 1 show a 3D rendering of a fully segmented PLLp with this population colored in magenta, and the remaining cells are colored in green.

We call these cells which lie in-between the skin and the protoneuromasts superficial primordium cells to reflect their superficial position in the primordium, and to distinguish them from the leading cells and the protoneuromast cells. These cells, like all cells in the PLLp, are labeled with the claudinb promoter, typically thought to reflect their epithelial lineage. However, these cells are clearly morphologically distinguishable from the cells that make up a protoneuromast as they are not obviously apico-basally polarized, and lack the clear apical constrictions of protoneuromast cells. Examples of leading, protoneuromast, and superficial cells can be seen in Figure 2—figure supplement 1.

From our segmentation, we can approximate the extent of the external surface contact area of the PLLp by cell type. We manually divided the PLLp into three cell types: leading cells, which occupy the very leading domain of the PLLp (13 cells), protoneuromast cells, cells which appeared apico-basally polarized and terminated in an apical constriction at the center of a protoneuromast (63 cells) and superficial cells, which contacted the overlying skin and were positioned superficial and lateral to the protoneuromast cells (45 cells). In this analysis, we ignored trailing cells which were depositing/deposited as interneuromast cells, as they are unlikely to contribute meaningfully to collective migration of the primordium.

In this segmented data set, we can quantify the external surface area of cells by measuring the number of surface pixels not in contact or close proximity to pixels belonging to other PLLp cells. In this reconstruction, of a total external surface area contact size of ~4400 square microns, approximately half (~2350 μm²) is made up of superficial primordium cells (exclusively making contact laterally and superficially) while protoneuromast cells (exclusively contacting the basal surface) make up only 1300 μm². Leading cells making up the remaining 750 μm² (Figure 2—figure supplement 2A). While almost all cells have some external contact, the superficial contacts are dominated by superficial cells, while the basal contacts are dominated by protoneuromast cells (Figure 2G–I, Figure 2—figure supplement 2A). In addition, while there are fewer superficial cells (43 vs 64), the average external contact area of superficial cells (54 μm²) is generally larger than the external surface area for protoneuromast cells (21 μm²), despite protoneuromast cells having more total surface area (Figure 2—figure supplement 2A,B). This is primarily due to the flatter morphology of superficial cells when compared to the elongated, apically constricted nature of protoneuromast cells (see Figure 2—figure supplement 1), where the basal ‘feet’ occupy a relatively smaller proportion of the cell surface area. This greater packing of the basal surfaces of the protoneuromast cells when compared to the extended, flat morphology of the superficial cells can be appreciated in renderings in Figure 2G,H, Figure 2—video 1, Figure 2—figure supplement 2A. These approximate proportions were similar both within a single embryo over time, and between multiple embryos analyzed (data not shown).

To assess the location of this population over time, we tracked the nuclei of all cells in a migrating primordium over the course of two hours. In this video, while there was significant movement of superficial nuclei around the periphery of the primordium (see Figure 2—video 2), cells in the superficial population maintained a more superficial and peripheral position, compared to the more centrally and basally located nuclei of the cells making up the protoneuromast, (Figure 2—figure supplement 2C–I). We did not observe any instance of superficial cells internalizing and integrating into a protoneuromast.

The presence of a cell population with minimal or nonexistent basal contact with the underlying tissue but significant surface area contact with the overlying skin suggested that these cells might contribute to the migration of the PLLp. Imaging suggested the presence of broad membrane protrusions from these cells which were oriented in the direction of migration (Figure 1D–F, Figure 1—video 1). To further examine these protrusions, we performed live imaging using Airyscan super-resolution microscopy in embryos where a BAC containing the cxcr4b regulatory elements was used to drive lifeact-citrine, which labels F-Actin (Colak-Champollion et al., 2019; Riedl et al., 2008), in the PLLp. The distribution of LifeAct-Citrine revealed the presence of transient Actin fibers extending within broader membrane protrusions, reminiscent of lamellipodia (Figure 3A, Figure 3—video 1). These protrusions were even more apparent when we imaged isolated cells by transplanting cells from TgBAC(cxcr4b:lifeact-citrine);Tg(cldnb:lyn-mscarlet) double transgenic embryos into Tg(cldnb:lyn-mscarlet) embryos (Figure 3C, Figure 3—video 2). Quantification of the direction of these protrusions in Airyscan super-resolution timelapse videos of the superficial surface of the PLLp showed that they were highly polarized, extending in the direction of migration (Figure 3B), with an average length of ~2.4 μm (green in Figure 6G), similar in orientation and length to previously described polarized migratory protrusions that extend from the basal aspect of cells that make up a protoneuromast (average length ~1.9 μm, green in Figure 6N). In contrast, the apical surface of most cells in a protoneuromast is tightly constricted, and we did not observe significant apical protrusive activity from these cells.

Figure 3 with 4 supplements see all

Download asset Open asset

In our timelapse videos of Tg(cxcr4b:lifeact-citrine) transgenic embryos, we noticed signs of significant retrograde actin flow in these protrusions. In migrating cultured cells, the rate of retrograde actin flow in lamellipodial protrusions has been shown to correlate with the traction force exerted by the migrating cell (Gardel et al., 2008). To assess the rate of retrograde actin flow, we took high-speed super-resolution Airyscan videos at 1-s intervals and used the movement of inhomogeneities in the signal to assess the rate of retrograde actin flow in these protrusions (Figure 3—video 3, example shown in Figure 6—figure supplement 1). Figure 3D shows one such inhomogeneity flowing backwards from the lamellipodial tip toward the cell body (arrowheads in Figure 3D). Quantification of the flow showed an average retrograde actin flow speed of 6.5 μm.min⁻¹ with a standard deviation of 1.8 μm.min⁻¹ (Figure 3E).

The orientation of migratory protrusions on the basal surface of the PLLp is known to be dependent on Fgf signaling, and directional polarization is lost when Fgf signaling is inhibited (Lecaudey et al., 2008). To test whether the orientation of the superficial protrusions, like those on the basal side of the PLLp, is also dependent on Fgf signaling, we treated embryos with either 20 μM of SU5402 or DMSO for 6 hr and measured the orientation of the superficial protrusions. After 6 hr of treatment with DMSO, superficial protrusions remained robustly polarized in the direction of migration (Figure 3F,G, Figure 3—video 4, upper left panel). However, this polarization was lost after Fgf inhibition by SU5402 treatment (Figure 3H,I, Figure 3—video 4, upper right panel), suggesting that, like the basal protrusions, the directionality of these apical protrusions was also dependent on Fgf signaling. For comparison, we performed the same analysis on basal protrusions. However, since the densely packed cell membranes comprising the basal surface of the PLLp make quantification of these basal protrusions challenging, we performed these experiments in chimeric embryos where we transplanted TgBAC(cxcr4b:lifeact-citrine);Tg(cldnb:lyn-mscarlet) donor cells into Tg(cldnb:lyn-mscarlet) host embryos. As expected, in DMSO-treated embryos the basal protrusions were primarily oriented in the direction of migration (Figure 3J,K, Figure 3—video 4, lower left panel). However, after 6 hr of treatment with SU5402, this polarity was completely abolished (Figure 3L,M, Figure 3—video 4, lower right panel).

These experiments demonstrate the existence of a population of cells that make significant contact with the overlying skin during migration. Furthermore, our analysis show that these cells extend protrusions reminiscent of lamellipodia against the skin, and that the orientation of these protrusions, like that of the basal cryptic lamellipodia, is sensitive to Fgf inhibition. This suggested a potential role for the overlying skin as a substrate for migration of the PLLp. To test whether the skin is necessary for PLLp migration, we removed the skin overlying the PLLp using a tungsten needle and then imaged the resulting behavior of the PLLp using regular (diffraction-limited) confocal microscopy. We performed these experiments both in tg(claudinb:lyn-egfp) embryos, in which the periderm (outer skin layer) is labeled with membrane localized Gfp, and in tg(claudinb:lyn-gfp; krt4:dsred) double transgenic embryos, in which both the periderm and basal layers of the skin are additionally labeled with cytoplasmically localized DsRed (O'Brien et al., 2012). Immediately following skin removal, the PLLp ceased migration and remained stationary.

In most cases, after skin removal, the skin rapidly healed over the PLLp with little to no damage to the PLLp itself (based on the minimal appearance dying cells). In these cases, after skin regrowth, the PLLp recovered normal migration and continued migrating along the length of the embryo (Figure 4—video 1). During this subsequent migration neuromast deposition appeared normal, and the primordium reached the tip of the tail after a slight delay (Figure 4—figure supplement 1). In a minority of cases, where large patches of skin covering a significant fraction of the trunk were removed such that the skin could not heal over the PLLp in the time course of our videos, we observed the PLLp for a period of several hours. In these cases, the PLLp did not recover forward migration, and the PLLp cells eventually died.

Figure 4A—C shows three still frames from a representative time-lapse video taken while the skin healed over the PLLp. Initially the PLLp is stationary, having stopped forward migration. The skin (magenta) is healing toward the PLLp (Figure 4A). Twenty minutes later, the skin has made contact with the PLLp, and the morphology of the leading cells has changed, becoming more stretched out, reminiscent of their morphology in intact primordia (Figure 4B). By 97 min, the skin has healed over the PLLp and normal forward migration has robustly resumed (Figure 4C). This dramatic behavior can be seen in kymographs of both the green PLLp membranes and the magenta krt4:DsRed-positive epidermal cells (Figure 4D–F). Initially, the PLLp is stationary, while uncovered by the skin (compare Figure 4E,F). However, as the skin heals over the PLLp, robust and continuous forward migration is recovered, as shown by diagonal lines in the kymograph (Figure 4D–F). We removed the skin from >10 independent embryos and, in all cases, migration was abolished until the skin healed over the PLLp.

Figure 4 with 2 supplements see all

Download asset Open asset

To quantify this migratory behavior, we crossed Tg(claudinb:lyn-egfp; krt4:dsred) fish to TgBAC(cxcr4b:h2a-mcherry) fish to visualize the deeper basal cells and more superficial periderm of the skin, as well as both the membranes and nuclei of the PLLp. We then removed the skin overlying the PLLp and tracked the movement of randomly selected cells distributed throughout the PLLp for each of three replicate embryos. For each cell, the time at which the skin covers the position of that individual cell was marked, and the average velocity in the normal direction of migration (along the rostral-caudal axis of the embryo) before and after this point was calculated. Figure 4G shows the paired measurements for each cell, with the top row representing the average velocity of cells before skin contact and the bottom row representing the average velocity after skin contact. Consistent with the results from bulk analysis of movement using kymographs, there was a dramatic increase in forward migration after skin contact. The velocity of cells before skin contact was distributed around 0, suggesting non-directional movement, and the velocities after skin contact are clustered around a mean of 0.82 μm.min⁻¹, close to the normal migration speed of a primordium. Aggregating the frame-to-frame velocities shows that the velocity before skin contact is normally distributed around a value of approximately zero (mean = 0.067 μm.min⁻¹, standard deviation = 0.51 μm.min⁻¹), whereas the velocities after skin contact are normally distributed around approximately 0.8 μm.min⁻¹ (mean = 0.807 μm.min⁻¹, standard deviation = 0.648 μm.min⁻¹). Taken together, these data suggest a profound inability of the PLLp to migrate without overlying skin, which is completely reversed after skin regrowth over the PLLp.

