Adult sox10+ Cardiomyocytes Contribute to Myocardial Regeneration in the Zebrafish.

During heart regeneration in the zebrafish, fibrotic tissue is replaced by newly formed cardiomyocytes derived from preexisting ones. It is unclear whether the heart is composed of several cardiomyocyte populations bearing different capacity to replace lost myocardium. Here, using sox10 genetic fate mapping, we identify a subset of preexistent cardiomyocytes in the adult zebrafish heart with a distinct gene expression profile that expanded after cryoinjury. Genetic ablation of sox10+ cardiomyocytes impairs cardiac regeneration, revealing that these cells play a role in heart regeneration.

Introduction

Cardiomyocyte (CM) renewal in the human heart is margihemical">nal ahemical">nd, after acute n class="Disease">myocardial infarction, millions of CMs are irreversibly lost and replaced by a fibrotic scar (Prabhu and Frangogiannis, 2016). Adult mammalian CMs have a poor capacity to proliferate, and an efficient contribution of an adult stem cell pool for myocardial replacement has not been demonstrated (Lerman et al., 2016). By contrast, zebrafish have an extraordinary capacity for heart regeneration after injury (González-Rosa et al., 2017, Kikuchi, 2014). Lineage tracing studies have revealed that preexisting CMs are the origin of de novo formed cardiac muscle (Jopling et al., 2010, Kikuchi et al., 2010). Upon injury, CMs adjacent to the injury lose their sarcomeric organization and exhibit a more immature phenotype. Concomitant with this structural change, CMs were described to express developmental genes, suggesting a reversion of their differentiated phenotype to a more embryonic-like state (Lepilina et al., 2006). Indeed, gata4, hand2, and tbx5a, which play key roles during heart development (Garrity et al., 2002, Grajevskaja et al., 2018, Kuo et al., 1997, Molkentin et al., 1997, Srivastava et al., 1997), are required for heart regeneration (Grajevskaja et al., 2018, Gupta et al., 2013, Schindler et al., 2014). Moreover, CMs contributing to heart regeneration activate gata4 and ctgfa enhancer elements upon injury (Gupta et al., 2013, Kikuchi et al., 2010, Pfefferli and Jaźwińska, 2017). While CMs that will contribute to heart regeneration upregulate a specific set of genes, it is unclear whether CM subsets contributing to regeneration can be distinguished by means of their expression profile in the uninjured heart.

Like mammals, zebrafish CMs derive from first ahemical">nd secohemical">nd heart field progenitors (de Pater et al., 2009, Mosimahemical">nn et al., 2015, Zhou et al., 2011). However, ihemical">n the n class="Species">zebrafish, the neural crest represents a third progenitor population that contributes to the developing heart. Cell transplantation and fluorescent dye tracing experiments suggested that cardiac neural crest cells incorporate not only into the areas of the outflow tract, as in mammals and birds, but also into the atrium and ventricle (Li et al., 2003, Sato and Yost, 2003). Moreover, genetic lineage tracing using sox10 as a neural crest cell marker revealed a cellular contribution of sox10+ cells to the zebrafish heart (Cavanaugh et al., 2015, Mongera et al., 2013) and suggested that sox10-derived CMs are necessary for correct heart development (Abdul-Wajid et al., 2018). Noteworthy, it is still unclear if a sox10+ neural crest population differentiates into CMs or if alternatively, a sox10+ CM subset is relevant for heart development.

Here, we assessed the contribution of sox10-derived cells to the adult n class="Species">zebrafish heart using sox10:CreERT2 fate mapping during homeostasis and regeneration. We found that embryonic sox10-derived cells contributed to significant portions of the adult heart. We also identified adult sox10+ CMs that expanded to a higher degree upon injury than other CMs and significantly contributed to cardiac regeneration. Their transcriptome differed from other CMs in the heart, and their genetic ablation impaired recovery from ventricular cryoinjury.

Results

sox10-Expressing Cells Contribute to the Regenerated Myocardium

sox10-derived cells were proposed to cohemical">ntribute to the adult n class="Species">zebrafish heart (Abdul-Wajid et al., 2018). Since a non-inducible sox10:Cre line was used for these studies, it was not clear whether embryonic sox10+ cells or adult sox10+ contributed to the adult zebrafish heart. To clarify the source of sox10-derived cells contributing to the adult heart, we used the line Tg(sox10:CreER) (Mongera et al., 2013), in which CreER is driven by the −4.9-kb sox10 promoter (Carney et al., 2006). This line was crossed with Tg(ubb:loxP-GFP-loxP-mCherry) (Mosimann et al., 2011), from now on named ubb:Switch (Figure 1A). In double-transgenic animals, 4-hydroxytamoxifen (4-OHT) administration leads to constitutive mCherry expression in cells expressing sox10:CreER at the time of recombination (Figure 1A). Recombination in adult sox10:CreER;ubb:Switch zebrafish led to the activation of mCherry expression in a few cardiac cells in the atrium, ventricle, and valves (Figures 1B–1G). We observed both sox10-derived CMs (mCherry+/MHC+) (Figures 1B′–1C″′, 1D′ –1D″′, and 1F′ –1F″″) and non-CMs (only mCherry+) (Figure 1B″), the latter mostly contributing to the valves. Thus, the adult heart bears a very small population of CMs present with an active sox10-promoter element.

Figure 1

Contribution of sox10-Derived CMs to Heart Regeneration

(A) Analysis of sox10-derived cells in the adult zebrafish heart. Two 4-OHT pulses were administered to adult sox10:CreER;ubb:loxP-GFP-loxP-mCherry zebrafish between 1 and 3 weeks before heart extraction to induce mCherry in sox10:CreER-expressing cells. mCherry (red), sox10 lineage. Anti-MHC marks CMs (green) in uninjured hearts (B–G) or hearts at 14 dpi (I–N).

(B–C″) Cryosections of uninjured hearts fixed 19 days after the last 4-OHT pulse. Upper row, mCherry channel, lower row, merged channels. Arrowheads, mCherry+ CMs; arrows, mCherry+ non-CMs within the atrioventricular valve.

(D–G) Whole-mount views (D) and sections (F and G) of uninjured hearts fixed 11 days after 4-OHT treatment. (D and E) Whole-mount immunostaining. Maximal projection of 352 μm, 44 z planes. (D′–D″′) Maximum projection of zoomed view of boxed area in D (7 μm, 7 z planes). (F and G) Immunostaining on cryosection of an uninjured heart. (F′–F″″) Zoomed view of boxed area in (F). Images are stitched high resolution acquisitions (9 images stitched together). sox10-derived CMs and non-CMs are observed in the subepicardial area of the ventricle. mCherry+ cells in the cortical myocardium (white arrowheads), mCherry+ trabecular CMs (yellow arrowheads).

(H) Analysis of sox10-derived cells in the regenerating heart. Two 4-OHT pulses were administered to adult sox10:CreER;ubb:loxP-GFP-loxP-mCherry 1 and 3 days after the injury to induce mCherry expression in sox10:CreER-expressing cells. Hearts were extracted at 14 dpi.

(I and J) Whole-mount view of injured heart at 14 dpi (594 μm, 73 z planes). Shown are the merged (I) and red (J) channel. sox10-derived CMs can be observed in the apical region of the heart. (J′–J″′) Zoomed view of boxed area in (J).

(K and L) z stack from heart in I (252 μm, 31 z planes), showing injured area. Image shown in (K) is comprised of four stitched high-resolution acquisitions. Shown are the merged (K) and the red (L) channels. Yellow arrowheads, mCherry+; MHC+ trabecular CMs. White arrowheads, mCherry+; MHC+ cortical CMs.

(M and N) Paraffin sections of injured hearts at 14 dpi. Note the distribution of sox10-derived CMs around the IA (yellow arrowheads). Shown are the merged (M) and the red channel (N). Zoomed view of boxed area in (M) is shown in (M′) and (M″). Note cortical (white arrowheads) and trabecular (yellow arrowheads) sox10-derived CMs adjacent to the IA.

(O) Percentage of the volume from mCherry+/MHC+ cells relative to all MHC+ cells in uninjured (n = 6) and injured hearts (n = 6) at 14 dpi. The volume of sox10-derived CMs expands after injury. Data are mean ± SD; p = 0.0022 (two-tailed non-parametric t test).

(P) The ventricles from whole-mount stained hearts were digitally sectioned in increments of 100 μm starting from the injury site. Shown is the percentage of the volume from mCherry+/MHC+ cells relative to all MHC+ cells within different heart segments. The percentage of mCherry+/MHC+ cells is high near to the injury site and decreases toward more distant ventricular regions (injured hearts, n = 6). Data are mean ± SD; ∗p < 0.05; ∗∗∗p < 0.001 (one-way ANOVA test followed by Tukey’s post hoc test).

(Q) The ventricle was digitally sectioned in increments of 100 μm starting from the apex. Shown is the percentage of the volume from mCherry+/MHC+ cells relative to all MHC+ cells within different heart segments. The percentage of mCherry+/MHC+ cells is similarly distributed through ventricular regions of the uninjured ventricle (uninjured hearts n = 6). Data are mean ± SD; p > 0.05 (one-way ANOVA followed by Tukey’s post hoc test).

4-OHT, 4-hydroxytamoxifen; At, atrium; CI, cryoinjury; CMs, cardiomyocytes; dpi, days post-injury; dpt, days post-treatment; IA, injured area; V, ventricle. Scale bars, 100 μm. See also Figure S1.

Contribution of sox10-Derived CMs to Heart Regehemical">nen class="Species">ration

(A) Analysis of sox10-derived cells ihemical">n the adult n class="Species">zebrafish heart. Two 4-OHT pulses were administered to adult sox10:CreER;ubb:loxP-GFP-loxP-mCherry zebrafish between 1 and 3 weeks before heart extraction to induce mCherry in sox10:CreER-expressing cells. mCherry (red), sox10 lineage. Anti-MHC marks CMs (green) in uninjured hearts (B–G) or hearts at 14 dpi (I–N).

(B–C″) Cryosections of uninjured hearts fixed 19 days after the last 4-OHT pulse. Upper row, mCherry chahemical">nnel, lower row, merged chahemical">nnels. Arrowheads, mCherry+ CMs; arrows, mCherry+ nohemical">n-CMs withihemical">n the n class="Disease">atrioventricular valve.

