Mechanosensory organ regeneration in zebrafish depends on a population of multipotent progenitor cells kept latent by Schwann cells

Abstract

Background: Regenerating damaged tissue is a complex process, requiring progenitor cells that must be stimulated to undergo proliferation, differentiation and, often, migratory behaviors and morphological changes. Multiple cell types, both resident within the damaged tissue and recruited to the lesion site, have been shown to participate. However, the cellular and molecular mechanisms involved in the activation of progenitor cell proliferation and differentiation after injury, and their regulation by different cells types, are not fully understood. The zebrafish lateral line is a suitable system to study regeneration because most of its components are fully restored after damage. The posterior lateral line (PLL) is a mechanosensory system that develops embryonically and is initially composed of seven to eight neuromasts distributed along the trunk and tail, connected by a continuous stripe of interneuromastic cells (INCs). The INCs remain in a quiescent state owing to the presence of underlying Schwann cells. They become activated during development to form intercalary neuromasts. However, no studies have described if INCs can participate in a regenerative event, for example, after the total loss of a neuromast.

Results: We used electroablation in transgenic larvae expressing fluorescent proteins in PLL components to completely ablate single neuromasts in larvae and adult fish. This injury results in discontinuity of the INCs, Schwann cells, and the PLL nerve. In vivo imaging showed that the INCs fill the gap left after the injury and can regenerate a new neuromast in the injury zone. Further, a single INC is able to divide and form all cell types in a regenerated neuromast and, during this process, it transiently expresses the sox2 gene, a neural progenitor cell marker. We demonstrate a critical role for Schwann cells as negative regulators of INC proliferation and neuromast regeneration, and that this inhibitory property is completely dependent on active ErbB signaling.

Conclusions: The potential to regenerate a neuromast after damage requires that progenitor cells (INCs) be temporarily released from an inhibitory signal produced by nearby Schwann cells. This simple yet highly effective two-component niche offers the animal robust mechanisms for organ growth and regeneration, which can be sustained throughout life.

Keywords: ErbB; Lateral line; Neuromast; Regeneration; Schwann cell; Zebrafish.

Fig. 1

Electroablation as a method for localized tissue injury in the posterior lateral line (PLL) of zebrafish larvae. a Trunk of a tg(cxcr4b:mCherry) larva showing red-labeled PLL cells, including the second, third, and fourth neuromasts of the PLL (L2, L3, and L4) connected by interneuromastic cells (INCs, white arrows). The image was captured 1 hour before injury (hbi). b The trunk of the same larva 4 hours post injury (hpi). The asterisk shows the damaged zone, where the L3 neuromast was located. c–e Higher magnifications of the injured area showing the process of neuromast regeneration. c At 1 hpi, we observed the gap generated between INCs (white arrows) flanking the injury zone (asterisk). d At 24 hpi, INCs located on both sides of the gap reconnected. e At 72 hpi, the L3 neuromast had regenerated (yellow arrowhead). At this stage, the secondary primordium (PrimII) deposited a secondary neuromast (white arrowhead). f–h Double transgenic tg(neurod:GFP; cxcr4b:mCherry) larvae, where the afferent lateral line neurons are labeled in green and neuromasts are labeled in red. f This image shows the L3 region 1 hbi. g Electroablation of L3 interrupts the lateral line nerve. h The nerve regenerates after 24 hpi. i–k Double transgenic tg(foxd3:GFP; cxcr4b:mCherry) larvae, showing the Schwann cells labeled in green (associated with the nerve) and neuromasts and INCs labeled in red. As occurs with the INCs, Schwann cells reconnected after 24 hpi (k). Scale bar: 50 μm

