1. Cell Biology
  2. Developmental Biology

Notch signaling restricts FGF pathway activation in parapineal cells to promote their collective migration

  1. Lu Wei
  2. Amir Al Oustah
  3. Patrick Blader
  4. Myriam Roussigné  Is a corresponding author
  1. Université de Toulouse, CNRS (UMR 5547), France
https://doi.org/10.7554/eLife.46275
Share this article
Cite this article
  1. Lu Wei
  2. Amir Al Oustah
  3. Patrick Blader
  4. Myriam Roussigné
(2019)
Notch signaling restricts FGF pathway activation in parapineal cells to promote their collective migration
eLife 8:e46275.
https://doi.org/10.7554/eLife.46275
  2. Download BibTeX
  3. Download .RIS

Abstract

Coordinated migration of cell collectives is important during embryonic development and relies on cells integrating multiple mechanical and chemical cues. Recently, we described that focal activation of the FGF pathway promotes the migration of the parapineal in the zebrafish epithalamus. How FGF activity is restricted to leading cells in this system is, however, unclear. Here, we address the role of Notch signaling in modulating FGF activity within the parapineal. While Notch loss-of-function results in an increased number of parapineal cells activating the FGF pathway, global activation of Notch signaling decreases it; both contexts result in defects in parapineal migration and specification. Decreasing or increasing FGF signaling in a Notch loss-of-function context respectively rescues or aggravates parapineal migration defects without affecting parapineal cells specification. We propose that Notch signaling controls the migration of the parapineal through its capacity to restrict FGF pathway activation to a few leading cells.

https://doi.org/10.7554/eLife.46275.001

eLife digest

Many animal cells must move through the body from the place they are born to where they are needed, for example, when embryos are developing or wounds are healing. Often cells migrate in groups, which helps them to navigate and co-ordinate more effectively. Cells typically migrate by sensing different signals across large areas, and groups of cells communicate with each other to co-ordinate their migration.

One group of cells that is studied to understand collective cell migration is found in the brain of the zebrafish. These cells are called parapineal cells and, in the developing fish, they move towards the left side of the brain under the influence of a signal called Fgf8. Although all parapineal cells can detect Fgf8, only those at the front of the migrating group respond to the signal. These are the cells that lead the migration. It was previously unclear how the capability of responding to Fgf8 is limited to only a few parapineal cells to ensure they all migrate correctly.

By studying parapineal cell migration, Wei et al. identified a different signaling system called Notch as a regulator of cell migration. When Notch signaling activity is artificially increased in the brain, parapineal cells do not respond to Fgf8 as efficiently as when levels of Notch are normal. Conversely, the number of parapineal cells that respond to Fgf8 increases when Notch signaling is lost. In both cases, migration of parapineal cells is affected. Therefore, Wei et al. showed that changing the balance of Notch signaling in the zebrafish brain modifies the ability of parapineal cells to respond to Fgf8 signal and stops parapineal cells from migrating correctly.

These results provide a general model for how cells migrate as a group. More studies are needed to see if similar mechanisms are involved in other examples of collective cell migration. This model of group migration could be applied to healthy processes such as embryonic development as well as examples of cell migration in illness such as cancer metastasis.

https://doi.org/10.7554/eLife.46275.002

Introduction

Coordinated migration of cell collectives is a widespread phenomenon, being seen predominantly during embryonic development but also during tissue repair in adults, for example. The molecular and cellular mechanisms underlying collective cell migration have been studied in vivo in different model organisms (Friedl and Gilmour, 2009; Ochoa-Espinosa and Affolter, 2012; Pocha and Montell, 2014; Theveneau and Mayor, 2013). Recent progress in the analysis of mechanical forces together with the development of in vitro models and in silico modeling have improved our understanding of coordinated cell migration. Such studies have highlighted the variability in mechanisms from one model to another, indicating that collective migration is a highly adaptive and plastic process (Haeger et al., 2015; Theveneau and Linker, 2017).

Members of the FGF family of secreted signals have been implicated in many models of cell migration. For example, FGF signaling is described to promote migration of cell collectives, potentially through chemotaxis (Kadam et al., 2012), through the modulation of cell adhesiveness (Ciruna et al., 1997; McMahon et al., 2010) or by increasing random cell motility (Bénazéraf et al., 2010). In the lateral line primordium, the FGF pathway is required for Notch-dependent formation of neuromast rosettes at the trailing edge of the migrating primordium (Durdu et al., 2014; Kozlovskaja-Gumbrienė et al., 2017; Lecaudey et al., 2008; Nechiporuk and Raible, 2008) and for a leading-to-trailing signaling that prevents splitting of the primordium (Dalle Nogare et al., 2014), with both of these processes being required for proper lateral line primordium migration. Despite the widespread and iterative role of the FGF pathway in cell migration models, however, it is not clear how the dynamics of FGF signaling correlate with cell behaviors and how this can be modulated by other signals.

The parapineal is a small group of cells that segregates from the anterior part of the pineal gland at the midline of the zebrafish epithalamus and migrates in an FGF-dependent manner to the left side of the brain (Concha et al., 2000; Duboc et al., 2015; Roussigne et al., 2012). To characterize the dynamics of FGF pathway activation during parapineal migration, we recently analyzed the temporal and spatial activation of a previously described FGF pathway reporter transgene, Tg(dusp6:d2GFP) (Molina et al., 2007; Roussigné et al., 2018). Using this reporter, we showed that the FGF pathway is activated in an Fgf8-dependant manner in only a few parapineal cells located at the migration front and that experimentally activating the FGF pathway in a few parapineal cells restores parapineal migration in fgf8-/- mutant embryos. Taken together, these findings indicate that the restricted activation of FGF signaling in the parapineal promotes the migration of the parapineal cell collective. While the parapineal can receive Fgf8 signals from both sides of the midline, focal pathway activation is primarily detected on the left (Roussigné et al., 2018). This asymmetry in FGF pathway activation requires the TGFβ/Nodal signaling pathway, which is activated on the left side of the epithalamus prior to parapineal migration (Bisgrove et al., 1999; Concha et al., 2000; Liang et al., 2000; Roussigné et al., 2018). Although the Nodal pathway appears to bias the focal activation of FGF signaling to the left, after a significant delay the restriction of FGF activity still occurs in the absence of Nodal activity and the parapineal migrates (Roussigné et al., 2018).

All parapineal cells appear competent to activate the FGF pathway begging the question as to how the activation of the pathway is restricted to only a few cells. In this study, we address whether Notch signaling might modulate the activation of FGF pathway in the parapineal. We show that while loss-of-function of Notch leads to expanded FGF pathway activation in the parapineal, activating the Notch pathway causes a strong reduction in the expression of the FGF reporter transgene, with both contexts leading to defects in parapineal migration. Loss or gain of function for Notch signaling also interferes with the specification of parapineal cell identity; loss-of-function results in a significant increase in the number of gfi1ab and sox1a expressing parapineal cells, whereas gain of Notch activity results in the opposite phenotype. In contrast, the number of parapineal cells expressing tbx2b, a putative marker for parapineal progenitors cells (Snelson et al., 2008), is not affected in either loss or gain of function for Notch. Pharmacological inhibition of Notch pathway activity suggests that the roles of Notch in the specification and migration of parapineal cells can be uncoupled. Finally, a global decrease or an increase in the level of FGF signaling can respectively rescue or aggravate the parapineal migration defect caused by Notch loss-of-function but without affecting the specification of parapineal cells. Our data indicate that the Notch pathway regulates the specification and migration of parapineal cells independently and that the role of Notch signaling in promoting parapineal migration, but not specification, depends on its ability to restrict FGF pathway activation to a few parapineal cells.

Results

The parapineal of mindbomb mutant embryos display expanded FGF pathway activation

In models of cell migration during sprouting of tubular tissues, Notch-Delta mediated cell-cell communication contributes to tip cell selection by restricting the ability of followers cells to activate RTK signaling (Ghabrial and Krasnow, 2006; Ikeya and Hayashi, 1999; Riahi et al., 2015; Siekmann and Lawson, 2007a). To address whether Notch signaling could similarly restrict FGF pathway activation in the freely moving group of parapineal cells, and thus promote its migration, we analyzed the expression of an FGF pathway activity reporter transgene, Tg(dusp6:d2EGFP) (Molina et al., 2007), in embryos mutant for the mindbomb (mibta52b) gene, a well-described loss-of-function context for the Notch pathway (Itoh et al., 2003). At 32 hours post-fertilization (hpf), we observed a larger number of Tg(dusp6:d2EGFP) expressing cells in the parapineal of mib-/- mutant embryos (7 ± 3 d2EGFP+ cells) compared to siblings (4 ± 2 d2EGFP+ cells; p-value=3.845e-05) (Figure 1A–1B’ and E). While the number of Tg(dusp6:d2EGFP)+ cells increases, we found that the mean intensity of Tg(dusp6:d2EGFP) does not change significantly in mib-/- mutants indicating that it is the restriction of FGF activation to few parapineal cells rather than the level of FGF activity that is affected in mib-/- mutants (Figure 1—figure supplement 1). The increase in Tg(dusp6:d2EGFP) expressing cells was not accompanied by an increase in the number of parapineal cells expressing sox1a, the earliest described parapineal specific marker (detected from 28 hpf in the parapineal, Clanton et al., 2013) (Figure 1C–1D and G), although we observed a slight increase in the total number of parapineal cells as determined using nuclear staining to visualize the parapineal rosette (22 ± 7 in mib-/- mutants compared to 19 ± 5 in sibling control embryos; p-value=0.037) (Figure 1F).

Figure 1 with 3 supplements see all
Download asset Open asset
Increased activation of FGF signaling in mindbomb mutants correlates with defects in migration and specification of parapineal cells.

(A–D) Confocal sections showing the expression of the Tg(dusp6:d2EGFP) transgene (Green, (A–B’) or sox1a (red, (C–D) in the epithalamia of 32 hpf in control (A, A’, n = 36 and C, n = 30) and in mib-/- mutant embryos (B, B’, n = 34 and D, n = 21). Confocal sections are merged with a nuclear staining (gray, A’, B’, C, D) that makes the epiphysis (white circle) and parapineal (yellow circle) visible. Embryo view is dorsal, anterior is up; scale bar = 10 µm. (E–G) Dot plots showing the number (E–G) of Tg(dusp6:d2EGFP) (E), Topro-3 (F) and sox1a (G) positive parapineal cells in control (blue dots) or mib-/- mutant embryos (orange dots) at 32 hpf, with mean ± SEM. Tg(dusp6:d2EGFP) FGF reporter is expressed in more parapineal cells in mib-/- than controls (E; 7 ± 3 d2EGFP + cells in mib-/- mutant compared 4 ± 2 in siblings; p-value=3.845e-05, Welch t-test) while the expression of sox1a is similar in both contexts (G); the average number of parapineal cells counted with a nuclear marker increases slightly (F; 22 ± 7 in mib-/- mutants versus 19 ± 5 in sibling control embryos; p-value=0.037, Welch t-test). (H–I’’) Confocal sections showing the expression of gfi1ab (red) at 48 hpf in control embryos (H; n = 40) and in three examples of mib-/- mutant embryos (I-I’’; n = 35). (J–K) Confocal sections showing the expression of tbx2b (red) at 48 hpf in control embryos (J; n = 39) and in one example of mib-/- mutants embryo with the parapineal on the right (K; n = 36). (L–N) Dot plots showing the mean position of parapineal cells expressing sox1a at 32 hpf (L) sox1a at 48 hpf (M) or gfi1ab (N) (in µm distance to the brain midline (x = 0)) with each dot representing an embryo. Parapineal migration is usually delayed in mib-/- mutants at 32 hpf (L). At 48 hpf, the parapineal of mib-/- mutant embryos either did not migrate (n = 12/35 with a parapineal mean position between −15 µM and +15 µM (shaded zone) relative to brain midline (Reference 0)) or migrates either to the left (n = 17/35) or to the right (n = 6/35) (N, orange dots), while it usually migrated to the left in control embryos (N, n = 39/40, blue dots); p-value<0.0001, Welch t-test on absolute values. (O–Q) Number of parapineal cells expressing sox1a (O), gfi1ab (P) and tbx2b (Q) at 48 hpf in control (blue dots) or in mib-/- mutant embryos (orange dots). The number of sox1a and gfi1ab+ positive parapineal cells at 48 hpf is increased in mib-/- mutant embryos (p-value<0.0001 in Welch t-test) compared with controls (O, P) while the number of tbx2b expressing cells is unchanged (Q). Mean ± SEM are indicated as long and short bars. *** p-value<0.0001; * p-value<0.05 in Welch t-test. Data are representative of three experiments (H–I’’, N, P) or two independent experiments (A–G, J–M, O, Q). See also Figure 1—figure supplement 1, Figure 1—figure supplement 2 and Figure 1—figure supplement 3. Source files used for dot plots and statistical analysis are available in Figure 1—source data 1.

https://doi.org/10.7554/eLife.46275.003
Figure 1—source data 1

Source files for data used to generate dot plots in Figure 1.

https://doi.org/10.7554/eLife.46275.007

To confirm that the phenotypes observed in mib-/- mutants are caused by Notch pathway loss-of-function, we analyzed the parapineal in embryos injected with a morpholino (MO; antisense oligonucleotide) blocking translation of the two zebrafish orthologs of su(H)/rbpj (Echeverri and Oates, 2007); Rpbj proteins are transcription factors required for canonical Notch signaling (Fortini and Artavanis-Tsakonas, 1994; Hsieh et al., 1996). As observed in mib-/- mutants, the proportion of Tg(dusp6:d2EGFP) expressing cells was increased in the parapineal of embryos injected with rbpja/b MO (4 ng) (10 ± 6 d2EGFP+ cells) compared to controls (6 ± 4 d2EGFP+ cells; p-value=0,0016) (Figure 1—figure supplement 2, C). Altogether, our results indicate that the FGF signaling pathway is activated in more parapineal cells when Notch signaling is abrogated.

Components of the Notch pathway are expressed in parapineal cells

To address whether the Notch pathway is active in the parapineal, we checked the expression of a well-characterized Notch reporter transgene, Tg(Tp1bglob:EGFP) (Corallo et al., 2013; Parsons et al., 2009), in the epithalamus at 30 hpf. Although rare, cells expressing GFP could be detected in the parapineal at 30 hpf (n = 3/19; Figure 1—figure supplement 3, A–B’).

In parallel, we analyzed the expression of genes encoding Notch ligands. We found that deltaB is expressed in one or two parapineal cells at both 28 hpf (n = 6/12) and 32 hpf (n = 17/23) (Figure 1—figure supplement 3, C–D’); no expression could be detected in the parapineal of the remaining embryos (n = 6/12 at 28 hpf or n = 6/23 at 32 hpf), suggesting that deltaB expression is highly dynamic. We found that deltaA mRNA could also be detected at 32 hpf in 1–2 parapineal cell, although its expression is less robust than deltaB (n = 3/10) (not shown). When detected in the parapineal, deltaB expression was found to overlap with Tg(dusp6:d2EGFP) expression in between half (n = 3/6 at 28 hpf) and two thirds of the embryos (n = 11/17 at 32 hpf).

Therefore, although the expression of Tg(Tp1:EGFP) Notch reporter transgene is not robust in the parapineal, the mosaic expression of deltaA and deltaB ligands in a few parapineal cells supports the existence of Notch signaling activity within the parapineal.

Loss of Notch signaling results in defects in parapineal migration

Interestingly, while in controls the parapineal had usually initiated migration at 32 hpf, it was still detected at the midline in most stage-matched mib-/- mutant embryos (Figure 1A–D and L). To address the effect of Notch inhibition on parapineal migration further, we examined mib-/- mutant embryos at a later stage, when the parapineal had unambiguously migrated in all controls. Analyzing sox1a (Clanton et al., 2013) expression at 2 days post-fertilization revealed that the parapineal failed to migrate in about a third of mib-/- embryos (defined by parapineal mean position within −15 µm and +15 µm of the midline (gray-shaded zone)) (Figure 1M). We confirmed this result by using another marker, gfi1ab (Dufourcq et al., 2004), whose expression is detected at a later stage (from 36 to 40 hpf) than sox1a (from 28 hpf) in parapineal cells. Using gfi1ab, we found that the parapineal had not migrated in 35% of mib-/- embryos (n = 12 of 35 embryos; Figure 1I” and N), while it migrated to the left in more than 95% of control embryos (n = 39/40, Figure 1H and N p-value<0.0001); the parapineal in mib-/- mutant embryos also migrated to the right more often than in controls (17%, n = 6/35; Figure 1I’ and N).