(D–G) Whole-mount views (D) and sections (F and G) of uninjured hearts fixed 11 days after 4-OHT treatment. (D ahemical">nd E) Whole-mount immunostaihemical">nihemical">ng. Maximal projectiohemical">n of 352 μm, 44 z plahemical">nes. (D′–D″′) Maximum projectiohemical">n of zoomed view of boxed area ihemical">n D (7 μm, 7 z plahemical">nes). (F ahemical">nd G) Immunostaihemical">nihemical">ng ohemical">n cryosectiohemical">n of ahemical">n uhemical">nihemical">njured heart. (F′–F″″) Zoomed view of boxed area ihemical">n (F). Images are stitched high resolutiohemical">n acquisitions (9 images stitched together). n class="Gene">sox10-derived CMs and non-CMs are observed in the subepicardial area of the ventricle. mCherry+ cells in the cortical myocardium (white arrowheads), mCherry+ trabecular CMs (yellow arrowheads).

(H) Analysis of sox10-derived cells ihemical">n the regehemical">nen class="Species">rating heart. Two 4-OHT pulses were administered to adult sox10:CreER;ubb:loxP-GFP-loxP-mCherry 1 and 3 days after the injury to induce mCherry expression in sox10:CreER-expressing cells. Hearts were extracted at 14 dpi.

(I and J) Whole-mount view of injured heart at 14 dpi (594 μm, 73 z planes). Shown are the merged (I) and red (J) channel. sox10-derived CMs cahemical">n be observed ihemical">n the apical regiohemical">n of the heart. (J′–J″′) Zoomed view of boxed area ihemical">n (J).

(M and N) Paraffihemical">n sections of ihemical">njured hearts at 14 dpi. n class="Chemical">Note the distribution of sox10-derived CMs around the IA (yellow arrowheads). Shown are the merged (M) and the red channel (N). Zoomed view of boxed area in (M) is shown in (M′) and (M″). Note cortical (white arrowheads) and trabecular (yellow arrowheads) sox10-derived CMs adjacent to the IA.

(O) Percentage of the volume from mCherry+/MHC+ cells relative to all MHC+ cells in uninjured (n = 6) and injured hearts (n = 6) at 14 dpi. The volume of sox10-derived CMs expahemical">nds after ihemical">njury. Data are meahemical">n ± SD; p = 0.0022 (two-tailed nohemical">n-parametric t test).

4-OHT, n class="Chemical">4-hydroxytamoxifen; At, atrium; CI, cryoinjury; CMs, cardiomyocytes; dpi, days post-injury; dpt, days post-treatment; IA, injured area; V, ventricle. Scale bars, 100 μm. See also Figure S1.

We next wanted to assess if sox10-derived cells cohemical">ntribute to heart regenen class="Species">ration. When 4-OHT treatment was performed shortly after ventricular cryoinjury (Figure 1H), we observed an expansion of sox10-derived cells (Figures 1I–1O). Immunostainings confirmed that sox10-derived cells were CMs, as they co-expressed the myocardial marker myosin heavy chain (MHC) (Figures 1J′–J″′ and 1M′–M″). In order to estimate the contribution of sox10-derived CMs during the regeneration process, we measured the ratio of sox10-derived CMs (mCherry+/ MHC+) versus other CMs (MHC+) in the entire heart (Figure 1O). When comparing the proportion of mCherry+ ventricular CMs from cryoinjured and control hearts, we observed that the portion of mCherry+ myocardium increased ∼100-fold at 14 days post-injury (dpi) (Figure 1O). We questioned whether sox10-derived cells expanded globally in the injured heart or whether there was a distinct contribution to the regenerating myocardium. To do this, we generated a distance distribution map for the sox10-derived CMs from the injury site (for injured hearts; Figure 1P) or apical myocardium (for uninjured hearts; Figure 1Q) to the basal myocardium. The closer to the injury area, the greater the contribution of sox10-derived CMs (Figure 1P; n = 6). This bias toward an apical region was not detected in uninjured hearts (Figure 1Q; n = 6).

The presence of adult sox10+ CMs was cohemical">nfirmed usihemical">ng the lihemical">ne Tg(−4.9n class="Gene">sox10:GFP) (Carney et al., 2006; Figure S1A). We were able to detect few sox10:GFP+ CMs in the ventricle of injured and uninjured hearts (Figures S1B–S1E″; n = 4 hearts). Moreover, RNAScope mRNA detection revealed sox10 expression in the heart (Figures S1F–S1I″). In uninjured hearts, we found expression both in the atrium (2 out of 3 hearts) as well as in the ventricle (1 out of 3 hearts). In injured hearts, sox10 signal was detected close to the cortical and trabecular myocardial boundaries, as well as at the borders of the injury area (n = 3 out of 3 hearts). This suggests that sox10 mRNA increases upon injury or that sox10+ cells accumulate at the site of injury. Surprisingly, we did not detect an expansion of sox10:GFP+ cells after cardiac injury (Figures S1A–S1E″; n = 4 hearts). The inconsistency between sox10 mRNA detection and sox10:GFP expression in the reporter line might be a consequence of partial silencing of the reporter leading to an incomplete recapitulation of endogenous sox10 expression. Alternatively, it might indicate that the sox10-derived cells in regenerating hearts were descendants from a small subset of sox10+ progenitor cells within the uninjured heart that turn off sox10 expression after cellular expansion and CM differentiation.

Preexistent sox10-Derived CMs Contribute to Cardiac Regeneration

To assess whether the increased number of sox10-derived CMs is a result of the expansiohemical">n of a small pool of preexistent n class="Gene">sox10+ cells, we induced recombination in adult zebrafish before cryoinjury (Figures 2 and S1J–S1T). 4-OHT treatments 2 weeks prior to cardiac injury allowed us to rule out that non-metabolized 4-OHT could be active and induce recombination after injury (Figure 2A). While in controls we again saw only few cells scattered throughout the ventricle, at 7 dpi, we indeed observed sox10-derived CMs close to the injury area (Figures 2B–2C′). More sox10-derived cells contributed to the ventricle after 7 dpi compared to uninjured hearts (Figure 2D). The percentage of sox10-derived CM was higher close to the injury area compared to the basal ventricle (Figure 2E). This preferential distribution of sox10-derived CMs toward the apical region was not found in uninjured hearts of siblings (Figure 2E). The proportion of sox10-derived CMs significantly expanded upon injury both in the trabecular and cortical myocardium regions (Figure S1S).

Figure 2

Preexistent sox10-Derived CMs Expand at the Injury Area and Contribute to Cardiac Regeneration

(A) Adult sox10:CreER;ubb:loxP-GFP-loxP-mCherry zebrafish were treated with 4-OHT on days 14 and 12 before cryoinjury. Hearts were fixed at 7, 14, or 210 dpi and processed for immunostaining with anti-mCherry+ (red, sox10-derived cells) and anti-MHC (green, myocardium).

(B) Whole mounts of uninjured hearts. Upper row, mCherry channel; lower row, myosin heavy chain (MHC) and mCherry merged channels. Arrowheads indicate sox10-derived cells.

(C and C′) Cryosection of an injured heart at 7 dpi. sox10-derived CMs are detected near the injured area (IA) and subepicardial regions of the myocardium. White arrowheads indicate mCherry+ CMs in the cortical myocardium, and yellow arrowheads indicate mCherry+ CMs in the trabecular myocardium.

(D) Quantification of the proportion of sox10-derived CMs in uninjured (n = 4) and injured hearts (n = 6) at 7 dpi. Each dot represents the value from one heart. Data are mean ± SD; p = 0.0095 (two-tailed non-parametric t test).

(E) Quantification of the distribution of ventricular mCherry+ CMs on uninjured and cryosectioned hearts at 7 dpi. The ventricle was digitally divided into increments of 100 μm starting from the injury site in injured hearts or apex in uninjured hearts. Shown is the percentage of the mCherry+/MHC+ area relative to the whole MHC+ area within different heart segments. The percentage of mCherry+/MHC+ is high near the IA and decreases toward ventricular regions distant from the IA (injured hearts n = 5; uninjured hearts, n = 4). Dashed lines represent statistical differences, and red bars represent injured hearts. The percentage of mCherry+/MHC+ cells is higher in every region of the injured heart ventricle (red bars) compared to uninjured heart ventricles (blue bars). Data are mean ± SD; ∗p < 0.05 (two-tailed non-parametric t test).

(F) Whole-mount image of a heart at 14 dpi. sox10-derived CMs are distributed around the IA and distant part of the ventricle.

(G–H′) Whole-mount view (G) and cryosections (H and H′) of regenerated hearts at 210 dpi. sox10-derived CMs can be observed in the apical region of the heart. White arrowheads show mCherry+ CMs in the cortical myocardium, and yellow arrowheads show mCherry+ CMs in the trabecular myocardium. Image shown in (H) is comprised of nine stitched high-resolution acquisitions.

(I) Quantification of the percentage of mCherry+ CMs in uninjured cardiac ventricles compared to ventricles at 14 and 210 dpi (whole-mount immunostained hearts). Shown are measurements of individual hearts (dots) as well as mean ± SD (two-tailed non-parametric t test; uninjured, n = 8; 14 dpi, n = 6; 210 dpi, n = 6).

(J) Quantification of the distribution of mCherry+ CMs in whole mount immunostained hearts. Distance calculation as shown in (E). Data are mean ± SD. ∗p < 0.05; ∗∗p < 0.01 (two-tailed non-parametric t test; n = 6).

At, Atrium; dpi, days post-injury; IA, injured area; MHC, myosin heavy chain; V, ventricle. Scale bars represent 100 μm (except for C′, where scale bars represent 200 μm). See also Figure S1.

Preexistent sox10-Derived CMs Expahemical">nd at the Injury Area ahemical">nd Cohemical">ntribute to Cardiac Regehemical">nen class="Species">ration

(A) Adult sox10:CreER;ubb:loxP-GFP-loxP-mCherry n class="Species">zebrafish were treated with 4-OHT on days 14 and 12 before cryoinjury. Hearts were fixed at 7, 14, or 210 dpi and processed for immunostaining with anti-mCherry+ (red, sox10-derived cells) and anti-MHC (green, myocardium).

(C and C′) Cryosection of an injured heart at 7 dpi. sox10-derived CMs are detected near the ihemical">njured area (IA) ahemical">nd subepicardial regions of the myocardium. White arrowheads ihemical">ndicate mCherry+ CMs ihemical">n the cortical myocardium, ahemical">nd yellow arrowheads ihemical">ndicate mCherry+ CMs ihemical">n the trabecular myocardium.

(D) Quantification of the proportion of sox10-derived CMs ihemical">n unihemical">njured (n = 4) ahemical">nd ihemical">njured hearts (n = 6) at 7 dpi. Each dot represehemical">nts the value from ohemical">ne heart. Data are meahemical">n ± SD; p = 0.0095 (two-tailed nohemical">n-parametric t test).

