The role of thyroid hormones in the development of coloration of two species of Neotropical cichlids

The family Cichlidae is one of the most hyperdiverse groups of freshwater fishes (Salzburger, 2018). It comprises more than 1700 species, but it is believed that many species have yet to be described (Salzburger, 2018; Fricke et al., 2023). An enormous phenotypic and ecological diversity, as well as high speciation rates and numerous examples of sympatric speciation, make cichlids one of the most actively studied model systems in evolutionary biology (Burress, 2015; Kornfield and Smith, 2000; Kocher, 2004; Seehausen, 2006; Powder and Albertson, 2016). The Neotropical cichlids (Cichlinae) are a numerous group including more than 600 species, most of which belong to the tribes Geophagini and Heroini distributed within South and Central America, respectively (Arbour et al., 2016; Fricke et al., 2023; López-Fernández et al., 2013; Říčan et al., 2013). The geophagine cichlids present multiple examples of adaptive radiation accompanied by rapid morphological differentiation in body shape. The heroine cichlids are characterized by an increased diversification rate of feeding apparatus, color vision and pigment pattern (Arbour and López-Fernández, 2016; Burress et al., 2018; Fan et al., 2012; Hauser et al., 2017; Henning et al., 2013; Maan and Sefc, 2013; Říčan et al., 2016; Torres-Dowdall et al., 2021).

Coloration plays an important role in the adaptation and speciation of cichlids in general (Elmer et al., 2009; Miyagi and Terai, 2013; Salzburger, 2009; Seehausen et al., 1999; Sefc et al., 2014; Urban et al., 2022), including the diversification via signal axes during the final stage of adaptive radiation (Ronco et al., 2021). Four types of pigment cells constitute the pigment pattern of cichlids: black/brown – melanophores; metallic iridescence – iridophores; yellow/orange – xanthophores; and red – erythrophores (Maan and Sefc, 2013; Miyagi and Terai, 2013). The composition and number of pigment cell populations change during ontogeny, and the majority of cichlids display distinct larval and adult colorations. In the Neotropical cichlids, the sequence and timing of pigment patterning are variable, and interspecific developmental discrepancies served as the basis for the ‘ontogenetic timing hypothesis’, which suggests four types of melanistic pigment patterning underlying their coloration diversity (Říčan et al., 2005, 2016).

To test this hypothesis, we induced developmental shifts in the pigment patterning of a model species of Neotropical cichlid, the convict cichlid Amatitlania nigrofasciata, via artificial manipulation of the level of thyroid hormones (THs) (Prazdnikov and Shkil, 2019a,b). THs are among the most important regulators of the timing and rate of a wide range of developmental processes in teleosts (Blanton and Specker, 2007; Lazcano and Orozco, 2018; Campinho, 2019), including the development of coloration (McMenamin et al., 2014; Prazdnikov, 2021; Salis et al., 2021; Saunders et al., 2019). These manipulations resulted in the occurrence of experimental phenotypes mimicking those in phylogenetically close and distant species. Our findings supported the ontogenetic timing hypothesis and provided grounds for the assumption of an involvement of the TH alterations in the developmental transformation of the Neotropical cichlid ornamentation. In particular, we considered the larva–juvenile transition of the pigment pattern as a metamorphic event and proposed that increased diversification rate of the Neotropical cichlid coloration is a consequence of evolutionarily relevant changes in the TH developmental dynamic (Prazdnikov and Shkil, 2019a).

Here, we continued our investigation of the role of THs in the pigment patterning of Neotropical cichlids. We studied the development of coloration and developmental dynamics of an active form of TH – triiodothyronine (T₃) – in two closely related heroine cichlids: the convict cichlid, Amatitlania nigrofasciata, and the blue-eye cichlid, Cryptoheros spilurus. We found species-specific features in hormonal developmental dynamics and pigment patterning. To test the role of THs in coloration development, we experimentally provoked hypothyroidism and hyperthyroidism at critical stages of pigment patterning and assessed their developmental and morphological consequences in both species.

All experimental procedures with fish were carried out following the recommendations described in the Guide for the Care and Use of Laboratory Animals (Garber et al., 2011) and were approved by the Ethics Committee of the Severtsov Institute of Ecology and Evolution of the Russian Academy of Sciences.

