Ambient Developmental Temperatures and Maternal Thyroid Supply Promote Early Life Diversification in Salmonids

ABSTRACT

The interplay between ambient temperature and thyroid signaling controls the morphogenesis of poikilotherms in the early stages of life. These two metabolic conditioning factors are crucial for the divergence of fish morphs undergoing adaptive diversification. Using Salvelinus malma, a model salmonid species with high phenotypic plasticity and a capacity for adaptive radiation, we conducted a series of experiments by rearing sympatric large predatory, mid‐sized generalized and dwarf insectivorous morphs from both lacustrine and riverine environments. First, we compared the developmental rates and thyroid hormone profiles of these morphs reared under morph‐specific and standard temperature regimes. Next, we reared the offspring of the generalized morph at temperatures imitating those at the spawning site of the morphs comprising the sympatric triplets. Finally, we reared the generalized morph under altered thyroid status at a standard temperature. Our findings revealed that morphs adapted to different developmental temperatures exhibit specific profiles of maternal thyroid hormone consumption, as well as discrepancies in the rate of life stage transition and cranial ossification. Changes in developmental temperature induce compensatory changes in the rate of thyroid hormone supply and heterochronic shifts in life stage transition and skull ossification. Lacustrine and riverine morphs demonstrate a similar developmental bias: fish with an elevated rate of thyroid consumption, developing in cold conditions follow a fast‐developing predatory morphogenesis pathway, whereas those with a reduced maternal thyroid endowment and forced to develop in warm water become dwarfs.

Lacustrine and riverine morphs of salmonid fish with similar feeding strategies demonstrate similar thermal preferences, parallel physiological adaptations and developmental specializations. The interplay between maternal thyroid supply and developmental temperature facilitates the adaptive radiation of the species.

1. Introduction

Teleosts provide numerous examples of adaptive radiation, in which uneven habitat utilization is one of the main drivers of lineage diversification (Schluter 2000; Todd Streelman and Danley 2003). Colonizing new habitats requires adjusting physiology and development to environmental cues that differ from those to which the ancestral lineage was adapted. Such tuning is achieved through a multivariant organismal response, with endocrine signaling being the key agent in regulating most physiological functions and development throughout life (Hulbert 2000; Lema 2020; Dufty et al. 2002; Blanton and Specker 2007; Salis et al. 2021). The interplay between environmental variables and endocrine axes ensures phenotypic plasticity, a source of selection, and guides ecomorphological specialization (Dufty et al. 2002; West‐Eberhard 2003; Lema 2020; Nijhout 2025). Despite the recurring pattern of diversification along the habitat axis in various teleosts, studying the interplay of which extrinsic cues and intrinsic regulators orchestrates each case of adaptive radiation remains an interesting challenge.

We studied the East Kamchatka charr Salvelinus malma, a cold–water species sensitive to thermal variation and known for its amazing phenotypic and ecological diversity. It forms numerous sympatric assemblages consisting of triplets of basic morphotypes: a benthivorous “normal,” an insectivorous dwarf with limited migratory abilities, and a large‐sized migratory piscivorous (Esin 2025). The main mechanism underlying morphological and ecological differentiation in the sympatric S. malma morphotypes was considered to be heterochronies (Esin et al. 2018, 2021a; 2021b). Given that morphotypes occupy spawning grounds with distinct temperature regimes (Markevich and Esin 2019; Markevich et al. 2021), and pronouncedly differ in thyroid hormone levels (Esin et al. 2021a; Woronowicz et al. 2024), the interplay between these factors seems an encouraging study object.

Most fishes are ectothermic, so their vital biological processes depend on ambient temperature (Beitinger and McCauley 1990; López‐Olmeda and Sánchez‐Vázquez 2011; Ozernyuk 2011; Kahilainen et al. 2014; Skoglund et al. 2015). This dependence prompts fishes actively choose environmental temperatures at which they can function effectively, that is, based on their temperature preference (Bjornn and Reiser 1991; Haesemeyer 2020). This preference is a species‐specific trait acquired as a result of adaptive adjustment of metabolic, physiological, and developmental processes, as well as behavior and life cycle, to the temperature of habitats (Ohlberger et al. 2013; Gilbert and Miles 2017; Andreassen et al. 2025). Occupying a new temperature niche requires a remodeling of thermal physiology, especially at the early developmental stages, when fishes are unable to compensate for the disadvantageous effects of non‐optimal temperature through their behavior. Such temperature‐induced tuning affects early developmental processes, often leads to phenotypic consequences, and is regarded as one of the main factors driving the diversification and adaptive specialization of fish in heterogeneous environments (Ohlberger et al. 2013; Esin et al. 2021b; Markevich et al. 2022).