(E) Quantification of the distribution of ventricular mCherry+ CMs ohemical">n uhemical">nihemical">njured ahemical">nd cryosectiohemical">ned hearts at 7 dpi. The ventricle was digitally divided ihemical">nto ihemical">ncrements of 100 μm startihemical">ng from the ihemical">njury site ihemical">n ihemical">njured hearts or apex ihemical">n uhemical">nihemical">njured hearts. Showhemical">n is the percentage of the mCherry+/MHC+ area relative to the whole MHC+ area withihemical">n different heart segments. The percentage of mCherry+/MHC+ is high near the IA ahemical">nd decreases toward ventricular regions distahemical">nt from the IA (ihemical">njured hearts n = 5; uhemical">nihemical">njured hearts, n = 4). Dashed lihemical">nes represent statistical differences, ahemical">nd red bars represent ihemical">njured hearts. The percentage of mCherry+/MHC+ cells is higher ihemical">n every regiohemical">n of the ihemical">njured heart ventricle (red bars) compared to uhemical">nihemical">njured heart ventricles (blue bars). Data are meahemical">n ± SD; ∗p < 0.05 (two-tailed nohemical">n-parametric t test).

(F) Whole-mount image of a heart at 14 dpi. sox10-derived CMs are distributed around the IA ahemical">nd distahemical">nt part of the vehemical">ntricle.

(G–H′) Whole-mount view (G) and cryosections (H and H′) of regenerated hearts at 210 dpi. n class="Gene">sox10-derived CMs can be observed in the apical region of the heart. White arrowheads show mCherry+ CMs in the cortical myocardium, and yellow arrowheads show mCherry+ CMs in the trabecular myocardium. Image shown in (H) is comprised of nine stitched high-resolution acquisitions.

We also analyzed the proportion of sox10-derived cells at later stages of regenen class="Species">ration to assess if the expansion is transient or whether these cells contribute to the regenerated heart (Figures 2F–2J and S1T). At 14 dpi, sox10-derived CMs remained around the injured area (Figures 2F and 2G). Even at 210 dpi, when regeneration is complete (González-Rosa et al., 2011), mCherry+/MHC+ CMs were still detected within the region presumably corresponding to the regenerated myocardium (Figures 2H and 2H′). Similar to the observation when tracing sox10-derived cells after cryoinjury, the volume of mCherry+ cells increased over 100-fold after injury when compared with uninjured hearts (Figure 2I). Quantification showed that consistent with the results at 7dpi, the mCherry+/MHC+ myocardial volume was significantly higher in injured hearts than in uninjured hearts at all regeneration stages analyzed (Figures 2J and S1T).

To understand the mechanisms of the accumulation of sox10-derived cells at the ihemical">njury area, we assessed cell prolifen class="Species">ration (Figure 3). Recombination was induced 2 weeks before cryoinjury, and bromodeoxyuridine (BrdU) was injected at 6 dpi (Figure 3A). Immunostaining against mCherry to detect the sox10 lineage, the myocardial marker MHC, and BrdU to label proliferating cells showed the presence of mCherry+/BrdU+ double-positive cells (Figures 3B–3G″″). Quantification revealed a statistically significant increase of ∼30% in the amount of proliferating sox10-derived CMs (n = 6) when compared to sox10-lineage negative CMs (n = 6) (Figure 3H). We also observed a higher degree of BrdU+ /mCherry+ cells compared to the rest of CMs at later stages of regeneration (Figures 3I–3K). Thus, CMs with an active sox10 promoter element in the uninjured heart divide at a higher rate than the rest of the CMs within the same anatomical region in response to injury.

Figure 3

Proliferation of sox10-Derived CMs

(A) Assessment of the proliferation index of sox10-derived CMs. Two pulses of 4-OHT were administered 14 and 12 days before the injury. BrdU was added at 6 dpi, and hearts were collected at 7 dpi.

(B–E) Heart section immunostained with anti-BrdU (white), anti-mCherry (red, sox10 lineage), and MHC (green). Nuclei were counterstained with DAPI (blue). Shown are merged channels (B), as well as single channels for MHC (C), mCherry (D), and BrdU (E) staining.

(F–G″″) Zoomed views of boxed regions in (B). Yellow arrowheads, mCherry+, BrdU+, MHC+ triple-positive cells. Shown are merged channels (F and G), as well as single channels for DAPI (F′ and G′), BrdU (F″ and G″), MHC (F″′ and G″′) and mCherry (F″″ and G″″).

(H) Quantification of mCherry+, BrdU+, MHC+ triple-positive cells versus all BrdU+, MHC+ double-positive cells in the 100 μm IA border zone. Shown are values for individual hearts as well as mean ± SD. ∗p < 0.05 (two-tailed non-parametric t test; n = 6).

(I) Assessment of proliferation at late stages of regeneration. 4-OHT was added at −12 and −14 days to sox10:CreER;ubb:loxP-GFP-loxP-mCherry zebrafish before cryonjury. BrdU was added at 6 and 29 dpi. Hearts were collected at 30 dpi.

(J) Immunofluorescence staining on heart at 30 dpi. MHC, green; mCherry, red; BrdU, white; nuclei are counterstained with DAPI (blue). Shown are merged images and single channels, and the zoomed region is highlighted with dotted lines.

(K) Quantification of BrdU+ CMs at 30 dpi. Data are mean ± SD (p = 0.029; two-tailed non-parametric t test).

At, Atrium, dpi, days post-injury; IA, injured area; MHC, myosin heavy chain; V, ventricle. Scale bars, 100 μm.

Proliferatiohemical">n of n class="Gene">sox10-Derived CMs

(A) Assessment of the proliferatiohemical">n ihemical">ndex of n class="Gene">sox10-derived CMs. Two pulses of 4-OHT were administered 14 and 12 days before the injury. BrdU was added at 6 dpi, and hearts were collected at 7 dpi.

(B–E) Heart section immunostained with anti-BrdU (white), ahemical">nti-mCherry (red, n class="Gene">sox10 lineage), and MHC (green). Nuclei were counterstained with DAPI (blue). Shown are merged channels (B), as well as single channels for MHC (C), mCherry (D), and BrdU (E) staining.

(F–G″″) Zoomed views of boxed regions in (B). Yellow arrowheads, mCherry+, BrdU+, MHC+ triple-positive cells. Shown are merged chahemical">nnels (F ahemical">nd G), as well as sihemical">ngle chahemical">nnels for DAPI (F′ ahemical">nd G′), n class="Chemical">BrdU (F″ and G″), MHC (F″′ and G″′) and mCherry (F″″ and G″″).

(H) Quantification of mCherry+, BrdU+, MHC+ triple-positive cells versus all n class="Chemical">BrdU+, MHC+ double-positive cells in the 100 μm IA border zone. Shown are values for individual hearts as well as mean ± SD. ∗p < 0.05 (two-tailed non-parametric t test; n = 6).

(I) Assessment of proliferatiohemical">n at late stages of regehemical">nen class="Species">ration. 4-OHT was added at −12 and −14 days to sox10:CreER;ubb:loxP-GFP-loxP-mCherry zebrafish before cryonjury. BrdU was added at 6 and 29 dpi. Hearts were collected at 30 dpi.

(J) Immunofluorescence staining on heart at 30 dpi. MHC, green; mCherry, red; BrdU, white; nuclei are counterstaihemical">ned with DAPI (blue). Showhemical">n are merged images ahemical">nd sihemical">ngle chahemical">nnels, ahemical">nd the zoomed regiohemical">n is highlighted with dotted lihemical">nes.

(K) Quantification of BrdU+ CMs at 30 dpi. Data are meahemical">n ± SD (p = 0.029; two-tailed nohemical">n-parametric t test).

Origin of sox10-Derived Cells Contributing to Regeneration

Next, we wahemical">nted to determihemical">ne the developmental time poihemical">nt at which this n class="Gene">sox10+ CM subset with capacity to contribute to heart regeneration appears, as well as elucidate a possible neural crest cell origin. We treated sox10:CreER;ubb:Switch embryos with 4-OHT between 12 and 48 h post-fertilization, the time window of neural crest cell addition to the developing heart (Abdul-Wajid et al., 2018, Cavanaugh et al., 2015, Mongera et al., 2013). We followed individual sox10:CreER;ubb:Switch animals in a longitudinal study (Figure S2A). First, we performed live imaging at 5 days post-fertilization (dpf). The sox10:CreER line additionally harbors a myocardial reporter cassette, a myosin light chain 7 (myl7) promoter element driving GFP expression in CMs, which was useful to mark myocardial cells. We detected mCherry+/myl7:GFP+ in the larval heart, indicating the presence of a subset of CMs derived from sox10-expressing cells (Figures S2B and S2C; n = 7). As reported for sox10:Cre genetic tracing (Abdul-Wajid et al., 2018), using the sox10:CreER line, the proportion of sox10-derived cells in the trabecular myocardium was higher than in the compact myocardium, which at this stage is composed exclusively of the primordial layer (Gupta and Poss, 2012; Figure S2D). After live imaging, larvae were grown separately to adulthood to follow the fate of mCherry+ cells. In the adult, we again detected mCherry+ cell clusters, showing that embryonic sox10-derived cells contribute to the adult heart (Figures S2E and S2F; n = 7) and confirming observations using a non-inducible sox10:Cre line (Abdul-Wajid et al., 2018). Embryonic sox10-derived cells contributed to 5% of the total ventricular volume of the adult heart (Figure S2G). Indeed, immunofluorescence staining and fluorescence activated cell (FAC) sorting confirmed that most of the mCherry+ cells were also myl7:GFP+ and MHC+ and thus CMs (Figures S2H–S2L). We observed mCherry+ cells in the ventricle and atrium in most of the animals analyzed. The mCherry signal was particularly evident in the basal and medial portion of the ventricle, presumably due to a higher cell density in these regions (Figure S2M; n = 19 hearts analyzed).

To determine whether the embryonic sox10-derived populatiohemical">n expahemical">nded ihemical">n response to ihemical">njury, we cryoihemical">njured ventricles from adult n class="Gene">sox10:CreER;ubb:Switch zebrafish recombined during embryogenesis and compared the percentage of mCherry+ cells in injured and uninjured hearts (Figure 4A). mCherry+ cells were detected in a similar proportion in uninjured hearts (Figures 4B–4E; n = 12) and hearts at 14 dpi (Figures 4F–4J, n = 9; and Figure 4K, n = 5). While a tendency of higher amount of mCherry+ signal was observed within the injury area, the increase did not reach significance, eventually due to the high degree of heterogeneity in injury response. Similarly, in uninjured hearts, no apicobasal region was identified with a preferred contribution of mCherry+ cells (Figure 4L; n = 5). These results imply that in the adult heart, CMs derived from sox10+ embryonic cells do not expand in response to injury and are not preferentially contributing to myocardium regeneration.