In embryos injected with rbpja/b MO (4 ng), the mean position of parapineal cells at 32 hpf was closer to the midline, suggesting a similar defect in parapineal migration as observed in mib-/- mutant embryos (Figure 1—figure supplement 2, D). Using gfi1ab as a marker at 48 hpf, we found that both parapineal migration per se and the orientation of migration were significantly affected in rbpja/b morphants (Figure 1—figure supplement 2, A–B’, E). While the parapineal migrated to the left in most uninjected controls (94%, n = 31/33), in rbpja/b morphants migration was blocked (13% of the embryos, n = 6/46; p-value=0.0001) or its orientation was partially randomized (left in 67% of the embryos, n = 31/46; right in 20% of the embryos, n = 9/46; p-value=0.0002) (Figure 1—figure supplement 2, A–B’, E).

Taken together, our data show that loss of Notch signaling results in defects both in parapineal migration per se (distance from the midline) and in the laterality of migration (left orientation), and that this correlates with the FGF signaling pathway being activated in more parapineal cells.

Loss of Notch signaling results in an increase in the number of certain parapineal cell subtypes

While quantifying parapineal mean position at 48 hpf, we observed an overall increased in the number of sox1a expressing cells in mib-/- mutants at this stage (Figure 1O) despite finding no significant change at 32 hpf (Figure 1G). Similarly, we found that the number of gfi1ab-positive cells at 48 hpf was significantly increased in mib-/- mutants (23 ± 5) compared to control embryos (17 ± 4; p-value=3.901e-07) (Figure 1H–1I’’ and P). In embryos injected with 4 ng of rbpj a/b MO, we also observed that the number of gfi1ab expressing parapineal cells at 48 hpf was significantly increased (19 ± 6) compared to control embryos (14 ± 3; p-value<0.0001) (Figure 1—figure supplement 2, A–B’, F). In contrast, the number of parapineal cells expressing tbx2b, a parapineal marker previously suggested to be required for the specification of parapineal cells (Snelson et al., 2008), was not increased in mib-/- mutants (Figure 1J–1K and Q).

Taken together, our data show that loss of Notch signaling results in defects in parapineal migration and, at later stages, an increase in the number of gfi1ab and sox1a expressing parapineal cells (putative differentiated parapineal cells) while the number of tbx2b expressing cells (putative parapineal progenitors) is unchanged.

The roles of Notch signaling in the specification and migration of parapineal cells can be uncoupled

Our data show that blocking Notch signaling leads to an expansion of FGF pathway activation in the parapineal, defects in parapineal migration and laterality, and an increase in the number of gfi1ab and sox1a expressing parapineal cells. With the aim of unraveling potential causative links between these different phenotypes, we used a pharmacological inhibitor of the γ-secretase complex, to block the Notch signaling pathway during different time windows (Romero-Carvajal et al., 2015; Rothenaigner et al., 2011); γ-secretase activity is required for the release of the intracellular domain of Notch, NICD, during activation of the canonical pathway (Geling et al., 2002).

We first treated wild-type embryos with LY411575 between 22 and 32 hpf, a time window corresponding to parapineal segregation from the pineal gland and the onset of its migration. While no change in gfi1ab expression was detected in embryos treated with 30 µM LY411575 (Figure 2C), a higher concentration of LY411575 (100 µM) resulted in an increase in the number of gfi1ab-positive cells in treated embryos (Figure 2A–2C); neither treatment resulted in defects in parapineal migration (Figure 2D). As this effect of LY411575 treatment was modest, we next treated embryos heterozygous for mib mutation (mib+/-) during the same time window thinking that this might provide a sensitized background for the drug. mib+/- embryos treated with the lower dose of LY411575 show a strong increase in the number of gfi1ab expressing parapineal cells (28 ± 5) compared to LY411575 treated wild-type controls (17 ± 3; p-value=1.0e-10) or DMSO-treated mib+/- (20 ± 2; p-value=2.2e-10) (Figure 2E–2H and I). As before, however, we did not detect a parapineal migration defect in LY411575-treated mib+/- (Figure 2E–2H and J), even when we increased the dose of LY411575 to 200 µM (Figure 2—figure supplement 1). Therefore, although LY411575 treatment from 22 to 32hpf can synergize with a mib+/- genetic background to promote an increase in the number of gfi1ab-positive parapineal cells, it does not affect parapineal migration. These data indicate that the role of Notch in controlling the specification of parapineal cells can be uncoupled from its function in parapineal migration.

Figure 2 with 3 supplements see all
Download asset Open asset
Uncoupled roles for the Notch pathway in the specification, laterality and migration of parapineal cells.

(A–B) Confocal sections showing the expression of gfi1ab (red) in embryos treated with DMSO (A; n = 24) or 100 µM LY411575 (B; n = 32) from 22 to 32 hpf and fixed at 48 hpf, merged with nuclear staining (gray). Embryo view is dorsal, anterior is up; epiphysis (white circle) and parapineal (yellow circle); scale bar = 10 µm. (C–D) Dot plots showing the number (C) and the mean position (D) of gfi1ab expressing parapineal (PP) cells at 48 hpf in embryos treated with DMSO (controls, blue dots, n = 47), with 30 µM LY411575 (light red dots, n = 31) or 100 µM LY411575 (red dots, n = 32) from 22 hpf to 32 hpf, with mean ± SEM; *** p-value=0.0003. (E–H) Confocal sections showing the expression of gfi1ab (red) merged with nuclear staining (gray) at 48 hpf, in wild-type (+/+) (E, F) or mib+/- embryos (G, H) treated with DMSO (E, n = 23 or G, n = 26) or with LY411575 (F, n = 25 and H, n = 34) from 22 to 32 hpf. (I–N) Upper panels show a schematic of the LY411575 treatment timeline (yellow box, 22 to 32 hpf for I-K or 8 to 22 hfp for dot plots L-N), and the time window corresponding to when the parapineal initiates its migration (blue box, 28 to 30 hpf). Dot plots showing the number (I, L) and the mean position (J, M) of gfi1ab expressing cells at 48 hpf, or the number of Tg(dusp6:d2EGFP) expressing cells at 32 hpf (K, N), in the parapineal of DMSO-treated wild-type (+/+, light blue dots; I-J, n = 23; L-M, n = 12; K, n = 28; N, n = 18), DMSO-treated mib+/- heterozygote (dark blue dots; I-J, ± = 26; L-M, n = 17; K, n = 21; N, n = 21), LY411575-treated wild-type ( light red dots; I-J, n = 25; L-M, n = 11; K, n = 23, N, n = 19) and LY411575-treated mib+/- (dark red dots, I-J, n = 34 or L-M, n = 16; K, n = 29, N, n = 26); each dot represents a single embryo. Mean ± SEM is shown in I, K, L, N; *** p-value<0.0001, in Wilcoxon test and Welch t-test. In J and M, there is no defect in migration per se (ns p-value in Welch t-test on absolute value) but LY411575 treatment from 8 to 22 hpf (M) triggers a laterality defect (increased number of embryos with a parapineal on the right). Data are representative of three (E–H, I–K) or two experiments (A–D and L–N). See also Figure 2—figure supplement 1, Figure 2—figure supplement 2 and Figure 2—figure supplement 3. Source files used for dot plots and statistical analysis are available in Figure 2—source data 1.

https://doi.org/10.7554/eLife.46275.008
Figure 2—source data 1

Source files for data used to generate dot plots in Figure 2.

https://doi.org/10.7554/eLife.46275.012

Finally, we saw no significant increase in the number of Tg(dusp6:d2EGFP) positive cells in mib+/- embryos treated with LY411575 during the 22 to 32 hpf time window (Figure 2K). This wild-type level of FGF pathway activation correlates with correct migration of the parapineal in LY411575-treated embryos and is consistent with our previous results suggesting that restricted FGF pathway activation is important for correct parapineal migration (Roussigné et al., 2018).

As the parapineal has usually started to migrate by 32 hpf, it is unlikely that the full parapineal migration defects observed in some mib-/- mutants is a consequence of a role of the Notch pathway after this stage; Notch signaling could nevertheless contribute to migration after 32 hpf. To address this possibility, we treated embryos with LY411575 during a later (32–36 hpf) or extended time window (22–48 hpf). We did not observe any effect of these treatments on the migration as assessed by gfi1ab expression at 48 hpf (Figure 2—figure supplement 2, A,B). Likewise, we did not observe changes in the number of gfi1ab+ cells upon these late treatments (Figure 2—figure supplement 2, C,D).

Notch activity is required early for unilateral activation of the Nodal pathway in the epithalamus

To determine whether the parapineal migration defects we observed in mib-/- mutant embryos and rbpja/b morphants might be an indirect consequence of an earlier role of Notch signaling, we also analyzed the epithalamus of embryos treated with LY411575 during an earlier 8 to 22 hpf time window. This early drug treatment did not interfere with the number of gfi1ab-positive cells (Figure 2L), or with migration in itself (Figure 2M), but led to a partial randomization of the direction of parapineal migration as shown by a significant increase in the number of embryos with a right parapineal (Figure 2M).

We hypothesized that the partial randomization of parapineal sidedness observed in mib-/- mutants, rbpja/b morphants or in embryos treated with LY411575 from 8 to 22 hpf could be caused by changes in the activation pattern of Nodal signaling in the epithalamus (Concha et al., 2000; Liang et al., 2000; Regan et al., 2009). To address this possibility, we analyzed the expression of pitx2c, a Nodal signaling target gene (Concha et al., 2000; Essner et al., 2000; Liang et al., 2000), in the different contexts of Notch loss-of-function. While pitx2c expression is detected in the left epithalamus in control embryos between 28 and 32 hpf (n = 26/28), we observed that its expression is bilateral in most mib-/- mutant embryos (n = 26/29) (Figure 2—figure supplement 3, A–C,D), in a majority of rbpja/b morphants (n = 13/23, Figure 2—figure supplement 3, E) and in approximately half of the embryos treated with LY411575 from 8 to 22 hpf, regardless of whether they were heterozygotes for the mib mutation (mib+/-, n = 6/10) or not (n = 4/13) (Figure 2—figure supplement 3, G). In contrast, the expression of pitx2c was indistinguishable from controls in embryos treated with LY411575 during the later time window (22–32 hpf, Figure 2—figure supplement 3, H) or in embryos expressing NICD from 26 hpf (Figure 2—figure supplement 3, F), which was expected given the absence of laterality defects in these contexts.

Despite triggering defects in parapineal laterality, the early time window of LY411575 treatment (8 to 22 hpf) did not affect parapineal migration per se and, as observed for the late time window (22–32 hpf), this correlates with no significant change in the Tg(dusp6:d2EGFP) expression pattern (Figure 2N). Our data indicate that the partial randomization of parapineal migration observed in mib-/- mutants or rbpj morphants is due to an early role of Notch pathway in restricting Nodal signaling to the left epithalamus and not to changes in the pattern of FGF activation.

Notch gain of function inhibits FGF pathway activation in the parapineal, and blocks the specification and migration of parapineal cells

When we abrogated Notch signaling, we observed an increase in the number of Tg(dusp6:d2EGFP)+ parapineal cells and correlated defects in parapineal migration (Figure 1E and L–M, Figure 1—figure supplement 2, C–E). To address further the role of the Notch pathway in modulating FGF activation, we analyzed the phenotypes associated with global activation of Notch signaling. For this, we used previously described transgenic lines, Tg(hsp70:gal4) and Tg(UAS:NICD-myc) (Scheer and Campos-Ortega, 1999), to induce widespread expression of the Notch Intracellular Domain (NICD) upon heat shock. In most embryos globally expressing NICD from 26 hpf, we observed a strong decrease in the number of Tg(dusp6:d2EGFP) expressing parapineal cells at 36 hpf (2.5 ± 2 cells) compared with the control embryos (5 ± 4; p-value=0.01) (Figure 3A–3B and G); the mean intensity of d2EGFP fluorescence was also significantly decreased in these embryos compared with the controls (p-value=0.0001) (Figure 3A–3B and H).

Figure 3 with 1 supplement see all
Download asset Open asset
Ectopic Notch signaling triggers decreased FGF activation and defects in migration and specification of parapineal cells.

(A–F) Confocal sections showing the expression of Tg(dusp6 :d2EGFP) (A-B, green), sox1a (C-D, red) or tbx2b (E-F, red) at 36 hpf, in control embryos (A, n = 16; C, n = 12; E, n = 5) or in Tg(hsp70l:Gal4), Tg(UAS:myc-notch1a-intra) embryos (B, n = 16; D, n = 17; F, n = 6) following a heat-shock (HS) at 26 hpf; sections are merged with nuclei staining (gray). (G–L) Dot plots showing the number (G) and the mean intensity fluorescence (H) of Tg(dusp6 :d2EGFP) expressing parapineal cells, the number of sox1a (I) and tbx2b (J) expressing parapineal cells, or the mean position of parapineal cells highlighted by Topro-3 nuclei staining (K) and tbx2b + parapineal cells (L) in controls (blue dots) or in NICD expressing embryos (orange dots) at 36 hpf following heat shock at 26 hpf. (M–R) Confocal sections showing the expression of sox1a (M–N), gfi1ab (O–P) or tbx2b (Q–R) (red) merged with nuclei staining (gray), at 48 hpf, in the epithalamia of control (M, n = 17; O, n = 27; Q, n = 17) or Tg(hsp70l:Gal4);Tg(UAS:myc-notch1a-intra) double transgenic embryos (N, n = 17; P, n = 25; R, n = 16), following heat-shock (HS) at 26 hpf. The expression of sox1a and gfi1ab is lost or decreased while tbx2b expression is unchanged in the parapineal of NICD expressing embryos. (S–X) Dot plots showing the number of sox1a (S), gfi1ab (T) and tbx2b (U) expressing parapineal cells at 48 hpf in controls (blue dots) or in embryos expressing NICD after heat shock at 26 hpf (orange dots) and the corresponding mean position of the cells (V–X) when expression was detected (number of sox1a + or gfi1ab+ cells > 0). In confocal sections, embryo view is dorsal, anterior is up; epiphysis (white circle) and parapineal gland (yellow circle); scale bar = 10 µm. Mean ± SEM is indicated on dot plots G-J and S-U; *** p-value<0.0001, * p-value<0.05 in Welch t-test and Wilcoxon test. For migration dot plots, p-value<0.01 (L) or p-value<0.001 (K and V–X) in pairwise Wilcoxon test and Welch t-test on absolute values. Data are representative of three (O, P, T, W) or two experiments (A–D, G–I, M–N, Q–R, S, U, V, X); data based on tbx2b expression at 36 hpf (E–F, J, L) represents one experiment. See also Figure 3—figure supplement 1. Source files used to generate dot plots and for statistical analysis are available in Figure 3—source data 1.

https://doi.org/10.7554/eLife.46275.013
Figure 3—source data 1

Source files for data used to generate dot plots in Figure 3.

https://doi.org/10.7554/eLife.46275.015

To assess a potential correlation between the inhibition of FGF pathway activation and defects in parapineal migration, we sought to analyze the mean position of sox1a or gfi1ab expressing cells at 36 or 48 hpf as we had done for Notch loss-of-function embryos. However, following heat shock at 26 hpf, sox1a expression was lost in most 36 hpf embryos (Figure 3C–D and I) and was strongly decreased at 48 hpf (Figure 3M–3N and S). Similarly, although the number of gfi1ab positive cells did not vary significantly in Tg(hsp70:gal4); Tg(UAS:NICD-myc) embryos heat shocked at 22 and 24 hpf (Figure 3—figure supplement 1, A–D,I–J), it was strongly decreased in embryos expressing NICD beginning from 26, 28 and 32 hpf (Figure 3O–P and T and Figure 3—figure supplement 1E–H,K–L); in embryos heat shocked at 26 hpf or 28 hpf, gfi1ab staining was often completely lost in the parapineal (n = 7/25 or n = 8/26, respectively) or detected in less than 4 cells (n = 11/25 or n = 14/26, respectively) (Figure 3O–P and T and Figure 3—figure supplement 1, E–F,K). However, nuclear staining indicated that the parapineal rosettes can be detected in most of the embryos expressing NICD (Figure 3A–F and M–R, yellow circle), suggesting that the parapineal does form despite global activation of Notch. Moreover, tbx2b expression was not affected in NICD expressing embryos either at 36 hpf (Figure 3E–F and J) or at 48 hpf (Figure 3Q–R and U). This result suggests that global Notch activation inhibits the specification/differentiation of gfi1ab+ or sox1a+ cells from tbx2b+ progenitors.