The convict cichlid, Amatitlania nigrofasciata (Günther 1867), and the blue-eye cichlid, Cryptoheros spilurus (Günther 1862) (Heroini; Cichlinae; Cichlidae; Teleostei) belong to the heroine cichlids and have an overlapping ranges. The former is widespread in Central America (Schmitter-Soto, 2007) and characterized by a low pigment pattern variability (Říčan et al., 2005; Prazdnikov and Shkil, 2019a; Prazdnikov, 2020). In contrast, the latter has a narrow range (Schmitter-Soto, 2007), but displays a high variability in the number of melanophore bars (Říčan et al., 2005). Both species are popular aquarium fish; therefore, techniques for their fertilization, incubation and rearing of larvae and juveniles are well developed (www.seriouslyfish.com).

For each species, we collected several ontogenetic series obtained from independent parent couples. Taking into account parental care typical of these species, the parents were removed from the breeding aquaria after fertilization. The staging of cichlid development was done following Prazdnikov and Shkil (2019a): (i) embryo (fertilization to hatching); (ii) early larva (hatching to complete yolk absorption); (iii) late larva (foraging onset to pelvic fin formation); (iv) juvenile; and (v) adult – fish with sexual dimorphism. As an additional temporal characteristic of the developmental timing in the conspecific series, we used days post-fertilization (dpf).

To describe pigment patterning, we used live specimens. The embryo, larvae and early juveniles were viewed and photographed with a Leica MS5 stereomicroscope equipped with an ocular-micrometer and a Canon EOS 100D digital camera. The juveniles and adults were photographed in the aquaria. The pigment cell type was identified by the color of the pigment. As A. nigrofasciata and C. spilurus are characterized by the same type of coloration ontogeny (Říčan et al., 2016), following our previously obtained data (Prazdnikov and Shkil, 2019a), we distinguished two lineages of melanophores: early larval melanophores (ELMs) and adult melanophores (AMs); and six stages of coloration development: embryonic (melanophores along the notochord and clusters on the yolk sac), early larval (forming two lateral ELM stripes), late larval (appearance of precursors of vertical bars), transition (segregation of AM populations into regions of bars), juvenile (completion of AM element development and disappearance of ELM stripes) and adult (formation of sexual dichromatism). For the melanistic pattern elements (bars, stripes and blotches), we used the classification developed by Říčan et al. (2005, 2016). Quantitative analysis of the pigment pattern was performed using ImageJ software (https://imagej.net/ij/index.html). We assessed the number and area of melanistic elements, as well as the area of interbar spaces on the body. Statistical significance was determined using the Kruskal–Wallis test with Dunn's multiple comparisons test and Bonferroni correction (α=0.05) when applicable, and Spearman's correlation.

In contrast to humans, who have a relatively massive and well-buffered circulating pool of prohormone thyroxin (T₄), teleosts have an extremely small and rapidly cleared pool of serum T₄ (Eales, 2019). Therefore, we assessed only total T₃ – an active form of TH – in the extract of whole fish. The extraction was performed from dry-frozen samples in accordance with the Holzer et al. (2017) extraction protocol. Total T₃ content was evaluated with commercially available ELISA kits (Monobind Inc.) with a range 0.04–7.5 ng ml⁻¹, and microstrip reader Stat Fax 303 plus (Awareness Technology) at 450 nm with reference wavelength at 620 nm in accordance with the manufacturer’s protocol. The optical density was measured in duplicate for each sample (technical replication), and the mean value of the optical density was compared with a standard curve to assess the actual hormonal concentration. Finally, hormone concentration was recalculated to the sample mass, which ranged from 0.1 g (in eggs and larvae) to 0.47 g (in adults) per sample.

In both species, the analysis of T₃ developmental dynamics was replicated (biological replication). For this, we used two clutches (crosses) obtained from different pairs of breeders. Samples were collected simultaneously with photography, with the exception of the last juvenile photograph because of the small number of remaining fish. To achieve the sample mass required for the analysis (≥0.1 g), for the first three developmental stages we used the following: (i) embryonic stage: 90–100 eggs per sample; (ii) early larval stage: 25–30 individuals per sample; and (iii) late larval stage: 11–15 individuals per sample. Thus, for these stages, in each cross we assessed a mean value of T₃ content only. At the early juvenile stage, in each cross, we collected 3–5 samples, each of which contained 2–3 fish. In juveniles and adult fish ≥0.1 g, T₃ content was assessed individually, with 5–14 individuals at each time point.