Fig. 2

Neuromast regeneration depends on interneuromastic cell accumulation. The L3 neuromasts of 3 days post fertilization tg(et20:GFP) larvae were electroablated or left uninjured as controls, and fixed at different time points after damage (hours post injury, hpi). a–c Detection of ET20:GFP-labeled cells after electroablation. d Quantification of GFP-labeled cells at the L3 position (n = 10). Initially, in electroablated fish, all accumulating cells expressed GFP but the percentage of GFP versus total cells diminished significantly between 24 and 48 hpi (## p < 0.01). At all stages after injury, L3 neuromasts of electroablated larvae had a much higher proportion of ET20:GFP cells in comparison with control larvae (***p < 0.001). e–h Immunodetection and quantification of Sox2-expressing cells (n = 10). At 6 hpi, few, if any, Sox2-expressing cells were seen in the injury zone but, after 24 hpi, the number of Sox2-expressing cells was approximately the same as in controls (h) (***p < 0.001) . Note the loss of Sox2 expression in the most centrally located cells at 48 hpi (g, yellow arrowhead). i Images extracted from a time-lapse sequence of a double tg(cxcr4b:mCherry;brn3c:GFP) electroablated larva. The sequence reveals the progressive appearance of GFP expression in centrally located hair cells. j In vivo quantification of the number of hair cells in control and injured larvae that regenerated their neuromasts at 2, 24, 48, and 72 hpi (n = 15); ^## and β indicate statistical differences within the same group, control or injured, comparing neighboring values (β p < 0.001, ## p < 0.01), while asterisks reflect statistical difference between control and injured at the same time points (***p < 0.001). Note that the ET20:GFP and Sox2 expression data corresponding to 6 and 24 hpi (shown in d and h, respectively), come from a mix of larvae committed and not committed to regenerate. This is because the samples had to be fixed at stages in which we could not distinguish between the outcomes. Scale bar a–g, i: 50 μm. Further details on replicates are provided in “Quantifications and statistical analysis” in the “Methods” section

Fig. 3

Regenerated neuromasts are chimeric structures derived from two interneuromastic cells (INCs). Transplanting cells from a tg(ubiquitin:RFP) blastula to a tg(et20:GFP) blastula occasionally resulted in fish with one or a few labeled INCs. A transplanted larva harboring a single labeled INC near L3 was selected 3 days post fertilization and subjected to electroablation of the L3 neuromast (a, b). The asterisc indicates the position of the ablated neuromast. The left panels (a, c, e, g; only the red fluorescent protein [RFP] channel is shown) show the behavior of the transplanted cell through time. The right panels (b, d, f, h) show ET20⁺ cells of host larvae (green) and the transplanted cells (pseudocolored in magenta). During regeneration, the single transplanted cell divided and its progeny differentiated into different cells types of the mature neuromast. At 72 hpi, a red-labeled cell (here in magenta) can be observed among the INCs, suggesting that at least one daughter cell maintained the original identity of the progenitor (g, h, yellow arrowhead). Note that the transplantation experiment randomly generated labeled cells of diverse lineages that did not participate in neuromast regeneration (see for example, e–h, white arrowhead). Scale bar: 50 μm

Fig. 4

Contribution of interneuromastic cells (INCs) to neuromast regeneration. a–d Complete elimination of all neuromasts and INCs between L3 and L8 was done 3 days post fertilization in tg(et20:GFP) larvae by electroablation (n = 20). Neuromasts were electroablated whereas INCs were ablated by mechanical displacement of the microelectrode through the skin. The white arrow in a shows the direction of the movement of the microelectrode. The asterisk in a shows the position of the L3 neuromast before electroablation. After injury, the behavior of INCs located proximal to the gap was examined at 11 hours post injury (hpi) (a), 30 hpi (b), 48 hpi (c), and 72 hpi (d). Starting at 30 hpi, INCs accumulated at the injury zone and organized to form a new neuromast. They also migrated, beginning at 48 hpi, extending caudally to create a new line of INCs. e–g An ectopic neuromast can appear de novo after electroablation. e The row of INCs between L2 and L3 was interrupted by electroablation at the position of the asterisk; the last remaining INCs are indicated by arrowheads. f At 21 hpi, INCs started to accumulate, reconnecting the line of cells. g At 72 hpi, a neuromast formed between L2 and L3 at a position where there was no preexisting neuromast (labeled ENm, ectopic new neuromast). At this stage, the secondary primordium (PrimII) was migrating close to L3 and had deposited secondary neuromast LII.3. Further details on replicates are provided in “Quantifications and statistical analysis” in the “Methods” section. Scale bar: 50 μm

Fig. 5

Neuromast regeneration success is inversely correlated with the size of the interneuromastic cells (INC) gap generated by electroablation. The L3 neuromast of tg(cxcr4b:mCherry;foxd3:GFP) larvae was electroablated. At 2 hours post injury (hpi), we individually injured larvae and measured the length of the gap between remaining INCs (labeled in red) and Schwann cells (SCs, labeled in green). At 72 hpi, we scored the regeneration of the L3 neuromast and compared the two outcomes (regeneration or no regeneration) after measuring the average gap size in each group. As is shown in a, larvae that could not regenerate the L3 neuromast presented a larger gap between INCs compared to those that regenerated (n = 64). We did the same comparison examining the size of the SC gap and found no effect in this case (b; n = 64). Calculating the ratio between INC and SC gap size again produced a significant difference when regenerating versus non-regenerating outcomes were compared (c; n = 64). Further, there was no difference in the ratio of the INC/SC gap between larvae that regenerated at 48 hpi versus those that regenerated at 72 hpi (d; n = 27). * p < 0.05; n.s. not significant. Further details on replicates are provided in “Quantifications and statistical analysis” in the “Methods” section