Among the primary intrinsic factors ensuring thermoregulation and energy homeostasis, as well as the temperature‐induced developmental and phenotypic changes in fishes, is the activity of the hypothalamic‐pituitary‐thyroid axis (Blanton and Specker 2007; Laudet 2011; Little et al. 2013; Iwen et al. 2018; Lema 2020; Zwahlen et al. 2024). This neuroendocrine axis provides a genomic response to environmental changes via translating ambient signals into humoral factors, thyroid hormones (Lema 2020). Experiments on model zebrafish have highlighted the maternal thyroid hormones' impact on the developmental trajectory of offspring via the heterochronic activity of numerous target genes (da Silva Rodrigues et al. 2025). The role of thyroid hormones in modulating genome expression suggests that the thyroid axis mediates transgenerational epigenetic inheritance (Seebacher and Little 2024).

Under these circumstances, if thyroid‐induced developmental and phenotypic changes in response to temperature changes are adaptive, they can persist for generations, driving the diversification and ecological specialization of fishes (Schluter 2000; Todd Streelman and Danley 2003; Gilbert et al. 2015; Esin et al. 2018; Nijhout 2025). Intra‐ and interspecific morphological divergence controlled by the activity of the thyroid axis during early ontogeny in numerous fish taxa (Harada et al. 2003; Кitano et al. 2010; McMenamin et al. 2014; Lorgen et al. 2015; Shkil et al. 2015; Bolotovskiy et al. 2018; Borisov and Shkil 2025) to some extent supports this assumption.

We hypothesize that the dynamics of temperature and the level of thyroid hormone in the earliest life stages are involved in the adaptive radiation of S. malma. In this context, there are three possible scenarios. Temperature affects early development and thyroid hormone dynamics of diversifying morphs; but the latter per se has a negligible effect on ontogeny (i). Alternatively, specific developmental vector is determined exclusively by thyroid hormone dynamics, which is morph‐specific and does not depend on the influence of temperature (ii). Finally, the integrated scenario implies ecological function of thyroid hormones (sensu Zwahlen et al. 2024): temperature affects the thyroid hormone dynamics, which in turn, affects the rate and timing of developmental processes, and thereby adjusts ontogeny to the current thermal regime (iii).

To test these scenarios, we first localized the spawning sites and examined the temperature regime of the early developmental stages for S. malma morphs: “normal” benthivorous; dwarf insectivorous; and large‐sized piscivorous, which comprise sympatric assemblages in the Kamchatka River and Lake Kronotskoe. Then, we obtained and delivered fertilized eggs of morphs to the laboratory and reared them at temperatures simulating natural ones, as well as at standard temperature. To evaluate the effects of temperature and thyroid hormone dynamics on the development, both separately and in interaction, we performed a series of experiments with the most biologically and ecologically generalized riverine morph. The early developmental divergence of the morphs was assessed via the rates of life stage transition and cranial ossification, alongside an evaluation of thyroid gland activity and temperature control in the experimental series.

2. Materials and Methods

2.1. The Ecosystems Under Study

The Kamchatka River is a large meandering river with numerous tributaries (basin area = 55,900 km²; 7700 tributaries). The “Dolly Varden” charr (DV) inhabits the entire river system. It feeds on various benthic organisms and prefers riffle habitats during the summer months. In September, DV migrates to spawn in the upper reaches of clean water streams (Tiller 2007). DV should be regarded as the most ecologically generalized morph in the study. The large piscivorous marble‐colored “stone” charr (SC) is an ambush hunter that spawns in streams with a plane riverbed and dark water flowing through coniferous thickets (Melnik et al. 2020). The small “goldish” charr (GC) inhabits spring brooks, specializes in feeding on insects, and spawns in the vicinity of large groundwater discharges. Spawning grounds of the three riverine morphs are spatially separated (Figure 1a,b).

Figure 1.

Map showing the spawning grounds (colored dots) in the Kamchatka River and Lake Kronotskoe basins (a); the studied morphs of S. malma from the Kamchatka River and Lake Kronotskoe basins (b and c, respectively); the muscle levels of T3 in adults of six morphs during the summer foraging period (Esin et al. 2021a; Woronowicz et al. 2024) (d, mean ± SD, min–max values and sample sizes are shown; *indicates a significant difference in SC and W values compared to GC and S values according to the van der Waerden test, p ≤ 0.005).