Interestingly, removal of the skin was associated with a significant increase in the average height of the primordium. Paired measurements of the average height of the PLLp before and after skin removal showed a ~ 18% increase in the height of the PLLp associated with removal of the overlying skin (Figure 4H), suggesting that the skin does in fact provide some compression to the PLLp.

We hypothesized that removing the skin caused a failure of the overlying cells to extend robust migratory protrusions and that this phenomenom contributes to the failure of collective migration. To assess this, we crossed TgBAC(cxcr4b:lifeact-citrine) fish to Tg(cldnb:lyn-m-scarlet) fish to generate double-transgenic embryos in which both Actin and membranes were labeled. The skin over the PLLp was then removed and the resulting Actin dynamics imaged using Airyscan super-resolution confocal microscopy.

After this manipulation, we noticed a dramatic change in the behavior of the overlying cells. In PLLp which were not covered by skin, we saw almost no polarized Actin-rich protrusions of the kind observed extending from overlying cells in intact embryos. Instead, cells appeared disorganized and rapidly extended and retracted protrusions reminiscent of membrane blebs. High-resolution imaging of these bleb-like structures showed that they initiate as a rapid membrane expansion devoid of cortical actin (Figure 5A–E, Figure 5—video 1). As the bleb expands, cortical Actin is recruited to the newly expanded membrane and can be detected by an increase in the LifeAct-Citrine signal, while the Lyn-mScarlet membrane marker remains constant in intensity (Figure 5E). After a short period, these blebs are retracted into the cell.

Figure 5 with 2 supplements see all

Download asset Open asset

Because the regrowth of the skin over the PLLp is associated with a rapid and robust recovery of forward migration, we imaged Actin dynamics in the PLLp during skin regrowth in TgBAC(cxcr4b:lifeact-citrine); Tg(cldnb:lyn-m-scarlet) embryos using Airyscan super-resolution confocal microscopy (Figure 5F–I, Figure 5—video 2). After skin removal -- but before skin regrowth over the PLLp -- we observed rapid membrane blebbing of superficial cells, as described above. However, as the skin heals over the PLLp, cells undergo a dramatic morphological change. The bleb-type morphology of overlying cells is abolished and cells begin to extend actin-rich protrusions in the direction of migration, reminiscent of those observed in superficial cells in intact, unperturbed embryos. This transition was so rapid that over the course of our timelapse video, we could simultaneously observe blebs from regions of the PLLp not yet covered by skin alongside actin-rich rod-like protrusions from cells that had been covered by the regrowing skin. Eventually, the entire PLLp is again covered by the skin, superficial cell protrusive activity is restored, and normal forward migration of the PLLp resumes.

These data suggested a rapid recovery of superficial cell protrusive activity after skin regrowth. To quantify this change, we performed high-resolution imaging of Actin dynamics in TgBAC(cxcr4b:lifeact-citrine);Tg(cldnb4.2:lyn-mscarlet) embryos under three conditions: unperturbed embryos in which the skin has not been removed, embryos in which the skin was removed over the PLLp, and embryos in which the skin had been removed and had subsequently healed over the PLLp. Figure 6A–F shows the results of this experiment, and examples of protrusions can be seen in Figure 6—figure supplement 1A. As expected, embryos where the skin had not been removed showed robust protrusions oriented in the direction of migration (Figure 6A,B, Figure 6—video 1, upper left panel), with an average length of ~2.4 um (Figure 6G, green). When the skin was removed, we again observed a profound loss of superficial Actin-rich protrusions, which were replaced by superficial membrane blebs of the type shown in Figure 5A–D (Figure 6C, Figure 6—video 1, upper middle panel). The few protrusions that remained were both shorter (Figure 6G, compare orange to green, mean skin intact = 2.45 μm, mean skin removed = 1.76 μm) and no longer oriented in the direction of migration (Figure 6D). However, as suggested by the timelapse analysis above, after the skin healed over the PLLp, superficial protrusions were again observed (Figure 6E, Figure 6—video 1, upper right panel), their orientation was again strongly polarized in the direction of migration (Figure 6F) and their length had recovered to almost unperturbed levels (Figure 6G, compare green and purple, mean intact = 2.45 μm, mean post-healing = 2.26 μm).

Figure 6 with 2 supplements see all

Download asset Open asset

Although the absence of these directional superficial cell protrusions was expected after removal of the overlying skin, we were surprised by the profound loss of migratory ability resulting from this manipulation, given that the basal surface of deeper protoneuromast cells were presumably still in contact with the underlying tissue. We wondered whether removal of the skin resulted in a broader loss of migratory protrusions in the cells of the PLLp. To assess this, we repeated the above experiment, this time using embryos in which we transplanted donor cells from Tg(BACcxcr4b:lifeact-citrine; cldnb:lyn-mscarlet) into Tg(cldnb:lyn-mscarlet) embryos to generate isolated clones in which we could quantify the directionality and length of basal protrusions.

In intact embryos, these basal protrusions, like the superficial cell protrusions, were highly polarized in the direction of migration, although they were on average slightly shorter than superficial cell protrusions (Figure 6H,I,N, average length = 1.95 μm, Figure 6—video 1, bottom left panel). When the skin was removed, surprisingly, we observed a profound loss of the normal directional orientation of these basal protrusions (Figure 6J,K, Figure 6—video 1, bottom middle panel), even though many of these cells had no contact with the overlying skin. This failure of directional orientation was accompanied, as with the superficial protrusions, by a decrease in the average length of the protrusions, (Figure 6N, compare green to orange, mean intact 1.95 μm, mean skin removed = 1.32 μm) and by the appearance of membrane blebs on the basal surface of these cells (arrowheads in Figure 6J). However, neither the loss of rod-like Actin-rich protrusions nor the appearance of membrane blebs were as dramatic on the basal surface as on the superficial surface.

When we imaged these protrusions after skin regrowth (Figure 6L, Figure 6—video 1, bottom right panel), we observed the same dramatic recovery of basal protrusions. The polarization of the basal protrusions in the direction of migration was recovered (Figure 6M) and their length was similar to that of protrusions in unperturbed embryos (Figure 6N, compare purple to green, mean intact = 1.95 μm, mean post healing = 1.83 μm).

Given the failure to observe superficial cell protrusions in primordia where the skin had been removed and embryos re-embedded in agarose, which is expected to provide some mechanical confinement, we wondered whether interactions with the extracellular matrix associated with overlying skin cells were necessary for these protrusions. To test this, we removed the skin over the PLLp in 36hpf embryos and embedded these embryos for imaging in 20 mg/mL Matrigel, a gelatinous artificial 3D medium derived from extracellular matrix, rather than agarose. We reasoned that by using undiluted 20 mg/mL Matrigel, we would provide the stiffest possible substrate for primordium migration.

Interestingly, we noted that healing of the skin was significantly inhibited in Matrigel, allowing us to image primordia in skin-free conditions for longer than is possible in agarose. In most cases, primordia either did not migrate, or migrated forward only short distances, less than the length of the primordium, and then stalled again and did not recover migration. Quantification showed that when skinned embryos were embedded in Matrigel, migration speed was significantly slower (~0.2 μm/min) than embryos with skin intact (~0.8 μm/min), suggesting that Matrigel is unable to completely replace the skin for normal primordium migration. When compared to skinned embryos in agarose, skinned embryos in Matrigel showed a slight, but not statistically significant, increase in migration speed (Figure 7E).

Figure 7

Download asset Open asset

Despite this failure to recover migration when skinned embryos are embedded in matrgiel, we observed a profound recovery in the morphology of the superficial cells, which no longer exhibited blebbing behavior, but instead extended actin-rich lamellipodial-like processes, qualitatively similar to those observed in unperturbed embryos (Figure 7A). However, when we quantified the directionality of these protrusions, we saw that they were oriented in all directions, unlike the highly polarized protrusions observed in intact embryos (Figure 7B). In contrast, when we embedded embryos in which we had removed the skin overlying the PLLp in Matrigel and examined the polarity of basal protrusions, we saw a robust recovery of the polarization of these protrusions (Figure 7C,D). This recovery was striking when compared to the loss of this polarization when skinned embryos were embedded in agarose.

In this study, we examined the Posterior Lateral Line primordium of zebrafish, a well-studied model of collective cell migration that migrates in a channel between the underlying somites and overlying epidermis. We show that, in addition to the previously described basal lamellipodia extended by epithelialized cells in this cluster (Haas and Gilmour, 2006), there is an additional population of cells lying superficially, covering these epithelial cells, which also extend migratory processes against the overlying skin. As with the directionality of basal cryptic lamellipodia extended from the epithelial cells that comprise a protoneuromast, the directionality of these superficial processes is abolished when Fgf signaling is inhibited by treatment with SU5402, suggesting that the requirement for Fgf signaling to maintain migration applies to both superficial and basal processes.

Supporting the idea that these migratory processes are in fact lamellipodia, we show retrograde Actin flow in these protrusions toward the cell body at approximately 6.5 μm.min⁻¹. Values reported from cultured cells in vitro vary significantly, from 0.12 to 1.5 μm.min⁻¹ in the distal lamellipodium of Ptk1 cells (Gardel et al., 2008), to 0.8 – 4.5 μm.min⁻¹ for spreading mouse embryonic fibroblasts (Dubin-Thaler et al., 2004) and 2.5–6 μm.min⁻¹ in migrating fish keratinocytes (Jurado et al., 2005). Retrograde actin flow in migrating cells in live embryos has not been widely quantified; however, a recent study examining Drosophila macrophages (Yolland et al., 2019) in vivo showed a retrograde flow rate of approximately 4 μm.min⁻¹. Explant cultures of Xenopus spinal cord neurons showed higher rates for neuron growth cones, from 9 to 12 μm.min⁻¹ depending on substrate (Nichol et al., 2016). Interestingly, when these rates were assessed in vivo, retrograde flow was slower, at around 7 μm.min⁻¹, which the authors speculate may be due to stronger F-actin clutching in the in vivo 3D environment. Our value of 6.5 μm.min⁻¹ is on the upper end of the values observed in cell culture, but within the bounds of the values reported in vivo.

In vitro, the rate of actin retrograde flow in Ptk1 cells has a biphasic relationship with traction stress on the substrate; Between 0 and 0.6 μm.min⁻¹, retrograde flow rate correlates positively with traction stress, but, between 0.6 and 1.8 μm.min⁻¹, it correlates negatively (Gardel et al., 2008). In the context of these observations in vitro, the high-retrograde flow rate of actin in superficial cell lamellipodia might suggest low traction stress exerted by these PLLp cells. However, we caution that the rates measured in vitro by Gardel et al are generally less than 1.2 μm.min⁻¹, significantly lower than our observed value of 6.5 μm.min⁻¹ in PLLp superficial cells. Therefore, how this relatively high-retrograde flow rate measured in vivo correlates with traction stress cannot be inferred with certainty.

