ABSTRACT
Basal stem cells of the epidermis continuously differentiate into keratinocytes and replenish themselves via self-renewal to maintain skin homeostasis. Numerous studies have attempted to reveal how basal cells undergo differentiation or self-renewal; however, this has been hampered by a lack of robust basal cell markers and analytical platforms that allow single-cell tracking. Here, we report that zebrafish integrin beta 4 is a useful marker for basal cell labelling, irrespective of the body region, stage and regenerative status. We employed Cre-loxP recombination in combination with live cell tracking of single basal clones in the caudal fin and investigated the embryonic origin and behaviour of basal cells during fish growth and homeostasis. Although most basal cells, including those in fins, became quiescent in the adult stage, genetic cell ablation showed that basal cells were reactivated to either self-renew or differentiate, depending on the injured cell type. Our study provides a simple and easy-to-use platform for quantitative in vivo imaging of basal stem cells at wider stages and under various conditions.
INTRODUCTION
The epidermis is the outermost tissue of the skin that protects internal tissues and organs throughout life. It has a flexible and multi-layered architecture that covers the entire body surface and prevents water loss and damage (Rawlings and Harding, 2004). Studies of the mammalian epidermis have shown that stem cells in the basal layer are responsible for maintaining epidermal integrity (Blanpain and Fuchs, 2007). Although basal stem cells differentiate into highly proliferative precursor cells, transit-amplifying cells (TACs) and keratinocytes (Potten, 1974; Clayton et al., 2007; Mascré et al., 2012; Koren et al., 2022), they also renew themselves to replenish the epidermal stem cell pool during animal growth and homeostasis.
In teleost fish such as zebrafish, the matured adult epidermis is covered by a thin mucosal barrier (Cabillon and Lazado, 2019) and mainly comprises three cell layers: basal cells of the epidermis in the innermost layer, suprabasal cells in the intermediate layer and differentiated keratinocytes in the outermost layer (Henrikson and Gedeon Matoltsy, 1967; Chang and Hwang, 2011). Similar to mammals, fish basal cells serve as stem cells for epidermal homeostasis (Lee et al., 2014) and cells in the suprabasal layer contain keratinocyte precursors, TACs (Guzman et al., 2013). Thus, by using transparent epidermal tissue, which is amenable to live imaging of fluorescent cells (Lee et al., 2014; Chen et al., 2016; Roan et al., 2021), the zebrafish epidermis provides an ideal system for analysing basal cell behaviour.
Despite similarities in epidermal structure and cellular turnover between fish and mammals, there is an apparent difference in their ability to repair injuries. Although fish perfectly regenerate the epidermal tissue after a severe injury such as fin amputation, mammals have limited regeneration ability and form a scar, an area of fibrous tissue rich in collagen, after severe injury (Marques et al., 2019). During zebrafish fin regeneration, regeneration-specific tissues, the wound epidermis (WE) and the blastema proliferating mesenchyme, which play essential roles in regeneration, are formed (Kawakami, 2010; Yoshinari and Kawakami, 2011; Shibata et al., 2016). Investigating the behaviour of basal stem cells in fish may open the way for improving skin regeneration ability in mammals.
We previously investigated the WE cell fate using the fibronectin 1b promoter and Cre-loxP-mediated cell tracking and suggested that the reconstitution of basal cells is key for epidermal regeneration (Shibata et al., 2018). However, most WE-derived cells, including basal cells, are not maintained in the epidermis in the long term; instead, they are replaced with new epidermal cells derived from the proximal non-WE region (Shibata et al., 2018).
The transcription factor ΔNp63 (also known as Tp63) is known to be a basal cell marker (Reischauer et al., 2009), and krtt1c19e promoter has also been used to label basal stem cells in zebrafish (Lee et al., 2014; Chen et al., 2016; Roan et al., 2021). However, comprehensive studies of the lineage and dynamics of basal cells throughout all life stages from embryo to adult has not been reported so far. To examine the fate and regulation of basal stem cells at a wider range of stages, tissue regions and regenerative statuses, robust basal cell promoters that can be used to label basal cells at wider stages and under various conditions are required.
In this study, we developed novel transgenic (Tg) tools that enable the marking and analysis of epidermal basal stem cells. Integrin beta 4 (itgb4) promoter drove the expression of fluorescent proteins in basal cells in all epidermal tissues and showed unbiased Cre recombination at all stages from 1 day post fertilisation (dpf) to adulthood. Keratin18 (krt18) promoter is active in basal cells during fin regeneration, but not in the homeostatic epidermis. Using these Tgs, we showed that basal cell precursors arise as early as 1 dpf during development and give rise to adult epidermal cells. Significantly, we also demonstrated, using genetic cell ablation, that quiescent basal cells were differentially reactivated to undergo differentiation or self-renewal depending on the injured cell types. Thus, this study showed that itgb4 and krt18 promoters and Tgs provide useful genetic tools for analysing and manipulating basal epidermal stem cells.
RESULTS
Generation of Cre transgenic lines with itgb4 or krt18 promoters
To identify markers of basal cells, we examined our unpublished RNA profiling data and found that itgb4 and krt18 genes are highly expressed in amputated and/or uncut zebrafish fins. The values of reads per kilobase of exon per million mapped reads (RPKM) of itgb4 at 12 h post amputation (hpa) were 87 in cut and 87 in uncut adult fins, and those of krt18 were 3465 and 104 in cut and uncut fins, respectively. Additionally, these genes were suggested to be expressed in basal cells of the epidermis (DeWard et al., 2014; Wehner et al., 2014; Khan et al., 2022; Human Protein Atlas database, https://www.proteinatlas.org). Indeed, in situ hybridisation (ISH) analysis showed that itgb4 expression was uniform in the basal layer of the epidermis in both uncut and cut fins (Fig. 1A). Unlike itgb4, krt18 was not detected in the epidermis of the uncut fin, but was induced in the WE by fin amputation. A weak krt18 expression was also observed in the blastema mesenchymal cells. (Fig. 1A).