Lake Kronotskoe is a large (surface area = 246 km²) and deep (up to 136 m) dimictic waterbody, home to the most diverse species flock of salmonid (Esin et al. 2024). Only three most numerous local morphs were selected for this study. The medium‐sized “nosed” morph (N) spawns in the lower reaches of tributaries and consumes benthic organisms (mainly amphipods) on the lake littoral. The upper course of the main tributary serves as a spawning place for the fast‐growing, robust‐bodied “widehead” morph (W), which migrates to the lake as a juvenile to feed and mature. This morph switches to piscivory with the growth. The small gracile‐bodied “smallmouth” morph (S) spawns on the lake slope and feeds on insects in the hypolimnion zone of the lake (Figure 1a,c).

The studied morphs of S. malma can be accurately distinguished by morphological characters. Their reproductive differentiation has been demonstrated through analysis of microsatellite loci variability, and hybrids are rare in both the Kamchatka River and Lake Kronotskoe ecosystems (Esin et al. 2020; Melnik et al. 2020). The riverine and lacustrine assemblages are similar in the pattern of muscle T3 differentiation across morphs (Esin et al. 2021a; Woronowicz et al. 2024): adults of slow‐growing insectivorous morphs are characterized by a lower mean T3 level, while fast‐growing piscivorous morphs—by a higher mean T3 level compared to the medium‐sized benthivorous morphs (Figure 1d).

2.2. Temperature Regime Evaluation

Long‐term study has shown that the morphs are characterized by a strict reproductive philopatry, returning to cast their eggs in the exactly river sections specified by the peculiar stream velocity, water chemistry, and temperature (Esin et al. 2021b; Markevich et al. 2021).

Temperature regime at the spawning sites of each morph was examined during the annual period 2015–2016. Starmon mini temperature loggers (Star Oddi) were installed in the surface layer of three separate spawning nests (redds) per spawning ground of each morph. Temperature was measured once per hour throughout the year (±0.05°C). Then, data from three loggers were averaged daily to obtain the temperature regime of the embryonic and early post‐embryonic stages for each morph.

The data collected by the loggers were smoothed with the characteristic time of the week without daily dynamics to suppress occasional temperature fluctuations. The spawning grounds of DV and N morphs were characterized by intense cooling by mid‐November, average winter temperatures of 1°C–2°C, and a gradual warming beginning in April. The spawning grounds of GC were defined by relatively cool summer and warm winter temperatures due to the influence of groundwater. The spawning grounds of S morph, located in the lake hypolimnion, were the warmest. The spawning grounds of SC and W morphs were cold in winter (< 1°C), with early warming at the SC site due to the dark color of the water, and a prolonged cold period at the W site due to its highland location (Figure 2).

Figure 2.

Daily‐averaged dynamics of the mean weekly temperatures at the bottom water level at the spawning grounds of Salvelinus malma morphs (labeled; the standard temperature regime of rearing in the experiment is 2.9°C ± 0.2°C). The black dots delimit the period of experimental development of the Dolly Varden morph at the corresponding temperatures from fertilization to the closure of the belly around the remnants of the yolk suck; the blue dots (with the letter E) show the time of eye pigmentation.

2.3. Embryo Collection

Spawners of each morph were caught at the peak of the mating period on the spawning grounds. Artificial crossing of each morph was realized by mixing the eggs collected from three medium‐sized females and sperm of 3–5 males of different size. Eggs were fertilized via the dry method (Ginsburg 1963) with a fertilization rate ~90% in all cases. In 48 h, fertilized eggs were delivered to the laboratory for experimental incubation and rearing.

2.4. Experimental Design

To evaluate the effects of temperature and thyroid hormone, both separately and in interaction, on the early development of S. malma morphs, four experiments were conducted.

2.4.1. Rearing of Different Morphs Under Specific Temperature Regimes

In the first experiment, we compared the dynamics of thyroid hormone levels and developmental rate (timing of developmental stage transition) in all morphs reared at temperatures simulating the natural morph‐specific temperature regimes. A temperature control unit (ERG‐01) with a relative sensitivity of 2 mK was placed in the room (3 × 6 m) with six water tanks, where the air temperature was maintained in the range of −3°C… + 2°C with an industrial air‐cooling system SM232 (Polair). The original software analyzed the dynamics of water tanks cooling and estimated the amount of heat required to maintain a given temperature (Esin et al. 2021b). The water was heated using a self‐made 120 W heating element. The temperature control system was duplicated for reliability. The rearing of each series continued for a month after the belly walls closure around the remains of the yolk suck.