Taking into account that we failed to find significant intraspecific differences, we used the combined data obtained by biological replication to calculate the average values, minimum/maximum value and standard deviation of T₃ content for each species. The hormonal profiles of the species under study were compared with the Kendall rank correlation coefficient (τ). For this, we ranked average values of T₃ content with age as a ranking coefficient.

Fertilized eggs were divided into equal groups (46 specimens per group for A. nigrofasciata and 40 specimens per group for C. spilurus) reared under the different hormonal regimes. The euthyroid fish were used as a control group. The hyperthyroid (TH-1/2/3) and hypothyroid (Tio-1/2/3) groups were consistent in the developmental timing of the start of treatment with the exogenous T₃ (concentration in water 0.15 μg ml⁻¹) and the goitrogen thiourea [CS(NH₂)₂; final concentration in water 0.03%], respectively. In the TH-1/Tio-1 group, treatment started from the early larva stage (4 dpf); in the TH-2/Tio-2 group, treatment started from the onset of the late larva stage (12 dpf); and in the TH-3/Tio-3 group, treatment started from middle of the late larva stage (20 dpf). In all experimental groups, treatment was aborted with the appearance of sexual dimorphism: for A. nigrofasciata groups: TH-1 85 dpf, TH-2 90 dpf and TH-3 110 dpf; Tio-1 (until complete cessation of development of the pigment pattern) 310 dpf, Tio-2 235 dpf and Tio-3 225 dpf; for C. spilurus groups: TH-1 62 dpf, TH-2 80 dpf, TH-3 95 dpf; Tio-1 (until complete cessation of growth and development of the pigment pattern) 250 dpf, Tio-2 340 dpf and Tio-3 335 dpf.

The concentrations of hormone and goitrogen used in this study were previously selected experimentally (Prazdnikov and Shkil, 2019a). Water (half the volume) from each aquarium was replaced daily. The necessary amounts of T₃ and goitrogen were administrated into the water up to the predetermined concentrations. Other conditions were the same as for the common-garden experiment. In each group, the total T₃ content was measured a week after the start of treatment. To compare the experimental phenotypes with the phenotypes of various cichlid fishes, we followed Říčan et al. (2005, 2016), a professional guide and atlas of cichlids (Baensch and Riehl, 1997; Boruchowitz, 2006) and the database FishBase (www.fishbase.org).

We have previously presented a detailed description of the pigment patterning of A. nigrofasciata (Prazdnikov and Shkil, 2019a). Here, we briefly characterize the main stages of coloration development in the control group with regards to the T₃ developmental dynamics. The first pigment cells, ELMs, emerge at the embryonic stage (Fig. 1). In the early larva stage (4–11 dpf), two lateral ELM stripes, sporadic ELMs on the lips, and the dispersed xanthophore and iridophore populations on the trunk appear (Fig. 1). This period is characterized by a rapid increase of T₃ content (Fig. 2).

Appearance of disperse assemblages of AMs on the trunk and on the base of unpaired fins, and a sharp decrease of T₃ concentration coincide with the beginning of the late larval stage (12–17 dpf) (Figs 1 and 2). Later (18–22 dpf), AMs emerge on the dorsal and anal fins, and form precursors of the adult vertical bars (pale dispersed AM assemblages) on the trunk (Fig. 1). This process is accompanied by the development of xanthophore populations on unpaired fins and iridophore populations on the operculum (Fig. 1).

The juvenile stage is characterized by the formation of adult AM vertical bars on the trunk and fins (38–40 dpf), the appearance of erythrophore populations on the unpaired fins and the disappearance of the ELM stripes (80–95 dpf) (Fig. 1). During the juvenile stage, the hormonal content fluctuates, but remains relatively low (Fig. 2).

A complete adult melanistic pattern develops by 100 dpf. It consists of eight postcranial vertical bars and the opercular spot (Fig. 1). The number of melanistic elements is constant (Fig. 3A). The sexual dimorphism associated with the increase of the xanthophore population in females appears at 135–146 dpf. T₃ content does not significantly change at this point (Fig. 2).

At the embryonic and early larval stages, C. spilurus displays pigment patterning similar to that of A. nigrofasciata (Fig. 4). T₃ content shows a steady elevation from the end of the early larval stage (Fig. 2).

Onset of the late larva stage in C. spilurus is accompanied by the appearance of AMs on the trunk between the lateral stripes and on the base of the unpaired fins. In the ventral part of the body, AMs form thin vertical lines extending from the first larval stripe to the base of the anal fin along the myomeres (Fig. 4). Then (15–18 dpf), populations of AMs, xanthophores and erythrophores emerge on the dorsal and anal fins, and iridophores appear on the operculum. The end of the stage (20–26 dpf) is characterized by the appearance of prospective AM vertical bars on the trunk. In both series, T₃ level significantly increases (Fig. 2).