As NICD expression led to a loss of Tg(dusp6:d2EGFP) and sox1a expression, we first relied on nuclear staining to assess parapineal cells mean position in this context at 36 hpf. We found that the parapineal was closer to the midline in embryos expressing NICD from 26 hpf, (−8.7 ± 6.3 µM; n = 16) compared to control embryos (−20.5 ± 8.3 µM; n = 16; p-value=0.001) (Figure 3A–F and K). This migration defect was also revealed by analyzing the mean position of tbx2b expressing parapineal cells at 36 hpf (p-value=0.0017) (Figure 3L). Defects in parapineal migration were confirmed at 48 hpf using sox1a, gfi1ab and tbx2b as markers to assess PP mean position (Figure 3M–R and V–X); for instance, when gfi1ab-positive cells were detected, their mean position was significantly closer to the midline in embryos expressing NICD from 26 hpf (−10.1 ± 14 µM; n = 18) compared to control embryos (−34.6 ± 7.5 µM; n = 27) (p-value<0.0001) (Figure 3O–P and W). Parapineal mean position was also affected in embryos with NICD induced just before parapineal formation (heat shock at 22 and 24 hpf), or after parapineal formation (heat shock at 28 and 32 hpf), although in the latter case, the penetrance varies (Figure 3—figure supplement 1, A–H,M–P).

Altogether, our data show that global activation of Notch signaling inhibits the migration of parapineal cells, and that this is correlated with a decrease in the level of FGF signaling detected in the parapineal. Ectopic Notch signaling also decreases the number of gfi1ab+ and sox1a+ parapineal cells, a phenotype opposite to that observed in Notch loss-of-function contexts.

Decreasing FGF signaling rescues the parapineal migration defects in loss of Notch context while increasing FGF signaling aggravates it

Inhibiting or activating the Notch pathway results in reciprocal effects on FGF pathway activation in the parapineal, as seen by an increase or decrease in the number of Tg(dusp6:d2EGFP) expressing parapineal cells, respectively. Both contexts are also associated with defects in parapineal migration, suggesting that Notch-dependent control of FGF activation in parapineal cells is important for their collective migration. To investigate the link between Notch and FGF signaling further, we analyzed whether the migration phenotype observed in mib-/- mutants could be rescued by decreasing FGF signaling, using a pharmacological inhibitor of the FGF pathway (Mohammadi et al., 1997). We have previously shown that treating wild-type embryos with 10 µM SU5402 interferes with parapineal migration (Regan et al., 2009) and with the expression of the Tg(dusp6:d2EGFP) FGF reporter (Roussigné et al., 2018). Treating embryos with 5 µM SU5402, however, does not affect parapineal migration (Figure 4A–4B and F); the parapineal migrates in all SU5402 treated embryos (n = 31/31) as well as in all DMSO-treated control embryos (n = 32/32) (Figure 4F, purple versus blue dots). Using this suboptimal dose, parapineal migration was partially rescued in mib-/- embryos, with 19% of SU5402-treated mib-/- embryos showing mean parapineal position between −15 and +15 µm (n = 8/41) compared to 52% of DMSO treated mib-/- embryos (n = 16/31) (Figure 4C–4D and F, yellow versus orange dots; p-value=0.01). Thus, decreasing the level of FGF signaling activity can partially restore parapineal migration in a context where FGF activation is expanded supporting the hypothesis that Notch promotes parapineal migration through restricting FGF pathway activation.

Figure 4
Download asset Open asset
Decreasing or increasing FGF signaling rescues or aggravates the parapineal migration defect in Notch loss-of-function.

(A–D) Confocal sections showing the expression of gfi1ab (red) merged with nuclei staining (gray) at 48 hpf in representative control sibling (A–B) or mib-/- mutant embryos (C–D) treated from 24 to 32 hpf with DMSO (A, C) or 5 µM SU5402 (B, D). Parapineal migration is not affected in controls embryos treated with 5 µM of SU5402 (A–B); in C-D, examples show a parapineal that failed to migrate in a DMSO treated mib-/- mutant embryos (C) or that migrated to the left in a SU5402 treated mib-/- mutant embryos (D). Embryo view is dorsal, anterior is up; epiphysis (white circle) and parapineal (yellow circle); scale bar = 10 µm. (E–F) Upper panel show a schematic of the SU5402 (or DMSO) treatment timeline (24–32 hpf) in control and mib-/- mutant embryos. Dot plots showing the number (E) and the mean position (F) of gfi1ab expressing parapineal (PP) cells at 48 hpf in control embryos treated with DMSO (light blue dots; n = 32) or with 5 µM SU5402 (dark blue dots, n = 31), and in mib-/- mutant embryos treated with DMSO (orange dots, n = 31) or with SU5402 (yellow dots, n = 41). The number of gfi1ab-positive cells is increased in mib-/- mutants embryos, regardless of whether they were treated with SU5402 or DMSO (DMSO control versus DMSO mib-/-, ** p-value=0.0042; SU5402 control versus SU5402 mib-/-, ** p-value=0.0086, Welch t-test). The parapineal fails to migrate in 52% of DMSO treated mib-/- mutant embryos (PP mean position between −15 µm to +15 µm, gray-shaded area), and this proportion decreases to 19% of SU5402 treated mib-/- mutant embryos; p-value=0.0139 on a Chi-square test and p=0.0103 on a Welch t-test on absolute value. (G) Upper panel shows a schematic of the SU5402 (or DMSO) treatment timeline (24–32 hpf) in control and mib-/- mutant embryos. Dot plot showing the number of Tg(dusp6:d2EGFP) expressing parapineal cells at 32 hpf in control embryos treated with DMSO (light blue; n = 20) or with 5 µM SU5402 (dark blue, n = 18), and in mib-/- mutant embryos treated with DMSO (orange, n = 17) or with SU5402 (yellow, n = 21). The number of Tg(dusp6:d2EGFP)+ cells is increased in DMSO treated mib-/- mutants versus controls (p-value=0.0077) but is decreased in SU5402 treated mib-/- mutants compared to DMSO treated mib-/- mutants (p-value=0.0248 in Welch’s t-test). (H–I) Upper panel in H shows a schematic of heat shock timeline in Tg(hsp70:ca-fgfr1) embryos injected with rbpja/b morpholinos (MO). Dot plots showing the number (H) and the mean position (I) of gfi1ab expressing parapineal cells in control embryos that carry (purple dots; n = 55) or do not carry the Tg(hsp70:ca-fgfr1) transgene (light blue dots; n = 54), or in rbpja/bMO injected embryos that carry (dark red dots; n = 73) or do not carry the Tg(hsp70:ca-fgfr1) transgene (orange dots; n = 67). In rbpja/b morphants expressing Tg(hsp70:ca-fgfr1), the parapineal failed to migrate in 25% of embryos (n = 18/73; p-value=0.0007 Welch’s t-test on absolute value and p-value=0.0232 in Chi-square test), a defect significantly higher than expected from merely adding the effects of activated receptor transgene and rbpja/b MO injections alone (p-value=0.0001 in Welch t-test on absolute value and p-value=0.0003 in Chi-square test). Data are representative of four (H–I), three (A–F) or two experiments (G). Source files used to generate dot plots and for statistical analysis are available in Figure 4—source data 1.

https://doi.org/10.7554/eLife.46275.016
Figure 4—source data 1

Source files for data used to generate dot plots in Figure 4.

https://doi.org/10.7554/eLife.46275.017

Suboptimal SU5402 treatment had no effect on parapineal cells specification, either in mib-/- mutant embryos or siblings (Figure 4E). As previously observed, we detected an increase in the number of gfi1ab expressing parapineal cells in DMSO-treated mib-/- embryos (18 ± 5) compared to controls (15 ± 2; p-value=0.0042). The number of gfi1ab positive cells at 48 hpf was also increased in SU5402 treated mib-/- embryos (19 ± 7) compared to SU5402-treated controls (15 ± 3; p-value=0.0086); however, there was no significant difference in the number of gfi1ab expressing cells between DMSO-treated mib-/- embryos and SU5402 treated mib-/- embryos (p-value=0.5687). These data support further that the role of Notch signaling in controling the number of parapineal cells is independent from its function in parapineal migration and does not involve Notch mediated modulation of FGF signaling. 

To address the connection between Notch and FGF signaling in an alternative way, we asked whether ectopic activation of the FGF pathway could elicit a more severe phenotype in a loss-of-function context for Notch. To achieve this, we used a transgenic line that expressed a constitutively activated Fgf receptor after heat shock, Tg(hsp70l:Xla.Fgfr1,cryaa:DsRed) (Marques et al., 2008), in embryos injected with rbpja/b MO (4 to 8 ng rbpja/b MO); we chose this Notch context as parapineal migration defects are more modest than in mib-/- mutants. As previously described (Roussigné et al., 2018), widespread expression of constitutively activated receptor (CA-FgfR1) prevented parapineal migration in a small number of embryos (7% of embryos, n = 4/55, with a parapineal mean position between −15 and +15 µm; p-value=0.0011), while the parapineal consistently migrates in heat shocked control embryos not carrying the transgene (n = 54/54 with n = 51/54 leftwards and n = 3/54 toward the right) (Figure 4I). In the absence of the CA-FgfR transgene, rbpja/b morphant embryos displayed a migration defect at low frequency (7%; n = 5/67) (Figure 4I). However, in rbpja/b morphants expressing the activated receptor, the frequency of embryos in which the parapineal failed to migrate increased significantly (25% of embryos; n = 18/73) (Figure 4I, red versus orange and purple dots). Notch and FGF pathways appear to interact in this context as the increase in parapineal migration defects observed in rbpja/b morphant embryos expressing the activated Fgf receptor is significantly higher than expected from adding the effects of activated receptor transgene and rbpja/b MO injections alone (p-value=0.0001 in Welch t test on absolute value and p-value=0.0003 in Chi-square test). The increased frequency of parapineal migration defects occurs in the absence of significant changes in parapineal cell-type specification (Figure 4H); the mean number of gfi1ab expressing parapineal cells did not vary significantly between heat shocked rbpja/b morphant embryos with or without the CA-FgfR transgene (14 ± 6 versus 13 ± 5; p-value=0.39).

Taken together, our loss-of-function, gain-of-function and epistasis experiments argue that the Notch pathway is required for parapineal migration, and that it acts by restricting FGF activation in parapineal cells.

Discussion

In this study, we address how activation of the FGF pathway is restricted within parapineal cells. We show that expression of the Tg(dusp6:d2EGFP) reporter transgene is broader in Notch loss-of-function contexts and is reduced following global activation of the Notch pathway. These changes in FGF activation correlate with defect in parapineal migration. Epistasis experiments show that decreasing or increasing FGF signaling can respectively rescue or aggravate parapineal migration defects in Notch loss-of-function contexts. We conclude that the Notch pathway participates in restricting the activation of FGF signaling to few cells of the parapineal cluster, thus promoting its migration.

Notch effects on parapineal cell specification can be uncoupled from migration

Studies addressing the link between cell specification and migration are rare and, as such, our knowledge of whether and how these two processes are coordinated is limited. In the lateral line primordium (LLP) model, proper morphogenesis of the future neuromast at the trailing edge of the primordium is required for migration (Durdu et al., 2014; Kozlovskaja-Gumbrienė et al., 2017; Lecaudey et al., 2008; Nechiporuk and Raible, 2008). In embryos treated with LY411575 from 22 to 32 hpf, the number of gfi1ab and sox1a expressing parapineal cells is strongly increased but neither parapineal migration nor Tg(dusp6:d2EGFP) expression are affected. Therefore, our data reveals that cell-type specification and migration can be uncoupled in the parapineal. Our results resemble previous observations describing uncoupling of specification and migration of cardiac cells during in heart development (Davidson et al., 2005).

Notch-mediated cell-cell communication is well described for its role on cell fate restriction and progenitors maintenance (Cau and Blader, 2009). In neural tissues or in the pancreas, for instance, inhibition of the Notch signaling pathway causes premature differentiation of the progenitor cells into mature differentiated cells (Li et al., 2015). In the lateral line system, Notch signaling is required to restrict sensory hair cell progenitor fate to a central cell in the forming pro-neuromast (Matsuda and Chitnis, 2010). Similarly, in the parapineal, the number of gfi1ab and sox1a expressing parapineal cells is affected in loss or gain of function for Notch. However, in both contexts, the parapineal rosette is formed and a normal number of tbx2b expressing parapineal cells is detected. Therefore, while in previous studies it was assumed that all described parapineal markers (tbx2b, sox1a, gfi1ab) label the same cells, our data indicate that the pool of tbx2b+ parapineal cells probably differs from a sox1a/gfi1ab+ pool. Our results also suggest that Notch signaling acts downstream of Tbx2b to control the transition from tbx2b expressing putative progenitors to differentiated parapineal cells expressing sox1a/gfi1ab.

Notch acts upstream of FGF signaling to promote parapineal migration

In most models describing cross-talk between the Notch and FGF pathways, Notch signaling is described to act downstream of the FGF pathway. In the zebrafish LLP, for instance, ectopic activation of Notch signaling can rescue the formation of neuromast rosettes in absence of FGF pathway activity, indicating that Notch signaling is required downstream of Fgf signals to promote apical constriction and rosette morphogenesis (Kozlovskaja-Gumbrienė et al., 2017). Notch signaling is also described to be a downstream effector of the FGF pathway for epithelial proliferation in the pancreas in mammalian embryos (Hart et al., 2003). Our epistasis experiments suggest that the Notch pathway acts upstream of FGF signaling in parapineal cells to restrict FGF pathway activation and promote migration. Therefore, although in both the parapineal and the lateral line system, Notch signaling plays a similar role in restricting the number of cells with a particular fate (as mentioned above), the crosstalk with the FGF pathway appears to differ between the two models; Notch signaling acts upstream or downstream of FGF pathway to promote parapineal migration or neuromast morphogenesis, respectively.

Notch signaling restricts activation of the FGF pathway to promote parapineal migration

Our data show that gain or loss of function for Notch respectively result in an increase or decrease in the number of Tg(dusp6:d2EGFP)+ cells in the parapineal, with both correlating with parapineal (PP) migration defects. The fact that the number of Tg(dusp6:d2EGFP)+ cells and not the mean d2GFP intensity of Tg(dusp6:d2EGFP)+ cells is affected in mib-/- mutants strongly suggests that it is the restriction of FGF activation to few parapineal cells that is important for correct migration rather than a specific absolute level of FGF activity in leading cells. These data, together with the fact that partial rescue of migration in SU5402-treated mib-/- mutants correlates with a decrease in the number of Tg(dusp6:d2EGFP)+ cells, indicate that the Notch pathway promotes migration by restricting the activation of FGF signaling to a few parapineal cells.

Notch signaling has previously been implicated in the migration of epithelial cells sheets (Riahi et al., 2015), trachea cells in Drosophila (Ghabrial and Krasnow, 2006; Ikeya and Hayashi, 1999) and in developing vertebrate blood vessels (Siekmann and Lawson, 2007b). In all these models of sheet or sprouting morphogenesis, migrating cells remains attached to the bulk of the tissue. The parapineal is thus the first described model of an isolated cluster of migrating cells in which Notch signaling modulates RTK signaling to define leading cells.

In both the tracheal and vascular systems, Notch-Delta signaling contributes to the selection of the tip cells by restricting the ability of follower cells to activate RTK signaling. The molecular mechanisms underlying Notch mediated restriction of RTK signaling in these two models are not clear but could possibly involve the transcriptional control of RTK ligands or receptors (Ghabrial and Krasnow, 2006; Ikeya and Hayashi, 1999; Siekmann and Lawson, 2007b). As FgfR4, the only Fgf receptor found to be expressed in the parapineal, does not display restricted expression in the parapineal (Regan et al., 2009), it is unlikely that the Notch pathway acts through the control of this receptor.

How Notch signaling could modulate FGF pathway activity in parapineal cells remains an open question. The fact that some rbpja/b morphants display parapineal migration defects similar to mib-/- mutants strongly supports a role for the canonical Notch pathway rather than a Notch-independent role of the Mindbomb ubiquitin ligase in parapineal migration, although we cannot exclude that Mindbomb also regulates migration independently of the Notch pathway, for instance through modulation of Rac1 (Mizoguchi et al., 2017). As all parapineal cells are competent to migrate and to activate the FGF pathway (Concha et al., 2003; Roussigné et al., 2018), a parsimonious mechanism would be that lateral inhibition based cell-cell communication between parapineal cells within the rosette would modulate the capacity of a cell to activate/maintain or to inhibit FGF pathway activity. As described for the C. elegans vulva (Berset et al., 2001; Yoo et al., 2004), for instance, Notch signaling could directly promote the transcription of Ras/MAPK pathway inhibitors. Alternatively, Notch signaling could be required outside the parapineal to trigger mosaic FGF pathway activation in parapineal cells. Finally, there could be both parapineal-intrinsic and extrinsic requirements for Notch. In any case, as the ectopic expression of the constitutive activated FGF receptor (CA-FgfR1) does not completely block parapineal migration in a wild-type background and can rescue migration in fgf8-/- mutants (Roussigné et al., 2018), it appears that the restriction mechanism acts downstream of the receptor rather than at the level of receptor gene expression.