2.4.2. Rearing of Different Morphs at a Standard Temperature

In the second experiment, we compared the dynamics of thyroid hormones and the developmental rate in four morphs (excluding SC and GC) reared at a standard temperature of ~2.9°C, which is considered optimal for the S. malma somatic growth and accelerated smoltification (Armstrong 1970; Elliott 1975). The standard temperature was maintained using external chillers (Hailea HC 500 A), with an accuracy of ±0.2°C. The experiment continued until the belly walls closure stage.

2.4.3. Rearing of the Generalized Morph Under Different Temperature Regimes

In the third experiment, we assessed the developmental rate and dynamics of thyroid hormones in the most generalized S. malma morph (DV), reared under temperature regimes simulating the natural ones characteristics of each morph (Figure 2; the segments of colored curves between black dots). The experiment continued until the belly walls closure stage.

2.4.4. Rearing of the Generalized Morph Under Altered Thyroid Status at a Standard Temperature

In the fourth experiment, we reared DV from the 30th day of development (the initiation of somitogenesis) until belly walls closure under two treatments. Chronic hypothyroidism was achieved by dissolution of thiourea (CAS 62‐56‐6, Solins Chem.; non‐specific goitrogen) in the aquarium water in 0.2 g L⁻¹ concentration (“Thio” series). Hyperthyroidism was induced by dissolving triiodothyronine (=T3; CAS 6893‐02‐3, Sigma; biologically active form of thyroid hormones) at a concentration of 1.0 μg L⁻¹ and adding iopanoic acid (CAS 96‐83‐3, Sigma; deiodinase inhibitor) at 0.003 μg per 1 g of fish (“T3” series). Administration of active agents was repeated once per week during water changes. The concentrations and frequency of agents' application were determined based on preliminary experiments on charrs (Esin et al. 2021a; Esin et al. 2018) and data on the dynamics of thyroid hormones in experimental fishes (Borisov and Shkil 2025). The series from clean water was used as a control (“Control”). The experimental charrs were kept at a standard temperature of ~2.9°C controlled with the external chillers.

2.4.5. Other Conditions of Incubation and Rearing

All other experimental conditions of rearing were identical. Each series of charrs was incubated and reared in an independent 250‐l tank. Fertilized eggs were incubated in UV‐treated soft water (150 ppm, pH = 7.8, oxygen ≥ 11 mg l⁻¹) with a filtration intensity of 900 L h⁻¹. The embryos were reared in darkness. To perform biological replication, the eggs were distributed across three trays. The newly hatched fish were transferred to mesh cages with a stocking density of 100 individuals per 0.004 m³ (25 000 _* m⁻³). Fish were resettled with a twice‐smaller density at the 15th week after hatching. At the start of feeding, the light regime was changed to 10 h day: 14 h night (Sun‐Glo lamps, Hagen). Fish were fed with standard weighted portions of the nauplii of brine shrimp Artemia salina during the first week and then with granulated food (Coppens Vital). The mortality was the same in all experimental series, reaching ~15% by the time of hatching and ~10% during the shift to external feeding.

2.5. Developmental Staging and Timing

Extended early ontogeny is characteristic of Salvelinus charrs (Balon 1985; Johnston 2002). To place experimental fish into discrete developmental stages, we identified several commonly used milestones: onset of somitogenesis; eye pigmentation; hatching; start of feeding; and the closure of belly walls over the remains of the yolk suck (belly closure). The completed milestone of 50% of individuals was used as a criterion for determining the timing of the transition to the next stage.

As varying temperatures cannot ensure proper comparison of the rate of charr development in the day*degree scale, we calculated D(τ) values of stage transitions using the equation:

$$D(\tau )={\int }_0^{\tau }A\cdot e^{\frac{-E_a}{R\cdot T\left(t\right)}}\mathit{dt}$$

where A = (8.57 ± 4.99) 10¹⁹ units and Ea = (1.16 ± 0.01) 10⁵kJ mol⁻¹ (for more details see Esin et al. 2021b). Specifically, the D(τ) value expresses the amount of developmental temperature required to overcome the developmental stage.

We did not use body size parameters to compare groups because the length and weight of early juveniles are strictly dependent on egg size, which differs between morphs and females.

2.6. Experimental Data Collection

To assess thyroid hormone content and the state of cranial bones ossification in the experimental series, we collected samples at each developmental milestone (Table 1). To determine the timing of endogenous thyroid hormone synthesis onset, we compared hormone levels and thyroid “gland” development in the morphs at different stages, including 1 month after 50% belly closure. In the fourth experiment, T3 was measured additionally after 50% belly closure followed by 3 days of starvation in clean water.