Typically, the PLLp has been conceptually separated into a leading mesenchymal-like domain, where cells are relatively flat and have no obvious apico-basal polarity, and a trailing epithelial domain, where cells become elongated and adopt a distinct epithelial morphology with apico-basal polarity. Despite this distinction, it has long been recognized that both domains contribute to collective migration, with epithelial cells extending cryptic lamellipodia in the direction of migration. In this study, we have defined the superficial primordium cells as those that occupy a superficial position above the protoneuromasts, and by definition exclude leading mesenchymal-like cells. However, it is likely that, rather than being an entirely separable population, these superficial cells simply represent an extension of leading mesenchymal-like cells over the entire PLLp and that these cells constitute a continuous migratory population. The Gilmour group has recently performed an elegant cell segmentation and morphological analysis study (Hartmann et al., 2020) using support vector classifiers to classify cells in the primordium based on morphology as central cells, leading cells, peripheral cells, as well as so-called inter-organ cells (which we suspect are similar to our superficial cells). While this work showed that PLLp cells populations can in fact be separated into morphologically distinct populations, it also showed significant overlap between these populations, with intermediate states common between, for example, leading and peripheral cells, again suggesting some degree of morphological continuity between distinct cell populations. In this context, in addition to conceptually dividing the PLLp along a longitudinal axis into a leading and trailing domain with distinct cellular morphologies, a second axis is radially arranged with apico-basally polarized epithelial cells incorporated into protoneuromasts at the core and non-protoneuromast, less apico-basally polarized cells at the periphery. This is an example of what Blanchard and co-authors have called ‘mesoscale heterogeneity’ in migrating systems (Blanchard et al., 2019).

Although this study has focused on the protrusions extended by these superficial cells against the overlying skin, we also note that this population also appears to cover the lateral edges of the primordium and the cells themselves seem to be relatively motile, often moving relative to the underlying epithelial cells which they cover. This suggests the possibility that migratory activity is not restricted to the deep or superficial surfaces of the primordium but is in fact circumferential. It is important to note that not only must the epithelial cells of the PLLp migrate effectively in this confined space, but while migrating, deeper cells must also undergo a complex series of coordinated morphogenetic events leading to the formation of apically-constricted protoneuromasts. In addition, previous work has shown that the apical constrictions of protoneuromasts house microlumen structures critical for sequestration of signaling molecules (Durdu et al., 2014). Damage to these microlumens causes a failure of Fgf signaling and delayed neuromast deposition. In this context, the presence of a layer of superficial cells between the skin and the protoneuromast cells may not only contribute to collective migration of the primordium, it might also contribute to robust formation of the underlying protoneuromasts during migration and protect their apical microlumens from sheer stress caused by migration against the overlying skin.

Consistent with the existence of migratory contacts on the superficial side of the PLLp closely apposed to the overlying skin, we show that when the skin is removed superficial lamellipodia are almost completely abolished and cells extend and retract rapid membrane blebs. The few Actin-rich rod-like protrusions that remain are no longer oriented in the direction of migration and are significantly shorter than those extended from the superficial surface of the PLLp when the skin is intact. Regrowth of the skin restores both the normal length and directionality of these protrusions with very little delay suggesting, again, a rapid switch in migratory ability and morphology between PLLp cells in the absence and presence of overlying skin. It should be noted that in this context the morphology of superficial cells with and without overlying skin is so dramatically different that even though we can in most cases quantify a small number of rod-like protrusions in primordia without skin, their relationship to the broad, flat lamellipodial protrusions seen in intact embryos is not clear. It may be that removal of the skin completely disrupts lamellipodia formation, and that a distinct class of actin-rod like protrusions are observed with low frequency in this context.

Intriguingly, we observe very similar changes in the basal cryptic lamellipodia of protoneuromast cells when the skin is removed, despite the fact that these cells, in most cases, do not directly contact the overlying skin. While Actin-rich protrusions are still observed at the basal surface of the PLLp these protrusions, like those observed superficially, become shorter and are no longer oriented in the direction of migration. Concurrent with this loss of directionality, we see the appearance of membrane blebs in the basal surface of these cells, although to a lesser extent than is observed superficially. This observation, along with the rapid recovery of directionality and length of basal protrusions after skin regrowth--despite the fact that these cells do not themselves directly contact the skin--raises the possibility that the skin is not only necessary to provide a substrate for the superficial migratory processes, but that it is provides some indirect support to basal migratory processes as well. Whether this is due to mechanical confinement, restriction of signaling molecules, provision of adhesive interactions, or some other unknown function of the overlying skin remains unclear.

While agarose provides some degree of mechanical confinement, it fails to provide other physiologically relevant factors such as adhesion sites. Matrigel is a complex protein mixture secreted by Engelbreth-Holm-Swarm mouse sarcoma cells and is commonly used as a 3D matrix for growing cultured cells. The Youngs modulus of gelled Matrigel can vary over twofold between batches, with typical values falling around 200–400 Pa (Soofi et al., 2009). The Youngs modulus of agarose is significantly higher. Although we have not directly measured the mechanical properties of agarose used in this study, a survey of values from the literature suggest that a value on the order of 10–15 kPa, (approximately 20-fold higher than Matrigel) is a reasonable estimate for the concentrations used (see for example Kazi et al., 2019).

When we embedded skin-removed embryos in Matrigel, we observed an almost-complete reversion of blebbing from superficial cells. However, this recovery is not associated with polarization of these superficial protrusions and provides only a modest recovery of cell migration speed. This inability to sustain collective migration is consistent with superficial polarized lamellipodia being required for migration. Furthermore, it suggests that while confinement under Matrigel and the potential to interact with its ECM components may be adequate for the formation of superficial cell lamellipodia, their polarization requires additional factors or environmental constraints that Matrigel is unable to provide. In the case of the basal protrusions, while protrusion directionality is not recovered by embedding in agarose and providing confinement alone, the directionality is recovered when embryos are embedded in the comparatively less stiff Matrigel. This argues against mechanical confinement alone as a requirement for polarized protrusion, instead suggesting that some component provided by Matrigel but not by agarose is required for effective migration. One possibility is that Matrigel is more effective at confining signaling molecules essential for migration than is agarose. While the major protein components of Matrigel appear to be extracellular matrix components such as Laminin, Collagen, and Entactin (Hughes et al., 2010), several growth factors have also been detected, and this has led to cautions about interpreting the results of experiments performed in Matrigel (Vukicevic et al., 1992). However, precisely why basal and not superficial polarity is recovered in Matrigel remains a mystery, and hints at distinct regulatory mechanisms underlying protrusive activity on the basal and superficial surfaces of the PLLp. In this context, it is important to note that the basal surfaces of cells remain in contact with their in vivo substrate, while the superficial surface is in contact with Matrigel. This may in part explain the recovery of protrusive polarization basally but not superficially. The failure to restore normal migration speed when basal, but not superficial, protrusive directionality is restored in this condition suggests that both superficial and basal migratory protrusions are required for robust migration. Future experiments will attempt to elucidate the determinants of the recovery observed when embryos are embedded in Matrigel, but not agarose, as well as the reason for the failure of Matrigel to rescue migration to wild-type levels.

How cells migrate through a confined environment is poorly understood, especially during collective migration. Unconfined cells in 2D culture can migrate by promoting attachment to a substrate through which force can be transmitted. However, in confined systems such as 3D scaffolds, the mechanical properties of the environment itself can force cells into significant contact with their extracellular migratory substrate. The rapid switch between a lamellipodial form of migration to unproductive blebbing and a failure to migrate is the primary consequence of removal of the skin from the PLLp. A similarly rapid switch between distinct cell morphologies has also been reported in early zebrafish gastrula cells placed in culture (Ruprecht et al., 2015). When unconfined in culture, these cells show a characteristic but unproductive blebbing, and fail to migrate. However, upon mechanical confinement in agarose, the cells adopted a pear-shaped morphology with a large stable leading protrusion and became highly migratory. This migration was associated with rapid retrograde cortical actin flow in the leading edge of the cell. While in this context the stable leading protrusion does not appear to be lamellipodial as are the protrusions we observe in the superficial cells of the primordium, the rapid switch between these two morphologies - unpolarized and unproductive blebbing to a highly polarized migratory protrusion - upon cell confinement is at least superficially similar. Whether or not these superficial similarities reflect deep mechanistic similarities in the mechanisms of migration in these two contexts will require further study.

Cells in confined environments can also employ so-called ‘chimneying’ mechanisms, where cells can push outwards against the surrounding environment and generate enough force to facilitate migration. Such a mechanism has been suggested to play a role in migration of leukocytes in an adhesion-independent manner (Malawista and de Boisfleury Chevance, 1997). A related but mechanistically distinct mechanism, flow-friction-driven force transmission, is hypothesized to transmit intracellular force from the cytoskeleton to the substrate by means of nonspecific friction between the cell and its environment (Hawkins et al., 2011; Ruprecht et al., 2015). In this context, retrograde flows of the actomyosin cortex, like those we observe in superficial protrusions in the PLLp, have been suggested to play a role in the generation of motile force (Paluch et al., 2016). However, the ability of Matrigel to recover superficial processes suggests that the presence of extracellular-matrix-like adhesion sites is critical for formation of these migratory processes. Furthermore, a recent report by the Nechiporuk group has shown phospho-paxilin immunoreactivity around the lateral and superficial sides of the PLLp, suggesting that a low-adhesion migration mechanism is unlikely Olson and Nechiporuk, 2020. However, the fact that in the vast majority of cases (>99%), we were able to remove the skin from the PLLp while retaining attachment of the PLLp to the underlying substrate suggests an asymmetry of adhesion between the superficial and basal surfaces of the PLLp, and that the basal surface of the PLLp is more tightly adhered to the underlying substrate than the superficial surface is to the overlying skin, despite the presence of superficial adhesions. The significance of this to primordium migration remains unclear and is a subject of further study.

Understanding how cells in the developing embryo move through diverse environments is critical for understanding both normal morphogenesis and a number of pathologies. Cell migration has been studied extensively in cell culture in the two-dimensional regime; however, there is growing evidence that 2D models emphasize specific modes and mechanisms of cell migration that might not apply to cells migrating in a three-dimensional confined environment. While multicellular systems acquire behaviors that correspond broadly to those observed in single cells, it is likely that the details about how they acquire those characteristics differ significantly. The development of in vivo model systems to study cell migration at high spatio-temporal resolution, as well as quantitative methods for analyzing such systems will be necessary to extend the study of 3D cell migration from the dish to the animal.

CSV files of raw measurements and notebooks containing all statistical analyses are available at Github: https://github.com/chitnislabnih/dallenogare2019 (copy archived at https://archive.softwareheritage.org/swh:1:rev:cf258e888577731586096fd464d34df1bd0b2c1a/).