Transition to the juvenile stage is associated with a sharp decline of T₃ content and the segregation of AM populations into the numerous vertical bars on the body (Figs 2 and 4). At 32–45 dpf, AM bar formation is complete and ELM elements disappear, while the density of xanthophore/erythrophore populations on the unpaired fins increases.

Development of the adult melanistic pattern, background coloration and pigment complex of the fins is completed by 80–90 dpf. The further development of coloration is the sexual dimorphism development. This stage is characterized by an increase of T₃ content (Fig. 2). In females (130–140 dpf), short bars or blotches appear on the dorsal fin, consisting of AMs and iridophores at the periphery. The number of vertical bars in adult fish varies from 7 to 10 (Fig. 3B) with the most common ‘eight postcranial bars’ phenotype. The background coloration is dominated by iridophores, which, in combination with other pigment cells, give shades from blue to gray. Subdominant individuals and/or individuals during spawning may develop a reticulate pattern on the trunk that masks the boundaries of the vertical bars (Fig. 9A).

In all groups treated with hormone, the total T₃ content was 4–5 times higher than in the control group. In the groups treated with goitrogen, total T₃ content was approximately 2 times lower than in the control group (Table 1). In comparison with the control groups, all experimental groups of A. nigrofasciata displayed significant changes in pigment patterning (Fig. 5B–G). Hyperthyroid groups displayed an accelerated development of coloration, whereas hypothyroid groups were characterized by pigment patterning retardation (Fig. 6).

The experimental treatments had pronounced consequences for the adult pigment pattern. In the TH-1 group, the trunk bars 3^a and 6 were absent (Fig. 7B). In the TH-2 group, bars 3^a, 5 and 6 were absent, while remaining bars were incomplete, appearing as blotches on the base of the dorsal fin and central part of the trunk. In many fish, these blotches coalesced and formed disrupted stripes and bars (Fig. 7C). In the TH-3 group, the number of bars varied from four to eight. Melanistic crossbars connected vertical bars in the central part of the trunk (usually dorsal and ventral parts of bars 3^p and 2) (Fig. 7D). Several individuals possessed a blotch at the site of bar 3^p. In all hyperthyroid groups, sexual dichromatism was more pronounced than in the control group (Fig. 7B–D).

In the Tio-1 group, the ELM elements disappeared only partially. The posterior part of the trunk lateral stripe, as well as ELMs on the lips, remained at the adult stage (Fig. 7E). In the Tio-2 and Tio-3 groups, the degradation of ELM elements proceeded somewhat more rapidly than in Tio-1 (Fig. 6), but their remnants were present until the end of the juvenile period, as well as in some adults (Fig. 7F,G). In the hypothyroid adults, the number and variability of melanistic elements increased in comparison with the control (Figs 3A and 7E–G). In the Tio-1 group, the melanistic pattern usually included 9–11 vertical bars and remnants of the ELM stripe in the caudal part of the trunk. Sexual dichromatism did not develop (Fig. 7E). In the Tio-2 and Tio-3 groups, the number of vertical bars varied from 8 to 12. The splitting bars sometimes formed a reticulate ornamentation in the ventral part of the trunk. Many fish possessed additional blotches and remnants of the ELM stripe. The development of sexual dichromatism in these groups was extremely retarded (Figs 6 and 7F,G).

The experimental groups displayed significant differences in the number of vertical bars (P<0.001; Fig. 3A), as well as the area of melanistic elements and interbar space (P<0.001; Fig. 8A,B). The number of bars was positively correlated with the area occupied by melanistic elements (Spearman r_s=0.527, P<0.001; Fig. 8C), but negatively with the area of the interbar space (Spearman r_s=−0.819, P<0.001; Fig. 8D).

Similar to A. nigrofasciata, all experimental groups of C. spilurus displayed changes in hormonal content, and timing and rate of coloration development (Table 1 and Fig. 6). In the hyperthyroid groups (TH-1 to 3), the rate of adult pigment pattern formation was accelerated to varying degrees (Figs 5I–K and 6). These developmental changes resulted in variability of the adult coloration. Thus, the TH-1 group lacked bars 3^p and 4, and possessed five to six bars (Fig. 9B). The number of vertical bars in the TH-2 group ranged from two to six. Bars 3 and 5 were always present, whereas other bars were absent or fused (Fig. 9C). In the TH-3 group, the number of bars varied from five to seven; the majority of individuals lacked bars 4 and 7 (Fig. 9D). The hyperthyroid groups developed a normal sexual dichromatism. Several TH-1 females demonstrated deviation in sexual coloration as a result of an absence of melanophore–iridophore blotches on the underdeveloped dorsal fin rays (Fig. 9B).

