Defects in ErbB-dependent establishment of adult melanocyte stem cells reveal independent origins for embryonic and regeneration melanocytes

Abstract

Adult stem cells are responsible for maintaining and repairing tissues during the life of an organism. Tissue repair in humans, however, is limited compared to the regenerative capabilities of other vertebrates, such as the zebrafish (Danio rerio). An understanding of stem cell mechanisms, such as how they are established, their self-renewal properties, and their recruitment to produce new cells is therefore important for the application of regenerative medicine. We use larval melanocyte regeneration following treatment with the melanocytotoxic drug MoTP to investigate these mechanisms in Melanocyte Stem Cell (MSC) regulation. In this paper, we show that the receptor tyrosine kinase, erbb3b, is required for establishing the adult MSC responsible for regenerating the larval melanocyte population. Both the erbb3b mutant and wild-type fish treated with the ErbB inhibitor, AG1478, develop normal embryonic melanocytes but fail to regenerate melanocytes after MoTP-induced melanocyte ablation. By administering AG1478 at different time points, we show that ErbB signaling is only required for regeneration prior to MoTP treatment and before 48 hours of development, consistent with a role in establishing MSCs. We then show that overexpression of kitla, the Kit ligand, in transgenic larvae leads to recruitment of MSCs, resulting in overproliferation of melanocytes. Furthermore, kitla overexpression can rescue AG1478-blocked regeneration, suggesting that ErbB signaling is required to promote the progression and specification of the MSC from a pre-MSC state. This study provides evidence that ErbB signaling is required for the establishment of adult MSCs during embryonic development. That this requirement is not shared with the embryonic melanocytes suggests that embryonic melanocytes develop directly, without proceeding through the ErbB-dependent MSC. Moreover, the shared requirement of larval melanocyte regeneration and metamorphic melanocytes that develops at the larval-to-adult transition suggests that these post-embryonic melanocytes develop from the same adult MSC population. Lastly, that kitla overexpression can recruit the MSC to develop excess melanocytes raises the possibility that Kit signaling may be involved in MSC recruitment during regeneration.

Figure 1. ErbB signaling is required for larval melanocyte regeneration.

(A) Cartoon of drug treatment timeline for assaying melanocyte ontogeny and regeneration. Arrow indicates when embryos were collected for photos and melanocyte counts, in this case at 168 hpf. (B) Quantitation of average dorsal melanocytes from somites 1–26 for each treatment in (A) for melanocyte ontogeny (gray) and melanocyte regeneration (red). Error bars represent standard deviation, * represents P<0.05 (Student t-test, N = 10). Photos of representative larvae at 168 hpf for ontogeny (C, E, G) and regeneration (D, F, H). WT larvae regenerate nearly completely (compare D to C). WT treated with AG1478 (E, F) and erbb3b mutants (G, H) largely fail to regenerate but have normal ontogenetic number of melanocytes. A few regeneration melanocytes are observed in the head and sporadically in parts of the trunk (black bars in F and H) but are mostly absent throughout the trunk (red bars in F and H).

Figure 2. Regeneration requires ErbB signaling before 48 hpf.

(A) Cartoon of drug treatment timeline. (B) Quantitation of average regenerated dorsal trunk melanocytes from somites 5–12 for each treatment in (A) for untreated (gray) and AG1478 treated (blue). Error bars represent standard deviation, * P<0.05, - P>0.05 (Student t-test, N = 10). Time shifts with AG1478 treatment reveal that it blocks regeneration when administered prior to 48 hpf. When AG1478 is added after 48 hpf regeneration is unaffected. The early effect is observed when MoTP treatment is withheld until after AG1478 washout at 48 hpf.

Figure 3. Temporal shifts of AG1478 reveal stem cell establishment occurs in a rostrocaudal progression.

(A) Cartoon of drug treatment timeline. (B) Quantitation of average regenerated dorsal melanocytes for each treatment in (A) in the trunk (somites 5–12, in gray) and in the tail (somites 16–21, in black). Error bars represent standard deviation, * P<0.05, - P>0.05 (Student t-test, N = 10). Larvae not treated with AG1478 show full regeneration in the trunk and in the tail. When larvae are treated with the full AG1478 treatment, from 9–48 hpf, they fail to regenerate in either the trunk or the tail. Larvae treated early with AG1478 from 9–30 hpf fail to regenerate in the trunk, but have normal regeneration in the tail. Later treatments of AG1478 from 30–48 hpf show more regeneration in the trunk than in the tail. (C) Larva with early treatment of AG1478 from 14–24 hpf showing a regeneration defect in the trunk but with normal regeneration in the head and tail.

Figure 4. ErbB signaling is not required for stem cell self-renewal.

(A) Cartoon of drug treatment timeline for single and double regeneration assays. (B) Quantitation of average regenerated dorsal melanocytes for each treatment in (A). Error bars represent standard deviation, - P>0.05 (Student's t-test, N = 10 for single regeneration, >5 for double regeneration). As previously shown (see Figure 2), when larvae are treated with AG1478 after 2 dpf, they regenerate normally following a single round of MoTP treatment. When animals are treated with two rounds of MoTP (from 2–4 dpf and from 5–7 dpf) they regenerate slightly less melanocytes than a single round by 9 dpf. To test whether AG1478 treatment would block stem cell self-renewal, we treated animals with MoTP and AG1478 from 2–4 dpf, and then with a second round of MoTP from 5–7 dpf. These animals regenerated melanocytes similar to non-AG1478 treated larvae for double regeneration (P = 0.9).