When and where is Notch signaling required?

Given that LY411575 does not block migration in any of the tested time windows, we cannot be sure about the timing of the Notch requirement for migration per se. However, the fact that ectopic Notch signaling at late stages (26–28 hpf) can efficiently block migration argues for a role of Notch at the time when the parapineal initiates its migration and not earlier in development. While expression of the Tg(Tp1:EGFP) Notch reporter transgene is not robustly detected in parapineal cells, the fact that the expression of the Notch ligand deltaB, and occasionally deltaA, can be detected in 1–2 parapineal cells after 28 hpf argues further for a role of Notch signaling within parapineal cells at the time migration starts.

If Notch signaling is indeed required at the time of migration initiation, it is unclear why LY411575 treatment (22 to 32 hpf) does not affect migration while it is able to trigger an increase in the number of gfi1ab and sox1a-positive parapineal cells. Classically, Notch signaling is thought to act through a lateral inhibition mechanism leading to a Notch-ON or Notch-OFF outcome between neighboring cells. Recent work, however, suggests that Notch acts in a level-dependent rather than an all-or-nothing manner (Ninov et al., 2012). In light of this, the differential effect of LY411575 on specification and migration of parapineal cells could reflect a different requirement in Notch signaling threshold, with a lower threshold of Notch signaling being required for migration and a higher one being required to control fate specification of parapineal cells. If LY411575 does not completely block Notch signaling, then residual Notch activity could be sufficient to promote the restriction of FGF activity and parapineal migration while not to limit the specification of gfi1ab and sox1a-positive cells. Consistent with the hypothesis that LY411575 treatment might be partially effective, we observed that the average number of Tg(dusp6:d2EGFP)+ cells increases slightly in LY411575 treated embryos from 22 to 32 hpf, without reaching statistical significance. Given the high variability observed in the number of Tg(dusp6:d2EGFP)+ cells in wild-type contexts, we might expect that parapineal migration is robust enough to tolerate a moderate increase in the number of Tg(dusp6:d2EGFP)+ cells.

Synergistic or parallel role of the Nodal and Notch pathways in restricting FGF pathway activation

The Notch signaling pathway has previously been implicated in regulating the expression of Nodal/TGF-β signal around the node and subsequently in the left lateral plate mesoderm (LPM) (Krebs et al., 2003; Raya et al., 2003). In zebrafish, expression of a nodal related gene (ndr3/southpaw) in the left LPM is required for the later expression of a second nodal gene (ndr2/cyclops) in the left epithalamus, which is required for left-biasing parapineal migration (Concha et al., 2000; Liang et al., 2000; Regan et al., 2009). In mib-/- mutants, rbpj morphant embryos or in embryos treated with LY411575 from 8 to 22 hpf, the Nodal pathway is activated on both sides of the epithalamus. This requirement for Notch signaling in unilateral activation of the Nodal pathway in the epithalamus is consistent with an early role of Notch signaling in establishing the initial left right asymmetry and explains the partial randomization of parapineal migration we observe in contexts of early Notch loss-of-function.

Our previous results indicate that Nodal signaling contributes to restricting FGF pathway activation, as well as biasing it to the left (Roussigné et al., 2018). Here, we show that the restriction of FGF activity requires a functional Notch pathway. As mentioned above, the Nodal pathway is bilateral in the epithalamus of embryos with compromised Notch signaling. However, it is unlikely that the role of Notch in restricting FGF pathway activation depends on its ability to control Nodal signaling. Indeed, in other contexts of bilateral Nodal signaling, such as in embryos injected with morpholinos against notail, we have shown that FGF pathway activation ultimately becomes restricted to a few parapineal cells and parapineal eventually migrates (Roussigné et al., 2018); Tg(dusp6:d2EGFP) reporter expression in this context is no longer left lateralized, which correlates with the parapineal migrating either to the left or the right. In absence of Nodal signaling, as in spaw morphants, Tg(dusp6:d2EGFP) expression is generally less restricted within the parapineal and this correlates with delayed parapineal migration, indicating that the restriction of FGF activity is influenced by Nodal signaling as well as by the Notch pathway. How these two pathways interact to restrict the activation of FGF signaling is not known and future investigations will be needed to address whether Nodal and Notch pathway act in a synergistic way or in parallel to restrict the FGF pathway.

Conclusion

Our study shows that Notch signaling is required for parapineal migration through its capacity to trigger cell state differences in FGF signaling and to restrict FGF activity to a few leading cells. As the function and cross-regulation of the FGF and Notch pathways might be conserved during the migration of invading cancerous cells, our results could help to understand these pathway interaction during metastasis.

Materials and methods

Key resources table
Reagent type
(species) or resource		DesignationSource or referenceIdentifiersAdditional
information
Genetic reagent (D. rerio)mibta52b
Itoh et al., 2003		RRID:ZFIN_ZDB-GENO-071218-1
Genetic reagent (D. rerio)fgf8ti282a
(ace)
Reifers et al., 1998		RRID:ZFIN_ZDB-GENO-980202-822
Genetic reagent (D. rerio)Tg(hsp70:Gal4)kca4
Scheer et al., 2001		RRID:ZFIN_ZDB-ALT-020918-6
Genetic reagent (D. rerio)Tg(UAS:myc-Notch1a-intra)kca3Scheer and Campos-Ortega, 1999RRID:ZFIN_ZDB-ALT-020918-8
Genetic reagent (D. rerio)Tg(hsp70:ca-fgfr1; cryaa:DsRed)pd3Marques et al., 2008
Neilson and Friesel, 1996		RRID:ZFIN_ZDB-GENO-090127-1
Genetic reagent (D. rerio)Tg(dusp6:d2EGFP)pt6Molina et al., 2007RRID:ZFIN_ZDB-GENO-071017-5
Genetic reagent (D. rerio)Tg(Tp1bglob:EGFP)ia12Corallo et al., 2013RRID:ZFIN_ZDB-ALT-130115-3
Sequence-based reagentrbpja/su(H)1; rbpjb/su(H)2Echeverri and Oates, 2007ZFIN ID : ZDB-FISH-150901–13686
AntibodyRabbit anti-GFPTorrey Pines BiolabsTP-401
RRID:AB_10013661		1/1000
AntibodyGoat anti-rabbit IgG Alexa 488-conjugatedMolecular probeA110341/1000
Chemical compound, drugLY411575MedChem ExpressHY-50752Rothenaigner et al., 2011
Chemical compound, drugSU5402Calbiochem572630Mohammadi et al., 1997
OtherFast RedSigma AldrichF46-48
OtherToPro-3Molecular probeT36051/1000

Fish lines

Request a detailed protocol

Embryos were raised and staged according to standard protocols (Westerfield, 2000). Embryos homozygous for mindbomb (mibta52b) (Itoh et al., 2003) and fgf8 mutations (fgf8 ti282a/acerebellar/ ace; Reifers et al., 1998) were obtained by inter-crossing heterozygous carriers. Carriers of the fgf8ti282a allele were identified by PCR as described previously (Roussigné et al., 2018). mibta52b+/- carriers were identified by PCR genotyping using primers 5’-GGTGTGTCTGGATCGTCTGAAGAAC-3’ and 5’-GATGGATGTGGTAACACTGATGACTC-3’ followed by enzymatic digestion with NlaIII. Tg(hsp70:Gal4)kca4 and Tg(UAS:myc-Notch1a-intra)kca3 transgenic lines have been described previously (Scheer et al., 2001; Scheer and Campos-Ortega, 1999) and identified by PCR genotyping following ZIRC genotyping protocols. Embryos carrying the Tg(hsp70:ca-fgfr1; cryaa:DsRed)pd3 transgene (Marques et al., 2008; Neilson and Friesel, 1996) were identified by the presence of DsRed expression in the lens from 48 hpf or, by PCR at earlier stages as described previously (Gonzalez-Quevedo et al., 2010). Tg(dusp6:d2EGFP)pt6 (Molina et al., 2007) and Tg(Tp1bglob:EGFP)ia12 (Corallo et al., 2013; Parsons et al., 2009) lines were used as reporters for the FGF and Notch pathways, respectively. Embryos were fixed overnight at 4°C in BT-FIX (Westerfield, 2000), after which they were dehydrated through an ethanol series and stored at −20°C until use.

Ethics statement

Request a detailed protocol

Fish were handled in a facility certified by the French Ministry of Agriculture (approval number A3155510). The project has received an agreement number APAFIS#3653–2016011512005922. Anaesthesia and euthanasia procedures were performed in Tricaine Methanesulfonate (MS222) solutions as recommended for zebrafish (0.16 mg/ml for anaesthesia, 0.30 mg/ml for euthanasia). All efforts were made to minimize the number of animals used and their suffering, in accordance with the guidelines from the European directive on the protection of animals used for scientific purposes (2010/63/UE) and the guiding principles from the French Decret 2013–118.

LY411575 treatment

Request a detailed protocol

Embryos collected from Tg(dusp6:d2GFP)pt6 carriers outcrossed with wild-type fish or with heterozygous mibta52b were dechorionated and treated from 8 to 22 hpf or 22 to 32 hpf with 30 µM, 100 µM or 200 µM of LY411575 (MedChem; Rothenaigner et al., 2011); control embryos were treated with an equal volume of DMSO diluted in E3 medium. Tg(dusp6:d2GFP) expressing embryos were incubated in LY411575 at 28°C and fixed at 32 hpf for immune-staining against EGFP; sibling embryos, not carrying the Tg(dusp6:d2GFP) transgene, were fixed at indicated time (32 hpf, 36 hpf or 48 hpf) for in situ hybridization against different parapineal markers.

SU5402 drug treatment

Request a detailed protocol

Embryos collected from in-crosses between mibta52b and Tg(dusp6:d2EGFP); mibta52b mutants were dechorionated and treated with 5 µM SU5402 (Calbiochem; Mohammadi et al., 1997) by diluting a 10 mM DMSO based stock solution in E3 medium; control embryos were treated with an equal volume of DMSO diluted in E3 medium. All embryos were incubated in SU5402 at 28°C from 24 hpf to 32 hpf. Tg(dusp6:d2EGFP) embryos were fixed at 32 hpf to analyze the d2EGFP expression pattern in both mibta52b mutants and sibling embryos; the remaining half of embryos were fixed at 48 hpf to analyze parapineal migration.

Morpholino injections

Request a detailed protocol

Morpholino oligonucleotides targeting both rbpja/su(H)one and rbpjb/su(H)2 (Echeverri and Oates, 2007) were dissolved in water at 3 mM. The resulting stock solution was diluted to working concentrations (0.3 mM, 2.5 ng/nl) in water and Phenol Red before injection of 1.5 nl (about 4 ng) or 3 nl (8 ng) into embryos at the one-cell stage. Embryos were subsequently fixed and processed for In situ hybridizations or antibody labeling.

Heat shock procedure

Request a detailed protocol

Ectopic expression of the intracellular domain of Notch receptor (NICD) was induced in Tg(hsp70:gal4); Tg(UAS:myc-Notch1a-intra) double transgenic embryos by incubating them at 39°C for 45 min starting at different time points (22 hpf, 24 hpf, 26 hpf, 28 hpf or 32 hpf). Embryos were then incubated at 28.5°C and fixed at 36 hpf to analyze Tg(dusp6:d2EGFP) expression or at 48 hpf for in situ hybridizations against indicated parapineal markers. Ectopic expression of CA-FgfR1 was induced in Tg(hsp70:ca-FgfR1; cryaa:DsRed)pd3 heterozygote embryos by performing a first heat shock at 25–26 hpf (39°C, 45 min) and a second short heat shock (39°C, 15 min) 3 hours later (28–29 hpf) in order to cover the entire period of parapineal migration. Control embryos are non-transgenic, or carry only Tg(UAS:myc-Notch1a-intra) or Tg(hsp70:Gal4) transgene, and were heat-shocked under the same conditions as the Tg(hsp70:ca-FgfR1) transgenic or Tg(hsp70:gal4); Tg(UAS:myc-Notch1a-intra) double transgenic embryos.

In situ hybridization and immunostaining

Request a detailed protocol

In situ hybridizations were performed as described previously (Roussigné et al., 2018), using antisense DIG labeled probes for gfi1ab (Dufourcq et al., 2004), sox1a (Clanton et al., 2013), pitx2c (Essner et al., 2000), tbx2b (Snelson et al., 2008), deltaA and deltaB (Haddon et al., 1998). In situ hybridizations were completed using Fast Red (from Sigma Aldrich) as an alkaline phosphatase substrate. Immunostainings were performed in PBS containing 0.5% triton using anti-GFP (1/1000, Torrey Pines Biolabs) and Alexa 488-conjugated goat anti-rabbit IgG (1/1000, Molecular Probes). For nuclear staining, embryos were incubated in ToPro-3 (1/1000, Molecular Probes) for 1 hr as previously described (Roussigné et al., 2018).

Image acquisition

Request a detailed protocol

Confocal images of fixed embryos were acquired on an upright Leica SP8 confocal microscope using the resonant fast mode and either an oil x63 (aperture 1.4) or x20 (aperture 1.4) objective. Confocal stacks were analyzed using ImageJ software. Figures were prepared using Adobe Photoshop software.

Quantification of the number and position of Tg(dusp6:d2EGFP) positive parapineal cells

Request a detailed protocol

The position and number of parapineal cells positive for the Tg(dusp6:d2GFP) transgene were analyzed using ROI Manager tool on ImageJ software as previously described (Roussigné et al., 2018). The mean intensity of the d2EGFP staining was quantified in an area corresponding to the cell nucleus and the same intensity threshold was used in the different experimental contexts to determine if a cell was Tg(dusp6:d2EGFP) positive or not. The total number of parapineal cells was estimated at 32 hpf by counting nuclei in the parapineal rosettes using Topro-3 nuclear staining.

Quantification of the number and position of gfi1ab, sox1a, tbx2b positive parapineal cells

Request a detailed protocol

The position and number of gfi1ab, sox1a or tbx2b positive parapineal cells were analyzed using the Multipoint tool on ImageJ software and determined as the centre of the cell nucleus detected with the Topro-3 nuclear staining as previously described (Roussigné et al., 2018). The position of each parapineal cell was measured relative to the brain midline (reference origin = 0) as determined by a line passing along the lumen of the epiphysis. For each embryo, we calculated the number of labeled parapineal cells and their mean position. To avoid any bias, data in Figure 4E–F were analyzed blind for DMSO versus SU5402 treatment and data describing LY411575 treatment (Figure 2I–N, Figure 2—figure supplement 1) were analyzed blind for mib genotypes (mib ± heterozygotes versus +/+ controls).

Statistical analysis

Request a detailed protocol

Statistical comparisons of datasets were performed using R Studio or GraphPad Prism software. For each dataset, we tested the assumption of normality with Shapiro-Wilks tests and variance homogeneity with F tests. As datasets on the number of parapineal cells were usually normal and often of unequal variances, they were compared using unpaired Welch t-tests; means (± SEM) are indicated as horizontal bars on dot plots. Data on parapineal mean position usually did not distribute normally (because of few embryos with the parapineal on the right) and were compared using the Wilcoxon rank sum non-parametric tests; we also compared parapineal mean position datasets with Welch t-tests using absolute values to discriminate between a defect in laterality (i.e. left-right randomization of parapineal migration) and a defect in migration per se (i.e. distance from the midline only). Most data are representative of at least two and more often three independent experiments and the number of biological replicates is mentioned in the figure legends for each graph. Means (± SD) are indicated in the text.

Data availability

All informations on datasets are provided in the eLife transparent reporting form and all data generated or analysed during this study are included in the manuscript and supporting files. Source data files required to reproduce dot plots and statistical analysis have been provided as an excel file for all main Figures 1-4.