Figure 4

Embryonic sox10-Derived CMs Do Not Expand after Cryoinjury

(A) 4-OHT was administered to sox10:CreER;ubb:loxP-GFP-loxP-mCherry zebrafish during embryogenesis (12–48 h post-fertilization), and adult uninjured hearts or hearts at 14 dpi were collected and imaged.

(B–I) Maximal intensity projection of a confocal z stack through an immunostained heart. sox10-derived cells are mCherry+ (red). The whole myocardium is MHC+ (green). mCherry+ cells are present in the ventricle (arrowheads) and atrium (arrows). Shown are merged channel views (C, E, G, and I) as well as mCherry (red) channel only (B, D, F, and H).

(J) Relative volume of embryonic sox10-derived CMs (mCherry+/ MHC+) compared to all CMs (MHC+) in uninjured (n = 12) and injured hearts at 14 dpi (n = 9). Shown are values from individual hearts as well as mean ± SD; p = 0.0955 (two-tailed non-parametric t test).

(K and L) Quantification of the distribution of mCherry+ CMs in whole mount immunostained hearts. The ventricle was digitally sectioned in increments of 100 μm starting from the injury site or apex. Shown is the relative volume from mCherry+/MHC+ cells versus all MHC+ cells within different heart segments. The amount of mCherry+/MHC+ cells does not differ between segments (one-way ANOVA). n = 5 hearts for 14 dpi (K) and n = 5 for uninjured condition (L). The atrium, ventricle, and injury area are outlined by dotted lines.

4-OHT, 4-hydroxytamoxifen; At, atrium; dpi, days post-injury; IA, injured area; CI, cryoinjury; CMs, cardiomyocytes; V, ventricle. Scale bars, 100 μm. See also Figure S2.

Embryonic sox10-Derived CMs Do n class="Chemical">Not Expand after Cryoinjury

(A) 4-OHT was admihemical">nistered to n class="Gene">sox10:CreER;ubb:loxP-GFP-loxP-mCherry zebrafish during embryogenesis (12–48 h post-fertilization), and adult uninjured hearts or hearts at 14 dpi were collected and imaged.

(B–I) Maximal intensity projection of a confocal z stack through an immunostained heart. sox10-derived cells are mCherry+ (red). The whole myocardium is MHC+ (green). mCherry+ cells are present ihemical">n the ventricle (arrowheads) ahemical">nd atrium (arrows). Showhemical">n are merged chahemical">nnel views (C, E, G, ahemical">nd I) as well as mCherry (red) chahemical">nnel ohemical">nly (B, D, F, ahemical">nd H).

(J) Relative volume of embryonic sox10-derived CMs (mCherry+/ MHC+) compared to all CMs (MHC+) ihemical">n unihemical">njured (n = 12) ahemical">nd ihemical">njured hearts at 14 dpi (n = 9). Shown are values from ihemical">ndividual hearts as well as meahemical">n ± SD; p = 0.0955 (two-tailed nohemical">n-parametric t test).

(K and L) Quantification of the distribution of mCherry+ CMs in whole mount immunostained hearts. The ventricle was digitally sectioned in increments of 100 μm starting from the injury site or apex. Shown is the relative volume from mCherry+/MHC+ cells versus all MHC+ cells within different heart segments. The amount of mCherry+/MHC+ cells does not differ between segments (ohemical">ne-way ANOVA). n = 5 hearts for 14 dpi (K) and n = 5 for uninjured condition (L). The atrium, ventricle, and injury area are outlined by dotted lines.

4-OHT, n class="Chemical">4-hydroxytamoxifen; At, atrium; dpi, days post-injury; IA, injured area; CI, cryoinjury; CMs, cardiomyocytes; V, ventricle. Scale bars, 100 μm. See also Figure S2.

To further fine-tune the characterization of sox10-derived cells, we performed n class="Chemical">4-OHT recombination in juveniles. We used 9-week-old sox10:CreER;ubb:Switch as well as sox10:CreER crossed to vmhcl:loxP-tagBFP-loxP-mCherry-NTR (vmBRN), a line that allows to specifically trace recombined vmhcl-expressing ventricular CMs by mCherry expression (Sánchez-Iranzo et al., 2018). After recombination, animals were raised and cryoinjured at 17 weeks of age (Figure 5A). We observed an increase in the sox10-derived cell area after injury (Figures 5B–5S). This was most evident when observing whole-mount hearts (Figures 5D and 5M). sox10-derived CMs were predominantly observed in the subepicardial regions of the ventricle (Figures 5H–5J and 5Q–5S). Quantification of mCherry signal within the ventricle indicated a difference between uninjured and injured groups, which did not reach statistical significance (Figure 5T). Also, no overall expansion of mCherry+ cells could be observed by quantification on the sectioned hearts (Figure 5U). However, we observed that within the different apicobasal regions, the proportion of mCherry+ cells was clearly increased when compared to uninjured hearts (Figure 5V). Similar to results obtained in adults, mCherry+ cells accumulated close to the injury site, suggesting that the sox10+ population is already present in the juvenile heart. Importantly, control experiments, with no 4-OHT, yielded no recombination, and the hearts were completely devoid of mCherry expression, both in juveniles and adults (Figures 5W–5Y). Thus, with sox10:CreER, we are fully controlling recombination events and can therefore faithfully trace the fate of cells with active sox10:CreER T2 expression at the time of 4-OHT addition.

Figure 5

Fate Mapping of Juvenile sox10+ Cells and Evidence of Non-leaky sox10:CreER Activity

(A) 4-OHT was administered to sox10:CreER; vmhcl:loxP-tagBFP-loxP-mCherry-NTR (sox10:CreER; vmBRN) zebrafish during juvenile stage (9 weeks post-fertilization), and adult uninjured hearts or hearts at 7 dpi were collected and imaged.

(B–D) Fluorescence stereomicroscope acquisition of a dissected uninjured heart showing sox10-derived CMs in red (mCherry+) and other ventricular CMs in blue (vmhcl:BFP+). (B) Merged channels (B), blue channel (C), and red channel (D) are shown.

(E–J) Cryosection of an uninjured heart. Shown are merged (E) as well as single green (MHC+ cells; F) and red (mCherry+ cells; G) channels. (H–J) Zoomed views of boxed region in (E). Again, merged (H) as well as single green (I) and red (J) channels are shown. sox10-derived CMs (mCherry+) are detected near the subepicardial regions of the myocardium.

(K–M) Fluorescence stereomicroscope acquisition of a dissected injured heart showing sox10-derived cells in red (mCherry+) and ventricular CMs in blue (vmhcl:BFP+). Merged channels (K), blue channel (L), and red channel (M) are shown.

(N–S) Cryosection of an injured heart at 7 dpi. sox10-derived CMs are detected near the subepicardial regions and close to the injured area. Shown are merged (N) as well as single green (MHC+ cells; O) and red (mCherry+ cells; P) channels.

(Q–S) Zoomed views of boxed region in (N). Shown are merged and single channels of MHC and mCherry signal. Again, merged (Q) as well as single green (R) and red (S) channels are shown.

(T and U) Percentage of juvenile sox10-derived CMs area (mCherry+/MHC+) compared to the area of all myocardial cells (MHC+) in uninjured (n = 4) and injured hearts at 7 dpi (n = 5). (T) Analysis performed on whole-mount images as shown in (B) and (K); (U) analysis performed on sectioned hearts as shown in (E) and (N). Shown are data for individual hearts as well as mean ± SD (two-tailed non-parametric t test).

(V) Quantification of sox10-derived cells in different apicobasal regions in injured and uninjured hearts. Data are mean ± SD. Statistical testing was performed using one-way ANOVA followed by Tukey’s honest significant difference test. ∗p < 0.05 when comparing percentage between injured at 100 μm with uninjured at 200–600 μm.

(W) sox10:CreER; vmBRN zebrafish as shown in (K)–(M) but without 4-OHT treatment. Note the lack of mCherry+ cells in whole-mount views (upper row) and on sections (lower row) (n = 2).

(X) EtOH was administered to sox10:CreER;ubb:loxP-GFP-loxP-mCherry zebrafish 10 and 7 days before injury, and hearts at 7 dpi were collected and imaged (n = 3).

(Y) Maximal intensity projection of a confocal z stack through a heart from condition as described in (X). sox10-derived cells are mCherry+ (red). The whole myocardium is MHC+ (green). Shown are merged and single channels. mCherry+ cells could not be detected. Image shown is comprised of nine stitched high-resolution acquisitions.

4-OHT, 4-hydroxytamoxifen; At, atrium; CI, cryoinjury; CMs, cardiomyocytes; dpi, days post-injury; EtOH, ethanol; IA, injured area; V, ventricle. Scale bar, 50 μm. See also Figure S2.

Fate Mapping of Juvenile sox10+ Cells ahemical">nd Evidehemical">nce of n class="Chemical">Non-leaky sox10:CreER Activity

(A) 4-OHT was admihemical">nistered to n class="Gene">sox10:CreER; vmhcl:loxP-tagBFP-loxP-mCherry-NTR (sox10:CreER; vmBRN) zebrafish during juvenile stage (9 weeks post-fertilization), and adult uninjured hearts or hearts at 7 dpi were collected and imaged.

(B–D) Fluorescence stereomicroscope acquisition of a dissected uninjured heart showing sox10-derived CMs ihemical">n red (mCherry+) ahemical">nd other ventricular CMs ihemical">n blue (vmhcl:BFP+). (B) Merged chahemical">nnels (B), blue chahemical">nnel (C), ahemical">nd red chahemical">nnel (D) are showhemical">n.

(K–M) Fluorescence stereomicroscope acquisition of a dissected injured heart showing sox10-derived cells ihemical">n red (mCherry+) ahemical">nd vehemical">ntricular CMs ihemical">n blue (vmhcl:BFP+). Merged chahemical">nnels (K), blue chahemical">nnel (L), ahemical">nd red chahemical">nnel (M) are shown.

(N–S) Cryosectiohemical">n of ahemical">n ihemical">njured heart at 7 dpi. n class="Gene">sox10-derived CMs are detected near the subepicardial regions and close to the injured area. Shown are merged (N) as well as single green (MHC+ cells; O) and red (mCherry+ cells; P) channels.

(Q–S) Zoomed views of boxed region in (N). Shown are merged ahemical">nd sihemical">ngle chahemical">nnels of MHC ahemical">nd mCherry sighemical">nal. Agaihemical">n, merged (Q) as well as sihemical">ngle greehemical">n (R) ahemical">nd red (S) chahemical">nnels are shown.