Fig. 6

Damage to Schwann cells is required for neuromast regeneration. tg(cxcr4b:mCherry;et20:GFP) larvae 3 days post fertilization (dpf) were treated with 100 μM CuSO₄ for 2 h to ablate all neuromasts without affecting Schwann cells. a A control (uninjured) larva showing the region between L2 and L4. The secondary primordium (PrimII) is seen migrating (white arrowhead in all images). b Three hours after copper treatment, all neuromasts of the lateral line system had been chemically ablated (gaps demarcated by yellow arrowheads). c At 24 hours post treatment (4 dpf), interneuromastic cells (INCs) had filled the gaps but no neuromast regeneration occurred. At this time, electroablation was carried out at the approximate position where L3 was (white box). d A new L3 neuromast formed only where electroablation took place at 48 hours post injury (hpi; at 6 dpf). Neither intervention (copper treatment or electroablation) impaired the migration of PrimII (white arrowhead) and deposition of secondary neuromasts (LII.3). The asterisk indicates the site of injury. e The graph shows the percentage of injured larvae that regenerated a neuromast after the different treatments: L3 neuromast electroablation (L3; n = 100); electroablation of INCs between the L2 and L3 neuromasts (INC; n = 100); L3 neuromast electroablation in larvae treated with 5 μM of AG1478 from 10 hpf until 58 hpf (5 μM AG1478 10–58 hpf; n = 75); L3 electroablation in larvae treated with 5 μM of AG1478 from 0 hpi until 72 hpi (5 μM AG1478 0–72 hpi; n = 75); 100 μM copper treatment (100 μM CuSO_4, n = 100); or copper treatment combined with electroablation of L3 (100 μM CuSO4 + Electroablation; n = 60). Scale bar a–d: 100 μm. Further details on replicates are provided in “Quantifications and statistical analysis” in the “Methods” section

Fig. 7

Adult caudal lateral line regeneration after electroablation. A 6-month-old tg(et20:GFP) fish was electroablated in the caudal fin lateral line by applying two pulses of 2 s duration and 25 μA current intensity (n = 60, three independent experiments). Images of the injury were taken beginning from 1-minute post injury (1 mpi) to 20 days post injury (dpi). The image in a was taken at 1 mbi to show the original position of the neuromasts. The asterisk and bracket in b show the extent of the damage. c, d The yellow arrowhead points to the rostral most interneuromastic cells located immediately proximal to the damage. The red arrowheads in each panel show the disappearance of the nearest neuromast located caudal to the injury. At 10 dpi, ET20⁺ cells accumulated (e, white arrowhead) and at 20 dpi had matured into a neuromast (g, white arrowhead). The position where regenerated neuromasts appeared did not recapitulate the original distribution of neuromasts before injury. Scale bar a–g: 100 μm

Fig. 8

Schematic representation of neuromast regeneration from interneuromastic cells (INCs). a A schematic representation of the four stages identified here during regeneration of a whole neuromast. I, after neuromast ablation, a remaining INC that has multipotent progenitor properties (green with blue nucleus, ET20:GFP⁺) becomes activated; II, the INC divides, and daughter cells accumulate where the original neuromast was and acquire expression of Sox2 (green with red nuclei; ET20:GFP⁺, Sox2⁺); III, mantle cells (brown) differentiate at the periphery of the cell cluster while central cells lose GFP expression (ET20:GFP⁻, Sox2⁺), becoming neuromast progenitor cells; IV, neuromast regeneration is achieved when hair cells (blue) differentiate. b Differentiation pathway of INCs during neuromast regeneration. It is unclear from our study whether INC progeny can differentiate directly into the mantle cells (MC) or whether MCs differentiate from a proliferative INC (pINC*) (an ET20⁺Sox2⁺Brn3c⁻ cell). The pINCs accumulate and can differentiate into a neuromast progenitor cell (PC), losing the ET20 marker. Finally, the acquisition of Brn3c expression (hair cells, HC)