1. 1. Aman A
   2. Nguyen M
   3. Piotrowski T

   (2011) Wnt/β-catenin dependent cell proliferation underlies segmented lateral line morphogenesis

   Developmental Biology 349:470–482.

   https://doi.org/10.1016/j.ydbio.2010.10.022

   • PubMed
   • Google Scholar
2. 1. Bindels DS
   2. Haarbosch L
   3. van Weeren L
   4. Postma M
   5. Wiese KE
   6. Mastop M
   7. Aumonier S
   8. Gotthard G
   9. Royant A
   10. Hink MA
   11. Gadella TW

   (2017) mScarlet: a bright monomeric red fluorescent protein for cellular imaging

   Nature Methods 14:53–56.

   https://doi.org/10.1038/nmeth.4074

   • PubMed
   • Google Scholar
3. 1. Blanchard GB
   2. Fletcher AG
   3. Schumacher LJ

   (2019) The Devil is in the mesoscale: mechanical and behavioural heterogeneity in collective cell movement

   Seminars in Cell & Developmental Biology 93:46–54.

   https://doi.org/10.1016/j.semcdb.2018.06.003

   • PubMed
   • Google Scholar
4. Software

   1. Blender Online Community

   (2018) Blender - a 3D modelling and rendering package, version 2.8

   Stichting Blender Foundation.

   https://www.blender.org/
5. 1. Boldajipour B
   2. Mahabaleshwar H
   3. Kardash E
   4. Reichman-Fried M
   5. Blaser H
   6. Minina S
   7. Wilson D
   8. Xu Q
   9. Raz E

   (2008) Control of chemokine-guided cell migration by ligand sequestration

   Cell 132:463–473.

   https://doi.org/10.1016/j.cell.2007.12.034

   • PubMed
   • Google Scholar
6. 1. Burns JM
   2. Summers BC
   3. Wang Y
   4. Melikian A
   5. Berahovich R
   6. Miao Z
   7. Penfold ME
   8. Sunshine MJ
   9. Littman DR
   10. Kuo CJ
   11. Wei K
   12. McMaster BE
   13. Wright K
   14. Howard MC
   15. Schall TJ

   (2006) A novel chemokine receptor for SDF-1 and I-TAC involved in cell survival, cell adhesion, and tumor development

   Journal of Experimental Medicine 203:2201–2213.

   https://doi.org/10.1084/jem.20052144

   • PubMed
   • Google Scholar
7. Preprint

   1. Chen J
   2. Ding L
   3. Viana MP
   4. Hendershott MC
   5. Yang R
   6. Mueller IA
   7. Rafelski SM

   (2018) The allen cell structure segmenter: a new open source toolkit for segmenting 3D intracellular structures in fluorescence microscopy images

   bioRxiv.

   https://doi.org/10.1101/491035

   • Google Scholar
8. 1. Chitnis AB
   2. Nogare DD
   3. Matsuda M

   (2012) Building the posterior lateral line system in zebrafish

   Developmental Neurobiology 72:234–255.

   https://doi.org/10.1002/dneu.20962

   • PubMed
   • Google Scholar
9. Report

   1. Cignoni P
   2. Callieri M
   3. Corsini M
   4. Dellepiane M
   5. Ganovelli F
   6. Ranzuglia G

   (2008)

   MeshLab: An Open-Source Mesh Processing Tool

   The Eurographics Association.

   • Google Scholar
10. 1. Clark AG
    2. Vignjevic DM

    (2015) Modes of Cancer cell invasion and the role of the microenvironment

    Current Opinion in Cell Biology 36:13–22.

    https://doi.org/10.1016/j.ceb.2015.06.004

    • PubMed
    • Google Scholar
11. 1. Colak-Champollion T
    2. Lan L
    3. Jadhav AR
    4. Yamaguchi N
    5. Venkiteswaran G
    6. Patel H
    7. Cammer M
    8. Meier-Schellersheim M
    9. Knaut H

    (2019) Cadherin-Mediated cell coupling coordinates chemokine sensing across collectively migrating cells

    Current Biology 29:2570–2579.

    https://doi.org/10.1016/j.cub.2019.06.061

    • PubMed
    • Google Scholar
12. 1. Coombs S
    2. Van Netten S

    (2005) The hydrodynamics and structural mechanics of the lateral line system

    Fish Physiology 23:103–139.

    https://doi.org/10.1016/S1546-5098(05)23004-2

    • Google Scholar
13. 1. Dalle Nogare D
    2. Somers K
    3. Rao S
    4. Matsuda M
    5. Reichman-Fried M
    6. Raz E
    7. Chitnis AB

    (2014) Leading and trailing cells cooperate in collective migration of the zebrafish posterior lateral line primordium

    Development 141:3188–3196.

    https://doi.org/10.1242/dev.106690

    • PubMed
    • Google Scholar
14. 1. Dambly-Chaudière C
    2. Cubedo N
    3. Ghysen A

    (2007) Control of cell migration in the development of the posterior lateral line: antagonistic interactions between the chemokine receptors CXCR4 and CXCR7/RDC1

    BMC Developmental Biology 7:23.

    https://doi.org/10.1186/1471-213X-7-23

    • PubMed
    • Google Scholar
15. 1. David NB
    2. Sapède D
    3. Saint-Etienne L
    4. Thisse C
    5. Thisse B
    6. Dambly-Chaudière C
    7. Rosa FM
    8. Ghysen A

    (2002) Molecular basis of cell migration in the fish lateral line: role of the chemokine receptor CXCR4 and of its ligand, SDF1

    PNAS 99:16297–16302.

    https://doi.org/10.1073/pnas.252339399

    • PubMed
    • Google Scholar
16. 1. Donà E
    2. Barry JD
    3. Valentin G
    4. Quirin C
    5. Khmelinskii A
    6. Kunze A
    7. Durdu S
    8. Newton LR
    9. Fernandez-Minan A
    10. Huber W
    11. Knop M
    12. Gilmour D

    (2013) Directional tissue migration through a self-generated chemokine gradient

    Nature 503:285–289.

    https://doi.org/10.1038/nature12635

    • PubMed
    • Google Scholar
17. 1. Dubin-Thaler BJ
    2. Giannone G
    3. Döbereiner HG
    4. Sheetz MP

    (2004) Nanometer analysis of cell spreading on matrix-coated surfaces reveals two distinct cell states and STEPs

    Biophysical Journal 86:1794–1806.

    https://doi.org/10.1016/S0006-3495(04)74246-0

    • PubMed
    • Google Scholar
18. 1. Durdu S
    2. Iskar M
    3. Revenu C
    4. Schieber N
    5. Kunze A
    6. Bork P
    7. Schwab Y
    8. Gilmour D

    (2014) Luminal signalling links cell communication to tissue architecture during organogenesis

    Nature 515:120–124.

    https://doi.org/10.1038/nature13852

    • PubMed
    • Google Scholar
19. 1. Farooqui R
    2. Fenteany G

    (2005) Multiple rows of cells behind an epithelial wound edge extend cryptic lamellipodia to collectively drive cell-sheet movement

    Journal of Cell Science 118:51–63.

    https://doi.org/10.1242/jcs.01577

    • PubMed
    • Google Scholar
20. 1. Friedl P
    2. Gilmour D

    (2009) Collective cell migration in Morphogenesis, regeneration and Cancer

    Nature Reviews Molecular Cell Biology 10:445–457.

    https://doi.org/10.1038/nrm2720

    • PubMed
    • Google Scholar
21. 1. Fuentes F
    2. Reynolds E
    3. Lewellis SW
    4. Venkiteswaran G
    5. Knaut H

    (2016) A plasmid set for efficient bacterial artificial chromosome (BAC) Transgenesis in zebrafish

    G3: Genes, Genomes, Genetics 6:829–834.

    https://doi.org/10.1534/g3.115.026344

    • PubMed
    • Google Scholar
22. 1. Gardel ML
    2. Sabass B
    3. Ji L
    4. Danuser G
    5. Schwarz US
    6. Waterman CM

    (2008) Traction stress in focal adhesions correlates biphasically with actin retrograde flow speed

    Journal of Cell Biology 183:999–1005.

    https://doi.org/10.1083/jcb.200810060

    • PubMed
    • Google Scholar
23. 1. Gerety SS
    2. Breau MA
    3. Sasai N
    4. Xu Q
    5. Briscoe J
    6. Wilkinson DG

    (2013) An inducible transgene expression system for zebrafish and chick

    Development 140:2235–2243.

    https://doi.org/10.1242/dev.091520

    • PubMed
    • Google Scholar
24. 1. Ghysen A
    2. Dambly-Chaudière C

    (2007) The lateral line microcosmos

    Genes, Development 21:2118–2130.

    https://doi.org/10.1101/gad.1568407

    • PubMed
    • Google Scholar
25. 1. Gompel N
    2. Cubedo N
    3. Thisse C
    4. Thisse B
    5. Dambly-Chaudière C
    6. Ghysen A

    (2001) Pattern formation in the lateral line of zebrafish

    Mechanisms of Development 105:69–77.

    https://doi.org/10.1016/S0925-4773(01)00382-3

    • PubMed
    • Google Scholar
26. 1. Grant KA
    2. Raible DW
    3. Piotrowski T

    (2005) Regulation of latent sensory hair cell precursors by Glia in the zebrafish lateral line

    Neuron 45:69–80.

    https://doi.org/10.1016/j.neuron.2004.12.020

    • PubMed
    • Google Scholar
27. 1. Guo M
    2. Li Y
    3. Su Y
    4. Lambert T
    5. Dalle Nogare D
    6. Moyle MW
    7. Duncan LH
    8. Ikegami R
    9. Santella A
    10. Rey-Suarez I
    11. Green D
    12. Beringer A
    13. Chen J
    14. Vishwasrao H
    15. Ganesan S
    16. Prince V
    17. Waters JC
    18. Annunziata CM
    19. Hafner M
    20. Shroff H

    (2020)

    Accelerating iterative deconvolution and multiview fusion by orders of magnitude

    Nature Biotechnology In press.