(T and U) Percentage of juvenile sox10-derived CMs area (mCherry+/MHC+) compared to the area of all myocardial cells (MHC+) ihemical">n uhemical">nihemical">njured (n = 4) ahemical">nd ihemical">njured hearts at 7 dpi (n = 5). (T) Analysis performed ohemical">n whole-mount images as showhemical">n ihemical">n (B) ahemical">nd (K); (U) ahemical">nalysis performed ohemical">n sectiohemical">ned hearts as showhemical">n ihemical">n (E) ahemical">nd (n class="Chemical">N). Shown are data for individual hearts as well as mean ± SD (two-tailed non-parametric t test).

(V) Quantification of sox10-derived cells ihemical">n different apicobasal regions ihemical">n ihemical">njured ahemical">nd uhemical">nihemical">njured hearts. Data are meahemical">n ± SD. Statistical testihemical">ng was performed usihemical">ng ohemical">ne-way ANOVA followed by Tukey’s honest significant difference test. ∗p < 0.05 when comparing percentage between injured at 100 μm with uninjured at 200–600 μm.

(W) sox10:CreER; vmBRn class="Chemical">N zebrafish as shown in (K)–(M) but without 4-OHT treatment. Note the lack of mCherry+ cells in whole-mount views (upper row) and on sections (lower row) (n = 2).

(X) EtOH was admihemical">nistered to n class="Gene">sox10:CreER;ubb:loxP-GFP-loxP-mCherry zebrafish 10 and 7 days before injury, and hearts at 7 dpi were collected and imaged (n = 3).

(Y) Maximal intensity projection of a confocal z stack through a heart from condition as described in (X). sox10-derived cells are mCherry+ (red). The whole myocardium is MHC+ (green). Showhemical">n are merged ahemical">nd sihemical">ngle chahemical">nnels. mCherry+ cells could not be detected. Image showhemical">n is comprised of nihemical">ne stitched high-resolutiohemical">n acquisitions.

4-OHT, n class="Chemical">4-hydroxytamoxifen; At, atrium; CI, cryoinjury; CMs, cardiomyocytes; dpi, days post-injury; EtOH, ethanol; IA, injured area; V, ventricle. Scale bar, 50 μm. See also Figure S2.

Collectively, these results show that sox10:CreER fate mappihemical">ng enables the detectiohemical">n of a subset of CMs present ihemical">n the adult heart that expahemical">nds ihemical">n response to ihemical">njury ahemical">nd cohemical">ntributes preferentially to myocardial regenen class="Species">ration in the zebrafish.

sox10-Derived CMs Reveal a Specific Gene Signature

To investigate if sox10-derived CMs differ ihemical">n their gene expressiohemical">n profile compared to other CMs, we performed transcriptome ahemical">nalysis (Figures 6A ahemical">nd 6B). To make the characterizatiohemical">n specific for the myocardium, we used the n class="Gene">sox10:CreER;vmBRN line, in which upon 4-OHT-induced recombination, sox10-derived CMs express mCherry and the rest of the ventricular CMs express blue fluorescent protein (BFP). Two pulses of 4-OHT were administered 2 weeks before injury and 3 weeks before heart dissection. mCherry+, BFP+, and double-positive mCherry+/BFP+ CMs were FAC sorted from sox10:CreER cn17;vmBRN uninjured hearts and hearts at 7 dpi and processed for RNA sequencing (RNA-seq). For bioinformatics analysis, we compared mCherry+ (comprising all samples that were mCherry+ or mCherry+/BFP+) with mCherry− (comprising samples that were only BFP+).

Figure 6

sox10-Derived CMs Reveal a Specific Transcriptome Signature

sox10:CreER zebrafish were treated with 4-OHT. A group of zebrafish was cryoinjured and collected at 7dpi and a control group was left uninjured. Hearts were disaggregated, sox10-derived CMs (mCherry+) and the rest of the CMs of the heart (mCherry−) were FAC sorted, and RNA-seq was performed on 26 samples consisting of 20 CMs each.

(A and B) Recombination and collection were carried out either on uninjured hearts (A) or hearts at 7 dpi (B). (A) Volcano plots representing differentially expressed genes (DEGs) in mCherry+ and mCherry− CMs from uninjured (A) or 7 dpi (B) hearts. DEGs upregulated in mCherry− CMs are shown in blue, and DEGs upregulated in mCherry+ CMs in red. These genes had an adjusted p value (adjp) ≤0.05 and a log2 fold change (LFC) ≥2. Genes with a false discovery rate (FDR) ≥0.05 are represented in gray. Note the high number of DEGs upregulated in mCherry+ cells at 7 dpi. Some differentially expressed genes previously linked to regeneration as well as sox10 are highlighted.

(C) Biological processes differentially enriched in mCherry+ versus mCherry− CMs from uninjured hearts (source for analysis, DEG list).

(D) Pathway analysis using gene enrichment analysis (GSEA) of genes with LFC ≥2 present in mCherry+ and mCherry− CMs of uninjured hearts. Representative biological processes were plotted and ordered according to the normalized enrichment score. Z score represents whether a specific function is increased or decreased. Red bar, enriched in mCherry+ CMs; blue bar, enriched in the rest of the CMs.

(E) Biological processes differentially enriched in mCherry+ versus mCherry− CMs from injured hearts (source for analysis, DEG list).

(F and G) Pathway analysis (F) and hallmarks (G) using GSEA of genes with LFC ≥2 present in mCherry+ and mCherry− CMs of injured hearts. Representative biological processes were plotted and ordered according to the normalized enrichment score. Z score represents whether a specific function is increased or decreased. Red bar, enriched in mCherry+ CMs; blue bar, enriched in the rest of the CMs.

4-OHT, 4-Hydroxytamoxifen; CI, cryoinjury; dpi, days post-injury; dpt, days post-treatment; GO, Gene Ontology. See also Figures S3–S5 and Tables S1, S2, S3, and S4.

sox10-Derived CMs Reveal a Specific Transcriptome Sighemical">nature

sox10:CreER n class="Species">zebrafish were treated with 4-OHT. A group of zebrafish was cryoinjured and collected at 7dpi and a control group was left uninjured. Hearts were disaggregated, sox10-derived CMs (mCherry+) and the rest of the CMs of the heart (mCherry−) were FAC sorted, and RNA-seq was performed on 26 samples consisting of 20 CMs each.

(A and B) Recombination and collection were carried out either on uninjured hearts (A) or hearts at 7 dpi (B). (A) Volcano plots representing differentially expressed genes (DEGs) in mCherry+ and mCherry− CMs from uninjured (A) or 7 dpi (B) hearts. DEGs upregulated in mCherry− CMs are shown in blue, and DEGs upregulated in mCherry+ CMs in red. These genes had an adjusted p value (adjp) ≤0.05 and a log2 fold change (LFC) ≥2. Genes with a false discovery rate (FDR) ≥0.05 are represented ihemical">n gray. Note the high number of DEGs upregulated in mCherry+ cells at 7 dpi. Some differentially expressed genes previously linked to regeneration as well as sox10 are highlighted.

4-OHT, n class="Chemical">4-Hydroxytamoxifen; CI, cryoinjury; dpi, days post-injury; dpt, days post-treatment; GO, Gene Ontology. See also Figures S3–S5 and Tables S1, S2, S3, and S4.

In uninjured hearts, 101 genes were upregulated and 129 genes were downregulated in sox10-derived (mCherry+) CMs compared to the rest of the ventricular CMs (mCherry−) (Figures 6A ahemical">nd 6B; Table S1). Gene enrichment ahemical">nalysis of mCherry+ ahemical">nd mCherry− transcriptomic profiles ihemical">n uhemical">nihemical">njured hearts revealed several metabolic differences ben class="Chemical">tween these two groups, including changes in oxidative phosphorylation and nucleic acid metabolism (Figures 6C, 6D, and S3; Table S1).

Notably, when comparihemical">ng mCherry+ ahemical">nd mCherry− groups after ihemical">njury, n class="Gene">sox10-derived cells were transcriptionally more active; 415 genes were upregulated in mCherry+ CMs, while only 30 genes were upregulated in mCherry− CMs (Figure 6B; Table S2). Importantly, sox10 mRNA was significantly upregulated in mCherry+ CMs from injured hearts at 7 dpi, showing that the sox10:CreER line allows tracing of endogenous sox10-expressing cells. The genes encoding the T-box transcription factors tbx20 and tbx5a were among the genes upregulated in sox10-derived CMs (Figure 6B; Table S2). These genes were previously shown to be expressed in CM populations involved in heart regeneration (Sánchez-Iranzo et al., 2018) and play an active role in this process (Grajevskaja et al., 2018, Xiang et al., 2016).

Gene enrichment analysis in injured conditions for mCherry+ and mCherry− CMs showed that Gene Ontology (GO) biological processes related to negative regulation of the cell cycle were inhibited in mCherry+ CMs. This suggests that mCherry+ CMs have a pro-regenerative profile (Figures 6E–6G ahemical">nd S3; Table S2). Furthermore, mCherry+ CMs were enriched for pathways ihemical">nvolved ihemical">n myocardial growth, ihemical">ncludihemical">ng CM differentiatiohemical">n, cardiac cell development, ahemical">nd n class="Disease">cardiac muscle contraction (Figures 6E–6G; Table S2). This result is consistent with a role for sox10-derived CMs in rebuilding the lost myocardium.

We next analyzed if mCherry+ cells respond equally to an injury as other CMs. For this, we compared on the one hand the chahemical">nges ihemical">n gene expressiohemical">n of mCherry+ CMs ihemical">n ihemical">njured with uhemical">nihemical">njured hearts (Figure S4) ahemical">nd ohemical">n the other the chahemical">nges ihemical">n gene expressiohemical">n of mCherry− CMs ben class="Chemical">tween both conditions (Figure S5). Upon injury, mCherry+ cells upregulated 767 genes, and only 26 genes were downregulated (Figure S4C; Table S3). Thus, overall, mCherry+ CMs respond to injury with an increase in gene expression. Biological processes related to cell proliferation, cell motility, and response to injury were enriched in mCherry+ CMs upon cryoinjury (Figure S4D; Table S3). Ingenuity pathway analysis (IPA) further confirmed the enrichment of canonical pathways related to proliferation in mCherry+ CMs of injured hearts (Figure S4E; Table S3).

When we analyzed the changes in gene expression of the rest of the CMs (mCherry−) in response to injury, we observed fewer differentially expressed genes (DEGs) than those detected for mCherry+ cells (Figure S5; Table S4). This indicated that sox10-derived cells reactivate gene expressiohemical">n to a larger extent when compared to the rest of the CMs.