References

    1. Bisgrove BW
    2. Essner JJ
    3. Yost HJ
    (1999)
    Regulation of midline development by antagonism of lefty and nodal signaling
    Development 126:3253–3262.
    1. Ciruna BG
    2. Schwartz L
    3. Harpal K
    4. Yamaguchi TP
    5. Rossant J
    (1997)
    Chimeric analysis of fibroblast growth factor receptor-1 (Fgfr1) function: a role for FGFR1 in morphogenetic movement through the primitive streak
    Development 124:2829–2841.
    1. Essner JJ
    2. Branford WW
    3. Zhang J
    4. Yost HJ
    (2000)
    Mesendoderm and left-right brain, heart and gut development are differentially regulated by pitx2 isoforms
    Development 127:1081–1093.
    1. Haddon C
    2. Smithers L
    3. Schneider-Maunoury S
    4. Coche T
    5. Henrique D
    6. Lewis J
    (1998)
    Multiple Delta genes and lateral inhibition in zebrafish primary neurogenesis
    Development 125:359–370.
    1. Ikeya T
    2. Hayashi S
    (1999)
    Interplay of notch and FGF signaling restricts cell fate and MAPK activation in the Drosophila Trachea
    Development 126:4455–4463.
    1. Liang JO
    2. Etheridge A
    3. Hantsoo L
    4. Rubinstein AL
    5. Nowak SJ
    6. Izpisúa Belmonte JC
    7. Halpern ME
    (2000)
    Asymmetric nodal signaling in the zebrafish diencephalon positions the pineal organ
    Development 127:5101–5112.
    1. Reifers F
    2. Böhli H
    3. Walsh EC
    4. Crossley PH
    5. Stainier DY
    6. Brand M
    (1998)
    Fgf8 is mutated in zebrafish acerebellar (ace) mutants and is required for maintenance of midbrain-hindbrain boundary development and somitogenesis
    Development 125:2381–2395.
    1. Scheer N
    2. Groth A
    3. Hans S
    4. Campos-Ortega JA
    (2001)
    An instructive function for notch in promoting gliogenesis in the zebrafish retina
    Development 128:1099–1107.
    1. Westerfield M
    (2000)
    A Guide for the Laboratory Use of Zebrafish (Danio Rerio
    The Zebrafish Book, A Guide for the Laboratory Use of Zebrafish (Danio Rerio, Fourth Edition, Eugene, University of Oregon Press.

Decision letter

  1. Marianne E Bronner
    Senior and Reviewing Editor; California Institute of Technology, United States
  2. Ajay B Chitnis
    Reviewer; Eunice Kennedy Shriver National Institute of Child Health and Human Development, United States

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

[Editors’ note: the authors were asked to provide a plan for revisions before the editors issued a final decision. What follows is the editors’ letter requesting such plan.]

Thank you for sending your article entitled "Notch signaling restricts FGF pathway activation in parapineal cells to promote their collective migration" for peer review at eLife. Your article is being evaluated by Marianne Bronner as the Senior Editor, a Reviewing Editor, and three reviewers.

Given the list of essential revisions, including new experiments, the editors and reviewers invite you to respond within the next two weeks with an action plan and timetable for the completion of the additional work. We plan to share your responses with the reviewers and then issue a binding recommendation.

In particular, the reviewers have several major concerns:

1) The reviewers wondered whether it was the levels of FGF activity rather than restricted FGF activity that may be playing an important role. To test this, they could us additional markers as a readout for FGF activity.

2) A crucial experiment is to determine in which cells Notch signaling is active. Without having this information, one cannot formulate a hypothesis of how Notch might be regulating FGF signaling.

3) While the study provides insights about when Notch activity is required, much less can be said about exactly where it is required. The authors can speculate in the discussion but must remain circumspect about this issue.

4) Although it would be interesting to know which Notch receptors and ligands are at play, this seems beyond the scope of what is necessary for this study.

Below are the full reviews from the three reviewers which we hope you find helpful.

Reviewer #1:

The manuscript describes how Notch signaling affects parapineal cell migration and cell type specification in the zebrafish brain. The manuscript is clearly written and illustrated, and the Discussion section is informative and thoughtful. The authors demonstrate that loss of Notch signaling leads to an expansion of the FGF pathway and defects in parapineal migration and laterality, and an increase in the number of gfi1ab and sox1a expressing parapineal cells. They further demonstrate that the roles of Notch signaling in cell type specification, cell migration and laterality of cell migration are uncoupled. Only the migration and laterality defects are FGF dependent, whereas the effect of Notch signaling on cell type specification is not. The authors also show that these different functions of Notch signaling occur at different time points of development. Cell migration to the left side (laterality) depends on Notch signaling before migration starts (8-22 hpf) by restricting Nodal signaling to the left side of the epithalamus. Whereas cell type specification occurs later. The main conclusion of the manuscript is that proper parapineal cell migration depends on Notch-dependent restriction of FGF signaling to just a few tip cells. The mechanisms by which Notch restrict FGF signaling is unknown.

The results are clearly illustrated and the majority of the data interpretation are sound. However, the manuscript would profit from a more detailed description of the different cell types. Also, the fact that suboptimal doses of SU5402 rescue the mib migration phenotype should be discussed/analyzed in more detail, as the results might suggest that it is not important that only the tip cells express FGF signaling but rather that these tip cells express the correct level of FGF signaling. Also, there is no description of where Notch signaling is active. As Notch reporter lines exist, the authors should add this information and discuss it.

Specific comments:

1) Subsection “The parapineal of mindbomb mutant embryos display expanded FGF pathway activation”: The number of parapineal cells is slightly increased but the number of sox1a positive cells is not increased (which is a marker for parapineal cells). Are there sox1a-negative parapineal cells or is the difference between Figure 1F and 1G due to biological noise?

2) Figure 1A: please indicate in the figure legend or figure what structures are outlined by yellow and white dotted lines.

3) Subsection “Loss of Notch signaling results in an increase in the number of certain parapineal cell subtypes”: A more detailed characterization of tbx2b, sox1a and gfi1a expressing cells is necessary. Are sox1a and gfi1ab expressed in the same cells?

4) Subsection “Loss of Notch signaling results in an increase in the number of certain parapineal cell subtypes” (and subsection “Notch effects on parapineal cell specification can be uncoupled from migration”) the authors conclude that Notch probably controls the transition from tbx2b-expressing putative progenitors to sox1a/gfi1ab-expressing differentiated cells. After loss of Notch signaling in mib mutants, the number of sox1a/gfi1ab positive cells increases, whereas the tbx2b positive cell number remains the same. How do the authors envision that the progenitor pool stays the same? Normally, defects in Notch dependent lateral inhibition should lead to the increase of one cell population at the expense of another. Do the progenitor cells divide asymmetrically and only the future sox1a/gfi1ab cells proliferate more? In this scenario tbx2b-expressing progenitor cells are independent of Notch signaling. In many other systems progenitor maintenance crucially depends on Notch signaling. Either the parapineal is different or the tbx2b positive cells are not self-renewing progenitors? Please discuss in more detail what is known about tbx2b, sox1a and gfi1ab positive cells.

5) Regarding the hs:NICD experiments: Are control embryos also heat shocked? Please label figure accordingly.

6) Figure 2I: The time line of treatment shown above the graph is very useful. It would facilitate the interpretation even further if it was indicated at what time point (32 hpf) migration begins.

7) The fact that ectopic Notch signaling at late stages (26-28 hpf) can efficiently block migration argues for a role of Notch at the time when the parapineal initiates its migration. If Notch signaling is indeed required just before migration, it is unclear why LY411575 treatment (22 to 32 hpf) does not affect migration. Maybe activation of Notch by inducing NICD until 28hpf persists much longer and NICD is still active after 28hpf. Therefore, embryos should be treated with LY411575 beyond 32hpf, as Notch might be required for migration during these later stages.

8) Subsection “Notch activity is required early for unilateral activation of the Nodal pathway in the epithalamus”: 'To determine whether the parapineal migration defects we observed (add: in mib mutants) might be a direct…. In sentence in the paragraph above the authors conclude that there was no migration defect in LY411575 treated embryos, which is confusing.

9) Subsection “Decreasing FGF signaling rescues the parapineal migration defects in loss of Notch context while increasing FGF signaling aggravates it”. The authors show that suboptimal doses of the FGF inhibitor SU5402 rescue to some degree the migration defect in mib mutants, which are characterized by FGF signaling activation in all migrating cells. This experiment actually argues that it is not important that the tip cells possess more FGF signaling then the follower cells but rather that FGF signaling needs to be expressed at the correct level. The authors should discuss the results of this experiment in more detail.

10) Subsection “Decreasing FGF signaling rescues the parapineal migration defects in loss of Notch context while increasing FGF signaling aggravates it”: 'Taken together, our.…experiments argue that the Notch pathway is required for parapineal migration, and that it acts by restricting FGF activation in parapineal cells.' However, in subsection “Notch activity is required early for unilateral activation of the Nodal pathway in the epithalamus” the authors argue that Notch and FGF act synergistically during migration, as rbpj morphant:cafgfr1a embryos display a more severe migration phenotype then CAfgfr1 embryos alone. Therefore, one has to conclude that Notch and FGF partially act in parallel on parapineal migration, correct?

11) Epistasis: Notch and FGF could also act in a feedback loop. Is Notch signaling affected after manipulation of FGF signaling?

Reviewer #2:

The paper by Lu et al., is an excellent follow up on previous work from Roussigne which demonstrated a role for restricted FGF signaling in collective migration of the parapineal cells. In this study the authors make good use of a combination of transgenic tools, mutants and chemical inhibitors to demonstrate the role of Notch signaling in a particular time window in restricting FGF signaling to a limited number of parapineal cells to determine asymmetric collective migration. They show that it is required after a relatively early stage when Notch is required to determine asymmetric Nodal signaling and after a stage is required to restrict the number of cells with gf1ab and sox1a positive cells. The experiments are presented in a logical order, the data is presented well, and its analysis supports the conclusions of the authors.

I have no major problems with the paper in its current form. My issues were only tangentially related to the study.

At various points in the Introduction and Discussion section the authors make comparisons with the role of FGF and Notch signaling in the lateral line system. I found those references confusing and potentially misleading. I would have liked to hear more about the role of Notch signaling in restricting migratory potential in tracheal cells and endothelial cells in vascular development. While the tracheal and vascular systems are referenced, the comparison could be elaborated on.

In the Introduction the authors suggest that the Dalle Nogare paper shows a role for FGF signaling in maintaining cluster cohesion. It is not clear what the authors have in mind here because the primary role of FGF shown in that paper is to suggest a role for FGFs released by leading cells in providing a guidance cue for collective migration of trailing cells. Their description of FGFs role in cluster cohesion is not essential for any point they wish to establish in this study and may only serve to establish a misleading impression about conclusions in the Dalle Nogare paper in the minds of readers for those not familiar with that paper.

Similarly, the authors contrast the role of Notch, downstream of FGF, in determining apical constriction in forming neuromasts in the migrating primordium, with the role of Notch in restricting the number of cells with FGF activity in the context of parapineal migration. The authors are not wrong here but they may give readers that not familiar with the Lateral Line system a simplistic idea about the role of Notch signaling in the primordium, where, as in the parapineal, it has multiple sequential roles. For example, first in restricting sensory hair cell progenitor fate to a central cell, and subsequently in determining Notch activation in neighboring cells to consolidate morphogenesis of epithelial rosettes. It is true that activation of Notch promotes apical constriction in the absence of FGF signaling and therefore in this context, downstream of FGF signaling and different in its epistatic relationship from the role of Notch in the parapineal. However, in the context of lateral inhibition and in restricting the number of cells with a particular fate/activity there may be important similarities in the role of Notch in the lateral line primordium and in parapineal cells. Without such a clarification, I felt the contrasting of the role of Notch in the parapineal and primordium could be potentially misleading.

A question that remained unanswered, which I was curious about and is not necessary in my mind for acceptance of the paper- which cells does Notch activity required in for its role in restricting migratory potential, which Notch ligands and receptors are required for this process and does their spatial distribution or that of downstream target genes provide insight about where Notch activity if required for restriction of FGF signaling in the parapineal.

Reviewer #3:

In this study by Wei et al., the authors investigate collective cell migration of parapineal cells during development using zebrafish as a model system. The authors found that global inhibition of Notch signaling, either by a chemical inhibitor or a MO-mediated knockdown, increased the number of cells that activate FGF signaling. As FGF is required for the parapineal migration and it is typically restricted to a few leading cells, this blocked the parapineal migration. Conversely, upregulating Notch signaling led to the loss of FGF signaling and also defects in its migration. These manipulations also affected a number of parapineal cells. Finally, the authors modulated FGF signaling in both Notch loss- and gain-of-function conditions. Decreasing FGF under this condition resulted in a partial rescue of the migration, whereas gain of FGF signaling worsened the migration phenotype. Interestingly, these manipulations did not affect the number of parapineal cells, indicating that these two roles of Notch signaling are distinct. Based on these data, the authors concluded that Notch restricts FGF signaling to a few leading parapineal cells, a process required for the proper migration. In addition, during later stages, Notch signaling controls the number of parapineal cells and this particular role of Notch can be uncoupled from its earlier role in restricting FGF signaling.

Overall, the conclusions are based on carefully done and controlled experiments (with a few exceptions). And while they provide some interesting insights into mechanisms of parapineal migration, the study is quite descriptive and does not go beyond the superficial analysis of cellular phenotypes. It is also not clear which cells express Notch ligands and receptors and whether effects on FGF signaling are direct. In summary, the overall findings are quite specialized and somewhat incremental and, thus, more suited for a specialized audience.

Major Comments:

Subsection “Loss of Notch signaling results in defects in parapineal migration”: Indicates that migration of the parapineal failed in 1/3 of mib-/-. I don't believe the authors discussed why this is the case/why the phenotype is partially penetrant in mib mutants? As well as why this phenotype was partially penetrant in the rbpj MO-injected embryos (with only 13% failing to migrate).

rbpj a/b MO: given specific guidelines for MOs (Stainier et al., 2017), these experiments do not seem to be appropriately controlled.

The results related to the inhibitor treatment seem odd. If the γ-secretase inhibitor, LY411575, is specific and Notch is playing a role in migration and specification of this system, I would expect to see a result with treatment from the 22-32 hpf time period. This drug and it has only been published in two zebrafish papers. Why not use the γ-secretase inhibitor DAPT that has been utilized in many other studies. Related to this, I am confused that 8-22 hour inhibitor treatment did not interfere with the Tg(dusp6:d2EGFP) transgene expression, neither did the 22-32 hour treatment. What is then the time window for Notch signaling that controls Tg(dusp6:d2EGFP) expression?

Throughout the manuscript, it is not clear where Notch signaling is active/ where Notch signaling components are expressed. It would be helpful to see the expression of where particular Notch signaling components are expressed during the time points of interest to further validate the reason behind studying Notch signaling. In other words, is Notch acting directly on parapineal cells?

[Editors’ note: formal revisions were requested, following approval of the authors’ plan of action.]

Further to my previous email, the editors and reviewers have considered your plan and invite you to proceed with your revisions as proposed. The asked us to pass on the following comments:

The authors should expand their discussion, add the discussion points raised by the reviewers and add relevant figure panels.

Specific comments:

In Figure 3 for reviewers they should add panels that only show the red signal.

As LY411575 does not inhibit migration at any time point, the authors should test if this drug is able to completely reduce Notch signaling. Is the signal in the TP Notch reporter line gone after treatment? Possibly, a small amount of Notch signaling is sufficient to allow for migration.

The expression data of Notch pathway members should be included and discussed in the manuscript.

The text describing the different cell types should be clarified, as it is currently unclear which genes mark possibly the same cells or which genes label all cells.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for submitting your article "Notch signaling restricts FGF pathway activation in parapineal cells to promote their collective migration" for consideration by eLife. Your article has been reviewed by Marianne Bronner as the Senior Editor, a Reviewing Editor, and two reviewers. 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. Please aim to submit the revised version within two months.

The reviewers felt that manuscript is much improved and addressed the main concerns to: (1) provide evidence that some Notch pathway signaling members are expressed in the parapineal during migration; (2) further investigate effects of the γ-secretase inhibitor treatment on the Notch pathway activation, and (3) clarify the dynamics of parapineal marker gene expression during PP development. However, the evidence that Notch pathway acts within parapineal cells remains weak: deltaB and the reporter expression are not robust enough to make that conclusion, and functional experiments testing this hypothesis are absent. Therefore, please tone down that conclusions and present alternative possibilities.

We will look forward to hearing from you with a revised article with tracked changes, and a response letter (uploaded as an editable file) describing the changes made in response to the decision and review comments.

https://doi.org/10.7554/eLife.46275.020

Author response

[Editors’ note: what follows is the authors’ plan to address the revisions.]

Given the list of essential revisions, including new experiments, the editors and reviewers invite you to respond within the next two weeks with an action plan and timetable for the completion of the additional work. We plan to share your responses with the reviewers and then issue a binding recommendation.

In particular, the reviewers have several major concerns:

1) The reviewers wondered whether it was the levels of FGF activity rather than restricted FGF activity that may be playing an important role. To test this, they could us additional markers as a readout for FGF activity.