    • Google Scholar
28. 1. Haas P
    2. Gilmour D

    (2006) Chemokine signaling mediates self-organizing tissue migration in the zebrafish lateral line

    Developmental Cell 10:673–680.

    https://doi.org/10.1016/j.devcel.2006.02.019

    • PubMed
    • Google Scholar
29. 1. Harris CR
    2. Millman KJ
    3. van der Walt SJ
    4. Gommers R
    5. Virtanen P
    6. Cournapeau D
    7. Wieser E
    8. Taylor J
    9. Berg S
    10. Smith NJ
    11. Kern R
    12. Picus M
    13. Hoyer S
    14. van Kerkwijk MH
    15. Brett M
    16. Haldane A
    17. Del Río JF
    18. Wiebe M
    19. Peterson P
    20. Gérard-Marchant P
    21. Sheppard K
    22. Reddy T
    23. Weckesser W
    24. Abbasi H
    25. Gohlke C
    26. Oliphant TE

    (2020) Array programming with NumPy

    Nature 585:357–362.

    https://doi.org/10.1038/s41586-020-2649-2

    • PubMed
    • Google Scholar
30. 1. Hartmann J
    2. Wong M
    3. Gallo E
    4. Gilmour D

    (2020) An image-based data-driven analysis of cellular architecture in a developing tissue

    eLife 9:e55913.

    https://doi.org/10.7554/eLife.55913

    • PubMed
    • Google Scholar
31. 1. Hawkins RJ
    2. Poincloux R
    3. Bénichou O
    4. Piel M
    5. Chavrier P
    6. Voituriez R

    (2011) Spontaneous contractility-mediated cortical flow generates cell migration in three-dimensional environments

    Biophysical Journal 101:1041–1045.

    https://doi.org/10.1016/j.bpj.2011.07.038

    • PubMed
    • Google Scholar
32. 1. Horstick EJ
    2. Jordan DC
    3. Bergeron SA
    4. Tabor KM
    5. Serpe M
    6. Feldman B
    7. Burgess HA

    (2015) Increased functional protein expression using nucleotide sequence features enriched in highly expressed genes in zebrafish

    Nucleic Acids Research 43:e48.

    https://doi.org/10.1093/nar/gkv035

    • PubMed
    • Google Scholar
33. 1. Hughes CS
    2. Postovit LM
    3. Lajoie GA

    (2010) Matrigel: a complex protein mixture required for optimal growth of cell culture

    Proteomics 10:1886–1890.

    https://doi.org/10.1002/pmic.200900758

    • PubMed
    • Google Scholar
34. 1. Hunter JD

    (2007) Matplotlib: a 2D graphics environment

    Computing in Science & Engineering 9:90–95.

    https://doi.org/10.1109/MCSE.2007.55

    • Google Scholar
35. Software

    1. Jones E
    2. Oliphant T
    3. Peterson P
    4. others

    (2001) SciPy, version 3.0.4

    Open Source Scientific Tools for Python.

    http://www.scipy.org/
36. 1. Jurado C
    2. Haserick JR
    3. Lee J

    (2005) Slipping or gripping? fluorescent speckle microscopy in fish keratocytes reveals two different mechanisms for generating a retrograde flow of actin

    Molecular Biology of the Cell 16:507–518.

    https://doi.org/10.1091/mbc.e04-10-0860

    • PubMed
    • Google Scholar
37. Software

    1. Kazhdan M
    2. Hoppe H

    (2013) Screened poisson surface reconstruction, version 8.0

    Association for Computing Machinery.

    http://www.cs.jhu.edu/~misha/Code/PoissonRecon/Version8.0/
38. 1. Kazi GAS
    2. Rahman KA
    3. Farahat M
    4. Matsumoto T

    (2019) Fabrication of single gel with different mechanical stiffness using three-dimensional mold

    Journal of Biomedical Materials Research Part A 107:6–11.

    https://doi.org/10.1002/jbm.a.36455

    • PubMed
    • Google Scholar
39. 1. Kimmel CB
    2. Ballard WW
    3. Kimmel SR
    4. Ullmann B
    5. Schilling TF

    (1995) Stages of embryonic development of the zebrafish

    Developmental Dynamics 203:253–310.

    https://doi.org/10.1002/aja.1002030302

    • PubMed
    • Google Scholar
40. 1. Koga A
    2. Cheah FS
    3. Hamaguchi S
    4. Yeo GH
    5. Chong SS

    (2008) Germline transgenesis of zebrafish using the medaka Tol1 transposon system

    Developmental Dynamics 237:2466–2474.

    https://doi.org/10.1002/dvdy.21688

    • PubMed
    • Google Scholar
41. 1. Lecaudey V
    2. Cakan-Akdogan G
    3. Norton WH
    4. Gilmour D

    (2008) Dynamic fgf signaling couples morphogenesis and migration in the zebrafish lateral line primordium

    Development 135:2695–2705.

    https://doi.org/10.1242/dev.025981

    • PubMed
    • Google Scholar
42. 1. Lewellis SW
    2. Nagelberg D
    3. Subedi A
    4. Staton A
    5. LeBlanc M
    6. Giraldez A
    7. Knaut H

    (2013) Precise SDF1-mediated cell guidance is achieved through ligand clearance and microRNA-mediated decay

    Journal of Cell Biology 200:337–355.

    https://doi.org/10.1083/jcb.201207099

    • PubMed
    • Google Scholar
43. 1. Lewiner T
    2. Lopes H
    3. Vieira AW
    4. Tavares G

    (2003) Efficient implementation of marching cubes' Cases with topological guarantees

    Journal of Graphics Tools 8:1–15.

    https://doi.org/10.1080/10867651.2003.10487582

    • Google Scholar
44. 1. Li Q
    2. Shirabe K
    3. Kuwada JY

    (2004) Chemokine signaling regulates sensory cell migration in zebrafish

    Developmental Biology 269:123–136.

    https://doi.org/10.1016/j.ydbio.2004.01.020

    • PubMed
    • Google Scholar
45. 1. López-Schier H
    2. Hudspeth AJ

    (2005) Supernumerary neuromasts in the posterior lateral line of zebrafish lacking peripheral Glia

    PNAS 102:1496–1501.

    https://doi.org/10.1073/pnas.0409361102

    • PubMed
    • Google Scholar
46. 1. Malawista SE
    2. de Boisfleury Chevance A

    (1997) Random locomotion and chemotaxis of human blood polymorphonuclear leukocytes (PMN) in the presence of EDTA: pmn in close quarters require neither leukocyte integrins nor external divalent cations

    PNAS 94:11577–11582.

    https://doi.org/10.1073/pnas.94.21.11577

    • PubMed
    • Google Scholar
47. Conference

    1. McKinney W

    (2010) Data structures for statistical computing in Python

    PROC. of the 9th PYTHON in SCIENCE CONF. (SCIPY 2010).

    https://doi.org/10.25080/Majora-92bf1922-00a

    • Google Scholar
48. 1. Meijering EH
    2. Niessen WJ
    3. Viergever MA

    (2001) Quantitative evaluation of convolution-based methods for medical image interpolation

    Medical Image Analysis 5:111–126.

    https://doi.org/10.1016/S1361-8415(00)00040-2

    • PubMed
    • Google Scholar
49. 1. Naumann U
    2. Cameroni E
    3. Pruenster M
    4. Mahabaleshwar H
    5. Raz E
    6. Zerwes HG
    7. Rot A
    8. Thelen M

    (2010) CXCR7 functions as a scavenger for CXCL12 and CXCL11

    PLOS ONE 5:e9175.

    https://doi.org/10.1371/journal.pone.0009175

    • PubMed
    • Google Scholar
50. 1. Nechiporuk A
    2. Raible DW

    (2008) FGF-dependent mechanosensory organ patterning in zebrafish

    Science 320:1774–1777.

    https://doi.org/10.1126/science.1156547

    • PubMed
    • Google Scholar
51. 1. Nichol RH
    2. Hagen KM
    3. Lumbard DC
    4. Dent EW
    5. Gómez TM

    (2016) Guidance of axons by local coupling of retrograde flow to point contact adhesions

    The Journal of Neuroscience 36:2267–2282.

    https://doi.org/10.1523/JNEUROSCI.2645-15.2016

    • PubMed
    • Google Scholar
52. 1. Nogare DD
    2. Nikaido M
    3. Somers K
    4. Head J
    5. Piotrowski T
    6. Chitnis AB

    (2017) In toto imaging of the migrating Zebrafish lateral line primordium at single cell resolution

    Developmental Biology 422:14–23.

    https://doi.org/10.1016/j.ydbio.2016.12.015

    • PubMed
    • Google Scholar
53. Software

    1. Nogare DED

    (2019) dallenogare2019, version cf258e8

    GitHub.

    https://github.com/chitnislabnih/dallenogare2019
54. 1. O'Brien GS
    2. Rieger S
    3. Wang F
    4. Smolen GA
    5. Gonzalez RE
    6. Buchanan J
    7. Sagasti A

    (2012) Coordinate development of skin cells and cutaneous sensory axons in zebrafish

    The Journal of Comparative Neurology 520:816–831.

    https://doi.org/10.1002/cne.22791

    • PubMed
    • Google Scholar
55. 1. Olson HM
    2. Nechiporuk AV

    (2020) Lamellipodia-like protrusions and focal adhesions contribute to collective cell migration in zebrafish

    Developmental Biology S0012-1606(20)30278-5.

    https://doi.org/10.1016/j.ydbio.2020.10.007

    • PubMed
    • Google Scholar
56. 1. Paluch EK
    2. Aspalter IM
    3. Sixt M

    (2016) Focal Adhesion-Independent cell migration

    Annual Review of Cell and Developmental Biology 32:469–490.

    https://doi.org/10.1146/annurev-cellbio-111315-125341

    • PubMed
    • Google Scholar
57. 1. Riedl J
    2. Crevenna AH
    3. Kessenbrock K
    4. Yu JH
    5. Neukirchen D
    6. Bista M
    7. Bradke F
    8. Jenne D
    9. Holak TA
    10. Werb Z
    11. Sixt M
    12. Wedlich-Soldner R

    (2008) Lifeact: a versatile marker to visualize F-actin

    Nature Methods 5:605–607.

    https://doi.org/10.1038/nmeth.1220

    • PubMed
    • Google Scholar
58. 1. Ruprecht V
    2. Wieser S
    3. Callan-Jones A
    4. Smutny M
    5. Morita H
    6. Sako K
    7. Barone V
    8. Ritsch-Marte M
    9. Sixt M
    10. Voituriez R
    11. Heisenberg CP

    (2015) Cortical contractility triggers a stochastic switch to fast amoeboid cell motility

    Cell 160:673–685.

    https://doi.org/10.1016/j.cell.2015.01.008

    • PubMed
    • Google Scholar
59. 1. Schindelin J
    2. Arganda-Carreras I
    3. Frise E
    4. Kaynig V
    5. Longair M
    6. Pietzsch T
    7. Preibisch S
    8. Rueden C
    9. Saalfeld S
    10. Schmid B
    11. Tinevez JY
    12. White DJ
    13. Hartenstein V
    14. Eliceiri K
    15. Tomancak P
    16. Cardona A

    (2012) Fiji: an open-source platform for biological-image analysis

    Nature Methods 9:676–682.

    https://doi.org/10.1038/nmeth.2019

    • PubMed
    • Google Scholar
60. 1. Soofi SS
    2. Last JA
    3. Liliensiek SJ
    4. Nealey PF
    5. Murphy CJ

    (2009) The elastic Modulus of Matrigel as determined by atomic force microscopy

    Journal of Structural Biology 167:216–219.

    https://doi.org/10.1016/j.jsb.2009.05.005

    • PubMed
    • Google Scholar
61. 1. Sun L
    2. Tran N
    3. Liang C
    4. Tang F
    5. Rice A
    6. Schreck R
    7. Waltz K
    8. Shawver LK
    9. McMahon G
    10. Tang C

    (1999) Design, synthesis, and evaluations of substituted 3-[(3- or 4-carboxyethylpyrrol-2-yl)methylidenyl]indolin-2-ones as inhibitors of VEGF, FGF, and PDGF receptor tyrosine kinases

    Journal of Medicinal Chemistry 42:5120–5130.

    https://doi.org/10.1021/jm9904295

    • PubMed
    • Google Scholar
62. 1. Szabó A
    2. Mayor R

    (2018) Mechanisms of neural crest migration

    Annual Review of Genetics 52:43–63.

