Post-embryonic nerve-associated precursors to adult pigment cells: genetic requirements and dynamics of morphogenesis and differentiation

Abstract

The pigment cells of vertebrates serve a variety of functions and generate a stunning variety of patterns. These cells are also implicated in human pathologies including melanoma. Whereas the events of pigment cell development have been studied extensively in the embryo, much less is known about morphogenesis and differentiation of these cells during post-embryonic stages. Previous studies of zebrafish revealed genetically distinct populations of embryonic and adult melanophores, the ectotherm homologue of amniote melanocytes. Here, we use molecular markers, vital labeling, time-lapse imaging, mutational analyses, and transgenesis to identify peripheral nerves as a niche for precursors to adult melanophores that subsequently migrate to the skin to form the adult pigment pattern. We further identify genetic requirements for establishing, maintaining, and recruiting precursors to the adult melanophore lineage and demonstrate novel compensatory behaviors during pattern regulation in mutant backgrounds. Finally, we show that distinct populations of latent precursors having differential regenerative capabilities persist into the adult. These findings provide a foundation for future studies of post-embryonic pigment cell precursors in development, evolution, and neoplasia.

Figure 1. Post-embryonic expression of embryonic neural crest and glial markers.

Shown are in situ hybridizations performed on transverse sections of 7–9 SSL larvae. (A) foxd3 transcript was detected in dorsal root ganglia (arrow) and scattered cells (e.g., arrowheads) within the myotome (m) and near the epidermis (e). (B) sox10 expression by cells adjacent to the neural tube (arrow) and within the myotome (arrowhead). (C) foxd3+ cells (e.g., arrowheads) at the base of the dorsal fin (f). (D) foxd3+ cells within the myotomes and near the epidermis. (E–G) foxd3+, sox10+, and crestin+ cells within the myotomes. (H–J) Cells expressed mitfa (arrowheads), within the horizontal myoseptum (H), at the surface of the myotome (I), and at the base of anal fin (J) (see text for details).

Figure 2. kitla misexpression induced ectopic melanophores within the myotomes.

(A) kitla was normally expressed within the epidermis and hypodermis (arrow) during post-embryonic development of wild-type fish. sc, spinal cord. m, myotome, e, epidermis. (B) In sibling Tg(hsp70::kitla) larvae, heat shock resulted in increased kitla transgene expression within the epidermis (arrowheads) as well as ectopic expression within the myotomes (arrow), spinal cord, and elsewhere. (C) The late melanophore lineage marker dopachrome tautomerase (dct), encoding an enzyme required for melanin synthesis , , was not expressed within the myotomes of wild-type fish. (D) However, dct was expressed by scattered cells within the myotomes (arrow) in larvae misexpressing kitla. (E,F) Newly differentiated ectopic melanophores (arrow) were found between myotubes (arrowheads; E) and these cells continued to express sox10 protein (F). (G) Vibratome section revealing ectopic melanophores (arrow) within the myotome of a Tg(hsp70::kitla) larva 48 h after the initiation of kitla misexpression. Melanophores deep within the myotome were found only in Tg(hsp70::kitla), though melanophores were occasionally found within the horizontal myoseptum of both transgenic and wild-type larvae [ectopic melanophores per larva, Tg(hsp70::kitla): mean±SE = 1.3±0.15, range = 0–7 cells, n = 80 larvae; non-transgenic siblings: mean±SE = 0±0, range = 0, n = 69]. Longer durations of kitla misexpression resulted in more ectopic melanophores per larva. Suggesting that ectopic melanophores differentiated in situ rather than migrated into the myotomes from the hypodermis, labeling of hypodermal cells by photoconversion of mitfa::Eos+ failed to reveal movement of cells away from enhanced kitla expression in the epidermis into the myotome (n = 10 larvae, 3–5 cells per individual). (H) In contrast to the wild-type, ectopic melanophores were significantly fewer in erbb3b; Tg(hsp70::kitla) mutants [Wilcoxon test, Z = 7.1, P<0.0001; ectopic melanophores per larva, erbb3b; Tg(hsp70::kitla): mean±SE = 0.04±0.03, range = 0–1, n = 50 larvae; non-transgenic siblings: mean±SE = 0±0, range = 0, n = 70] and in Tg(hsp70::kitla) larvae treated with AG1478 during the ErbB embryonic critical period [Wilcoxon test, Z = 2.9, P<0.005; ectopic melanophores per larva, AG1478-treated Tg(hsp70::kitla): mean±SE = 0.7±0.2, range = 0–4, n = 45 larvae; untreated Tg(hsp70::kitla) siblings: mean±SE = 1.6±0.2, range = 0–5, n = 35].