Our data show that gain or loss of function for Notch results respectively in an increase or decrease in the number of Tg(dusp6:d2EGFP)+ cells in the parapineal, with both resulting in parapineal (PP) migration defects. These data, together with the fact that partial rescue of migration (Figure 4F) correlates with a decrease in the number of Tg(dusp6:d2EGFP)+ cells (Figure 4G), led us to conclude that FGF signaling needs to be restricted within PP cells. Reviewer 1 proposed that a correct level of FGF signaling in the leading cell might be required rather than or in addition to a restriction of the pathway activation to a few cells. This alternative interpretation is indeed a possibility. As explained below, our capacity to analyze the absolute level of FGF signaling is limited by the complexity of the FGF pathway and the tools at our disposal. However, here we provide a fuller set of arguments in favor of a requirement for the restriction in activation over a specific level of FGF activity being necessary for correct parapineal migration.

We analyzed the expression of additional markers as readout for FGF activity prior to our original submission. The activation of the FGF pathway leads to the transcription of various target genes that either encode positive effectors, such as members of the PEA3 subfamily of ETS domain transcription factor (Etv5, ERM and Pea3), or feedback inhibitors (Sprouty 1,2,4 and sef) of the FGF pathway (Tsang and Dawid, 2004). As part of the so-called fgf8 syn-expression group, these genes have been described in the literature or the ZFIN database to be expressed in the diencephalon in a pattern similar to Fgf8 and Fgf3, but their detailed expression pattern in the epithalamus had not been addressed. We analyzed the expression pattern of these genes in the epithalamus at cellular resolution and found that erm and pea3 are expressed broadly in the epithalamus, in both the parapineal and presumptive habenulae, while etv5 is more specifically detected in the PP and enriched at the migration front (see Author response image 1 for reviewers). We also found that the Tg(sprouty1:YFP) transgene is expressed in some PP cells, although we could not detect endogenous sprouty1 expression; the expression of sprouty 2,4 or sef is only detected in the habenulae but not in the PP.

Author response image 1
Download asset Open asset
Expression of others FGF target genes in the epithalamus.

Confocal sections of the head of Tg(dusp6:d2EGFP) embryos (A-C’’’) at 36 hpf after a whole-mount in situ hybridization against erm (red, A-A’’’), pea3 (red, B-B’’’), etv5a (red, C-C’’’) or of Tg(sprouty1:YFP) embryos at 30 hpf (D-D’) or 36 hpf (E, E’) after immunostaining against GFP (Green; A-E’); the pictures are merged with cell nuclei staining (Topro-3 in grey) that makes brain structures visible and specially the parapineal (yellow circle). A’’-C’’’, D’-E’ show Zoom in of merge pictures. Embryos are viewed dorsally with anterior up.

These data show that FGF target genes other than dusp6 are focally expressed in the parapineal (etv5 and perhaps sprouty1) and enriched at the migration front. Nonetheless, their expression does not systematically overlap with Tg(dusp6:d2EGFP) reporter transgene expression, highlighting the complexity of the FGF signaling pathway: although the classic FGF target genes are all expressed in the same territories as Fgf8 (the epithalamus), their expression pattern are somewhat distinct at a cellular level. As such, we do not think that performing in situ against these other markers will allow us to quantify absolute level of FGF activity better than we have already managed in the Tg(dusp6:d2EGFP) line, and thus to fully address the alternative hypothesis raised by reviewer 1.

The complexity revealed by the expression pattern of other FGF target genes led us to choose to focus on Tg(dusp6:d2EGFP) transgene expression only as we have previously shown that its expression reflects an Fgf8 dependent activation of one branch of the FGF pathway that is required for parapineal migration (Roussigné et al., 2018). To avoid any bias in counting the number of Tg(dusp6:d2EGFP)+ cells, in our study, we quantified the mean intensity of d2EGFP expression for each PP cell and used, for each experiment, the same intensity threshold to determine if a cell was Tg(dusp6:d2EGFP) positive or not in the various contexts (see Material and Methods). While not originally designed to address the level of FGF pathway activation per se, these data can be used to analyze the mean intensity of Tg(dusp6:d2EGFP)+ cells in addition to the number of Tg(dusp6:d2EGFP). As such, we revisited the data concerning mib-/- mutants versus wildtype (Figure 1) or in SU5402 versus DMSO treated wild-type and mib-/- (Figure 4) relative to the general level of d2GFP expression. Treatment with the suboptimal dose of SU5402 (5µM) in mib-/- context triggers a partial rescue of PP migration together with a decrease in both the number (Figure 4G) and the mean intensity of Tg(dusp6:d2EGFP)+ cells (Author response image 2E, F): although the p-value is higher than the 0,05 significance threshold (p=0,59), we observe a clear tendency for a decrease in the mean intensity of Tg(dusp6:d2EGFP)+ cells upon SU5402 treatment (See Author response image 2E). Therefore, in this experiment, we cannot conclude if the SU5402 mediated partial rescue is due to an overall decrease in the number or in the intensity of Tg(dusp6:d2EGFP) FGF reporter. However, when we compare the mean intensity of Tg(dusp6:d2EGFP)+ cells in mib-/- versus wild-type embryos, both in data from the original Figure 4 and Figure 1, we clearly observe that the number of Tg(dusp6:d2EGFP)+ increases (See Author response image 2C, D, F for Reviewers) while the mean intensity does not change significantly in mib-/- mutants (See Author response image 2A, B, E for Reviewers). Therefore, PP migration defects in mib-/- mutants correlate with an increase in the number rather than the mean d2GFP intensity of Tg(dusp6:d2EGFP)+ cells. This strongly suggests that, rather than the level of FGF activity, it is the restriction of FGF activation to few parapineal cells that is important for correct migration.

Author response image 2
Download asset Open asset
The number but not the mean intensity of Tg(dusp6:d2EGFP) expressing cells is increased in mib-/- mutants.

(A-B) Dot plots showing the mean GFP fluorescence intensity (A, B) of Tg(dusp6:d2EGFP) expressing parapineal cells in the first (A) and second experiment (B) in controls (blue dots; A, n=21; B, n=15) or in mib-/-mutant embryos (orange dots; A, n=21; B, n=13) at 32 hpf stage. (C-D); we chose to present data for the 2 experiments separately as the mean intensity threshold chosen to define Tg(dusp6:d2EGFP)+ cells differ between both experiments. (C-D) Dot plots showing the number of Tg(dusp6:d2EGFP) expressing parapineal cells in the first (C) and second experiment (D); merged data for both experiments are presented in the Manuscript as Figure 1E. (E-F) Dot plots showing the mean fluorescence intensity of Tg(dusp6:d2EGFP) expressing parapineal in control embryos treated with DMSO(light blue; A, n=20) or with 5 µM SU5402 (dark blue; A, n=18), and in mib-/-mutant embryos treated with DMSO (orange; A, n=17) or with SU5402 (yellow; A, n=21) at 32 hpf stage. Mean ± SEM is indicated; ns, p-value>0,05; * p-value<0,05; ** p-value<0,01 in Welch t-test.

We would be happy to mention this ‘level of FGF’ alternative interpretation in the discussion of a revised manuscript and present the d2GFP intensity data in a Supplement Figure to support our interpretation. Alternatively, we could include these data as new panels in Figure 1 and Figure 4 if the reviewers think it is appropriate.

2) A crucial experiment is to determine in which cells Notch signaling is active. Without having this information, one cannot formulate a hypothesis of how Notch might be regulating FGF signaling.

To address whether the Notch pathway is active in the parapineal, we checked the expression of a well-characterized Notch reporter transgene, Tg(TP1:EGFP) (Parson et al., 2009; Corallo et al., 2013), at 30 and 36 hpf. We found that, despite being expressed at very high level in the epiphysis, this transgene was not robustly detected in the parapineal. Indeed, we rarely detected parapineal cells expressing GFP at 30hpf (n=3/19) and no expression was observed in the parapineal at 36hpf (n=0/10).

In parallel, we analyzed the expression of several known target genes of the Notch pathway, including her2 (Quillien et al., 2011), her4 (Chapouton et al., 2011), her5 (Ninkovic et al., 2005), her6 (Chapouton et al., 2011), her9 (Chapouton et al., 2011), her12 (Nikolaou et al., 2009), her15 (Webb et al., 2009). Among the her genes analyzed, we found that two of them, her6 and her9, are robustly expressed in parapineal cells in a mosaic pattern (n=22 for her6 and n=8 for her9), in addition to other brain regions (see Author response image 3). The expression of both her6 and her9 appear reduced in many brain regions of mib-/- mutants but, in the parapineal, we found that only the number of her6 expressing cells (n=21) but not the number of her9 expressing cells (n=10) was decreased (see Author response image 3). This result could be added in a revised version of our manuscript as a Figure supplement for Figure 1.

Author response image 3
Download asset Open asset
Notch target genes her6 and her9 are mosaïcally expressed in parapineal cells.

(A-D) Confocal sections showing the expression of her6 (A, B) and her9 (C, D) (red) at 32 hpf in control embryos (A, C) and in mib-/- mutant embryos (B, D) merged with a cell nuclei staining (grey). (E-F) Dot plots showing the number of her6 (E) and her9 (F) in control (Blue marks) and mib-/- mutant embryos (Orange marks). Embryo view is dorsal, anterior is up; epiphysis (white circle), parapineal gland (yellow circle); scale bar=10 µm. Mean ± SEM are indicated as long and short bars. ** P-value<0.01; * P-value<0.05 in welsh t test (Wilcoxon test); data corresponds to one experiment.

3) While the study provides insights about when Notch activity is required, much less can be said about exactly where it is required. The authors can speculate in the discussion but must remain circumspect about this issue.

Our LY411575 treatment gives us a good idea of when Notch is required to control the laterality of parapineal migration (early time window 8-22 hpf) and the overall number of PP cells (22-32 hpf). Given that the LY411575 does not block migration in any of these time windows however, we agree that we cannot be fully sure about the timing of Notch requirement for migration per se; some possible reasons for this were already presented in the original discussion.

While our loss of function (LOF) experiments are not informative concerning migration, our gain of function (GOF) experiments strongly suggest that Notch acts at the time the PP initiates migration. As suggested by reviewer 1 in one of his/her specific comment, it is possible that NICD persists long after the heat shock and, in that case, Notch could act after 32 hpf. As the PP has usually migrated by 32 hpf, however, it is likely that the Notch pathway acts before this stage to explain the full PP migration defects observed in some mib-/- mutants. We nonetheless agree with Reviewer 1 that Notch could contribute to migration after 32hpf. To address this possibility, we treated embryos with LY411575 during a later (32-36 hpf) or extended time window (22-48 hpf). We did not observe any effect of these treatment on the migration as assessed by gfi1a expression at 48hpf (Author response image 4); likewise, we did not observe changes in the number of gfi1a+ cells upon these late treatment.

Author response image 4
Download asset Open asset
Late LY41 treatment does not affect parapineal migration.

(A-B) Upper panels show a schematic of the LY411575 treatment timeline, from 32 to 36 hfp (A) or 22 to 48 hpf (B) for dot plots. Dot plots showing, for each embryo, the mean position of gfi1ab expressing cells at 48 hpf in the parapineal of DMSO treated wild-type (+/+, light blue dots; A, n=12; B, n=15), DMSO treated mib+/- heterozygote (dark blue dots; A, n=15; B, n=17), LY411575 treated wild-type (light red dots; A, n=12; B, n=23) and LY411575 treated mib+/- embryos (dark red dots, A, n=13; B, n=17). No migration defect was observed in either context; data corresponds to one experiment.

We could modify the Discussion section of a revised manuscript to modulate our conclusion: whereas our original text read ‘The fact that ectopic Notch signaling at late stages (26-28 hpf) can efficiently block migration argues for a role of Notch at the time when the parapineal initiates its migration’, it could be replaced by ‘Given that LY411575 does not block migration in any of the tested time window, we cannot be absolutely sure about the timing of the Notch requirement for migration per se. However, our GOF experiments strongly suggest that Notch acts around the time that the PP initiates migration and not earlier in development.’

4) Although it would be interesting to know which Notch receptors and ligands are at play, this seems beyond the scope of what is necessary for this study.

We agree that it would be interesting to know whether component of the Notch pathways are indeed expressed in the PP during the appropriate time, and we had already begun to address this question over the course of our study. Among the genes encoding Notch ligands, we analyzed the expression of deltaA and deltaB in the epithalamus. We found that deltaB is expressed in one or two parapineal cells at both 28 hpf (n=6/12) and 32 hpf (n=17/23) (See Author response image 5); no expression could be detected in the PP of the remaining embryos (n=6/12 at 28 hpf or n=6/23 at 32 hpf), suggesting that deltaB expression is highly dynamic. We found that deltaA mRNA could also be detected at 32 hpf in 1-2 PP cell, although its expression is less robust than deltaB mRNA (n=3/10) (not shown).

Author response image 5
Download asset Open asset
Notch ligand deltaB is mosaically expressed in the parapineal.

(A-B’’) Confocal sections showing the expression of Tg(dusp6:d2EGFP) transgene (green, A, B) or deltaB (red, A’-B’) and merge (A’’-B’’) in the epithalamia of two representative embryos at 32 hpf. deltaB mRNA is detected in only one or two parapineal cells in most embryos (n=17/23); staining is either co-expressed with Tg(dusp6:d2EGFP) (n=11/23) or detected in Tg(dusp6:d2EGFP) negative cells (n=6/23); in n=6/23, deltaB was not detected in the parapineal. Embryo view is dorsal, anterior is up; epiphysis (white circle), parapineal gland (yellow circle); scale bar=10 µm. Data representative of 2 experiments.

When detected in the PP, deltaB expression was found to overlap with Tg(dusp6:d2EGFP) expression in between half (n=3/6 at 28hpf) to two thirds of the embryos (n=11/17 at 32 hpf). If a lateral inhibition model is at play downstream of Notch signaling to restrict FGF pathway, we expected to find perfect co-expression of at least one Delta with the Tg(dusp6:d2EGFP) transgene, and non overlapping expression of Notch targets, such as the her genes. Although deltaB expression is most often found co-expressed with Tg(dusp6:d2EGFP) (in about 2/3 of the embryos at 32hpf), this is not always the case. Similarly, the expression of her6 and her9 can sometime overlap with Tg(dusp6:d2EGFP) expressing cells. These results might be explained by the differential dynamic that exist in the detection of a GFP transgene versus mRNA. Alternatively, the classical lateral inhibition model might be too simplistic to recapitulate the complexity of interactions between the Notch and FGF pathways.

Altogether, our results show that both components and read-outs of the Notch pathway can be detected mosaically in the PP, supporting a role of Notch signaling within the PP in restricting FGF activation to a few cells. We agree with reviewer 2 and the Editor that addressing further the role of Notch components and target genes is beyond the scope of our study and this is the reason why these data were not included in our original manuscript. However, as suggested for her6 and her9 Figure, we could easily add our data on deltaB expression in a future supplement Figure if the reviewers and the editor think it would be pertinent.

Below are the full reviews from the three reviewers which we hope you find helpful.

Reviewer #1:

The manuscript describes how Notch signaling affects parapineal cell migration and cell type specification in the zebrafish brain. The manuscript is clearly written and illustrated, and the Discussion section is informative and thoughtful. The authors demonstrate that loss of Notch signaling leads to an expansion of the FGF pathway and defects in parapineal migration and laterality, and an increase in the number of gfi1ab and sox1a expressing parapineal cells. They further demonstrate that the roles of Notch signaling in cell type specification, cell migration and laterality of cell migration are uncoupled. Only the migration and laterality defects are FGF dependent, whereas the effect of Notch signaling on cell type specification is not. The authors also show that these different functions of Notch signaling occur at different time points of development. Cell migration to the left side (laterality) depends on Notch signaling before migration starts (8-22 hpf) by restricting Nodal signaling to the left side of the epithalamus. Whereas cell type specification occurs later. The main conclusion of the manuscript is that proper parapineal cell migration depends on Notch-dependent restriction of FGF signaling to just a few tip cells. The mechanisms by which Notch restrict FGF signaling is unknown.

The results are clearly illustrated and the majority of the data interpretation are sound.

We thank reviewer 1 for his/her positive comments.

However, the manuscript would profit from a more detailed description of the different cell types.

As it is a small structure, about 15 cells, cell diversity in the parapineal has probably been underestimated until now; in previous papers, it has always been assumed that all described markers label the same cells in the PP. Our results showing a difference in the behavior of tbx2b+ PP cells versus sox1a and gfi1a+ cells suggest that the pool of tbx2b+ parapineal cells differs from a sox1a/gfi1a+ pool. We have preliminary data, by double fluorescent in situ, confirming that tbx2b and gfi1a indeed label two distinct pools of PP cells. However, we are reticent to include these data in our manuscript as our study mainly focuses on the role of Notch signaling in controlling the collective migration of parapineal cells through its capacity to restrict FGF pathway activation to a few leading cells and not the potential changes to different PP cell types. In that context, we feel that adding data on the characterization of PP cells subtypes might dilute our main message.