    https://doi.org/10.1146/annurev-genet-120417-031559

    • PubMed
    • Google Scholar
63. 1. van der Walt S
    2. Schönberger JL
    3. Nunez-Iglesias J
    4. Boulogne F
    5. Warner JD
    6. Yager N
    7. Gouillart E
    8. Yu T
    9. scikit-image contributors

    (2014) scikit-image: image processing in Python

    PeerJ 2:e453.

    https://doi.org/10.7717/peerj.453

    • PubMed
    • Google Scholar
64. 1. Venkiteswaran G
    2. Lewellis SW
    3. Wang J
    4. Reynolds E
    5. Nicholson C
    6. Knaut H

    (2013) Generation and dynamics of an endogenous, self-generated signaling gradient across a migrating tissue

    Cell 155:674–687.

    https://doi.org/10.1016/j.cell.2013.09.046

    • PubMed
    • Google Scholar
65. 1. Vukicevic S
    2. Kleinman HK
    3. Luyten FP
    4. Roberts AB
    5. Roche NS
    6. Reddi AH

    (1992) Identification of multiple active growth factors in basement membrane matrigel suggests caution in interpretation of cellular activity related to extracellular matrix components

    Experimental Cell Research 202:1–8.

    https://doi.org/10.1016/0014-4827(92)90397-q

    • PubMed
    • Google Scholar
66. 1. Wang X
    2. Enomoto A
    3. Asai N
    4. Kato T
    5. Takahashi M

    (2016) Collective invasion of Cancer: perspectives from pathology and development

    Pathology International 66:183–192.

    https://doi.org/10.1111/pin.12391

    • PubMed
    • Google Scholar
67. Software

    1. Waskom M
    2. Botvinnik O
    3. Hobson P
    4. Cole JB
    5. Halchenko Y
    6. Hoyer S
    7. Miles A
    8. Augspurger T
    9. Yarkoni T
    10. Megies T
    11. Coelho LP
    12. Wehner D
    13. cynddl EZ
    14. Zaytsev YV
    15. Hoppe T
    16. Seabold S
    17. Cloud P
    18. Allan D

    (2014) diego0020: seaborn, version V0.5.0

    Zenodo.

    https://doi.org/10.5281/zenodo.12710
68. 1. Wong M
    2. Newton LR
    3. Hartmann J
    4. Hennrich ML
    5. Wachsmuth M
    6. Ronchi P
    7. Guzmán-Herrera A
    8. Schwab Y
    9. Gavin AC
    10. Gilmour D

    (2020) Dynamic buffering of extracellular chemokine by a dedicated scavenger pathway enables robust adaptation during directed tissue migration

    Developmental Cell 52:492–508.

    https://doi.org/10.1016/j.devcel.2020.01.013

    • PubMed
    • Google Scholar
69. 1. Wu Y
    2. Wawrzusin P
    3. Senseney J
    4. Fischer RS
    5. Christensen R
    6. Santella A
    7. York AG
    8. Winter PW
    9. Waterman CM
    10. Bao Z
    11. Colón-Ramos DA
    12. McAuliffe M
    13. Shroff H

    (2013) Spatially isotropic four-dimensional imaging with dual-view plane illumination microscopy

    Nature Biotechnology 31:1032–1038.

    https://doi.org/10.1038/nbt.2713

    • PubMed
    • Google Scholar
70. 1. Yolland L
    2. Burki M
    3. Marcotti S
    4. Luchici A
    5. Kenny FN
    6. Davis JR
    7. Serna-Morales E
    8. Müller J
    9. Sixt M
    10. Davidson A
    11. Wood W
    12. Schumacher LJ
    13. Endres RG
    14. Miodownik M
    15. Stramer BM

    (2019) Persistent and polarized global actin flow is essential for directionality during cell migration

    Nature Cell Biology 21:1370–1381.

    https://doi.org/10.1038/s41556-019-0411-5

    • PubMed
    • Google Scholar

Acceptance summary:

This beautiful imaging study shows that the posterior lateral line primordium in zebrafish is covered by a sheath of cells on its apical side that extend lamellipodial-like protrusion underneath the skin, very similar to the basal cells that extend lamellipodial-like protrusion on top of the muscle. Removal of the skin changes the protrusions – they become bleb-like and disoriented – and the primordium ceases to migrate directionally. Replacing the skin with Matrigel or sealing the wound partly or completely restore protrusions and migration.

Decision letter after peer review:

Thank you for submitting your article "A sheath of motile cells supports collective migration of the Zebrafish posterior lateral line primordium under the skin" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Didier Stainier as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

We would like to draw your attention to changes in our revision policy that we have made in response to COVID-19 (https://elifesciences.org/articles/57162). Specifically, when editors judge that a submitted work as a whole belongs in eLife but that some conclusions require a modest amount of additional new data, as they do with your paper, we are asking that the manuscript be revised to either limit claims to those supported by data in hand, or to explicitly state that the relevant conclusions require additional supporting data.

Our expectation is that the authors will eventually carry out the additional experiments and report on how they affect the relevant conclusions either in a preprint on bioRxiv or medRxiv, or if appropriate, as a Research Advance in eLife, either of which would be linked to the original paper.

Summary:

This is a beautiful imaging study that identifies a previous overlooked cell population in the posterior lateral line primordium in zebrafish – a model for collective cell migration. This cell population is sandwiched between the two-layered skin of the zebrafish embryo and the more basally located main cell population of the primordium. Similarly, to the basal cell population which extends protrusion on the basal side of the primordium, this new cell population also extends protrusions to the apical side between the skin and the primordium. Beautiful microsurgery experiments combined with high resolution imaging shows that skin removal causes the cells of the primordium to stall and change their protrusion type from lamellipodia-like to bleb-like, suggesting that the skin is required to maintain the correct chemical milieu and/or physical pressure to support the migration of the primordium. The identification of this cell population and the requirement of the skin for migration are important contribution to our understanding of tissue migration.

Overall, this manuscript is well written and compelling. We have the following suggestions to improve the manuscript.

Essential revisions:

1) Confinement of the primordium. The reviewers agree that the height of the primordium cells with and without skin should be measured to see if the primordium cells are higher along the medial-lateral axis after skin removal. If the cells are higher, this would be in agreement with the mechanical confinement idea. However, all reviewers feel that the skin removal experiments do not show that confinement is essential for primordium migration and this claim should be softened and other possibilities should be discussed (changes in the environment, wounding response etc).

2) Sheath cells.

a) The reviewers agree that it is unclear (and unlikely) that the sheath cells are a separate cell population. The authors should rename this cell population using a more descriptive term such as apical cells.

b) The claim that the sheath cells are supporting collective migration should be toned down since the authors do not direct evidence for this.

c. The claim that the sheath cells are mesenchymal should be softened since they are labeled by a epithelial promoter.

3) Discussion. The Discussion should be more focused and concise as laid out in the individual reviewer's comments.

Reviewer #1:

The lateral line primordium (LLP) of zebrafish has long been a premier system to study the molecular and cellular mechanisms that underly collective cell migration. In this manuscript, Dalle Nogare et al. identify a new cell population that is essential for the movement of the LLP. Specifically, they describe a superficial sheet of cells that lies between the main primordium and the skin. These cells extend polarized protrusions in the direction of tissue movement and appear to work in concert with the protrusions made by deeper cells to allow the LLP to generate tractions forces against both the overlying skin and the underlying somites simultaneously. They further show that, like the deeper cells, the superficial cells require FGF signaling to orient these protrusions. They also perform a clever set of experiments in which they remove sections of the skin and then allow it to heal back over the LLP to show a strong correlation between the presence of the skin, the protrusive behavior of both the deep and superficial cells, and the migratory ability of the LLP as a whole.

Overall, I think this paper is a strong candidate for publication in eLife. Although the study is largely descriptive in nature, the discovery of this new cell population is important, as it fundamentally changes the way we think about how motive forces are generated in this collective. Moreover, the skin healing experiments are elegant and quite revealing of the underlying biology.

1) I am unconvinced that the newly described cell population is mesenchymal, as proposed by the authors. Since a claudin promoter is used to label the cells, it seems more likely that they are epithelial.

2) Results – The authors point out that the rate of retrograde actin flow correlates with traction forces in migrating cells. Please provide some context for the flow rate measured in this paper. How does it compare to other cell types and how can this information help us better understand the underlying biology?

3) Figure 6C – Wild-type control data should be added to this figure.

Reviewer #2:

This insightful manuscript shows that the zebrafish posterior lateral line primordium (PLLp) contains protruding cells around its entire circumference, including on the overlying skin. These cells are structurally distinct from mesenchymal leading cells and epithelialized neuromast-forming cells. Technically impressive skin removal experiments with high-resolution imaging demonstrate that the skin has a previously undescribed role in PLLp protrusion formation and migration, demonstrating that the cluster of migrating cells walks on the "ceiling" as well as on the "floor". These striking results will certainly have an impact on the field. However, we feel that some textual revision is necessary for the conclusions to reflect the data presented. In particular, these experiments do not definitively show that the skin's role depends on confinement. We recommend concluding further explanation that the skin's role may be mediated by confinement, adhesion, tissue integrity signaling, or a combination thereof (see comment #1). Our other suggestions could strengthen the paper but are not required.

1) The skin removal experiments require a more nuanced interpretation of the variety of mechanism(s) by which skin may contribute to PLLp migration. We ask that the authors address the following points in the Discussion and Abstract:

a) The term "confinement" needs to be used more precisely. In each condition of the skin removal experiments, the primordium is in contact with a different material with unique mechanical and chemical properties. We caution the authors from referring to embryos embedded in agarose as "unconfined", as agarose often acts as a form of confinement for cultured cells. Perhaps the PLLp experiences weaker forces or geometric constraints while embedded in agarose versus contained within the skin, but the authors should more clearly acknowledge these unknown mechanical parameters that are inherent to the experiment (Results and Discussion).

The authors could additionally choose to strengthen the mechanical comparison by measuring the height of the PLLp under each condition--one would expect that taller cells are less confined. This measurement should be possible by further analysis of image data that the authors have already collected.

b) The possible role of adhesions should be elevated. We find this model interesting, and the slow migration under Matrigel does not technically prove that adhesion is insufficient for PLLp migration, as the authors correctly note in the Discussion (eighth paragraph). It is possible that growth factors in Matrigel alter chemotaxis, counteracting the pro-migratory effect of lamellipodial formation.

If experimental follow-ups were feasible, we would propose observing transplanted basal cells under Matrigel confinement, in order to determine whether superficial lamellipodia rescue basal cryptic lamellipodia with this perturbation.

c) Tissue integrity should be discussed as a potential contributor to the skin effect. There are multiple signalling pathways triggered by breaking the skin barrier, including calcium signaling, reactive oxygen species, purinergic signaling, and cell swelling (c.f. Enyedi and Niethammer (2015). Trends in Cell Biology 25(7): 398-407), any of which could potentially also interfere with PLLp migration. Because protrusions resume on cells near the front of the primordium as soon as they come into contact with the skin, but directional movement does not actually start up again until the skin is fully healed, it seems plausible that the cells are detecting both the adhesive contact and also some other property of tissue integrity.

The experiments shown in Figure 4 could shed light on this hypothesis if the authors were to additionally measure cell speed after the skin has entirely closed (and presumably skin integrity restored) and compare that with speeds before/after cell contact with skin (Figure 4G).