Figure 3. Extra-hypodermal cells expressing mitfa::GFP, foxd3, and sox10 in wild-type larvae and their deficiency in erbb3b mutants.

All images are from early metamorphic (6.2–8.0 SSL) wild-type larvae except for B, L, and M, from erbb3b mutant larvae. (A–C) Transverse confocal projections (collapsing ∼1.5 mm of trunk along the anterior-posterior axis) showing GFP+ cells in living larvae (left side of each larva is shown). Images correspond to Videos S1, S2, S3. (A) mitfa::GFP+ cells in a wild-type fish occur in the hypodermis, between the epidermis (e) and the myotome (m), within the the myotome itself (arrow), and above the spinal cord (sc). Arrowhead, location of the horizontal myoseptum. (B) In erbb3b mutants, most mitfa::GFP+ cells were missing. This image is intentionally overexposed compared to A, revealing faint reflected fluorescence from iridescent iridophores in the hypodermis (arrowhead), which are present in wild-type larvae as well. (C) sox10::GFP+ cells were found in extra-hypodermal locations of wild-type Tg(−4.9sox10:egfp)^ba2 larvae . (D–M) Immunohistochemical analyses of fixed specimens. (D) Co-expression of mitfa::GFP (green) and sox10 (red). v, vertebral column. (E) mitfa::GFP+ cells aligned with mbp+ glia (red) along ventral root motor fibers. Arrow, mitfa::GFP+ cells did not co-express mbp. (F) Lateral view showing mitfa::GFP+ cells and foxd3+ cells (red) between Hu+ neurons (blue) of dorsal root ganglia. mitfa::GFP+ and foxd3+ cells were often found close to one another (e.g., arrows) whereas other cells co-expressed mitfa::GFP and foxd3 (arrowhead). (G) Lateral view with superimposed brightfield and fluorescence images showing mitfa::GFP+ and foxd3+ cells along a peripheral nerve fiber stained for acetylated alpha tubulin (blue) within the myotome. Arrows, adjacent mitfa::GFP+ and foxd3+ cells. (H). A nerve plexus near the base of the caudal fin harbored numerous mitfa::GFP+ and foxd3+ cells. (I) Along a peripheral nerve within the myotome some cells co-expressed mitfa::GFP and foxd3 (arrowhead), whereas cells expressing either mitfa::GFP+ or foxd3+ were often juxtaposed (arrows). (J,K) Transverse sections through the dorsal trunk showing sox10+ cells (J) and foxd3+ cells (K) n the hypodermis, within the myotomes, and near the spinal cord. Arrow, lateral line nerve. (L,M) In erbb3b mutant larvae, very few sox10+ (L) or foxd3+ (M) cells were found.

Figure 4. Missing extra-hypodermal precursor cells in erbb3b mutants, co-expression of molecular markers, and temporally regulated proliferation.

(A) Occurrence of cells in sections from the mid-trunk of wild-type and erbb3b mutant larvae. Each class of cells was reduced in erbb3b mutants (all P<0.0001). (B,C) Cell frequencies in wild-type and erbb3b mutants across stages. pre-met, pre-metamorphosis (4.9–5.3 SSL); early met, early metamorphosis (6.2–8.0 SSL); late met, late metamorphosis (9.0–13.0 SSL). (B) The frequency of foxd3+; mitfa::GFP+ cells was greatest in wild-type larvae during early pigment pattern metamorphosis (doubly vs. singly labeled cells, difference among stages: χ ² = 15.7, d.f. = 2, P<0.0005; N = 1217 total cells examined). Doubly labeled cells tended to be rarer and delayed in erbb3b mutants (N = 83 total cells examined). (C) The frequencies of EdU+ cells differed significantly among stages, with more foxd3+ and sox10+ cells labeled with EdU during the pre-metamorphic period, and more mitfa::GFP+ cells labeled with EdU during early metamorphosis (EdU labeling frequency variation among stages, foxd3+: χ ² = 11.3, d.f. = 2, P<0.005, N = 450 cells; sox10+: χ ² = 140.7, d.f. = 2, P<0.0001; N = 1679 cells; mitfa::GFP+: χ ² = 14.4, d.f. = 2, P<0.001, N = 927 cells). In erbb3b mutants, EdU labeling frequencies were reduced in comparison to wild-type for foxd3+ cells (χ ² = 3.4, d.f. = 1, P = 0.06; N = 44 cells) and sox10+ cells (χ ² = 11.4, d.f. = 1, P<0.001; N = 77 cells), though not significantly so for mitfa::GFP+ cells (P>0.1; N = 59 cells). Asymmetric confidence intervals in B and C, Bayesian 95% upper and lower bounds.