Generation of Cre Tg lines with itgb4 or krt18 promoters. (A) Whole-mount ISH analysis of itgb4 and krt18 expressions in adult zebrafish fin. dpa, days post amputation. Respective longitudinal sections are shown on the right side. itgb4 is expressed in the basal layer of amputated and uncut fin, whereas krt18 is only expressed in the regenerating fin. Arrowheads show the basal layer of the epidermis. Note that the ISH signal of krt18 is also seen in the blastema (bl). Scale bars: 100 µm (whole mount); 30 µm (section). (B) Diagram of the Tol2 transposon-based vector used to generate the Tgs. cryaa, crystalline alpha a promoter. (C) Scheme illustrating the Cre-mediated lineage tracking. The double Tgs of Cre-expressing line and the loxP reporter line, Tg(Olactb:loxP-dsred2-loxP-egfp) (RSG), were treated with tamoxifen (TAM) or fulvestrant (ICI) to induce the recombination at loxP sites. The resulting EGFP expression persists in all progeny cells. pA, polyadenylation signal sequence. (D,E) EGFP expression in amputated and uncut fins of Tg(krt18:creERt2) (D) and Tg(itgb4:creERt2) (E). EGFP expression is also seen in a fraction of mesenchymal cells within the fin ray of Tg(krt18:creERt2). TAM was added at 3 dpa (Amputated) and all images were taken at 4 days post labelling (dpl). Dotted line, amputation plane. Dashed line, basement membrane. Scale bars: 300 µm (left); 50 µm (right). White arrows indicate the posterior side (P). Experimental procedures are schematically shown below each image.
Generation of Cre Tg lines with itgb4 or krt18 promoters. (A) Whole-mount ISH analysis of itgb4 and krt18 expressions in adult zebrafish fin. dpa, days post amputation. Respective longitudinal sections are shown on the right side. itgb4 is expressed in the basal layer of amputated and uncut fin, whereas krt18 is only expressed in the regenerating fin. Arrowheads show the basal layer of the epidermis. Note that the ISH signal of krt18 is also seen in the blastema (bl). Scale bars: 100 µm (whole mount); 30 µm (section). (B) Diagram of the Tol2 transposon-based vector used to generate the Tgs. cryaa, crystalline alpha a promoter. (C) Scheme illustrating the Cre-mediated lineage tracking. The double Tgs of Cre-expressing line and the loxP reporter line, Tg(Olactb:loxP-dsred2-loxP-egfp) (RSG), were treated with tamoxifen (TAM) or fulvestrant (ICI) to induce the recombination at loxP sites. The resulting EGFP expression persists in all progeny cells. pA, polyadenylation signal sequence. (D,E) EGFP expression in amputated and uncut fins of Tg(krt18:creERt2) (D) and Tg(itgb4:creERt2) (E). EGFP expression is also seen in a fraction of mesenchymal cells within the fin ray of Tg(krt18:creERt2). TAM was added at 3 dpa (Amputated) and all images were taken at 4 days post labelling (dpl). Dotted line, amputation plane. Dashed line, basement membrane. Scale bars: 300 µm (left); 50 µm (right). White arrows indicate the posterior side (P). Experimental procedures are schematically shown below each image.
Using the promoters of itgb4 and krt18, we generated transgenic lines (Tgs) expressing the CreERt2 fusion protein (Fig. 1B,C). When the respective lines were crossed with the Cre reporter Tg(Olactb:loxP-dsRed2-loxP-egfp), hereafter termed RSG (Yoshinari et al., 2012), and Cre recombination with 4-OH tamoxifen (TAM) was induced, many EGFP+ cells appeared in the WE of the krt18:creERt2 line and the basal cells of the uncut itgb4:creERt2 line (Fig. 1D,E). In the krt18:creERt2 line, highly efficient Cre labelling was observed, but the overall fates of labelled cells were nearly the same as those observed by fn1b:creERt2, and they mostly disappeared by 3 weeks post amputation (Shibata et al., 2018).
In the Tg(itgb4:creERt2) line, the distribution of EGFP+ cells was observed throughout the body, including in the trunk and fin regions, at all stages of fish life after 1 dpf (Fig. S1A,B). However, in both cre Tg lines, EGFP+ cells that emerged in the absence of TAM were rare (Fig. S1C). Six and 42 EGFP+ cell clones were observed in ten juvenile and ten adult fish, respectively, in the Tg(itgb4:creERt2), and only 1 EGFP+ cell clone in ten adult fish of the Tg(krt18:creERt2), indicating that the TAM-independent recombination rarely occurs in the basal cells. Thus, these data demonstrate that these Cre Tgs are useful for highly specific and unbiased labelling of epidermal basal cells under homeostatic and/or regenerative conditions.
itgb4+ cells at 1 dpf are the origin of adult basal stem cells
In teleost fish, the enveloping layer, a monolayer of cells surrounding the embryo in the blastula, gives rise to the outermost epidermal layer termed the periderm, which covers the body surface during embryonic stage (Bouvet, 1976; Kimmel et al., 1990). As the periderm does not contribute to the adult epidermis (Kimmel et al., 1990), the basal cell precursors are expected to emerge within non-periderm epithelial cells.
We sought to investigate the time at which basal stem cell precursors emerge during development. ISH analysis of krt4 (Chen et al., 2011), krt18 and itgb4 during embryonic development showed that krt4 and krt18 expression was observed in entire epithelial cells at 50% epiboly stage, but krt18 expression was downregulated from the somite stage onward. Complementary to krt18, itgb4 expression became evident only after 24 h post fertilisation (hpf) (Fig. 2A). To examine gene expression in individual cells, we generated a series of Tgs to visualise gene expression (Fig. S2). During early embryonic stages before 10 hpf, the expression of krt4 and krt18 was observed both in the periderm and non-periderm cells (Fig. S2A). In subsequent stages, krt18+ or krt4+ cells were confined to the upper epithelial layer by 24 hpf (Fig. S2B,C; Le Guellec et al., 2004; Shibata et al., 2018), where the cells were mosaic in their expression of krt4 and krt18 (Fig. 2B,C). In contrast, the deeper epithelial layer was occupied by itgb4+ cells, which were negative for krt4 and krt18 (Fig. 2B,C). By 3 dpf, nearly all upper cells became krt4+/krt18− (Fig. 2C; Fig. S2C).