2) When defining the sheath of motile cells, we suggest including a more thorough analysis of the entire PLLp population, rather than just the cells above apical constrictions (Figure 2). Specifically, the authors define the relevant cell population as those which are in direct physical contact with the skin and also whose nuclei are above a rosette. Why are both criteria necessary? Another perspective on the functional roles of the two distinct populations of actively motile cells in migration of the primordium might explore more completely the relative contributions of those that are walking on the "floor" versus the "ceiling". For example, what proportion of cells are touching the skin versus the basal substrate? Of these populations, how many cells have apical constrictions? The authors could better contextualize these results by citing any previous ultrastructural work on the PLLp, or other work that would shed light on the arrangement of cells within and around the PLLp. Also, the authors should give more prominence to their striking observation that the skin was removed without the PLLp 95% of the time. This result suggests functional importance of differential adhesion among the different cell types. Analogous findings on the significance of differential cell adhesion have been reported in other developmental contexts, for example Tsai et al., bioRxiv 803635.

3) While there are a lot of interesting issues raised in the Discussion, we found it to be unfocused and confusing. We suggest focusing on two key trends in order to consolidate the studies in the Discussion and provide a more cogent summary of how their new results fit in with previous work. First, in both this and previous studies, an environment with more restrictive geometry and/or more adhesivity promotes cell polarity and persistence (Ruprecht et al., 2015; Liu et al., 2015; Bergert et al., 2012; Ramalingam et al., 2015 Nat Comm 6:8496; Logue et al., 2015 eLife 4:e08314; Graziano et al., 2019 PLoS Biol 17(10): e3000457). The second key trend is that protrusion type is determined by the interplay of adhesion, confinement, contractility, and polymerization. In agreement with this study, Ruprecht et al., Liu et al., Bergert et al., show that low adhesivity promotes blebs over lamellipodia. However, two Dictyostelium studies show a protrusion type trend that is less consistent with this study, because Dicty produces more blebs under greater confinement (Ibo et al., 2016; Srivastava et al., 2020 PNAS 117 (5) 2506-2512). The authors note this, but do not explicitly address differential adhesion. It might be helpful for the authors to express their opinions about the limitations of comparing different cell types, particularly in the context of the ~billion year separation between cellular slime molds and vertebrate animals. We are not convinced by the statement that the results of Bergert et al., 2012, are "less consistent" with the observations in the primordium than the other works cited. In particular, the result that the primordium produced more lamellipodia on Matrigel than on agarose is consistent with the Bergert et al. experiment discussed by the authors, which shows that Walker 256 cells produce more lamellipodia on adhesive regions as compared to non-adhesive regions (Figure 5 in Berget et al.).

In the present version of the manuscript, the Abstract emphasizes the importance of confinement, but the results are also consistent with a more nuanced interpretation where confinement acts along with adhesion and signaling in determining cell behavior.

Reviewer #3:

The manuscript by Dalle Nogare and colleagues presents two interesting ideas. The first is that a collectively migrating tissue, the zebrafish PLLp, uses a 'circumferential' or 'chimneying' motility mechanism, where traction is generated on underlying muscles and overlying skin. This novel idea is supported with cutting-edge, high resolution live imaging data. The second is that the 'confinement' of the tissue between muscles and skin regulates the type of protrusions that these motile cells generate in vivo, an idea supported with technically-impressive skin removal experiments. The study is well written and presented and the findings are potentially of sufficient importance and novelty to justify publication in eLife. However, neither conclusion is sufficiently supported by the data presented. If the authors can strengthen EITHER conclusion with additional experimental support, they will make an important contribution to the field.

1) Stunning time-lapse data reveal that cells on the apical surface of the PLLp display protrusions that are consistent with the idea that they are migrating on the overlying skin layer. However, imaging data alone are insufficient to conclude that this motile 'sheath' of cells supports the collective migration of the PLLp, as stated in the title. The only experimental test of this idea – removal of the overlying skin – results in a very strong defect in protrusions on both sides of the primordium, and thus does not reveal a specific requirement for the newly discovered apical motility.

Recommendation 1: Provide more direct evidence for the proposal that apical cells support PLLp migration by interacting with the overlying skin. Ideally, the authors could use their genetic mosaic technique to target these cells directly (rather than disrupting the skin). Alternatively, it may be possible to apply correlative approaches on pre-existing imaging data to further support the suggestion that the dynamic behavior of sheath cells has an influence on tissue migration. For example, do changes in sheath cell behavior, perhaps when they move from one skin cell to the next, have any effect on the migration of the PLLp (e.g. shearing) as might be expected from a chimneying model?

Comment: It may be argued that the imaging data, which clearly show the motile behavior of apical sheath cells on the skin, are sufficient to demonstrate the existence of circumferential motility mechanism. In that case, the authors should tone down their claim that these cells 'support' collective migration.

2) The other major claim is that the study 'highlights the broad role confinement plays in the formation of polarized lamellipodia'. While 'confinement' is not clearly defined, the discussion centres around proposed roles for mechanical confinement in regulating protrusion types that have emerged from in vitro model systems. Here, the support for a role for mechanical confinement comes from skin removal experiments. The issue is that this experiment does more than reduce mechanical confinement, it also severely disrupts what is normally a closed and controlled chemical environment. It's hardly surprising that cells that normally migrate in the presence of patterned diffusible cues and organized matrices suddenly arrest when their closed and controlled environment is forced open.

Recommendation 2: In order to support the proposed role for mechanical confinement in the formation of lamellipodia the authors should rule out that the observed changes aren't the result of chemical effects such as osmotic changes. Do cells of the rosette structures – that don't normally show apical protrusions – start to bleb apically after skin removal? Alternatively, more direct evidence for the proposed role for mechanical confinement in regulating protrusion type could come from experiments that emulate in vitro studies where 'purely mechanical' changes can be applied (by placing cells under agarose pads, for example). The finding that the authors can replace some aspects of the skin with Matrigel is a promising step in this direction. Did the authors consider modulating the elastic modulus of the Matrigel via different concentrations to address the impact of altered mechanics more specifically? The choice of Matrigel concentration is not explained in the Materials and methods section. Better still, the authors could consider replacing Matrigel with agarose to alter mechanical confinement in a way analogous to the under agarose in vitro experiments.

Dear Editors and Reviewers,

Thank you for your thorough consideration and review of our paper. To address the concerns of the reviewers (listed under the “essential revisions”), we have made the following modifications. In some cases, we have performed additional experiments, however due to the restricted work schedule we are under because of COVID-19, we could not experimentally address all of the reviewers' criticisms. However, we hope the revised manuscript satisfies the majority of the reviewers’ concerns.

Additional experiments and the role of superficial cells In addition to the experiment described above to examine confinement of the PLLp, we have also performed the experiment suggested by reviewer #2 looking at basal protrusions in skinned but matrigel embedded embryos. We report that, unlike superficial protrusions, the directionality of basal protrusions is rescued when embryos are embedded in matrigel. This surprising result further argues against a purely mechanical confinement mechanism (as recovery of polarization is not observed after agarose embedding), and so we have, as discussed above, softened the relevant conclusions. However, despite recovery of polarized basal protrusions, directed primordium migration is not recovered in matrigel embedded embryos. We feel that this failure to migrate, when basal protrusions are polarized but superficial protrusions are not, provides stronger support for the necessity of polarized superficial protrusions in migration. We have discussed this experiment in the relevant portions of the results and discussion. In addition, we would like to address the following comments by the reviewers directly .

Reviewer #1

Actin flow rates: We have added some discussion of actin flow rates in a variety of other contexts to better situate our measurements. As we note, measurement of Actin retrograde flow rates in vivo are very rare, and often significantly different from those observed in in vitro studies, and so we refrain from interpreting too much from our measurement, other than situating it in context of other measured rates (both in vivo and in vitro).

Reviewer #2

As discussed above, we have made significant changes to the way we discuss “confinement” in the manuscript, including softening conclusions about the role of confinement, given the new results reported in Figure 7. We agree with the reviewer that we could have done a more thorough job in describing the nature of this population. To this end, we have added a section in the results describing external contacts of each cell population, which is further explained and quantified in a new Figure S2. In addition, we have provided two examples of the morphology of leading, protoneuromast, and superficial cells in a new Figure 2 – figure supplement 1. We feel that these new figures and results text will do a better job giving context to the reader. We have rewritten parts of the results and discussion to increase clarity, and included a supplementary table showing the number of embryos and movies used for each quantification. In addition, we have added a new supplementary figure with examples of each type of quantification used (actin rods in a variety of contexts, retrograde flow kymographs). We have also added a discussion of these measurements to the main text.

Reviewer #3

We have made all of the modifications in the text and figures suggested by this reviewer. We hope that the additional experiments described help to alleviate their major concerns, however we regret that we were unable to perform some of the experiments suggested. We have, in these cases, toned down the relevant conclusions. Some experiments (for example modulating the elastic modulus of matrigel) we plan to perform in upcoming studies. We agree with the reviewer that figure S1 showing a track of 5 nuclei is underwhelming. We have replaced Figure S1 with a more robust figure and movie (Figure 2 – figure supplement 2, Movie S3). In this figure, we segment, 3D reconstruct and track every nucleus in a primordium over two hours. This data is used to construct a new figure showing the averaged nuclear position of all nuclei in a 120 micron cross section of the PLLp (centered around a neuromast). We have also included a reconstruction movie of all nuclei in the PLLp over this time course, color coded by cell type, and showing the cross section and position heatmaps as movie S3, and the area used to generate the position heatmap is indicated in this movie. We also agree that we do not have complete enough data to make firm statements about the ultimate fate of these cells. Although we strongly suspect that they become interneuromast cells, we have not conclusively demonstrated this, and so have removed this claim from the paper.

Essential revisions:

1) Confinement of the primordium. The reviewers agree that the height of the primordium cells with and without skin should be measured to see if the primordium cells are higher along the medial-lateral axis after skin removal. If the cells are higher, this would be in agreement with the mechanical confinement idea.

While we could extract some values from our existing data, we felt the better approach was to perform an additional experiment to quantify this potential change. We have therefore performed additional paired measurements of the height of the PLLp in primordia before and after skin removal. These data are discussed in the Results section, and presented in Figure 4. Removal of the skin results in an almost 20% increase in average primordium height (p = 6e^-7, paired t-test). Note that for this measurement, we could not robustly measure individual cell height paired between skin intact and skin removed, and so we have devised an average measure of primordium height, measured at 4 equally spaced points along the length of the primordium, as detailed in the Materials and methods.

However, all reviewers feel that the skin removal experiments do not show that confinement is essential for primordium migration and this claim should be softened and other possibilities should be discussed (changes in the environment, wounding response etc).

Despite the results described above suggesting that the primordium is in fact under mechanical confinement by the skin, we agree with the reviewers critique that our experiments do not adequately demonstrate that mechanical confinement, per se, is essential for migration. We have toned down the relevant conclusions and expanded discussion of additional ways confinement by the skin could contribute to migration.