Figure 5. Proliferative extra-hypodermal cells revealed by post-embryonic EdU incorporation.

All images from transverse sections of wild-type larvae staged as in Figure 3 (see Figure S4 for comparisons with erbb3b mutant larvae). (A,B) Merged images showing cells (arrowheads) within the myotomes labeled for either mitfa::GFP (red in A) or sox10::GFP (red in B) as well as EdU (green). (C,D) Cells within the lateral myotomes (m) and near the hypodermis (C) or at the base of the anal fin (f in D) labeled for mitfa::GFP (red), foxd3 (blue), and EdU (green). Merged views show fluorescence images or fluorescence images with brightfield overlays. Arrowheads, triple-labeled cells. m, myotome. e, epidermis.

Figure 6. DiI-labeling showed extra-hypodermal contributions to metamorphic melanophores and iridophores.

(A–C) DiI labeled tissues imaged immediately after injection into the base of the dorsal fin (A), the vicinity of the horizontal myoseptum and lateral line nerve (B), and the inner myotome (C). Each site yielded hypodermal DiI+; mitfa::GFP+ cells or DiI+ melanophores (12 of 30 larvae, 3 of 30 larvae, 15 of 87 larvae, respectively). (D–F) DiI+ cells that expressed either mitfa::GFP (D) or contained melanin (E,F) found within the lateral hypodermis 4 d following injection into the base of the dorsal fin (D, F) or the inner myotome (E). (G) DiI-labeling was observed for additional cells including iridophores. Although the frequencies with which DiI labeled pigment cells were found differed between injection sites, each site gave rise to DiI+ iridophores at a frequency indistinguishable from that of DiI+ mitfa::GFP+ cells (χ ² = 0.6, d.f. = 1, P = 0.4). We did not observe DiI-labeled xanthophores.

Figure 7. Ex vivo time-lapse imaging revealed extra-hypodermal origins and morphogenetic behaviors of hypodermal mitfa::GFP+ cells and melanophores.

All panels show lateral views of larval trunks and are derived from time-lapse movies of Tg(mitfa::GFP) larvae. Elapsed time (min) at lower right of each panel. (A) mitfa::GFP+ cells differentiated into melanophores (e.g., arrowhead). (B–D) mitfa::GFP+ cells entered the hypodermis during the larval-to-adult transformation. (B) A cell at the dorsal margin of the myotomes extended a long process (arrowhead) into the hypodermis and interacted with processes of a second cell. *, cell body. (C) A long process (arrowhead) preceded emergence of the cell body (*) from the level of the horizontal myoseptum (dotted line). This cell subsequently interacted with a neighboring cell, extended a processes ventrally, and moved in that direction. (D) A cell initially deep within the myotome (*) emerged into the hypodermis. The focal plane changes across panels, from deep within the myotome to the surface of the myotome and hypodermis, where other cells are found already. (E) Death of mitfa::GFP+ cell (*) revealed by fragmentation and cellular debris (arrowheads). (F,G) mitfa::GFP+ cells (F) and melanophores that retain some residual GFP expression (G) proliferating within the hypodermis. Melanophores in G are imaged in a kita mutant (see text for details). See Videos S4, S5, S6, S7, S8, S9, S10, S11, S12, S13, S14, S15, S16.

Figure 8. Sources of mitfa::GFP+ cells that entered the hypodermis.