Epidermal cells become mosaic at 15-somite stage. (A) ISH analysis of krt4, krt18 and itgb4 expressions. krt4 was detected in all embryonic stages. krt18 was expressed in early embryonic stage, but the expression declined after the somite stage. itgb4 was only detected after 24 hpf. At the somite stage onward, epithelial cells have a mosaic expression of krt4 and krt18. Scale bars: 50 µm. (B) Confocal images of upper and lower layers of epithelial cells of Tg(krt4:tagBFP-nfsB), Tg(krt18:mcherry) or Tg(itgb4:mcherry) in the trunk region at 24 hpf. Scale bar: 20 µm. (C) Confocal images of the cross section of Tg(krt18:mcherry;itgb4:tagBFP-nfsB) (upper panels) and Tg(krt4:tagBFP-nfsB;itgb4:mcherry) (lower panels) in the middle trunk region at 24 hpf and 72 hpf. Fluorescence of krt4, krt18, itgb4 and SYTO11 are shown in blue, red, green and white pseudocolours, respectively. The expression of itgb4 is localised in the basal layer. Scale bar: 20 µm. Boxed areas (broken lines) are shown in the lower panels. Curved dashed lines indicate basement membrane.
Epidermal cells become mosaic at 15-somite stage. (A) ISH analysis of krt4, krt18 and itgb4 expressions. krt4 was detected in all embryonic stages. krt18 was expressed in early embryonic stage, but the expression declined after the somite stage. itgb4 was only detected after 24 hpf. At the somite stage onward, epithelial cells have a mosaic expression of krt4 and krt18. Scale bars: 50 µm. (B) Confocal images of upper and lower layers of epithelial cells of Tg(krt4:tagBFP-nfsB), Tg(krt18:mcherry) or Tg(itgb4:mcherry) in the trunk region at 24 hpf. Scale bar: 20 µm. (C) Confocal images of the cross section of Tg(krt18:mcherry;itgb4:tagBFP-nfsB) (upper panels) and Tg(krt4:tagBFP-nfsB;itgb4:mcherry) (lower panels) in the middle trunk region at 24 hpf and 72 hpf. Fluorescence of krt4, krt18, itgb4 and SYTO11 are shown in blue, red, green and white pseudocolours, respectively. The expression of itgb4 is localised in the basal layer. Scale bar: 20 µm. Boxed areas (broken lines) are shown in the lower panels. Curved dashed lines indicate basement membrane.
To examine the timing of fate decision of basal stem cells, we used the Cre Tgs that have krt4, krt18 or itgb4 promoters and tracked their respective cell fates. When labelling with krt4:cre or krt18:cre was performed at the pre-somite stage (TAM 0-10 hpf), EGFP+ cells gave rise to epidermal cells in larvae and adult fish (Fig. 3A), indicating that embryonic epithelial cells before 10 hpf contain the future basal cells. However, when labelling was performed at the somite stage (TAM 10-24 hpf) or later (TAM 1-2 dpf), EGFP+ cells labelled with krt4:cre or krt18:cre disappeared from the epidermis without contributing to the adult epidermis (Fig. 3B,C). The results are consistent with the data reported by Lee et al. (2014) and further suggest that the fate decision to basal cells occurs by 10 hpf.
Fate segregation of basal cells from embryonic epithelial cells. (A) Tracking of epithelial cells that are krt4+ or krt18+ in the pre-somite stage. Tamoxifen (TAM) (1 µM) treatment was carried out between 0 and10 hpf. In both Tgs, progenies of EGFP-labelled cells were maintained in the adult fin epidermis including the basal cells and keratinocytes. (B) Tracking of epithelial cells that are krt4+ or krt18+ in the somite stage. EGFP+ cells were induced at 3 dpf (left). In contrast to A, progenies of labelled cells disappeared by the adult stage in both Tgs, except a few remaining surface keratinocytes. In krt18:cre-labelling, a group of mesenchymal cells that were maintained until the adult stage were observed (arrows). (C) Tracking of epithelial cells that are krt4+ or itgb4+ during 1-2 dpf. The progenies of itgb4+ cells contribute to the adult epidermis including the fins (left panels), whereas the progeny of krt4+ cells mostly disappeared by the adult stage (right panels). White arrows indicate the posterior side (P). Respective experimental procedures are schematically shown. Scale bars: 100 µm (2-4 dpf larva and adult fin); 50 µm (section). Straight dashed lines indicate the place of the optical section (right panels). Curved dashed lines indicate basement membrane.
Fate segregation of basal cells from embryonic epithelial cells. (A) Tracking of epithelial cells that are krt4+ or krt18+ in the pre-somite stage. Tamoxifen (TAM) (1 µM) treatment was carried out between 0 and10 hpf. In both Tgs, progenies of EGFP-labelled cells were maintained in the adult fin epidermis including the basal cells and keratinocytes. (B) Tracking of epithelial cells that are krt4+ or krt18+ in the somite stage. EGFP+ cells were induced at 3 dpf (left). In contrast to A, progenies of labelled cells disappeared by the adult stage in both Tgs, except a few remaining surface keratinocytes. In krt18:cre-labelling, a group of mesenchymal cells that were maintained until the adult stage were observed (arrows). (C) Tracking of epithelial cells that are krt4+ or itgb4+ during 1-2 dpf. The progenies of itgb4+ cells contribute to the adult epidermis including the fins (left panels), whereas the progeny of krt4+ cells mostly disappeared by the adult stage (right panels). White arrows indicate the posterior side (P). Respective experimental procedures are schematically shown. Scale bars: 100 µm (2-4 dpf larva and adult fin); 50 µm (section). Straight dashed lines indicate the place of the optical section (right panels). Curved dashed lines indicate basement membrane.
In contrast to krt4 and krt18, itgb4:cre-labelled cells efficiently contributed to the adult epidermis, including basal stem cells (Fig. 3C), and were maintained throughout life (Fig. S3). The itgb4:cre-labelled cells gradually increased in the epidermis and occupied over 90% of the epidermal cells by 42 dpf, whereas the krt4:cre-labelled cells, which decreased in an inverse proportion (Fig. S4), did not contribute to adult epidermis. These observations showed that the itgb4 promoter is specific to basal cell precursors, offering a powerful tool for labelling and tracking basal stem cells at various stages and regions.
Identity of basal stem cells is evident immediately after fate commitment
Next, to investigate cellular identities, we isolated itgb4+ basal cell precursors, krt4+ upper epithelial cells and adult basal cells by fluorescence-activated cell sorting (Fig. S5A). The krt4+ cells were isolated from Tg(krt4:creERt2;RSG), in which krt4-expressing upper cells were labelled with TAM at 10-24 hpf (Fig. S5A). This was performed to precisely sort cells expressing krt4 during the 10-24 hpf stage.