Altogether, these data indicate that sox10 promoter expressiohemical">n defihemical">nes a group of myocardial cells ihemical">n the adult uhemical">nihemical">njured n class="Species">zebrafish heart with a unique gene expression signature and pro-regenerative transcriptomic profile in response to injury.

sox10+ Cells Are Necessary for Cardiac Regeneration

The accumulation of sox10-derived cells ihemical">n the regenen class="Species">rated myocardium strongly suggests that these cells contribute to the replacement of injured myocardium. To determine the function of this population during heart regeneration, we genetically ablated sox10-derived cells using the transgenic line sox10:CreER;β-actin:loxP-mCherry-loxP-DTA (Wang et al., 2011), which allows cell ablation upon diptheria toxin (DTA) overexpression. We administered 4-OHT to adults 3 days and 1 day before cryoinjury (Figure S6A). At 21 dpi, the injured area was larger in sox10+-cell-depleted animals than in the control group (Figures S6B–S6D; n = 13 without 4-OHT and n = 8 with 4-OHT), suggesting that sox10-derived cells are necessary for heart regeneration. There was no significant reduction in animal survival or cardiac function in sox10-cell ablated animals (Figures S6E–S6I; n = 6). During development, sox10-derived cells contribute to the peripheral nervous system, which plays an important role in heart regeneration (Mahmoud et al., 2015, White et al., 2015). Thus, the impaired regeneration in animals with sox10-ablated cells might be a consequence of compromised cardiac innervation. Yet, the β-actin promoter used in the transgenic line has been reported to be strongly expressed in CMs but weak in other cardiac cells (Kikuchi et al., 2010). Nonetheless, we assessed a possible change in innervation upon induced Diphtheria toxin A (iDTA)-based sox10+ cell ablation by anti-tyrosine hydroxylase immunostaining. This experiment yielded no evidence of a reduction in innervation in sox10+ cell ablated hearts when compared to controls (Figures S6J–S6N), suggesting that the ablation of sox10-derived CMs is responsible for the phenotype.

For confirmation, we genetically ablated sox10-derived cells ihemical">n ventricular CMs usihemical">ng n class="Gene">sox10:CreER cn17;vmBRN zebrafish (Figure 7A). With this double transgenic line, ventricular sox10+-derived CMs can be genetically ablated upon addition of metronidazole (Mtz), which induces cell death in nitroreductase (NTR)-expressing cells (Curado et al., 2008; Figure 7A). We ablated mCherry+ cells 1 week before cryoinjury and confirmed the efficiency of ablation by comparing the amount of mCherry+ cells with Mtz-nontreated animals (Figures 7B and 7C; n = 26 treated and 20 untreated). A significant difference in the amount of sox10-derived CMs could be observed between untreated (n = 5) and treated (n = 8) adult zebrafish (Figure 7D), confirming the efficiency of ablation. Consistent with the results using the DTA ablation model, the survival rate of the different groups was similar (Figure 7E). To assess if loss of adult sox10-derived ventricular CMs affects heart regeneration after cryoinjury, we collected hearts at 30 dpi and assessed fibrotic tissue deposition (Figures 7F–7H). Fish in which ventricular sox10-derived CMs had been ablated revealed a persistent fibrotic scar and a larger injury area compared to the control groups sox10:CreER cn17;vmBRN without Mtz administration and sox10:CreER cn17;ubb:loxP-mCherry with Mtz (Figures 7F–7I). This indicates that ablation of the small pool of sox10-derived ventricular CMs in the uninjured adult zebrafish heart, comprising less than 1% of total myocardial volume, affects subsequent heart regeneration upon cryoinjury.

Figure 7

Genetic Ablation of sox10+ CMs Impairs Cardiac Regeneration

(A) sox10:CreER;vmBRN or sox10:CreER;ubb:Switch zebrafish were treated with 4-OHT 14 and 12 days before cryoinjury. They were treated with Mtz on days 10 and 7 before injury. A control group of sox10:CreER;vmBRN was not treated with Mtz. Cryoinjured hearts were collected at 30 dpi.

(B and C) Whole-mount view of a confocal 3D projection of z stacks through a sox10:CreER;vmBRN heart after 4-OHT and with (B) or without (C) Mtz treatments and 30 dpi.

(D) Percentage of the volume from mCherry+ CMs relative to all CMs: mCherry+; myosin heavy chain (MHC)+ versus all MHC+ cells (p = 0.006; two-tailed non-parametric t test).

(E) Survival rate of animals from the different groups as described in (A). No difference in mortality was observed among groups according to a Fisher’s exact test (p = 1.000).

(F–H) AFOG histological staining on sagittal sections of cryoinjured hearts at 30 dpi. (F) 4-OHT- and Mtz-treated sox10:CreER;vmBRN heart section. (G) 4-OHT-treated sox10:CreER;vmBRN heart section. (H) 4-OHT- and Mtz-treated sox10:CreER;ubb:Switch heart section.

(I) Quantification of IA in the three conditions shown in (F-H). IA versus total ventricular myocardial area was measured. Shown are values for individual hearts as well as mean ± SD (statistical analysis by non-parametric t test: sox10:CreER;vmBRN Mtz-treated versus sox10:CreER;vmBRN Mtz-untreated, p = 0.0078; sox10:CreER;vmBRN Mtz-treated versus sox10:CreER;ubb:Switch Mtz-treated, p = 0.0072; sox10:CreER;vmBRN Mtz-untreated versus sox10:CreER;ubb:Switch Mtz-treated, p = 0.1020).

4-OHT, 4-hydroxytamoxifen; At, atrium; CI, cryoinjury; dpi, days post-injury; IA, injured area; Mtz, Metronidazole V, ventricle. Scale bars represent 100 μm (B and C) and 200 μm (F–H). See also Figure S6.

Genetic Ablation of sox10+ CMs Impairs Cardiac Regehemical">nen class="Species">ration

(A) sox10:CreER;vmBRn class="Chemical">N or sox10:CreER;ubb:Switch zebrafish were treated with 4-OHT 14 and 12 days before cryoinjury. They were treated with Mtz on days 10 and 7 before injury. A control group of sox10:CreER;vmBRN was not treated with Mtz. Cryoinjured hearts were collected at 30 dpi.

(B and C) Whole-mount view of a confocal 3D projection of z stacks through a sox10:CreER;vmBRn class="Chemical">N heart after 4-OHT and with (B) or without (C) Mtz treatments and 30 dpi.

(E) Survival rate of ahemical">nimals from the differehemical">nt groups as described ihemical">n (A). n class="Chemical">No difference in mortality was observed among groups according to a Fisher’s exact test (p = 1.000).

(F–H) AFOG histological staining on sagittal sections of cryoinjured hearts at 30 dpi. (F) 4-OHT- ahemical">nd Mtz-treated n class="Gene">sox10:CreER;vmBRN heart section. (G) 4-OHT-treated sox10:CreER;vmBRN heart section. (H) 4-OHT- and Mtz-treated sox10:CreER;ubb:Switch heart section.

(I) Quantification of IA in the three conditions shown in (F-H). IA versus total ventricular myocardial area was measured. Shown are values for individual hearts as well as mean ± SD (statistical analysis by non-parametric t test: sox10:CreER;vmBRn class="Chemical">N Mtz-treated versus sox10:CreER;vmBRN Mtz-untreated, p = 0.0078; sox10:CreER;vmBRN Mtz-treated versus sox10:CreER;ubb:Switch Mtz-treated, p = 0.0072; sox10:CreER;vmBRN Mtz-untreated versus sox10:CreER;ubb:Switch Mtz-treated, p = 0.1020).

4-OHT, n class="Chemical">4-hydroxytamoxifen; At, atrium; CI, cryoinjury; dpi, days post-injury; IA, injured area; Mtz, Metronidazole V, ventricle. Scale bars represent 100 μm (B and C) and 200 μm (F–H). See also Figure S6.

Discussion

CMs can proliferate ihemical">n the adult n class="Species">zebrafish (Wills et al., 2008), and this can represent a basis for the high regenerative capacity observed in the injured heart. Although little is known about CM populations that contribute to regeneration, studies showed that they activate gata4 and ctgfa regulatory elements in response to injury (Kikuchi et al., 2010, Pfefferli and Jaźwińska, 2017). Recent clonal analysis studies using pan-myocardial lineage tracing have suggested that distinct CM subsets can contribute to heart regeneration in the zebrafish as well as mouse (Gupta et al., 2013, Sereti et al., 2018). Here, we report genetic fate mapping with a specific promoter element of preexistent CMs, which are present in the adult heart and expand more than the rest of the CMs in response to injury. Our study suggests that this small sox10-derived CM population is essential for regeneration, since its genetic ablation prior to cryoinjury impairs cardiac regeneration. Unfortunately, it was technically not possible to perform an ablation of an equally small random CM population. Thus, we could not fully evaluate if the effect on regeneration was specific to of the effect due to ablation of the sox10-derived CM population.

Adult neural crest stem cells were proposed as the source of progenitor cells during adult pigment cell regeneratiohemical">n ihemical">n the n class="Species">zebrafish (Iyengar et al., 2015). Moreover, in rodents, neural crest stem cells were suggested to participate in repair mechanisms after myocardial infarction (Tamura et al., 2016). A neural crest cell origin of the sox10+ population contributing to heart regeneration is plausible but should be confirmed with genetic fate mapping using additional neural crest marker genes. Alternatively, the zebrafish heart might harbor a small pool of sox10+ CMs that efficiently expand in response to cryoinjury. It will be interesting to elucidate if such an expansion also occurs in other injury setups such as ventricular resection or genetic ablation. Adult sox10+ CMs reveal a unique gene signature both in uninjured hearts and upon injury. Interestingly, sox10 transcripts were also detected in CMs in a recently published single-cell transcriptome of zebrafish embryos (Wagner et al., 2018), further supporting our findings of a sox10+ CM population in the zebrafish heart. Moreover, here, we also report sox10 expression in sox10-derived CMs, supporting that our driver lines recapitulate expression of the endogenous gene. The biological pathways enriched in sox10-derived CMs compared to the rest of the ventricular myocardium were related to developmental processes, metabolism, and cell proliferation. This gene signature could be key for their increased contribution to the regenerated myocardium. The fact that sox10-derived cells upregulate more genes in response to injury than other ventricular CMs could suggest that they are epigenetically less repressed and therefore more sensitive to injury response and prone to contribute to heart regeneration.

An active sox10 promoter might represent a particular state of CMs. Alterhemical">natively, CMs with ahemical">n active n class="Gene">sox10 promoter might represent a distinct CM population in the zebrafish heart with high regenerative capacity. Understanding if cardiac sox10+ cells are unique to this species or shared in mammals might help us understand the basis of regenerative capacity.

STAR★Methods

Key Resources Table

Lead Contact and Materials Availability

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Nadia Mercader (nadia.mercader@ahemical">na.unibe.ch).