Shown are the relative frequencies of cells entering the hypodermis in time-lapse movies of trunks derived from wild-type larvae (N = 127 larvae, 1220 mitfa::GFP+ total cells examined) during different periods of pigment pattern metamorphosis (very early, 6.4–6.6 SSL; early, 6.6–6.8 SSL; middle, 6.8–7.6 SSL). Cells newly arrived in the hypodermis (n = 644) were classified as migrating over the dorsal or ventral margins of the myotomes (DV), emerging from the vicinity of the horizontal myoseptum (HM) and lateral line, or emerging from within the myotomes (M), tyically along vertical myosepta. Individuals were binned by stage and bar widths are proportional to the average numbers of newly appearing hypodermal mitfa::GFP+ cells per larva, normalized to cells mm⁻² day⁻¹ (mean±SE above bars). The relative frequencies of cells arising from different sources differed between stages (χ ² = 20.0, d.f. = 4, P<0.0001), with cells increasingly likely to emerge from within the myotomes as compared to migrating over the dorsal or ventral margins of the myotomes.

Figure 9. Genetic controls over the origins, differentiation, and morphogenesis of hypodermal mitfa::GFP+ cells.

Shown are analyses of time-lapse movies for trunks derived from wild-type and mutant larvae (N = 281 larvae, 5241 total cells examined). (A) Total numbers of newly arising hypodermal mitfa::GFP+ cells differed among genotypes (square root transformed data, F _5,273 = 30.2, P<0.0001). Shown are least squares means (±95% confidence intervals) after controlling for significant differences among stages (F _2,273 = 3.9, P<0.0001) and normalized to cells mm⁻² day⁻¹. Letters above bars indicate means that differed significantly (P<0.05) by Tukey-Kramer post hoc comparisons. Numbers within bars indicate numbers of larval trunks examined. (B) The origins of new hypodermal mitfa::GFP+ cells differed among genotypes (χ ² = 145.6, d.f. = 10, P<0.0001; n = 1582 total new cells). Bar widths are proportional to the total numbers of new hypodermal cells observed in each genotype (shown in A). DV, cells entering the hypodermis after migrating over the dorsal or ventral myotome margins; HM, cells entering from the vicinity of the horizontal myoseptum. M, cells entering from within the myotomes. The sources of mitfa::GFP+ cells did not differ significantly across stages overall (χ ² = 0.003, d.f. = 4, P = 1), though different genotypes exhibited stage-dependent variation (stage x genotype interaction: χ ² = 46.3, d.f. = 20, P<0.0001; not shown). (C) Frequencies of differentiation, death and proliferation differed among genotypes. Bar widths are proportional to the total numbers of hypodermal mitfa::GFP+ cells and melanophores observed per larva, after controlling for area and duration of imaging, and normalized to cells mm⁻² day⁻¹ (larva means±SE: wild-type, 116±9; erbb3b, 13±8; tuba8l3a, 62±9; kita, 72±10; csf1r, 78±14; ednrb1, 61±16). Differentiation, The likelihood of mitfa::GFP+ cells acquiring melanin during imaging differed among genotypes (χ ² = 100.6, d.f. = 5, P<0.0001; n = 335 total differentiating cells): the relatively few erbb3b and tuba8l3a mutant cells were especially likely to differentiate whereas very few kita mutant cells differentiated. Additional effects were attributable to stage (χ ² = 30.3, d.f. = 2, P<0.0001) and a stage x genotype interaction (χ ² = 27.3, d.f. = 10, P<0.0001; not shown). Death, The incidence of mitfa::GFP+ cells dying during imaging differed among genotypes (χ ² = 116.9, d.f. = 5, P<0.0001; n = 507 total dying cells) with particularly high rates of death in kita and ednrb1 mutants. Additional variation was attributable to differences among stages (χ ² = 23.5, d.f. = 2, P<0.0001) and a stage x genotype interaction (χ ² = 20.4, d.f. = 10, P<0.05) resulting from an increased likelihood of ednrb1 mutant cells dying at later stages (not shown) Division, The incidence of mitfa::GFP+ cells dividing differed significantly among genotypes (χ ² = 23.6, d.f. = 5, P<0.0001; n = 142 total dividing cells). Asymmetric confidence intervals, Bayesian 95% upper and lower bounds.

Figure 10. Extra-hypodermal precursors were deficient in tuba8l3a mutant larvae.