Also, the fact that suboptimal doses of SU5402 rescue the mib migration phenotype should be discussed/analyzed in more detail, as the results might suggest that it is not important that only the tip cells express FGF signaling but rather that these tip cells express the correct level of FGF signaling.

See our reply concerning the major issues raised.

Also, there is no description of where Notch signaling is active. As Notch reporter lines exist, the authors should add this information and discuss it.

See our reply concerning the major issues raised.

Specific comments:

1) Subsection “The parapineal of mindbomb mutant embryos display expanded FGF pathway activation”: The number of parapineal cells is slightly increased but the number of sox1a positive cells is not increased (which is a marker for parapineal cells). Are there sox1a-negative parapineal cells or is the difference between Figure 1F and 1G due to biological noise?

Yes, some cells are negative for sox1a but seems to be part of the rosette highlighted by staining nuclei.

2) Figure 1A: please indicate in the figure legend or figure what structures are outlined by yellow and white dotted lines.

We apologize for this oversight and will add this information in the Figure.1 legend.

3) Subsection “Loss of Notch signaling results in an increase in the number of certain parapineal cell subtypes”: A more detailed characterization of tbx2b, sox1a and gfi1a expressing cells is necessary. Are sox1a and gfi1ab expressed in the same cells?

See our reply above.

4) Subsection “Loss of Notch signaling results in an increase in the number of certain parapineal cell subtypes” (and subsection “Notch effects on parapineal cell specification can be uncoupled from migration”) the authors conclude that Notch probably controls the transition from tbx2b-expressing putative progenitors to sox1a/gfi1ab-expressing differentiated cells. After loss of Notch signaling in mib mutants, the number of sox1a/gfi1ab positive cells increases, whereas the tbx2b positive cell number remains the same. How do the authors envision that the progenitor pool stays the same? Normally, defects in Notch dependent lateral inhibition should lead to the increase of one cell population at the expense of another. Do the progenitor cells divide asymmetrically and only the future sox1a/gfi1ab cells proliferate more? In this scenario tbx2b-expressing progenitor cells are independent of Notch signaling. In many other systems progenitor maintenance crucially depends on Notch signaling. Either the parapineal is different or the tbx2b positive cells are not self-renewing progenitors? Please discuss in more detail what is known about tbx2b, sox1a and gfi1ab positive cells.

Nothing is known about the lineage of PP cells. This question is indeed interesting, but we have not addressed it further as it is not the main message of our paper.

5) Regarding the hs:NICD experiments: Are control embryos also heat shocked? Please label figure accordingly.

Yes, control embryos are non-transgenic or single Tg(UAS:NICD) or Tg(hsp70:Gal4) transgenic embryos that were heat-shocked under the same conditions as the double transgenic embryos. We can mention this in the Materials and methods section or in the Figure Legend of a revised version of our manuscript.

6) Figure 2I: The time-line of treatment shown above the graph is very useful. It would facilitate the interpretation even further if it was indicated at what time point (32 hpf) migration begins.

The initiation of parapineal migration is variable but usually starts between 28hpf and 30 hpf; at 32 hpf parapineal position is obviously displaced from the midline. We can add a migration box (28-30hpf) on the time-line.

7) The fact that ectopic Notch signaling at late stages (26-28 hpf) can efficiently block migration argues for a role of Notch at the time when the parapineal initiates its migration. If Notch signaling is indeed required just before migration, it is unclear why LY411575 treatment (22 to 32 hpf) does not affect migration. Maybe activation of Notch by inducing NICD until 28hpf persists much longer and NICD is still active after 28hpf. Therefore, embryos should be treated with LY411575 beyond 32hpf, as Notch might be required for migration during these later stages.

See our reply concerning the major issues raised and Author response image 4.

In our original discussion, we suggest that LY411575 might not completely block Notch signaling and that the differential effect of LY411575 on specification and migration of parapineal cells could reflect a different requirement in Notch signaling threshold. To keep our original manuscript short, we did not include other speculative hypotheses.

8) Subsection “Notch activity is required early for unilateral activation of the Nodal pathway in the epithalamus”: 'To determine whether the parapineal migration defects we observed (add: in mib mutants) might be a direct…. In sentence in the paragraph above the authors conclude that there was no migration defect in LY411575 treated embryos, which is confusing.

Indeed, although LY411575 treatment from 22 to 32hpf promotes an increase in the number of gfi1ab positive parapineal cells, it does not affect parapineal migration. As suggested by the reviewer, we can add ‘in mib-/- mutants and rbpj morphant embryos’ in the previous sentence to remind the reader that PP migration defects were observed in these two contexts only.

9) Subsection “Decreasing FGF signaling rescues the parapineal migration defects in loss of Notch context while increasing FGF signaling aggravates it”. The authors show that suboptimal doses of the FGF inhibitor SU5402 rescue to some degree the migration defect in mib mutants, which are characterized by FGF signaling activation in all migrating cells. This experiment actually argues that it is not important that the tip cells possess more FGF signaling then the follower cells but rather that FGF signaling needs to be expressed at the correct level. The authors should discuss the results of this experiment in more detail.

We have replied to this point above. We can add a sentence in the discussion to mention this alternative hypothesis.

10) Subsection “Decreasing FGF signaling rescues the parapineal migration defects in loss of Notch context while increasing FGF signaling aggravates it”: 'Taken together, our.…experiments argue that the Notch pathway is required for parapineal migration, and that it acts by restricting FGF activation in parapineal cells.' However, in subsection “Notch activity is required early for unilateral activation of the Nodal pathway in the epithalamus” the authors argue that Notch and FGF act synergistically during migration, as rbpj morphant:cafgfr1a embryos display a more severe migration phenotype then CAfgfr1 embryos alone. Therefore, one has to conclude that Notch and FGF partially act in parallel on parapineal migration, correct?

By using ‘synergistic’, our goal was not to suggest that FGF and Notch act in parallel but rather that the two pathways appear to interact. Indeed, our data suggest that Notch acts upstream of the FGF pathway as LOF and GOF for Notch trigger an increase or decrease in FGF signaling in the PP. However, we understand from reviewer 1's comment that the wording chosen might be confusing. We can change the sentence ‘The interaction detected… ‘to ‘Notch and FGF pathways appears to interact in this context as the increase in parapineal migration defects observed in rbpja/b morphant embryos expressing the activated FGF receptor is significantly higher than expected from simply adding the effects of activated receptor transgene and rbpja/b MO injections alone.’

11) Epistasis: Notch and FGF could also act in a feedback loop. Is Notch signaling affected after manipulation of FGF signaling?

It is an interesting possibility. We could check her9 and her6 expression in fgf8 mutants or in SU5402 treated embryos. This said however, we feel it might dilute our message as this question is beyond the main focus of our study.

Reviewer #2:

The paper by Lu et al., is an excellent follow up on previous work from Roussigne which demonstrated a role for restricted FGF signaling in collective migration of the parapineal cells. In this study the authors make good use of a combination of transgenic tools, mutants and chemical inhibitors to demonstrate the role of Notch signaling in a particular time window in restricting FGF signaling to a limited number of parapineal cells to determine asymmetric collective migration. They show that it is required after a relatively early stage when Notch is required to determine asymmetric Nodal signaling and after a stage is required to restrict the number of cells with gf1ab and sox1a positive cells. The experiments are presented in a logical order, the data is presented well, and its analysis supports the conclusions of the authors.

I have no major problems with the paper in its current form. My issues were only tangentially related to the study.

We thank reviewer 2 for his enthusiasm.

At various points in the Introduction and Discussion section the authors make comparisons with the role of FGF and Notch signaling in the lateral line system. I found those references confusing and potentially misleading. I would have liked to hear more about the role of Notch signaling in restricting migratory potential in tracheal cells and endothelial cells in vascular development. While the tracheal and vascular systems are referenced, the comparison could be elaborated on.

The interaction between Notch and FGF/RTK pathways in the tracheal and vascular systems is not well understood but we could discuss it more if required.

In the Introduction the authors suggest that the Dalle Nogare paper shows a role for FGF signaling in maintaining cluster cohesion. It is not clear what the authors have in mind here because the primary role of FGF shown in that paper is to suggest a role for FGFs released by leading cells in providing a guidance cue for collective migration of trailing cells. Their description of FGFs role in cluster cohesion is not essential for any point they wish to establish in this study and may only serve to establish a misleading impression about conclusions in the Dalle Nogare paper in the minds of readers for those not familiar with that paper.

By writing ‘a role of FGF in cluster cohesion’, we were referring to a function of FGF leading to trailing signaling in preventing splitting of the primordium. Therefore, we fully agree with reviewer 2 on the interpretation of the data from Dalle Nogare et al. We understand that referring to ‘cluster cohesion’ to summarize these data is misleading and we can re-word this sentence.

Similarly, the authors contrast the role of Notch, downstream of FGF, in determining apical constriction in forming neuromasts in the migrating primordium, with the role of Notch in restricting the number of cells with FGF activity in the context of parapineal migration. The authors are not wrong here but they may give readers that not familiar with the Lateral Line system a simplistic idea about the role of Notch signaling in the primordium, where, as in the parapineal, it has multiple sequential roles. For example, first in restricting sensory hair cell progenitor fate to a central cell, and subsequently in determining Notch activation in neighboring cells to consolidate morphogenesis of epithelial rosettes. It is true that activation of Notch promotes apical constriction in the absence of FGF signaling and therefore in this context, downstream of FGF signaling and different in its epistatic relationship from the role of Notch in the parapineal. However, in the context of lateral inhibition and in restricting the number of cells with a particular fate/activity there may be important similarities in the role of Notch in the lateral line primordium and in parapineal cells. Without such a clarification, I felt the contrasting of the role of Notch in the parapineal and primordium could be potentially misleading.

To keep the manuscript‘s length within reason, we indeed decided to focus our discussion on the differences that exist in the interaction between Notch and FGF signaling in the PP (Notch upstream FGF) versus the LLP (FGF upstream Notch) rather than also including text on the role of Notch alone. We agree that, in both models (PP and LLP), the Notch pathway plays a similar role in restricting the number of differentiated cells. Nonetheless, we don’t feel we have the space to describe in detail the LLP system. We can add a sentence to mention the existence of similarities in the role of Notch between these two models, as suggested by reviewer 2 if requested.

A question that remained unanswered, which I was curious about and is not necessary in my mind for acceptance of the paper- which cells does Notch activity required in for its role in restricting migratory potential, which Notch ligands and receptors are required for this process and does their spatial distribution or that of downstream target genes provide insight about where Notch activity if required for restriction of FGF signaling in the parapineal.

See our reply above and Author response image 3 and Author response image 5.

Reviewer #3:

In this study by Wei et al., the authors investigate collective cell migration of parapineal cells during development using zebrafish as a model system. The authors found that global inhibition of Notch signaling, either by a chemical inhibitor or a MO-mediated knockdown, increased the number of cells that activate FGF signaling. As FGF is required for the parapineal migration and it is typically restricted to a few leading cells, this blocked the parapineal migration. Conversely, upregulating Notch signaling led to the loss of FGF signaling and also defects in its migration. These manipulations also affected a number of parapineal cells. Finally, the authors modulated FGF signaling in both Notch loss- and gain-of-function conditions. Decreasing FGF under this condition resulted in a partial rescue of the migration, whereas gain of FGF signaling worsened the migration phenotype. Interestingly, these manipulations did not affect the number of parapineal cells, indicating that these two roles of Notch signaling are distinct. Based on these data, the authors concluded that Notch restricts FGF signaling to a few leading parapineal cells, a process required for the proper migration. In addition, during later stages, Notch signaling controls the number of parapineal cells and this particular role of Notch can be uncoupled from its earlier role in restricting FGF signaling.

Overall, the conclusions are based on carefully done and controlled experiments (with a few exceptions). And while they provide some interesting insights into mechanisms of parapineal migration, the study is quite descriptive and does not go beyond the superficial analysis of cellular phenotypes. It is also not clear which cells express Notch ligands and receptors and whether effects on FGF signaling are direct. In summary, the overall findings are quite specialized and somewhat incremental and, thus, more suited for a specialized audience.

We thank reviewer 3 for acknowledging the quality of our work but we disagree with him/her on the fact that our findings will be more suited for specialized journal.

As mentioned in our cover letter to the editor, our work will be of interest to the community working on left-right brain asymmetry as it sets a framework to address how Nodal signaling provides a leftward bias to the Notch dependent restriction of FGF signaling that is required for parapineal laterality and subsequent habenular asymmetries. However, we believe that our study will also be of interest to groups interested in collective cell migration in general. Indeed, our findings potentially apply to other models of FGF dependent cell migration and, in the future, they would set the stage for studies investigating how Notch can restrict the activation of FGF pathway and how other factors or signaling pathways modulate this Notch dependent restriction in normal or pathological contexts of collective cell migration, including metastasis. For these reasons, we believe that our study will be of considerable interest to the broad audience readership of eLife.

Major comments:

Subsection “Loss of Notch signaling results in defects in parapineal migration”: Indicates that migration of the parapineal failed in 1/3 of mib-/-. I don't believe the authors discussed why this is the case/why the phenotype is partially penetrant in mib mutants? As well as why this phenotype was partially penetrant in the rbpj MO-injected embryos (with only 13% failing to migrate).

Why a single gene mutation results in a partially penetrant phenotype (expressivity) is a common question in genetics, and probably depend on various parameters such as individual variation in the genetic background, a redundancy of the mechanisms at play or intrinsic properties of the system that make it more or less robust.

In the case of parapineal migration, the number of Tg(dusp6:d2EGFP)+ cells is quite variable suggesting that parapineal migration tolerates variation in the degree of mosaicism of FGF pathway activation and/or in the level of FGF signaling. The general robustness of parapineal migration might also reflect that migration is very short in terms of distance travelled (measured in tens of µm) compared to, for instance, the migration of the lateral line primordium (LLP; measured in many hundreds of µm). As a consequence of the scale, intermediate migration phenotypes that can be observed for LLP migration might not be easily visible for PP migration. Indeed, in our study, we chose to define a parapineal as non-migrating when its mean position was found between -15 μm and +15 μm of the midline (the standard width of the epiphysis). One can notice, however, that in the mib-/- mutant embryos where the parapineal migrates further than -15µm, the parapineal often migrates less far from the midline than in wildtype (Figure 1N); we thus do observe a range of phenotypes that goes from no migration to delayed migration rather than a black or white phenotype.

rbpj a/b MO: given specific guidelines for MOs (Stainier et al., 2017), these experiments do not seem to be appropriately controlled.

These MO have been previously validated as their injection phenocopy the mib-/- mutant phenotype in neurogenesis and somites boundary (Echeverri and Oates, 2007). Furthermore, in our study it is only used as a second approach to confirm a role of the Notch pathway in parallel to our use of midbomb mutants. As the phenotypes are similar, we do not see a need to control as strictly as if we were only using the morpholinos.

The results related to the inhibitor treatment seem odd. If the γ-secretase inhibitor, LY411575, is specific and Notch is playing a role in migration and specification of this system, I would expect to see a result with treatment from the 22-32 hpf time period. This drug and it has only been published in two zebrafish papers. Why not use the γ-secretase inhibitor DAPT that has been utilized in many other studies. Related to this, I am confused that 8-22 hour inhibitor treatment did not interfere with the Tg(dusp6:d2EGFP) transgene expression, neither did the 22-32 hour treatment. What is then the time window for Notch signaling that controls Tg(dusp6:d2EGFP) expression?

We decided to use LY411575 as it is understood to be more powerful compared to the previously described γ-secretase inhibitor DAPT. For instance, in Rothenaigner et al., 2011, the authors compared the γ-secretase inhibitor DAPT with a dose range of its second-generation derivative, LY411575 and noted that ‘LY411575 (10 µM) proved as efficient as 100 µM DAPT in interfering with Notch signaling. As such, since its first description in 2007 (Fauq et al., 2007) and its first use in zebrafish (Rothenaigner et al., 2011), LY411575 has become the gold standard in the literature at the expense of DAPT; although numbers are difficult to assess, there are definitively more studies describing the use of LY411575 in zebrafish than just the two papers one can find referenced on Pubmed when using ‘zebrafish’ and ‘LY411575’ as searching keywords. For instance, recently, Kozlovskaja-Gumbrienė et al., (eLife 2017) or Thomas et al., (eLife 2019) used LY411575 to block Notch pathway in the migrating lateral line primordium or in neuromasts and these studies are not among the publication easily found in Pubmed with searching keywords.