All unique/stable reagents generated ihemical">n this study are available from the Lead Cohemical">ntact with a completed Materials Transfer Agreemehemical">nt.

Experimental Model and Subject Details

Experiments were conducted with zebrafish embryos, juveniles ahemical">nd adults aged 6–18 mohemical">nths, raised at maximal 5 fish/l ahemical">nd maihemical">ntaihemical">ned uhemical">nder the same envirohemical">nmental cohemical">nditions: 27.5-28°C, 650-700μs/cm, pH 7.5, the lightihemical">ng cohemical">nditions were 14:10 hours (light: dark) ahemical">nd 10% of water exchahemical">nge a day. Experiments were approved by the Commuhemical">nity of Madrid “Dirección General de Medio Ambiente” ihemical">n Spaihemical">n, the Lahemical">ndesamt für Verbraucherschutz Thürihemical">ngen, Germahemical">ny ahemical">nd the “Amt für Lahemical">ndwirtschaft uhemical">nd Natur” from the Canton of Bern, Switzerland. All animal procedures conformed to EU Directive 86/609/EEC and Recommendation 2007/526/EC regarding the protection of animals used for experimental and other scientific purposes, enforced in Spanish law under Real Decreto 1201/2005. Experiments in Switzerland were conducted under the licenses BE95/15 and BE64/18. For longitudinal experiments, the selected animals were grown together with Casper (White et al., 2008) zebrafish until heart collection at the density as explained above.

Method Details

Generation of Tg(sox10:CreERT2)cn17

In order to remove the reporter myl7:GFP from the lihemical">ne Tg(n class="Gene">sox10:CreER,myl7:GFP) (Mongera et al., 2013) we injected the guide RNA sgRNA eGFP1 GGCGAGGGCGATGCCACCTA targeting GFP (Auer et al., 2014) together with Cas9 protein into 1-cell stage embryos and raised offspring without GFP expression in the heart. Germline transmission was evaluated by crossing F0 into ubi:Switch and inducing recombination with 4-OHT administration. This line was used for the experiments shown in Figures 5, 6, and 7.

4-Hydroxytamoxifen administration

4-hydroxytamoxifen (n class="Chemical">4-OHT; Sigma H7904) was administered at the indicated times and treatments were performed overnight. Prior to administration, the 10 mM stock (dissolved in 99.8% ethanol) was heated for 10 minutes at 65°C (Felker et al., 2016). For genetic labeling in Tg(sox10:CreER; ubb:Switch) embryos, 4-OHT was administered at 5-10 μM from 12 to 48 hours post-fertilization (hpf). For lineage tracing studies in adult fish, 4-OHT was administered at 10 μM overnight.

Cryoinjury and analysis of the injured area

Cryoinjury was performed as previously described (González-Rosa and Mercader, 2012). Adult fish were anesthetized with 0.032% tricaine (Sigma, St Louis, MO, USA) ahemical">nd their pericardial cavity on class="Gene">pened with microdissection scissors to expose the heart. A copper filament cooled in liquid nitrogen was placed on the ventricular surface of the heart until thawing. After surgery, animals were revived by gently directing water to their gills using a plastic Pasteur pipette.

For analysis of regeneratiohemical">n, ahemical">nimals were euthahemical">nized at different times post-ihemical">njury by immersiohemical">n ihemical">n 0.16% n class="Chemical">Tricaine (Sigma, St Louis, MO, USA), and hearts were dissected in media containing 2 U/ml heparin and 0.1 M KCl. For quantification of injured area on paraffin sections as shown in Figures 1F and 1G, color deconvolution tool and color threshold tool (ImageJ Software) were used to segment and measure the injured and uninjured myocardium in μm2.

Metronidazole administration

For genetic ablation using Metronidazole (Mtz; Sigma, M3761), Mtz was diluted ihemical">n fish water at 10 mM with n class="Chemical">DMSO at 0.2% and administered overnight.

BrdU administration

Animals were injected intraperitoneally either at 6 dpi or at 7 dpi and 29 dpi with 30 μl of 2.5 mg/ml of 5-Bromo-2-deoxyuridine (n class="Chemical">BrdU, B5002-1G, Sigma). Hearts were collected and processed for analysis at 7 dpi and 30 dpi. To calculate the proliferation index, cryosections were immunostained with anti-BrdU, anti-RFP, and anti-MHC antibodies as described below. At least 3 ventricular sections were imaged for each heart. MHC+/mCherry+/BrdU+ CMs compared to MHC+/mCherry-/BrdU+ CMs were counted manually using ImageJ software.

Histological staining

Hearts were fixed in 2 or 4% paraformaldehyde (n class="Chemical">PFA) in phosphate-buffered saline (PBS) overnight at 4°C. Samples were then washed in PBS, dehydrated through graded alcohols, washed in Xylol and embedded in paraffin wax. All histological staining were performed on 7 μm paraffin sections cut on a microtome (Leica and Reichert-Jung), mounted on Superfrost slides (Fisher Scientific), and dried overnight at 37°C. Sections were deparaffinized in xylol, rehydrated and washed in distilled water. Connective tissue was stained using Acid Fuchsine Orange G (AFOG) (González-Rosa et al., 2014). ImageJ software was used to quantify cryoinjured area in uninjured and injured hearts.

Cardiac imaging by echocardiography

Animals were anaesthetized by immersion for approx. 5 min in a combined solution of 60 mM Tricaine/3 mM n class="Chemical">Isoflurane dissolved in fish tank water (González-Rosa et al., 2014). Individual fish were placed ventral side up on a custom-made sponge in a Petri dish filled with the anesthetic solution. Two-dimensional (2D) high-resolution real-time in vivo images were obtained with the Vevo2100 Imaging System through a RMV708 (22-83 MHz) scanhead (VisualSonics, Toronto, Canada). Imaging and image analysis were performed as described (González-Rosa et al., 2014).

Immunofluorescence on sections

Heart sections were deparaffinized, rehydrated ahemical">nd washed ihemical">n distilled water. Epitope recovery was carried out by boilihemical">ng ihemical">n n class="Chemical">citrate buffer (pH 6.0) for 20 min in a microwave at full power. Sections were permeabilized with Triton X-100 0.5% for 15 min. Non-specific binding sites were saturated by incubation for 1 hour in blocking solution (5% BSA, 5% goat serum, 0.1% Tween-20). Endogenous biotin was blocked with the avidin-biotin blocking kit (Vector, Burlingame, CA, USA). For tyramides amplification, slides were blocked in 3% H202-PBS for 20 minutes. Slides were incubated overnight with the following primary antibodies at 4°C: anti-myosin heavy chain (MF20, DSHB; diluted 1:20), anti-tropomyosin (CH1, DSHB; diluted 1:20), anti-RFP (Abcam, diluted 1:150), anti-BrdU (BD PharMingen diluted 1:100), anti-mCherry (16D47, Thermo Fisher Scientific, diluted 1:150), anti-Tyrosine Hydroxylase (Sigma, diluted 1:150), biotinylated anti-rabbit (Thermo Fisher Scientific, 1:150), HRP (DAKO, 1:250). Antibody signals were detected with streptavidin- or Alexa (488, 568, 633)-conjugated secondary antibodies (Invitrogen; each diluted 1:250) and streptavidin-Cy3 (Molecular Probes, SA1010) after incubation for 1 hour at room temperature. For tyramides amplification (Merck, TSA Plus Cyanine 3 System, Cat# NEL744001KT), Cy3 was conjugated for 4 minutes. Nuclei were stained with 4’,6-Diamidino-2-phenylindole dihydrochloride (DAPI) (1:1000, Merck) and slides were mounted in DAKO fluorescent mounting medium (DAKO). Images were analyzed and processed using ImageJ.

Cryosections were prepared as above with the following modifications: heart sections were incubated for 30 min in PBS at 37°C to remove the gelatin. They were washed two more times with PBS at room temperature ahemical">nd then immunofluorescence proceeded as for paraffihemical">n sections.

Whole mount heart imaging and image processing

Hearts were incubated with block solution (5% BSA, 5% got serum, 0.1% Tween-20) for two days ahemical">nd ihemical">ncubated ihemical">n primary ahemical">ntibody for three days at 4°C ihemical">n rockihemical">ng n class="Disease">agitation. We washed hearts three times with PBS Tw 0.1% for 1 hour. Secondary and tertiary antibody incubations were performed for two days at 4°C in rocking agitation. Again, we washed hearts three times with PBS Tw 0.1% for 1 hour. Then, hearts were fixed in 2% PFA overnight. We used CUBIC reagent (Susaki et al., 2015) for tissue clearing. Hearts were incubated in CUBIC reagent 1 for three days at 37°C, washed in 0.1% PBS/Tween 20 three times for 20 minutes, whole mount immunofluorescence was performed and samples were incubated afterward in CUBIC reagent 2 for three further days at room temperature. Hearts were mounted on a glass bottom culture dish (MatTek Corporation) for confocal acquisition. Whole heart images were obtained with Zeiss LSM 780, Zeiss LSM 880, and Leica TCS SP8 confocal microscopes with a 10 dry, 20 × dry and 40x water-dipping lenses. Images were recorded at 512 × 512, 1024 × 1024 and 2048x 2048 resolution. Tile scan and z stack of each heart was acquired. The proportion of mCherry+/MHC+ versus all MHC+ CMs was evaluated with Imaris software 8.2 (BITPLANE). A distance transformation algorithm (Imaris software 8.2) was used to study the distance of mCherry+/MHC+ CMs to the injured or apex area. Adult zebrafish heart representing embryonic sox10-derived CMs in the Figures S2E and S2F. lsm raw data file was converted to. ims (Imaris file extension) by Imaris software 8.2 (BITPLANE). The Imaris image was saved as TIFF, with a larger field of view than the original file. This corresponds to the images shown in the panels.

Imaging of larvae in vivo

Double transgenic larvae Tg(sox10:CreER;ubb:Switch) were transferred to E3 medium cohemical">ntaihemical">nihemical">ng 0.2 mg/ml n class="Chemical">tricaine and 0.0033% PTU and immobilized using 0.7% agarose (Bio-Rad Low Melting agarose, #Cat 161-3111) in a glass bottom microwell-dish (MatTek Corporation). Zebrafish hearts were scanned using bidirectionally acquisition with SP5 confocal microscope (Leica SP5) using a 20x glycerol lens. Larvae were carefully removed from the agarose embedding and were grown to adults in fish tanks together with Casper fish. Adult zebrafish hearts were collected, fixed in 2% PFA overnight and scanned with LSM 700 Confocal microscope using 20x dry lens (Zeiss). 3D reconstruction and analysis were done using Imaris Software 8.2.