By comparing the transcripts enriched in krt4+ and itgb4+ cells at 1 dpf, we identified 147 enriched genes in itgb4+ basal precursors and 381 enriched genes in krt4+ upper epithelial cells (Fig. S5B; Tables S1, S2). Among the 147 genes in basal precursors, 32% (47 genes) were significantly expressed in adult basal cells, but not in krt4+ epithelial cells (Fig. S5B). Among the 381 enriched genes in krt4+ upper cells, 78% (297 genes) showed a low expression in the embryonic and adult basal cells. Gene ontology (GO) analysis indicated that the 47 genes specific to basal precursors were characteristic of basal cell functions, such as focal adhesion, cell-matrix adhesion and integrin binding (Ghorbani et al., 2022) (Fig. S5C), whereas many of the krt4+ cell-specific genes had differentiation-associated profiles (Fig. S5D). Taken together, these data suggest that itgb4+ basal precursors already possess a basal cell-like identity immediately after fate segregation at 1 dpf.
Basal cell regulation during growth and homeostasis
Next, we used Tg(itgb4:creERt2) to look at the behaviour of basal precursors during fish growth and homeostasis. As Cre labelling by TAM produced too many EGFP+ cells to track individual cells (Fig. S6), we used the oestrogen receptor antagonist fulvestrant (ICI-182780, hereafter referred to as ICI) to sparsely label itgb4+ basal cells of the epidermis (Shibata et al., 2018; Fig. S6).
We firstly labelled itgb4+ cells at the juvenile stage (1 month old) and tracked the cell fate. Cell behaviour observed in a 1-week time window was categorised into three proliferation patterns: fast-proliferating (more than two cell divisions, 29.2%), slow-proliferating (one or two cell divisions, 35.3%), and quiescent (no cell division, 35.5%) (Fig. 4A-C). Confocal optical sections at 9 days post labelling (dpl) showed that the fast-proliferating clones were composed of only non-basal cells. This is interpreted as EGFP-labelled basal cells were differentiated into proliferating keratinocyte precursors, TACs, in the suprabasal layer. In contrast to the fast clones, the slow-proliferating and quiescent clones contained only basal cells (Fig. 4B,C). As far as we examined, we did not observe a regional difference of distribution of quiescent, fast and slow clones.
Tracking of the basal cell behaviour in juvenile fish. (A) Tracking of the itgb4:cre-labelled cells at a single cell resolution in the 1-month-old juvenile fish fin. Fish were treated with 0.5 µM ICI to sparsely label the itgb4+ basal cells. dpl, days post labelling. The same region of the tissue was imaged in a 1-week time window (2-9 dpl). The EGFP+ cells displayed three proliferation patterns: quiescent (1 cell, blue arrowheads), slow-proliferating (2-4 cells, purple arrowheads) and fast-proliferating (>4 cells, orange arrowheads). White arrows indicate the posterior end (P) of fins. Experimental procedure is shown on the left. Scale bar: 200 µm. (B) Confocal images and their optical sections of fast-proliferating and slow-proliferating clones in the juvenile fish fin. Note that the fast-proliferating clone only contains non-basal cells, whereas the slow-proliferating clone only contains the basal cells. White arrows indicate the posterior side (P). Scale bars: 20 µm. (C) Ratios of clones with respective proliferation patterns (n=312 clones from five juvenile zebrafish). (D) Experimental scheme of long-term tracking of clonal cell fates. Time-lapse images were taken at intervals of 7 days. (E) Representative image of fast-proliferating clones. They quickly proliferated, but finally disappeared in the long run, suggesting that such clones did not contain the basal stem cells. Scale bar: 100 µm. (F) Representative image of quiescent clone. These clones kept quiescent for more than 79 dpl. Scale bar: 50 µm. (G) Representative image of slow-proliferating clone, which slowly expanded the clone size and maintained for more than 79 days. Scale bar: 30 µm. White arrowheads indicate examples of tracked clones.
Tracking of the basal cell behaviour in juvenile fish. (A) Tracking of the itgb4:cre-labelled cells at a single cell resolution in the 1-month-old juvenile fish fin. Fish were treated with 0.5 µM ICI to sparsely label the itgb4+ basal cells. dpl, days post labelling. The same region of the tissue was imaged in a 1-week time window (2-9 dpl). The EGFP+ cells displayed three proliferation patterns: quiescent (1 cell, blue arrowheads), slow-proliferating (2-4 cells, purple arrowheads) and fast-proliferating (>4 cells, orange arrowheads). White arrows indicate the posterior end (P) of fins. Experimental procedure is shown on the left. Scale bar: 200 µm. (B) Confocal images and their optical sections of fast-proliferating and slow-proliferating clones in the juvenile fish fin. Note that the fast-proliferating clone only contains non-basal cells, whereas the slow-proliferating clone only contains the basal cells. White arrows indicate the posterior side (P). Scale bars: 20 µm. (C) Ratios of clones with respective proliferation patterns (n=312 clones from five juvenile zebrafish). (D) Experimental scheme of long-term tracking of clonal cell fates. Time-lapse images were taken at intervals of 7 days. (E) Representative image of fast-proliferating clones. They quickly proliferated, but finally disappeared in the long run, suggesting that such clones did not contain the basal stem cells. Scale bar: 100 µm. (F) Representative image of quiescent clone. These clones kept quiescent for more than 79 dpl. Scale bar: 50 µm. (G) Representative image of slow-proliferating clone, which slowly expanded the clone size and maintained for more than 79 days. Scale bar: 30 µm. White arrowheads indicate examples of tracked clones.
We also performed long-term cell tracking for 3 months and observed that many of the fast-proliferating clones disappeared from the epidermis (Fig. 4D,E), confirming the above observation that the fast clones did not contain basal stem cells. This observation is consistent with the view that basal cells differentiate without self-renewing and retaining labelled basal cells (Jones and Watt, 1993; Suzuki and Senoo, 2012). Although the quiescent clones did not proliferate until 79 dpl (Fig. 4F), the slow-proliferating clones gradually produced more basal cells over time (Fig. 4G).
In contrast to the juvenile stage, as many as 70.5% of the labelled basal cells were quiescent and only 3.5% were fast-proliferating within a 2-week time window in the adult fin epidermis (>3 months old) (Fig. 5). The turnover rate of differentiated epidermal cells (∼37 days from Fig. 4E, n=5 fish) was comparable with that of juvenile fish (∼38 days as shown in Fig. S4, n=5 fish) and mammals (28-45 days; Hoath and Leahy, 2003). Thus, Tg(itgb4:creERt2) in combination with sparse Cre recombination by ICI enables long-term tracking of individual basal stem clones.