Hypothyroid groups (Tio-1 to 3) were characterized by drastic changes in the timing of pigment patterning (Fig. 6). In Tio-1, the development of coloration stopped at the onset of the larva–juvenile transition (Fig. 5L). Fish belonging to Tio-2 and 3 demonstrated severe pigment patterning retardation, resulting in the appearance of the alternative coloration phenotypes (Fig. 5M,N). In both groups, the number of melanistic elements increased, and background coloration was gray–yellow shades as a result of the iridophore deficiency (Fig. 9E,F). In the Tio-2 and Tio-3 groups, the melanistic pattern included 10–12 and 9–11 vertical bars, respectively, as well as remnants of ELM stripes in the caudal part of the trunk. Sexual dichromatism was present, but was postponed and weakly pronounced (Fig. 6).

Experimental groups demonstrated significant differences in the number of vertical bars (P<0.001; Fig. 3B), as well as the area of melanistic elements and interbar space (P<0.001; Fig. 8E,F). The number of bars was positively correlated with the area occupied by melanistic elements (Spearman r_s=0.411, P<0.001; Fig. 8G), but negatively with the area of interbar space (Spearman r_s=−0.528, P<0.001; Fig. 8H).

Ignoring some insignificant species-specific peculiarities, both cichlids, A. nigrofasciata and C. spilurus, have a common I-type pigment patterning sensuŘíčan et al. (2016). Their coloration development is a complex process with six distinct stages: embryonic, early larval, late larval, transition, juvenile and adult. Each stage is characterized by a specific composition of pigment cell lineages and pattern elements (Figs 1 and 4). The embryonic stage is the period of appearance of ELMs on the yolk sac, head and trunk. The early larval stage is characterized by the development of lateral stripes composed of ELMs and the appearance of dispersed xanthophores and iridophores. The emergence of AMs indicates the beginning of the late larval stage. The transition from larval to juvenile coloration is a rapid and concerted process during which the replacement of ELMs by AMs occurs. The juvenile stage is characterized by the development of the melanistic elements (vertical bars and blotches) composed of AMs, propagation of the xanthophore and iridophore populations, and the appearance and increase of the erythrophore population. The development of sexual dichromatism indicates the transition to the adult state.

At the same time, hormonal developmental dynamics differ between species. Amatitlania nigrofasciata has a pronounced but short surge of T₃ content at the early larval stage (Fig. 2). In C. spilurus, the T₃ developmental profile is more complex and displays two peaks: at the late larval stage and at the end of the juvenile stage. In the context of pigment patterning, in A. nigrofasciata, a surge of hormonal content precedes the appearance of AMs, and xanthophore and iridophore populations. In C. spilurus, the first hormonal surge occurs simultaneously with the appearance of AMs, and xanthophore and iridophore populations, and continues up to the beginning of the juvenile melanistic pattern development. The second rise of T₃ content coincides with the development of sexual dichromatism. Differences in the developmental dynamics of T₃ content are confirmed by the Kendall rank correlation coefficient (τ=−0.03; P≤0.01).

However, in both species, experimental exposure to hormone and goitrogen at different stages resulted in similar changes in the timing and rate of coloration development. Hyperthyroidism induced the precocious disappearance of ELMs and accelerated transition to the adult pigment pattern. Conversely, hypothyroidism led to the prolonged presence of ELMs and a drastic retardation in the development of adult pigmentation. In some cases, hypothyroidism caused an interruption in the transition to the adult coloration. Thus, the complex irreversible changes in composition and localization of pigment cell populations, which normally occur during the larva–juvenile transition, depend on THs.