Figure 5. Overexpression of kitla after 4 dpf results in proliferation of melanocytes.

(A) Cartoon of the pT2hsp70:kitla expression construct used in heatshock experiments. The heatshock promoter, hsp70 drives kitla expression, allowing for expression of kitla after heatshocking the injected embryos at 37°C. (B) Cartoon of experimental protocol. Following injection of pT2hsp70:kitla, larvae were heatshocked for 1 hour at 4 and 5 dpf in the presence of BrdU (from 4 to 8 dpf). (C) Quantitation of dorsal melanocytes represented as a percentage of wild type, and the percentage of BrdU labeled melanocytes is represented in red for each treatment. Larvae injected with pT2hsp70:kitla and heatshocked develop approximately 30% more melanocytes than uninjected larva. These larvae also show an increased percentage of BrdU labeled melanocytes (∼27%) that is comparable to the number of excess melanocytes. Representative examples of melanocytes in WT (D, F, H) and pT2hsp70:kitla (E, G, I), showing melanocyte nuclei stained with DAPI for WT (F) and pT2hsp70:kitla (G), and BrdU staining for WT (H) and pT2hsp70:kitla (I). Melanocytes considered having BrdU+ nuclei are labeled with arrows and BrdU- nuclei are labeled with arrowheads.

Figure 6. The regeneration specific drug, ICI-118,551, reveals that excess melanocytes in PT2hsp70:kitla larvae arise from the MSC lineage.

(A) Cartoon of drug treatment timeline. (B) Quantitative data for ontogeny and regeneration represented as percentage of WT melanocyte numbers. Error bars show standard deviations, * P<0.05, - P>0.05 (Student's t-test, N>7). (C) WT larvae treated with ICI-118,551 develop faintly melanized ontogenetic melanocytes, in contrast to (D) failure to develop melanized melanocytes when challenged to regenerate in the presence of ICI-118,551. (E) In situ analysis reveals regeneration proceeds to the dct+ melanoblast stage (arrows) in the presence of ICI-118,551. (F) ICI-118,551 treated PT2hsp70:kitla embryos develop similar numbers of faintly melanized melanocytes as ICI-118,551 treated embryos shown in (C). Differences in iridophore appearance between (C) and (F) are results of slightly different illumination conditions.

Figure 7. kitla melanocyte proliferation is enhanced by AG1478.

(A) Cartoon of the PT2hsp70:kitla construct in transgenic animals. The constitutive cmv promoter drives expression of kitla. (B) Schematic of drug treatment timeline with early treatment of AG1478. (C) Quantitation of average ontogenetic dorsal melanocytes for each treatment in (A) in the trunk (somites 5–12) and in the tail (somites 16–21). Error bars represent standard deviation, * P<0.05, - P>0.05 (Student t-test, N = 10). As reported previously (see Figure 1) WT larvae show no effect in ontogenetic melanocyte number with AG1478 treatment compared to untreated WT. PT2hsp70:kitla animals have significantly more melanocytes in both the trunk and the tail. Treating PT2hsp70:kitla animals with AG1478 early, from 9–30 hpf, produces a significantly greater number of melanocytes in the trunk. In contrast, the tail region of AG1478-treated PT2hsp70:kitla larvae shows a significant, but much smaller increase than the trunk, in the number of excess melanocytes. (D) Trunk dorsal stripes of WT larvae are typically two melanocytes wide. (E) PT2hsp70:kitla, however, have trunk dorsal stripes that are about 4 cells wide. (F) AG1478-treated PT2hsp70:kitla have trunk dorsal stripes ∼6 melanocytes wide.

Figure 8. Model for the parallel establishment of the zebrafish embryonic melanocyte lineage and the adult melanocyte stem cell lineage.

Our results indicate that there are two distinct melanocyte lineages that develop in the zebrafish embryo: the embryonic or ontogenetic melanocyte lineage and the regenerative and metamorphic melanocyte lineage. In embryonic development, neural crest cells (squares) give rise to dct+ melanoblasts (triangles), which later develop into embryonic melanocytes (black star). The regeneration and metamorphic lineage develops in parallel to embryonic melanocytes, prior to 48 hpf and presumably from the neural crest. The establishment of melanocyte stem cells (MSCs, dark circle) require ErbB signaling to progress from the pre-MSC state (light circle). The MSC can self-renew (circular arrow) and give rise to melanoblasts and melanocytes during metamorphosis or regeneration, when inhibition from embryonic melanocytes (block arrow) is relieved. The MSC–derived melanocytes are sensitive to developmental blockade by ICI-118,551 after the dct+ melanoblast stage. Ectopic expression of kitla can recruit the MSC lineage by either inducing MSCs or pre–MSCs resulting in overproliferation of melanocytes.