Tracking of the basal cell behaviour in adult fish. (A) Tracking of the itgb4:cre-labelled cells at a single cell resolution in adult fish fin (3 months old). The images were taken from the same fin region of the fish. White arrows indicate the posterior side (P). Experimental procedure is shown at the top. Arrowheads indicate representative corresponding cells between these panels. Most of the labelled clones were quiescent in a time window between 2-14 dpl. Scale bar: 200 µm. (B) Ratios of clones with respective proliferation patterns in the adult fin epidermis (n=897 clones from six adult zebrafish). (C) Confocal optical sections of itgb4:cre-labelled cells in the adult fin epidermis. Basal cells were labelled with tamoxifen (TAM) treatment. The basal cells did not produce the non-basal cells. Scale bar: 50 µm. Dashed lines indicate basement membrane.
Tracking of the basal cell behaviour in adult fish. (A) Tracking of the itgb4:cre-labelled cells at a single cell resolution in adult fish fin (3 months old). The images were taken from the same fin region of the fish. White arrows indicate the posterior side (P). Experimental procedure is shown at the top. Arrowheads indicate representative corresponding cells between these panels. Most of the labelled clones were quiescent in a time window between 2-14 dpl. Scale bar: 200 µm. (B) Ratios of clones with respective proliferation patterns in the adult fin epidermis (n=897 clones from six adult zebrafish). (C) Confocal optical sections of itgb4:cre-labelled cells in the adult fin epidermis. Basal cells were labelled with tamoxifen (TAM) treatment. The basal cells did not produce the non-basal cells. Scale bar: 50 µm. Dashed lines indicate basement membrane.
Basal cell differentiation or self-renewal by injuries in keratinocytes or basal cells
Next, we investigated how basal cell self-renewal and differentiation are regulated in response to injury. To this end, we injured keratinocytes or basal stem cells using Tgs expressing the Escherichia coli nitroreductase nfsB under the control of krt4 or itgb4 promoters. Bacterial nitroreductase catalyses the reduction of the innocuous prodrug metronidazole (Mtz), thereby producing a cytotoxic product that induces cell death (Curado et al., 2008; Ando et al., 2017). Basal cells were labelled 4 days before ablation, and Mtz treatment was repeated at 2-day intervals (Fig. 6A). The apoptosis of the respective cells was successfully induced without causing fish death (Fig. S7).
Induction of basal cell differentiation by non-basal keratinocyte injury. (A) Experimental procedure of basal cell labelling and krt4+ cell injury. Mtz, metronidazole treatment. (B) Confocal images and their optical sections of the EGFP-labelled basal cells in the adult fin before (left panel) and after (6 dpt, right panel) non-basal cell injury. Thin white arrows indicate posterior side (P). Dashed lines indicate basement membrane. After injury of krt4+ cells, the basal cells often became the krt4+ keratinocytes at 6 dpt. Note that such clones did not contain the basal cells, indicating that the basal cells differentiated into non-basal cells (thick white arrows). Scale bar: 20 µm. (C) Quantification of the basal cell reactions in response to non-basal cell injury. The number of clones was counted on confocal images (n=156 clones at 0 dpt, n=276 clones at 6 dpt from four adult fish fins). Respective proliferative patterns are schematically shown above the graph. Error bars denote mean±s.e.m. ***P<0.001 (paired two-tailed Student's t-test). dpt, days post treatment.
Induction of basal cell differentiation by non-basal keratinocyte injury. (A) Experimental procedure of basal cell labelling and krt4+ cell injury. Mtz, metronidazole treatment. (B) Confocal images and their optical sections of the EGFP-labelled basal cells in the adult fin before (left panel) and after (6 dpt, right panel) non-basal cell injury. Thin white arrows indicate posterior side (P). Dashed lines indicate basement membrane. After injury of krt4+ cells, the basal cells often became the krt4+ keratinocytes at 6 dpt. Note that such clones did not contain the basal cells, indicating that the basal cells differentiated into non-basal cells (thick white arrows). Scale bar: 20 µm. (C) Quantification of the basal cell reactions in response to non-basal cell injury. The number of clones was counted on confocal images (n=156 clones at 0 dpt, n=276 clones at 6 dpt from four adult fish fins). Respective proliferative patterns are schematically shown above the graph. Error bars denote mean±s.e.m. ***P<0.001 (paired two-tailed Student's t-test). dpt, days post treatment.
Upon injury to krt4+ non-basal cells, differentiation of basal cells into non-basal cells and their rapid proliferation were induced (Fig. 6B,C). However, basal cell self-renewal was rarely observed. In contrast to non-basal injury, basal cell injury induced basal cell self-renewal, whereas basal cell differentiation was rarely observed (Fig. 7A,B). These results indicate that injury to non-basal keratinocytes or basal cells causes different signals for basal cells to adopt either self-renewal or differentiation.
Induction of basal cell self-renewal by basal cell injury. (A) Confocal images and their optical sections of the EGFP-labelled basal cells in the adult fin before (left panel) and after (4 dpt, right panel) basal cell injury. Experimental procedure is shown at the top. Dashed lines indicate basement membrane. After the injury of itgb4+ basal cells, many of the labelled basal cells stayed in the basal layer. They either remained quiescent or slowly proliferated to self-renew the basal cells (thick white arrows). Scale bar: 20 µm. (B) Quantification of A. The number of clones was counted on confocal images (n=460 clones at 0 dpt, n=445 clones at 4 dpt from four adult fish fins). Error bars denote mean±s.e.m. ***P<0.001 (paired two-tailed Student's t-test). (C) Induction of krt18 expression after basal cell injury but not by the keratinocyte injury as revealed by Tg(krt18:mcherry) (upper panels) and ISH analyses (longitudinal fin sections, lower panels). Experimental procedure is shown at the top. Mtz, metronidazole treatment. All images were obtained from 1 dpt fin. krt18 expression was only induced by the basal cell injury. Scale bars: 100 µm (upper panels); 20 µm (lower panels). Arrowhead indicates krt18+ cells in the basal layer. (D) Experimental procedure of basal cell injury and the following krt18+ cell labelling. (E) Tracking of krt18-expressed cells in the fin that were induced by basal cell injury. Confocal 3D images were taken at 3, 5, 7 dpt from the same location of the caudal fin. Scale bar: 20 µm. (F) Relative ratio of the type of clones that are labelled by krt18:cre. The number of clones, either self-renewal or differentiation, was counted on the confocal images at 7 dpt (n=75 clones from five fish fins). Most of the labelled cells underwent self-renewal to replenish the itgb4+ basal cells. Error bars denote mean±s.e.m. ***P<0.001 (unpaired two-tailed Student's t-test). Thin white arrows indicate posterior side (P). dpt, days post treatment.