Concerning the second point, treating with LY411575 between 8-22 hpf or 22-32 hpf does not block parapineal migration and we speculated in the discussion about potential reasons. However, the fact that normal migration correlates with no significant changes in the pattern of the Tg(dusp6:d2EGFP) FGF reporter transgene for both windows treatment support the conclusion of our study.

Throughout the manuscript, it is not clear where Notch signaling is active/ where Notch signaling components are expressed. It would be helpful to see the expression of where particular Notch signaling components are expressed during the time points of interest to further validate the reason behind studying Notch signaling. In other words, is Notch acting directly on parapineal cells?

See our reply concerning the major issues raised and Figure 3 and Figure 5 for reviewers.

[Editors’ notes: the authors’ response after being formally invited to submit a revised submission follows.]

Further to my previous email, the editors and reviewers have considered your plan and invite you to proceed with your revisions as proposed. The asked us to pass on the following comments:

The authors should expand their discussion, add the discussion points raised by the reviewers and add relevant figure panels.

As mentioned above, we have added 3 Figure supplements and discussed them to address the concerns raised by the reviewers. We have also made the changes to the Introduction, Results section, Discussion section, Materials and Methods section and Figure legends requested by the reviewers (See at the end below).

Specific comments:

In Figure 3 for reviewers they should add panels that only show the red signal.

As our Author response image 3 was based on a single experiment, we performed 2 more experiments to confirm her6 and her9 expression in the parapineal and the decrease of her6 in the parapineal of mib-/-mutants. The expression of her6 and her9 appear to be globally decreased in mib-/- mutants, as we observed previously, but also in LY411575 treated embryos. While we have managed to confirm the expression of both of these genes in the parapineal, we could not confirm a significant decrease in the number of her6+ cells in the parapineal of mib-/- mutants, despite the high number of embryos analyzed (n=43 controls and n=41 mib-/- embryos). One explanation is probably that fluorescent in situ hybridization are not quantitative enough to allow us to detect modest decreases in the level of expression or in the number of her expressing parapineal cells. In addition, her6 and her9 expression as detected by fluorescent in situ hybridizations is often weak and/or punctate, which makes the quantification difficult. While it remains possible that the expression of both her6 and/or her9 are targets of Notch pathway in the parapineal, as they appear to be elsewhere, we haven't managed to reproduce our initial results and have decided not to include the expression of her6 and her9 in our revised manuscript.

We have, on the other hand, included a Figure1—figure supplement 3 showing the expression of deltaB that supports the existence of active Notch signaling in the parapineal. We have also included 2 examples of embryos with Tg(Tp1:GFP) Notch reporter expression detected in the parapineal (n=3/19). The reason for the lack of robustness of expression of the Notch reporter line is unclear. Also, as Tg(Tp1:GFP) expression was too rarely detected in the parapineal, we have not analyzed it in mib-/- mutants or LY411575 treated embryos.

As LY411575 does not inhibit migration at any time point, the authors should test if this drug is able to completely reduce Notch signaling. Is the signal in the TP Notch reporter line gone after treatment? Possibly, a small amount of Notch signaling is sufficient to allow for migration.

As mentioned above and in the revised version of manuscript presented here, we did not detect robust expression of the Tg(Tp1:EGFP) reporter line in the parapineal. As such, we have not analyzed Tg(Tp1:GFP) expression in LY411575 treated embryos. To test our LY411575 treatment regime, we chose to analyze the expression of her6 and her9 (expressed in the parapineal of most embryos) and her4, a well-described consensus Notch target gene. We found that her4 expression is completely lost in both mib-/- mutants and upon LY411575 treatment (100µM from 22 to 32 hpf); see below our Figure for Reviewers. We also observed a global decrease in the expression of her6 and a slight reduction of her9 expression in LY411575 treated embryos as seen in mib-/- mutant embryos. Our data confirm previous work showing that the abundance of her6 transcripts is significantly reduced, but not lost, in the dorsal diencephalon and hindbrain of mib-/-mutants (Cunliffe, 2004). Previous studies have also already suggested that late her9 expression would be partly dependent on Notch signaling (while its early expression is not) (Bae, 2005). Thus, while our LY411575 treatment does not affect parapineal migration, it appears to block or reduce the expression of the 3 tested her genes as previously described in mib-/- mutants.

The expression data of Notch pathway members should be included and discussed in the manuscript.

As mentioned above, we have not included the expression of her6 and her9 in parapineal cells as we haven't managed to detect a robust decrease of their expression in mib-/- mutants and, thus, cannot confirm that they are Notch pathway components in the context of the parapineal. In the new version of our manuscript, however, we have included the expression data for deltaB, which can be found in one or two parapineal cells in about half of the embryos analyzed (Figure 1—figure supplement 3).

The text describing the different cell types should be clarified, as it is currently unclear which genes mark possibly the same cells or which genes label all cells.

As mentioned in our previous reply, parapineal cells diversity has yet to be thoroughly characterized in the literature, with all parapineal markers assumed to label all parapineal cells. Our data suggest that the pool of tbx2b+ parapineal cells differs from a sox1a/gfi1a+ pool. Our preliminary data from double fluorescent in situs shows that tbx2b and gfi1a indeed label two distinct pools of cells. However, while these findings are interesting, we fear that adding data on the characterization of parapineal cell subtypes might dilute the main message of our study that focuses on the role of Notch signaling in controlling the collective migration of parapineal cells through a restriction of FGF pathway activation. In our revised manuscript, we have nonetheless added/modified the text to clarify what is described in the literature and what is suggested by our study. This includes:

Subsection “Loss of Notch signaling results in defects in parapineal migration”: We added the sentence “We confirmed this result by using another parapineal specific marker, gfi1ab (Dufourcq et al., 2004), whose expression is detected at a later stage (from 36-40 hpf) than sox1a (from 28 hpf) in parapineal cells.”

Subsection “Loss of Notch signaling results in an increase in the number of certain parapineal cell subtypes”: We modified the sentence to read “In contrast, the number of parapineal cells expressing tbx2b, a parapineal marker previously suggested to be required for the specification of parapineal cells (Snelson et al., 2008), was not increased in mib-/- mutants (Figure 1J-1K, 1Q)."

Subsection “Notch effects on parapineal cell specification can be uncoupled from migration”: We added the sentence “Therefore, while in previous studies it was assumed that all described parapineal markers (tbx2b, sox1a, gfi1ab) label the same cells, our data indicate that the pool of tbx2b+ parapineal cells probably differs from a sox1a/gfi1a+ pool.”

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for submitting your article "Notch signaling restricts FGF pathway activation in parapineal cells to promote their collective migration" for consideration by eLife. Your article has been reviewed by Marianne Bronner as the Senior Editor, a Reviewing Editor, and two reviewers. 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. Please aim to submit the revised version within two months.

The reviewers felt that manuscript is much improved and addressed the main concerns to: (1) provide evidence that some Notch pathway signaling members are expressed in the parapineal during migration; (2) further investigate effects of the γ-secretase inhibitor treatment on the Notch pathway activation, and (3) clarify the dynamics of parapineal marker gene expression during PP development. However, the evidence that Notch pathway acts within parapineal cells remains weak: deltaB and the reporter expression are not robust enough to make that conclusion, and functional experiments testing this hypothesis are absent. Therefore, please tone down that conclusions and present alternative possibilities.

We will look forward to hearing from you with a revised article with tracked changes, and a response letter (uploaded as an editable file) describing the changes made in response to the decision and review comments.

We are grateful that the reviewers found our revised manuscript is an improvement and we would like to thank them again for their constructive comments.

Concerning the final request of the reviewers, as we detect mosaic expression of deltaB Notch ligand in the parapineal of a majority of embryos (1 or 2 cells in more than 2/3 of embryos at 32 hpf) and the expression of some Notch target genes (her6 and her9), our preferred hypothesis is that Notch signaling acts in the parapineal to restrict FGF pathway. However, we fully agree that we have not provided concrete proof of this, and have moderated our conclusion accordingly. For this, we have added sentences to our discussion mentioning that Notch signaling could be required in the tissue surrounding the parapineal rather than in the parapineal itself or could possibly act simultaneously within and outside parapineal cells. We believe that our modifications correspond to your expectations but would be happy to consider any specific suggestions to improve further this point.

https://doi.org/10.7554/eLife.46275.021

Article and author information

Author details

  1. Lu Wei

    Centre de Biologie Intégrative (CBI), Centre de Biologie du Développement (CBD), Université de Toulouse, CNRS (UMR 5547), Toulouse, France
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing—original draft
    Competing interests
    No competing interests declared

  2. Amir Al Oustah

    Centre de Biologie Intégrative (CBI), Centre de Biologie du Développement (CBD), Université de Toulouse, CNRS (UMR 5547), Toulouse, France
    Contribution
    Validation, Investigation
    Competing interests
    No competing interests declared

  3. Patrick Blader

    Centre de Biologie Intégrative (CBI), Centre de Biologie du Développement (CBD), Université de Toulouse, CNRS (UMR 5547), Toulouse, France
    Contribution
    Supervision, Funding acquisition, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3299-6108

  4. Myriam Roussigné

    Centre de Biologie Intégrative (CBI), Centre de Biologie du Développement (CBD), Université de Toulouse, CNRS (UMR 5547), Toulouse, France
    Contribution
    Conceptualization, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing
    For correspondence
    myriam.roussigne@univ-tlse3.fr
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-4240-4105

Funding

Fondation pour la Recherche Médicale (DEQ20131029166)

  • Patrick Blader

Fondation ARC pour la Recherche sur le Cancer (PJA 20131200173)

  • Patrick Blader

Agence Nationale de la Recherche (ANR-16-CE13-0013-01)

  • Patrick Blader

China Scholarship Council (CSC No.201504910807)

  • Lu Wei

Centre National de la Recherche Scientifique

  • Myriam Roussigné

Université Toulouse III - Paul Sabatier

  • Myriam Roussigné

Inserm

  • Patrick Blader

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We are grateful to all members of the Blader lab, especially Elise Cau and Julie Batut for critical reading of the manuscript and Aurélie Quillien for insightful discussions. We also thank Eric Theveneau and Matthias Carl for critical reading of the manuscript, as well as Xiaobo Wang and Steve W Wilson for their input. We thank Brice Ronsin from the Toulouse RIO Imaging platform and Aurore Laire for fish care. This work was supported by the Centre National de la Recherche Scientifique (CNRS), the Institut National de la Santé et de la Recherche Médicale (INSERM), Université de Toulouse III (UPS), the Fondation pour la Recherche Médicale (FRM; DEQ20131029166), the Fondation ARC (PJA 20131200173), the Agence nationale de la recherche (ANR-16-CE13-0013-01) and the China Scholarship Council (CSC No.201504910807) for PhD funding for Lu Wei.

Ethics

Animal experimentation: Fish were handled in a facility certified by the French Ministry of Agriculture (approval number A3155510). The project has received an agreement number APAFIS#3653-2016011512005922. Anaesthesia and euthanasia procedures were performed in Tricaine Methanesulfonate (MS222) solutions as recommended for zebrafish (0,16mg/ml for anaesthesia, 0,30 mg/ml for euthanasia). All efforts were made to minimize the number of animals used and their suffering, in accordance with the guidelines from the European directive on the protection of animals used for scientific purposes (2010/63/UE) and the guiding principles from the French Decret 2013-118.

Senior and Reviewing Editor

  1. Marianne E Bronner, California Institute of Technology, United States

Reviewer

  1. Ajay B Chitnis, Eunice Kennedy Shriver National Institute of Child Health and Human Development, United States

Publication history

  1. Received: February 20, 2019
  2. Accepted: July 18, 2019
  3. Version of Record published: September 9, 2019 (version 1)

Copyright

© 2019, Wei et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,196
    Page views
  • 141
    Downloads
  • 3
    Citations

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

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Lu Wei
  2. Amir Al Oustah
  3. Patrick Blader
  4. Myriam Roussigné
(2019)
Notch signaling restricts FGF pathway activation in parapineal cells to promote their collective migration
eLife 8:e46275.
https://doi.org/10.7554/eLife.46275

Categories and tags

Research organism

Further reading

    1. Cell Biology

    Insights into cargo sorting by SNX32 and its role in neurite outgrowth

    Jini Sugatha, Amulya Priya ... Sunando Datta
    Research Article Updated

    Sorting nexins (SNX) are a family of proteins containing the Phox homology domain, which shows a preferential endo-membrane association and regulates cargo sorting processes. Here, we established that SNX32, an SNX-BAR (Bin/Amphiphysin/Rvs) sub-family member associates with SNX4 via its BAR domain and the residues A226, Q259, E256, R366 of SNX32, and Y258, S448 of SNX4 that lie at the interface of these two SNX proteins mediate this association. SNX32, via its PX domain, interacts with the transferrin receptor (TfR) and Cation-Independent Mannose-6-Phosphate Receptor (CIMPR), and the conserved F131 in its PX domain is important in stabilizing these interactions. Silencing of SNX32 leads to a defect in intracellular trafficking of TfR and CIMPR. Further, using SILAC-based differential proteomics of the wild-type and the mutant SNX32, impaired in cargo binding, we identified Basigin (BSG), an immunoglobulin superfamily member, as a potential interactor of SNX32 in SHSY5Y cells. We then demonstrated that SNX32 binds to BSG through its PX domain and facilitates its trafficking to the cell surface. In neuroglial cell lines, silencing of SNX32 leads to defects in neuronal differentiation. Moreover, abrogation in lactate transport in the SNX32-depleted cells led us to propose that SNX32 may contribute to maintaining the neuroglial coordination via its role in BSG trafficking and the associated monocarboxylate transporter activity. Taken together, our study showed that SNX32 mediates the trafficking of specific cargo molecules along distinct pathways.

    1. Cell Biology
    2. Stem Cells and Regenerative Medicine

    Atf3 defines a population of pulmonary endothelial cells essential for lung regeneration

    Terren K Niethamer, Lillian I Levin ... Edward E Morrisey
    Research Article

    Following acute injury, the capillary vascular bed in the lung must be repaired to reestablish gas exchange with the external environment. Little is known about the transcriptional and signaling factors that drive pulmonary endothelial cell (EC) proliferation and subsequent regeneration of pulmonary capillaries, as well as their response to stress. Here, we show that the transcription factor Atf3 is essential for the regenerative response of the mouse pulmonary endothelium after influenza infection. Atf3 expression defines a subpopulation of capillary ECs enriched in genes involved in endothelial development, differentiation, and migration. During lung alveolar regeneration, this EC population expands and increases the expression of genes involved in angiogenesis, blood vessel development, and cellular response to stress. Importantly, endothelial cell-specific loss of Atf3 results in defective alveolar regeneration, in part through increased apoptosis and decreased proliferation in the endothelium. This leads to the general loss of alveolar endothelium and persistent morphological changes to the alveolar niche, including an emphysema-like phenotype with enlarged alveolar airspaces lined with regions that lack vascular investment. Taken together, these data implicate Atf3 as an essential component of the vascular response to acute lung injury that is required for successful lung alveolar regeneration.

    1. Cell Biology
    2. Microbiology and Infectious Disease

    Profiling the bloodstream form and procyclic form Trypanosoma brucei cell cycle using single-cell transcriptomics

    Emma M Briggs, Catarina A Marques ... Keith R Matthews
    Research Article Updated

    African trypanosomes proliferate as bloodstream forms (BSFs) and procyclic forms in the mammal and tsetse fly midgut, respectively. This allows them to colonise the host environment upon infection and ensure life cycle progression. Yet, understanding of the mechanisms that regulate and drive the cell replication cycle of these forms is limited. Using single-cell transcriptomics on unsynchronised cell populations, we have obtained high resolution cell cycle regulated (CCR) transcriptomes of both procyclic and slender BSF Trypanosoma brucei without prior cell sorting or synchronisation. Additionally, we describe an efficient freeze–thawing protocol that allows single-cell transcriptomic analysis of cryopreserved T. brucei. Computational reconstruction of the cell cycle using periodic pseudotime inference allowed the dynamic expression patterns of cycling genes to be profiled for both life cycle forms. Comparative analyses identify a core cycling transcriptome highly conserved between forms, as well as several genes where transcript levels dynamics are form specific. Comparing transcript expression patterns with protein abundance revealed that the majority of genes with periodic cycling transcript and protein levels exhibit a relative delay between peak transcript and protein expression. This work reveals novel detail of the CCR transcriptomes of both forms, which are available for further interrogation via an interactive webtool.