RNAscope 2.5HD Detection Reagent (RED) - Immunofluorescence method

All the hearts were fixed at room temperature (RT) for 24h ihemical">n 10% n class="Chemical">Neutral Formalin Buffer (NFB). After fixation, samples were washed 3 times for 10 minutes in 1x PBS. Dehydration process was performed using a standard ethanol series (10 minutes each), followed by two xylol washes (5 minutes each) and embedding of tissues was carry out.

Paraffin blocks were cut with microtome (Microm) at 6μm thickness per section, collected in the water bath with SuperFrost slides and baked slides in a dry oven for one hour at 60°C. After that, slides were dewaxed by incubating 2 times for 5 minutes in xylol, 2 times for 2 mihemical">nutes ihemical">n 100% n class="Chemical">ethanol and dried in a dry oven for 5 minutes at 60°C. They were then, permeabilized with hydrogen peroxide (ACD#322381) for 10 minutes at RT and washed 2 times for 1 minute in distilled water.

Target retrieval (ACD, #322000) was performed for 15 minutes at 100°C. Slides were then washed for 15 s in distilled water and 100% ethanol for 3 mihemical">nutes ahemical">nd dried for 5 mihemical">nutes at 60°C. Afterward PAP n class="Gene">Pen (Vector, #H-4000) was used to create a hydrophobic barrier for each section. After that, slides were incubated with protease Plus (ACD#322381) for 5 minutes at 40°C and washed in distilled water twice.

Two probes were designed for the experiment (Dr-Sox10 ahemical">nd negative cohemical">ntrol probe-DapB). Hybridizatiohemical">n was dohemical">ne by ihemical">ncubatihemical">ng the sections with the the probes for 2 hours at 40°C, followed by 2 washes of 2 mihemical">nutes ihemical">n washihemical">ng buffer (ACD, #310091).

Finally, signal detection was done by using the RNAscope 2.5 n class="Disease">HD Detection Reagent-RED (ACD, #322360) as follows: incubation with AMP1 for 30min at 40°C; washing with wash buffer AMP2 for 15min at 40°C, wash buffer AMP3 for 30min at 40°C, wash buffer AMP4 for 15min at 40°C, wash buffer AMP5 for 30min at 40°C, wash buffer AMP6 for 15min at 40°C, wash buffer RED working solution for 10min at RT, and wash in distilled water twice for 5 minutes.

Disaggregation of zebrafish hearts, cardiomyocytes sorting and RNA-Seq library production

Uninjured and injured recombined adult zebrafish n class="Gene">sox10:CreER;vmhcl:loxP-tagBFP-loxP-mCherry-NTR hearts were collected 12 days after the final 4-OHT pulse or 7 dpi respectively, and processed according to previous protocols (Sánchez-Iranzo et al., 2018, Tessadori et al., 2012). Atrium and bulbus arteriosus were removed to obtain only the ventricle. A total of 15 uninjured and 18 injured hearts were used to create 5 pools (comprised of 3 hearts each) of uninjured hearts and 6 pools (each again comprised of 3 hearts) for injured hearts. From each pool 20 CMs were FAC-sorted.

mCherry+, mCherry;BFP+ and BFP- CMs were sorted in 0.2 mL tubes in lysis buffer using Synergy 4L Cell Sorter and immediately frozen at −80°C. Smart-Seq2 RNA library prepan class="Species">ration was performed according to previous protocols (Picelli et al., 2014). An Agilent Bioanalyzer was used to measure quality of library preparation. Library concentration was measured using the Qubit fluorometer (ThermoFisher Scientific). Final libraries concentration was 10 nM. Libraries were sequenced using Illumina NextSeq 500.

Bioinformatics analysis

BCL files were converted to FastQ files, using bcl2fastq2 (v2.20.0.422 – Illumina). Reads were mapped to the reference genome (Ensembl build 11, release 94) using Hisat2, version 2.1.0 (Kim et al., 2015) and counting was performed using featureCounts, version 1.6.0 (Liao et al., 2014). Multiple quality control features were measured and observed using both FastQC, version 0.11.5 (Andrews, 2010) and RseQC, version 2.6.4 (Wang et al., 2012). From 44 sequenced samples 18 were discarded as they did not pass the quality control. For bioinformatics analysis, we compared mCherry+ (mCherry+, mCherry+BFP+) with mCherry- (BFP+) pools. The comparison was performed for samples extracted from uninjured and 7 dpi hearts. Downstream analysis was performed in R, version 3.5.1 (R Core Team, 2018).

Counts were normalized and differential expression between desighemical">n groups was tested usihemical">ng package DESeq2 v.1.20.00 with no log2 fold chahemical">nge shrihemical">nkage (default betaPrior optiohemical">n for the latest versions of the tool). Prihemical">ncn class="Chemical">ipal component analysis (PCA) plots, volcano plots and heatmaps were generated using the ggplot2 package, version 3.0.0 (H. Wickham. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York, 2016.).

Further analyses were performed with DESeq2 results. For the enrichment we selected all the Gene Stable IDs and translated to Mus musculus Gene Stable IDs ahemical">nd obtaihemical">ned the ENTREZIDs and SYMBOLs using biomaRt package (Durinck et al., 2005). With the genes translated, the top differentially expressed genes (DEG) that passed FDR < 0.05 for Gene Ontology (GO) (Ashburner et al., 2000) over representation analysis (ORA) using clusterProfiler package (Yu et al., 2012).

Afterwards, we performed deeper analysis for overall gene expression with gene set enrichment analysis (GSEA). The differential gene expression results from DESeq2 were sorted by Log2FoldChange value. Kyoto Encyclopedia of Genes and Genomes (KEGG) (Kanehisa and Goto, 2000) and Molecular Signature Database (MsigDB) (Liberzon et al., 2015) using the Hallmarks collection were used for biological insight. For the GSEA analyses clusterProfiler and FGSEA packages (Sergushichev, 2016) were used with KEGG and MsigDB Hallmarks gene set respectively. For data representation only those with adjusted p value < 0.05 were considered significant of the results obtained.

Ingenuity pathway analysis core analysis (IPA, QIAGEn class="Chemical">N Inc., https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis) was used to identify canonical pathways, functions and diseases related to our differentially expressed genes in uninjured and injured conditions.

Quantification and Statistical Analysis

Statistical analysis was performed using GraphPad. Shown are means and error bars represent SD in all graphs. The specific test used, exact value of n, explanation of what n represents, definition of center, dispersion and precision measures are indicated in each figure legend or citation in the main text. Normal distributiohemical">n was tested to decide if a parametric or nohemical">n-parametric test needed to be applied.

Data and Code Availability

RNA-seq data were deposited at Gehemical">ne OmnibusDatabase: Gehemical">ne Omnibus Database with referehemical">nce n class="Chemical">GSE 133571. Raw data of images as well as statistical analysis has been uploaded at Mendeley Database: https://doi.org/10.17632/5h7z68ck98.2

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal anti MF20	DSHB	Cat# MF 20, RRID:AB_2147781
Mouse monoclonal anti tropomyosin	DSHB	Cat# CH1, RRID:AB_2205770
Mouse monoclonal anti BrdU	BD PharMingen	Cat# 347583, RRID:AB_400327
Rat monoclonal anti mCherry	Thermo Fisher Scientific	Cat# M11217, RRID:AB_2536611
Rabbit polyclonal anti-Tyrosine Hydroxylase	Sigma Aldrich	Cat# T8700, RRID:AB_1080430
Rabbit polyclonal anti-RFP	Abcam	Cat# ab34771, RRID:AB_777699
Biotin-SP-conjugated AffiniPure anti-rabbit IgG (H+L)	Jackson Immuno Research Laboratories	Cat# 713-065-003, RRID:AB_2340715
Goat anti-Rabbit IgG (H+L) Secondary Antibody, Alexa Fluor® 568 conjugate	Thermo Fisher Scientific	Cat# S-11226, RRID:AB_2315774
Goat anti-Mouse IgG1 Secondary Antibody, Alexa Fluor® 488 conjugate	Thermo Fisher Scientific	Cat# A-21121; RRID:AB_2535764
Goat anti-Mouse IgG1 Secondary Antibody, Alexa Fluor® 568 conjugate	Thermo Fisher Scientific	Cat# A-21124; RRID:AB_2535766
Goat anti-Mouse IgG1 Secondary Antibody, Alexa Fluor® 647 conjugate	Thermo Fisher Scientific	Cat# A-21240; RRID:AB_2535809
Goat anti-Mouse IgG2b Secondary Antibody, Alexa Fluor® 488 conjugate	Thermo Fisher Scientific	Cat# A-21242, RRID:AB_253581
Chemicals, Peptides, and Recombinant Proteins
4’,6-Diamidino-2-phenylindole dihydrochloride (DAPI)	Thermo Fisher Scientific	Cat# D3571
5-bromo-2′-deoxyuridine (BrdU)	Sigma Aldrich	Cat# 59-14-3
4-hydroxytamoxifen	Sigma Aldrich	Cat#T5648
2,3-butanedione monoxime	Sigma-Aldrich	Cat#B0753
Metronidazole	Sigma-Aldrich	Cat#M3761
Software and Algorithms
Fiji	NIH	SCR_002285
GraphPad Prism 7	GraphPad Software	SCR_002798
Imaris software 8.2	Bitplane	SCR_007370
Ingenuity Pathway Analysis	QIAGEN	SCR_008653
DESeq2 v1.20.00	SciCrunch	SCR_015687
R Project for Statistical Computing	SciCrunch	SCR_001905
Gene Ontology	SciCrunch	SCR_002811
KEGG	SciCrunch	SCR_012773
clusterProfiler	SciCrunch	SCR_016884
MSigDB	SciCrunch	SCR_016863
biomaRt	SciCrunch	SCR_002987
Experimental Models: Organisms/Strains
Zebrafish: Tg(−4.9sox10:egfp)ba2	(Carney et al., 2006)	ZDB-ALT-050913-4
Zebrafish: Tg(sox10:CreERT2,myl7:GFP)t007	(Mongera et al., 2013)	ZDB-ALT-130322-3
Zebrafish: Tg(-3.5ubi:loxP-EGFP-loxP-mCherry)cz1701	(Mosimann et al., 2011)	ZDB-TGCONSTRCT-110124-1
Zebrafish: Tg(actb2:loxP-mCherry-loxP-DipTox)pd36	(Wang et al., 2011)	ZDB-ALT-110914-1
Zebrafish: Tg(vmhcl:loxP-myctagBFP-STOP-loxP-NTR-mCherry)cn5	(Sánchez-Iranzo et al., 2018)	ZDB-ALT-170711-2
Zebrafish: Tg(sox10:CreERT2)cn17	This manuscript	n/a