Induction of basal cell self-renewal by basal cell injury. (A) Confocal images and their optical sections of the EGFP-labelled basal cells in the adult fin before (left panel) and after (4 dpt, right panel) basal cell injury. Experimental procedure is shown at the top. Dashed lines indicate basement membrane. After the injury of itgb4+ basal cells, many of the labelled basal cells stayed in the basal layer. They either remained quiescent or slowly proliferated to self-renew the basal cells (thick white arrows). Scale bar: 20 µm. (B) Quantification of A. The number of clones was counted on confocal images (n=460 clones at 0 dpt, n=445 clones at 4 dpt from four adult fish fins). Error bars denote mean±s.e.m. ***P<0.001 (paired two-tailed Student's t-test). (C) Induction of krt18 expression after basal cell injury but not by the keratinocyte injury as revealed by Tg(krt18:mcherry) (upper panels) and ISH analyses (longitudinal fin sections, lower panels). Experimental procedure is shown at the top. Mtz, metronidazole treatment. All images were obtained from 1 dpt fin. krt18 expression was only induced by the basal cell injury. Scale bars: 100 µm (upper panels); 20 µm (lower panels). Arrowhead indicates krt18+ cells in the basal layer. (D) Experimental procedure of basal cell injury and the following krt18+ cell labelling. (E) Tracking of krt18-expressed cells in the fin that were induced by basal cell injury. Confocal 3D images were taken at 3, 5, 7 dpt from the same location of the caudal fin. Scale bar: 20 µm. (F) Relative ratio of the type of clones that are labelled by krt18:cre. The number of clones, either self-renewal or differentiation, was counted on the confocal images at 7 dpt (n=75 clones from five fish fins). Most of the labelled cells underwent self-renewal to replenish the itgb4+ basal cells. Error bars denote mean±s.e.m. ***P<0.001 (unpaired two-tailed Student's t-test). Thin white arrows indicate posterior side (P). dpt, days post treatment.
We further found that krt18 expression was induced by injury to basal cells, but not to non-basal keratinocytes (Fig. 7C). Moreover, we labelled and tracked krt18-expressing cells using Tg(krt18:creERt2), in which krt18+ cells were labelled 1 day after basal cell injury (Fig. 7D). Tracking revealed that most of the krt18-expressing cells self-renewed to replenish basal cells (Fig. 7E,F), indicating that basal cells go through the krt18+ intermediate state to reproduce new basal cells.
DISCUSSION
In this study, we established new Tgs with itgb4 and krt18 promoters to label epidermal basal cells and studied their fate and behaviour. Tracking of itgb4+ cells showed that basal cells are derived from itgb4+ precursors at 1 dpf, and that cells derived from itgb4+ precursors expel embryonic epithelial cells during the larval and juvenile stages. Although the basal stem cells became mostly quiescent in adult fish, we found that injury to keratinocytes or basal cells induces different basal cell reactions, keratinocyte differentiation or basal cell self-renewal (Fig. S8).
Previously, studies have used p63 and krtt1c19e as epidermal basal cell markers. However, comprehensive studies of the lineage and dynamics of basal cells at all stages from embryo to adult have not been reported so far. We identified in this paper that itgb4 is specifically and uniformly expressed in the basal layer of the epidermis throughout life. In contrast to itgb4, krt18 is only induced in the regenerating epidermis, including basal cells, but the WE cells can be efficiently labelled by the Cre Tg. These promoters and Tgs are useful tools for investigating basal cell behaviour in both homeostatic and regenerative situations.
Despite accumulating knowledge on developmental processes, little is known about how embryonic cells give rise to adult tissues in many organs. In this study, we showed that basal cell fate decision occurs as early as 1 dpf in a population of embryonic epithelial cells. Our observation using itgb4:cre Tg is largely consistent with previous observations using Tg(krtt1c19e:cre) (Lee et al., 2014), but further demonstrated that basal cell fate is determined before 1 dpf as an itgb4+ cell population. The mechanism that induces the basal stem cell fate within the embryonic epithelial cells is intriguing, but it remains a topic for future research.
Using itgb4:cre Tg, we also performed live lineage tracking of the basal cell clones to examine their individual clonal fates during growth and homeostasis. Tracking suggested that itgb4+ cells adopt three different cell proliferative statuses: fast-proliferating, slow-proliferating and quiescent. These basal stem cell behaviours are consistent with those observed in mice (Clayton et al., 2007; Li and Clevers, 2010; Damen et al., 2021; Koren et al., 2022). Thus, our study indicated that not only the epidermal tissue architecture but also the behaviour of basal stem cells is well conserved between fish and mammals. Taking advantage of tissue transparency and live imaging, zebrafish serves as a useful model for revealing the regulatory mechanisms of self-renewal and differentiation of basal stem cells. Despite numerous studies in mice, it remains controversial whether basal stem cells are a heterogeneous population of cells with different differentiation (fast cycling) or self-renewal (slow cycling) potentials, or whether they are homogeneous in their potential as stem cells (Sada et al., 2016; Dekoninck et al., 2020). Histone H2B-EGFP Tet-ON Tg mice (Tumbar et al., 2004; Fuchs and Horsley, 2011) have been used to track basal cell fate. However, by taking advantage of long-term live tracking in combination with the sparse Cre-labelling in zebrafish, individual clonal fates can be directly assessed using Tgs like the itgb4:cre Tg.
In this study, we used genetic manipulation to further investigate the basal cell response to epidermal injury. We revealed that the self-renewal or differentiation of basal cells is induced by injury to different cell types. This suggests that signals originating from different epidermal cell types regulate the self-renewal or differentiation of basal cells. Moreover, we found that krt18 expression was re-expressed in basal cells undergoing self-renewal. It is speculated that krt18 could be a marker of dedifferentiated basal cells that acquire a proliferative status. Studies in mice have suggested that the self-renewal of basal cells is driven by the differentiation of neighbouring basal cells (Mesa et al., 2018). It has also been suggested that signalling pathways, such as Delta-Notch (Lowell et al., 2000) and BMP-FGF (Zhu et al., 2014) are essential for controlling the self-renewal and differentiation of epidermal basal cells. Brock et al. (2019) suggested that dead basal cells release wnt8a into apoptotic bodies to stimulate epidermal cell proliferation. The Tgs and method reported in this study offer a simple and excellent platform for elucidating the mechanisms of basal cell regulation during growth, homeostasis and injury response.