A similar role of THs in pigment patterning has been revealed in different teleosts, including zebrafish (McMenamin et al., 2014; Saunders et al., 2019), Endler's livebearer (Prazdnikov, 2021) and clownfish (Salis et al., 2021). Based on the data obtained in the experiments with various fish models, one can conclude that THs have a pleiotropic influence on pigment cell biology and ecology. In particular, THs drive AMs into a terminally differentiated state and stop their proliferation, and also intensify the proliferation of xanthophores and iridophores and their accumulation of pigments, which makes the pattern elements composed by them more pronounced. In addition, THs can regulate AM migration and affect the formation of melanophore and xanthophore pattern elements indirectly through iridophores because of changes in intercellular interaction (Saunders et al., 2019; Prazdnikov, 2022). At the same time, pattern formation may be associated with additional autonomous interactions between melanophores and xanthophores, the behavior of which is described by Turing's model with modifications (Kondo et al., 2021).

Multiple aspects of skin development, a pigment cell environment, are under the control of THs (Campinho, 2019; Aman et al., 2021). Taking into account that pigment patterning is a complex cascade of inductive events between different skin components (including pigment cells), each with its own response on THs, hormonal alterations are likely to desynchronize the timing of inductive signals and competence, and, as a result, change the duration or disrupt interactions between inducing and responding elements. These developmental alterations are likely to affect adult coloration. However, the influence of THs on the activity and distribution of various types of pigment cells, as well as the relationship with other hormonal systems and the possible involvement of Turing's mechanism in pattern formation in cichlid fishes, remains to be explored in the course of further research.

In A. nigrofasciata and C. spilurus, manipulations of hormonal status had various morphological consequences, such as a change in the number, shape and severity of pigment pattern elements, and can be considered an example of experimental heterochrony (Dobreva et al., 2022). All hyperthyroid groups possessed a reduced number of melanistic elements and overdeveloped sexual dichromatism. The highest variability in the melanistic pattern was inherent in groups treated with the hormone from the late larva stage. In the group treated with the hormone from the early larval stage, the melanistic pattern variability was much lower. The hypothyroid groups displayed an increased number of melanistic elements, but weakly pronounced or no sexual coloration. Similar to hyperthyroid fish, groups treated with goitrogen at the late larval stage demonstrated the highest variability, whereas in groups treated from the early larva stage, the melanistic pattern variability was lower. These findings indicate that TH responsiveness is stage specific, and probably determined by numerous factors, the study of which requires additional experiments. A similar relationship was observed in the two species of experimental cichlids between the change in the number of bars and the area of melanistic elements, as well as the interbar space.

There were some species-specific differences in the reaction to experimental treatment. In A. nigrofasciata, the shape of the melanistic elements changed, while in C. spilurus, it did not. Moreover, in A. nigrofasciata, sexual dichromatism was present in varying degrees in all experimental groups, except for the Tio-1 group. Cryptoheros spilurus treated with goitrogen from the early larval stage did not develop adult coloration at all. Despite changes in coloration, adult fish in all experimental groups (except Tio-1) demonstrated reproductive behavior and breed.

It is noteworthy that many experimental fish possess phenotypes resembling the phenotypes of the phylogenetically close and distant Neotropical cichlids. For example, hyperthyroid fish possess features of cichlids belonging to the genera Mesoheros, Tomocichla, Paraneetroplus, Rocio and Theraps. The hypothyroid groups are characterized by morphological traits inherent in the genera Caquetaia, Mayaheros, Tomocichla and Parachromis. It should also be noted that, in addition to changes in pigment pattern, experimental fish displayed TH-induced shifts in various adaptive and taxonomically significant characters, such as the morphology of the neurocranium and jaws, body shape, pharyngeal and jaw dentition, squamation, paired and unpaired fin skeleton, etc. These morphological traits and their development under hormonal influence are beyond the scope of the present study, but seem to be a promising model for future research.

Summarizing, we conclude that two species of Neotropical cichlids display species-specific differences in T₃ developmental dynamics but a similar type of pigment patterning, as well as a similar pronounced response of pigment cell lineages to experimental hyperthyroidism and hypothyroidism. These findings somewhat contradict each other and the presumed leading role of thyroid axis activity in the determination of the type of coloration ontogeny in cichlids. However, they indicate that an increase of TH level is a necessary but not sufficient signal inducing the transition from larval to juvenile ornamentation, and is a component of a complex, concerted endocrine cascade that drives pigment patterning as a part of skin development.

We are grateful to the reviewers for many valuable comments and suggestions that helped improve the quality of the manuscript. The study was conducted within the framework of the Severtsov Institute of Ecology and Evolution, Russian Academy of Sciences and Koltzov Institute of Developmental Biology, Russian Academy of Sciences research programs 0089-2021-0003 and 0088-2021-0019, respectively.