(A) Wild-type larvae exhibited mitfa::GFP+ cells (green) associated with mbp+ glia (red) of peripheral nerves (arrow). sc, spinal cord; m, myotome; e, epidermis; ll, lateral line nerve. (B) mitfa::GFP+ cells (green), foxd3+ cells (red; arrow), and doubly labeled mitfa::GFP+; foxd3+ cells (arrowhead) were associated with nerve fibers stained for acetylated alpha tubulin (blue). (C) In tuba8l3a mutants, regions deficient for mbp+ glia were also deficient for mitfa::GFP+ cells. (D) Peripheral nerves were often defasciculated (arrow) and were deficient for foxd3+ and mitfa::GFP+ cells.

Figure 11. Limited regeneration of adult hypodermal melanophores following genetic ablation.

(A) Time-course of temperature shifts, with letters corresponding to images in B–H. Final sample size at 698 days post-fertilization: n = 5. (B) A young adult kita; csf1r^TS mutant at 33°C lacked melanophores as in kita; csf1r^j4e1 mutants (some melanized cellular debris resulting from melanophore death is evident dorsally). (C) Temperature downshift to 24°C allowed recovery of a hypodermal melanophore (e.g., arrow) complement initially indistinguishable from kita single mutants . Arrowhead, the dorsal flank is initially devoid of scale-associated melanophores, as is typical of kita mutants. (D,E) Additional rounds of ablation and recovery yield progressively fewer hypodermal melanophores (arrow), though some melanophores develop on the dorsal scales (arrowhead). Hypodermal xanthophores and iridophores were depleted as well (data not shown). (F) Dorsal flank of another individual showing hypodermal melanophores (arrow) and scale melanophores (arrowhead). (G) Detail of scale-associated melanophores. Iridescent iridophores and yellow-orange xanthophores are apparent as well. (H) Detail of hypodermal melanophores viewed through an overlying scale containing a concentration of xanthophores and iridophores (outlined in red).

Figure 12. Model for establishment and maintenance of adult pigment cell precursors and their recruitment during development and regeneration.

(A) Hypothesized lineage relationships, showing neural crest (NC) cells in the early embryo that give rise to Schwann cells and pigment cells of the early larva as well as erbb3b-dependent progenitors to metamorphic glial and pigment cell lineages (mGP). mGP are maintained in association with peripheral nerves and ganglia, express foxd3, and their population expands (multiple arrowheads) in a tuba8l3a-dependent manner. During pigment pattern metamorphosis (met), some mGP differentiate as metamorphic Schwann cells (S) and the expansion of this lineage likely requires erbb3b (not shown). Other mGP become specified for metamorphic pigment cell lineages, as marked by mitfa::GFP expression. Some mitfa::GFP+ cells will give rise to melanophores or iridophores and are initially extra-hypodermally located in peripheral nerves and ganglia (M/I-e) but then migrate to the hypodermis (M/I-h). The expansion of this population requires ednrb1 . Some M/I-h will differentiate as metamorphic iridophores (I), other M/I-h expand their population in a tuba8l3a- and kita-dependent manner and ultimately differentiate as metamorphic melanophores (M). Individual M/I-e or M/I-h may be bipotent for melanophore and iridophore fates, as in embryos , or their respective populations may harbor precursors already specified for either the melanophore or iridophore fate. csf1r-dependent metamorphic xanthophores (X) presumably arise from a different precursor population (dashed line) and promote the survival of metamorphic melanophores (orange arrow) , , . Some mGP persist into the adult and have a limited re-population potential. (B) Schematic of metamorphic larva illustrating sources and migratory pathways of metamorphic melanophore and iridophore precursors. Shown are mGP and M/I-e (colors as in A) associated with nerves and beneath the dorsal fin (f). M/I-e enter the hypodermis (arrows) from the dorsal or ventral margins of the myotomes (m), or after migrating along nerves associated with the the vertical myosepta (vm) or the horizontal myoseptum (hm). Others may arise from the lateral line nerve (ll). Once in the hypodermis, these cells differentiate as melanophores (green cell with heavy black outline) or iridophores (not shown). sc, spinal cord. Additional populations of precursors that may give rise to LM melanophores and scale melanophores are not shown (see text).