MATERIALS AND METHODS
Fish husbandry and established strains
Fish were maintained in a recirculating water system in a 14 h day/10 h night photoperiod at 28.5°C. The wild-type zebrafish (Danio rerio) strain used in this study was originally derived from the Tubingen strain and has been maintained in our facility for more than 10 years by inbreeding. Tg(krt4:mcherry-t2a-creERt2)tyt214 and Tg(Olactb:loxP-dsRed2-loxP-egfp)tyt201 (RSG) were established in our laboratory (Yoshinari et al., 2012; Shibata et al., 2018).
All surgeries were performed under 0.002% tricaine (3-aminobenzoic acid ethyl ester, Sigma-Aldrich) anaesthesia, and every effort was made to minimise suffering. All animal experiments were performed in strict accordance with the recommendations of the Act on Welfare Management of Animals in Japan and the Guide for the Care and Use of Laboratory Animals from the National Institutes of Health. All the animals were handled in accordance with the Animal Research Guidelines of the Tokyo Institute of Technology. The study protocol was approved by the Committee on Ethics of Animal Experiments at Tokyo Institute of Technology.
Generation of transgenic (Tg) lines
The Cre constructs, pTol2(cryaa:egfp;itgb4:mcherry-t2a-creERt2) and pTol2(cryaa:egfp;krt18:mcherry-t2a-creERt2) were made by replacing the krt4 promoter of pTol2(krt4:mcherry-t2a-creERt2) (Shibata et al., 2018) with −4.5 kb itgb4 or −3.8 kb krt18 promoters at SfiI and FseI sites. The crystallin alpha A (cryaa):egfp cassette was inserted into the EcoNI site, upstream of the promoter. pTol2(itgb4:mcherry) and pTol2(krt18:mcherry) were prepared by replacing the sp7 promoter of pTol2(sp7:mcherry) (Ando et al., 2017) with the −4.5 k itgb4 and −3.8 k krt18 promoter sequences, respectively, at SfiI and AgeI sites. The pTol2(itgb4:tagBFP-nfsB) and pTol2(krt4:tagBFP-nfsB) were made by replacing the mcherry-t2a-creERt2 sequence of pTol2(hsp70l:mcherry-t2a-creERt2) (Yoshinari et al., 2012) with tagBFP-nfsB, and then further replacing the promoter sequence with −4.5 kb itgb4 and −4.3 kb krt4 promoter sequences, respectively, at SfiI and AgeI sites. The Tgs generated in this paper, Tg(itgb4:creERt2)tyt233, Tg(itgb4:mcherry)tyt234, Tg(itgb4:BFP-nfsB)tyt235, Tg(krt18:creERt2)tyt236, Tg(krt18:mcherry)tyt237 and Tg(krt4:BFP- nfsB)tyt238, were deposited in the National Bioresource Project Zebrafish (https://shigen.nig.ac.jp/zebra/).
The engineered constructs were injected into one-cell-stage zebrafish embryos with 25 ng/μl transposase mRNA at a concentration of 25-40 ng/μl. The F0 founders were either incrossed or outcrossed to screen for their F1 careers. The following primers were used to amplify the promoter sequences from the genome. The underlined portions of the primers denote the restriction enzyme recognition sites for cloning. Krt4-pro fw, 5′-ggccagatgggccTGCTGATTCTGGCCCGGTAGCT-3′; Krt4-pro rv, 5′-accggtGATGCCTGTGTCTTTGAGTTGCTGA-3′; Krt18-pro fw, 5′-ggccagatgggccCAGGACATCTGCCCTCCAGCAC-3′; Krt18-pro rv, 5′-ggccggccGGTGTAAGTGAGCAGACGAGTG-3′; Krt18-pro rv, 5′-accggtGGTGTAAGTGAGCAGACGAGTG-3′; Itgb4-pro fw, 5′-ggccagatgggccGTGTTTGTGTGTGTACCTGTGCATGCATGTGCAG-3′; Itgb4-pro rv, 5′-ggccggccCAGTCTGAAACACAAGAGCGAGATCAC-3′; Itgb4-pro rv, 5′-accggtCAGTCTGAAACACAAGAGCGAGATCAC-3′.
Whole-mount ISH
Whole-mount ISH was performed according to the standard protocols (Thisse and Thisse, 2008). The DNA templates for the respective probes were amplified from cDNA using KOD Plus Neo (Toyobo) and the primers listed below. The underlined portions of the primers denote the T7 promoter. RNA probes were synthesised using the Dig RNA Labeling Kit (Sigma-Aldrich). Itgb4 fw, 5′-ATGGGAGGATGGGCGATACGTCTGGCTGTGGGGAT-3′; Itgb4 rv, 5′-taatacgactcactatagggGTCAGCCTGGTTGACCTTACACACGGGCTCTTCTC-3′; Krt18 fw, 5′-ATGTCCTGTGCTCTGCTCTGCTGTGTTACTGAGC-3′; Krt18 rv, 5-taatacgactcactatagggTTCAGCTCCAGCTTGCTGTTGGCCTGCTCCAGGA-3′; Krt4 fw, 5′-ATGTCAACCAGGTCTATCACCTACTCCAGCGGTGG-3′; Krt4 rv 5′-taatacgactcactatagggATAGCGTTTACTGCTGACGGTGGTGACACTGG-3′.
Cell tracking and quantification
Tg strains carrying the itgb4:creERt2 or krt18:creERt2 were crossed with the RSG reporter Tg (Yoshinari et al., 2012) to generate double Tg lines. Cre-loxP recombination was induced by treating double Tgs with 4-OH tamoxifen (TAM, Sigma-Aldrich) or ICI (Sigma-Aldrich). The recombination conditions are shown in the figure legends. Generally, for single-cell tracking, recombination was induced with 0.5 μM ICI for 30 min. For maximum labelling, recombination was induced with 1 μM TAM for 24 h. The treated fish were kept in fresh water in an aquarium for at least 4 days and used for cell tracking.