2) Sheath cells.

a) The reviewers agree that it is unclear (and unlikely) that the sheath cells are a separate cell population. The authors should rename this cell population using a more descriptive term such as apical cells.

We agree with the reviewers that this population need not represent a separate cell population with a distinct fate. It could, for example, represent the persistence of mesenchymal-like cells that are present in the leading domain. We have rewritten the relevant sections to provide better context for this population, and now refer to these cells as “superficial primordium cells”, rather than “sheath cells”, which inadvertently suggested a distinct cell type. We have, in a small number of places, retained the use of the word “sheath” to describe the overall structure as we feel this is a good intuitive analogy to their position wrapping around the top and sides of the primordium.

b) The claim that the sheath cells are supporting collective migration should be toned down since the authors do not direct evidence for this.

c) The claim that the sheath cells are mesenchymal should be softened since they are labeled by a epithelial promoter.

With regards to the mesenchymal/epithelial nature of these cells, we agree and have changed the text to reflect the ambiguity about the nature of these cells, and included a caveat that they are labeled by an epithelial promoter. We also note that the leading cells of the primordium, commonly referred to in the literature of “mesenchymal” or “mesenchymal-like” are also labelled by the cldnB promoter. We have changed text referring to these types of cells as “mesenchymal-like” rather than “mesenchymal” and discuss them now in the context of polarized epithelial cells incorporated into rosettes as opposed to cells in the leading domain and superficial cells that are not incorporated into a neuromast.

3) Discussion. The Discussion should be more focused and concise as laid out in the individual reviewer's comments.

We have significantly rewritten the discussion and hope that is more precise and clearer than the initially submitted manuscript. In addition, to address the reviewers concerns about the legibility of the figures, we have made several changes to font sizes etc to enhance readability. To this end, we have split Figure 5 (which had panels A-Zz) into a new Figure 5 and Figure 6 to allow for larger panels and a re-designed figure.

Reviewer #1:

Overall, I think this paper is a strong candidate for publication in eLife. Although the study is largely descriptive in nature, the discovery of this new cell population is important, as it fundamentally changes the way we think about how motive forces are generated in this collective. Moreover, the skin healing experiments are elegant and quite revealing of the underlying biology.

1) I am unconvinced that the newly described cell population is mesenchymal, as proposed by the authors. Since a claudin promoter is used to label the cells, it seems more likely that they are epithelial.

We have changed the use of “mesenchymal” in the paper, and have added a section noting that these cells are labelled with the claudinB promoter

2) Results – The authors point out that the rate of retrograde actin flow correlates with traction forces in migrating cells. Please provide some context for the flow rate measured in this paper. How does it compare to other cell types and how can this information help us better understand the underlying biology?

We have added some discussion of other flow rates measured in a variety of contexts to this part of the Results section and attempted to do a better job contextualizing this flow rate

3) Figure 6C – Wild-type control data should be added to this figure.

We have added wild-type (skin intact) data, from embryos embedded both in agarose and matrigel, to this figure.

Reviewer #2:

[…] We recommend concluding further explanation that the skin's role may be mediated by confinement, adhesion, tissue integrity signaling, or a combination thereof (see comment #1). Our other suggestions could strengthen the paper but are not required.

1) The skin removal experiments require a more nuanced interpretation of the variety of mechanism(s) by which skin may contribute to PLLp migration. We ask that the authors address the following points in the Discussion and Abstract:

a) The term "confinement" needs to be used more precisely. In each condition of the skin removal experiments, the primordium is in contact with a different material with unique mechanical and chemical properties. We caution the authors from referring to embryos embedded in agarose as "unconfined", as agarose often acts as a form of confinement for cultured cells. Perhaps the PLLp experiences weaker forces or geometric constraints while embedded in agarose versus contained within the skin, but the authors should more clearly acknowledge these unknown mechanical parameters that are inherent to the experiment (Results and Discussion).

We agree with the reviewer and have modified the text to more precisely use the word “confinement”.

The authors could additionally choose to strengthen the mechanical comparison by measuring the height of the PLLp under each condition--one would expect that taller cells are less confined. This measurement should be possible by further analysis of image data that the authors have already collected.

We have performed experiments to measure this, as discussed earlier, and now show that the primordium is ~15-20% taller on average when the skin is removed than when it is in intact. As we feel these data add more to the narrative than the current Figure 4H (which showed data to support Figure 4G but didn’t add anything new), we have made these height measurements Figure 4H and removed the old Figure 4H.

b) The possible role of adhesions should be elevated. We find this model interesting, and the slow migration under Matrigel does not technically prove that adhesion is insufficient for PLLp migration, as the authors correctly note in the Discussion (eighth paragraph). It is possible that growth factors in Matrigel alter chemotaxis, counteracting the pro-migratory effect of lamellipodial formation.

We have added discussion of the potential role of adhesions and other potential matrigel-related affects to the Discussion.

If experimental follow-ups were feasible, we would propose observing transplanted basal cells under Matrigel confinement, in order to determine whether superficial lamellipodia rescue basal cryptic lamellipodia with this perturbation.

We have performed these experiments, and now show that basal protrusions are in fact recovered after matrigel embedding. Surprisingly, we also show that these basal protrusions (unlike the superficial protrusions) are in fact polarized in matrigel. This may, in part, explain the limited recovery of migratory ability in these primordia when compared to primordia embedded in agarose. The reason for this asymmetric recovery is unclear.

c) Tissue integrity should be discussed as a potential contributor to the skin effect. […] The experiments shown in Figure 4 could shed light on this hypothesis if the authors were to additionally measure cell speed after the skin has entirely closed (and presumably skin integrity restored) and compare that with speeds before/after cell contact with skin (Figure 4G).

2) When defining the sheath of motile cells, we suggest including a more thorough analysis of the entire PLLp population, rather than just the cells above apical constrictions (Figure 2). […] Analogous findings on the significance of differential cell adhesion have been reported in other developmental contexts, for example Tsai et al., bioRxiv 803635.

We have now performed a more thorough analysis of the nature of this cell population from segmented data, and answered several of these questions. We have redesigned Figure 2—figure supplement 1 to include some of these data on cell morphology and external contacts.

3) While there are a lot of interesting issues raised in the Discussion, we found it to be unfocused and confusing. […] In particular, the result that the primordium produced more lamellipodia on Matrigel than on agarose is consistent with the Bergert et al. experiment discussed by the authors, which shows that Walker 256 cells produce more lamellipodia on adhesive regions as compared to non-adhesive regions (Figure 5 in Berget et al.).

We have rewritten the Discussion to attempt to address these points.

In the present version of the manuscript, the Abstract emphasizes the importance of confinement, but the results are also consistent with a more nuanced interpretation where confinement acts along with adhesion and signaling in determining cell behavior.

We agree with the reviewer, and have modified the Results and Discussion to better discuss other interpretations consistent with the data.

Reviewer #3:

1) Stunning time-lapse data reveal that cells on the apical surface of the PLLp display protrusions that are consistent with the idea that they are migrating on the overlying skin layer. However, imaging data alone are insufficient to conclude that this motile 'sheath' of cells supports the collective migration of the PLLp, as stated in the title. The only experimental test of this idea – removal of the overlying skin – results in a very strong defect in protrusions on both sides of the primordium, and thus does not reveal a specific requirement for the newly discovered apical motility.

Recommendation 1: Provide more direct evidence for the proposal that apical cells support PLLp migration by interacting with the overlying skin. Ideally, the authors could use their genetic mosaic technique to target these cells directly (rather than disrupting the skin). Alternatively, it may be possible to apply correlative approaches on pre-existing imaging data to further support the suggestion that the dynamic behavior of sheath cells has an influence on tissue migration. For example, do changes in sheath cell behavior, perhaps when they move from one skin cell to the next, have any effect on the migration of the PLLp (e.g. shearing) as might be expected from a chimneying model?

Comment: It may be argued that the imaging data, which clearly show the motile behavior of apical sheath cells on the skin, are sufficient to demonstrate the existence of circumferential motility mechanism. In that case, the authors should tone down their claim that these cells 'support' collective migration.

We have tried experiments similar to the ones suggested by the reviewer, but have been unable to target these cells with the precision necessary for these experiments, without also damaging surrounding tissues. However, we also now demonstrate that when re-embedded in matrigel basal, but not superficial, protrusion directionality is restored, while migration is still profoundly affected. We argue that this supports our conclusion that polarized superficial protrusive activity is necessary for migration. However, we recognize the limitations of the evidence provided and have we have toned down our claim that these cells support collective migration and added some necessary caveats.

2) The other major claim is that the study 'highlights the broad role confinement plays in the formation of polarized lamellipodia'. While 'confinement' is not clearly defined, the discussion centres around proposed roles for mechanical confinement in regulating protrusion types that have emerged from in vitro model systems. Here, the support for a role for mechanical confinement comes from skin removal experiments. The issue is that this experiment does more than reduce mechanical confinement, it also severely disrupts what is normally a closed and controlled chemical environment. It's hardly surprising that cells that normally migrate in the presence of patterned diffusible cues and organized matrices suddenly arrest when their closed and controlled environment is forced open.

Recommendation 2: In order to support the proposed role for mechanical confinement in the formation of lamellipodia the authors should rule out that the observed changes aren't the result of chemical effects such as osmotic changes. […] Better still, the authors could consider replacing Matrigel with agarose to alter mechanical confinement in a way analogous to the under agarose in vitro experiments.

We have considered these experiments, but felt that a more systematic exploration of the mechanical properties of artificial substrates would be better described in a separate manuscript that focuses on these parameters. We have added a sentence describing the rationale for using 20mg/mL matrigel (this was the undiluted concentration, and we reasoned that we would like to provide as stiff a substrate as possible for primordium migration) in the appropriate place in the Results.

Author details

1. Damian E Dalle Nogare

   Section on Neural Developmental Dynamics, Division of Developmental Biology, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, United States

   Contribution

   Conceptualization, Data curation, Software, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review and editing

   For correspondence

   damian.dallenogare@nih.gov

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-9615-9299

2. Naveen Natesh

   Section on Neural Developmental Dynamics, Division of Developmental Biology, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, United States

   Present address

   Center for Genomics and Computational Biology and Department of Biomedical Engineering, Duke University, Durham, United States

   Contribution

   Investigation, Writing - review and editing

   Competing interests

   No competing interests declared

3. Harshad D Vishwasrao

   Advanced Imaging and Microscopy Resource, National Institutes of Health, Bethesda, United States

   Contribution

   Resources, Software, Writing - review and editing

   Competing interests

   No competing interests declared

4. Hari Shroff

   1. Advanced Imaging and Microscopy Resource, National Institutes of Health, Bethesda, United States
   2. Laboratory of High Resolution Optical Imaging, National Institute of Biomedical Imaging and Bioengineering, National Institutes of Health, Bethesda, United States

   Contribution

   Resources

   Competing interests

   No competing interests declared

5. Ajay B Chitnis

   Section on Neural Developmental Dynamics, Division of Developmental Biology, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, United States

   Contribution

   Resources, Supervision, Funding acquisition, Investigation, Project administration, Writing - review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-6423-6624

• 2,886

  Page views

• 234

  Downloads

• 10

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.