To capture images, fish were anaesthetised with tricaine, and images were taken using fluorescence microscopy (M205FA, Leica) or confocal microscopy (LSM780, Carl Zeiss). For confocal 3D imaging, the fish fins or embryos were embedded in 1.5% low melting temperature agarose gel and observed under a 40× water-immersion objective lens (212.55 μm×212.55 μm with 512×512 pixel-density). The assessment of cell location was performed on the reconstructed z-stack images (50-140 μm depth, 1 μm z-step size) (ZEN software, Carl Zeiss). EGFP-labelled clones were examined by confocal 3D image (the Ortho function of the ZEN software, Zeiss) to distinguish the quiescent, self-renewing or differentiating basal cells according to the criteria schematically shown in Fig. 6C.
Fluorescence-activated cell sorting and RNA sequencing
mCherry+ or EGFP+ cells were isolated from Tgs using the fluorescence-activated cell sorter SH800S (Sony) according to the standard protocol (Manoli and Driever, 2012). Briefly, 500-1000 deyolked embryos (Link et al., 2006) or two fins were digested in 0.5% collagenase in phosphate buffered saline (PBS) at 37°C for 40 min. The cell suspension was filtered through a strainer (40 μm mesh size, Greiner) into the L-15 buffer (1.5% Leibovitz's L-15, 50 U/ml penicillin, 10% bovine serum albumin). After dissociation, the cells were stained with RedDot2 (Biotium) to exclude the dead cells.
The sorted cells were directly collected in a lysis buffer and used for RNA extraction using the RNeasy Plus Micro Kit (Qiagen). RNA quality assessment was performed using a 2100 Bioanalyzer (Agilent Technologies). High-quality RNA samples (RIN>7) were used for RNA sequencing. RNA sequencing was performed at the Seibutsu-Giken Bioengineering Lab (Kanagawa, Japan) on the DNBSEQ-G400 instrument (MGI Tech). High-quality reads were mapped to the zebrafish reference genome sequence (Danio rerio GRCz11) using Hisat2 ver.2.1.0 and counted using featureCounts ver.2.0.3. The raw data were deposited in the Gene Expression Omnibus under accession number GSE241757.
The differentially expressed genes (DEGs) (Table S1 and S2) were identified according to the following criteria: ratio of transcripts per million (TPM) (itgb4+ cells at 1 dpf versus krt4+ cells at 1 dpf) ≥5, either of TPM values in krt4+ cells or itgb4+ cells ≥10. Heatmaps of the DEGs were plotted using Microsoft Excel 2013.
The selected DEGs were used to perform GO analysis using the DAVID Bioinformatics Resource (https://david.ncifcrf.gov) (Huang et al., 2009; Sherman et al., 2022) and mapped with the GO Database. Enriched terms were ranked according to the number of enriched genes.
Cell ablation
To specifically ablate the krt4+ keratinocytes or itgb4+ basal cells in the adult epidermis, 5 mM Mtz (Tokyo Chemical Industry) was dissolved in fish water (0.3% artificial sea salt, 0.0001% methylene blue) and treated once with Tg(krt4:tagBFP-nfsB) for 4 h or Tg(itgb4:tagBFP-nfsB) for 3 h. Mtz treatment was repeated every 2 days to increase cell ablation without causing fish death.
To evaluate cell ablation efficiency, acridine orange (AO; Sigma-Aldrich) staining was performed (Tucker and Lardelli, 2007). After Mtz treatment, the fish were placed in 10 μg/ml AO solution for 5 min at room temperature in the dark and washed with fresh water for 3 h. After staining, fish were anaesthetised with tricaine and observed under a microscope.
Immunostaining and histological analysis
The collected tissues were fixed in 4% paraformaldehyde overnight at 4°C and washed with PBTx (0.1% Triton X-100 in PBS) three times for 5 min each. For whole-mount staining, samples were stained with DAPI (1:1000; Dojin Chemical) or SYTO11 (1:500; Thermo Fisher Scientific) for staining the nucleus. The cryosections were stained with an anti-EGFP antibody (1:1000; Nacalai Tesque, 04404-26) and an anti-rat secondary antibody with Alexa Fluor 488 (1:1000; Thermo Fisher Scientific, A-11006). After washing, the samples were mounted in 80% glycerol containing 2.5 mg/ml 1, 4-diazabicyclo [2.2.2] octane as an antifade reagent and observed by confocal microscopy.
Statistical analysis
No statistical method was used to determine the sample size. Sample sizes were chosen based on previous publications, and experiment types are shown in the figure legends. After selecting embryos or fish with wild-type morphology, the clutch mates were randomised into different groups for each experiment. No animal or sample was excluded from analysis unless the animal died during the procedure. Most assessments of phenotypes and expression patterns were repeated in at least three independent experiments, except for cell sorting and RNA sequencing analyses (two replicates for embryonic samples, one for adult fin). Whenever possible, conditions were masked during data collection and analysis. In some experiments, when masking was not possible, the same investigator processed the samples and collected data. Sample sizes are indicated in the figures or legends.
Statistical analyses of the quantified data were performed using Microsoft Excel 2013 and GraphPad Prism 6. All statistical values are presented as the mean±standard error of the mean (s.e.m.).
Acknowledgements
We thank the Open Facility Center for Life Science and Technology at the Tokyo Institute of Technology for providing the sequencing and imaging support. We also thank the National BioResource Project (NBRP) for maintenance of zebrafish strains.
Footnotes
Author contributions
Conceptualization: A.K.; Methodology: Z.L., A.K.; Validation: Z.L.; Formal analysis: Z.L.; Investigation: Z.L., Y.M., A.I.; Data curation: Z.L., A.K.; Writing - original draft: Z.L.; Writing - review & editing: A.K.; Supervision: A.K.; Project administration: A.K.; Funding acquisition: A.K.
Funding
This work was supported by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Scientific Research (B) (19H03232) and from the Japan Society for the Promotion of Science (JSPS), Challenging Exploratory Research (19K22417 and 22K19306) to A.K. Z.L. was supported by the Tsubame Scholarship for Doctoral Students from the Tokyo Institute of Technology.
Data availability
RNA sequencing data have been deposited in the Gene Expression Omnibus under accession number GSE241757.
Peer review history
The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.202315.reviewer-comments.pdf
References
Competing interests
The authors declare no competing or financial interests.