Table 1.

The number of samples used for the analysis of T3/T4 level (ossification progress) in four experiments.

Series	Fertilization	Somitogenesis	Eye pigmentation	Hatching	Start of feeding	Belly closure	Fry (+ one month)
1. Development of six Salvelinus malma morphs under temperature regimes that imitated the natural morph‐specific temperature dynamics
SC	1/4 (0)	4/0 (0)	4/0 (0)	12/0 (10)	10/3 (10)a	12/0 (10)	9/3 (10)a
DV	2/6 (0)	5/0 (0)	6/0 (0)	7/0 (10)	6/3 (10)	6/0 (10)	6/6(10)
GC	1/4 (0)	7/0 (0)	7/0 (0)	6/0 (10)	6/3 (10)a	6/0 (10)	6/5 (10)a
W	1/4 (0)	7/0 (0)	6/0 (0)	6/0 (10)	6/3 (10)a	6/0 (10)	6/4 (10)a
N	1/4 (0)	6/0 (0)	6/0 (0)	6/0 (10)	6/3 (10)	6/0 (10)	6/4 (10)
S	1/4 (0)	5/0 (0)	6/0 (0)	6/0 (10)	6/3 (10)a	6/0 (10)	6/1 (10)a
2. Development of four S. malma morphs at a standard temperature
DVst	0/0 (0)	4/0 (0)	7/0 (0)	6/0 (10)	6/0 (10)	6/0 (10)
Wst	0/0 (0)	5/0 (0)	5/0 (0)	5/0 (10)	5/0 (10)	6/0 (10)
Nst	0/0 (0)	4/0 (0)	4/0 (0)	4/0 (10)	4/0 (10)	5/0 (10)
Sst	0/0 (0)	4/0 (0)	6/0 (0)	5/0 (10)	5/0 (10)	5/0 (10)
3. Development of the Dolly Varden morph under temperature regimes that imitated the natural temperature dynamics of the six morphs
TSC	1/3 (0)	6/0 (0)	6/0 (0)	10/0 (10)	7/0 (10)	6/0 (10)
TDV	1/3 (0)	6/0 (0)	6/0 (0)	7/0 (10)	6/0 (10)	6/0 (10)
TGC	1/3 (0)	6/0 (0)	6/0 (0)	6/0 (10)	6/0 (10)	6/0 (10)
TW	1/3 (0)	6/0 (0)	6/0 (0)	6/0 (10)	6/0 (10)	6/0 (10)
TN	1/3 (0)	6/0 (0)	6/0 (0)	6/0 (10)	6/0 (10)	6/0 (10)
TS	1/0 (0)	6/0 (0)	6/0 (0)	6/0 (10)	6/0 (10)	6/0 (10)
4. Development of the Dolly Varden morph under altered thyroid status at a standard temperature, compared with the control series
T3			3/0 (0)	3/0 (12)	3/0 (12)	3/0 (12)
Thyo			0/0 (0)	3/0 (12)	3/0 (12)	3/0 (12)
Control			3/0 (0)	3/0 (12)	3/0 (12)	3/0 (12)

^a Three specimens per morph were used to prepare histological sections of the thyroid gland.

2.7. Analytical Protocols

2.7.1. Thyroid Hormone Content

Total T3 (triiodothyronine, bioactive form of thyroid hormone) and T4 (thyroxine, prohormone) concentrations were assessed in the extract of whole fish by enzyme‐linked immunosorbent assay (ELISA). To obtain the sample weight sufficient for analysis (> 0.025 g), we used 3–4 fish per sample. The extraction was done with a commonly used barbiturate‐extraction protocol, which has been designed for the assessment of thyroid hormones in whole‐body samples of small fish. This protocol effectively eliminates interfering matrix substances via the sequential centrifugation rounds, subsequent drying of the obtained supernatant, and final extraction of thyroid hormones with the phosphate buffer saline (Holzer et al. 2017). Hormone content was measured using commercially available ELISA kits (Monobind Inc.) with a sensitivity of 0.04 ng mL⁻¹ (tT3) and 3.2 pg mL⁻¹ (T4). Such a combination of the barbiturate‐extraction with the ELISA kit has proven to be highly robust and reliable for tracking relative changes and profiles of THs in whole‐body fish samples (Holzer et al. 2017; Esin et al. 2021a; Salis et al. 2021; Esin et al. 2023; Borisov and Shkil 2025; Woronowicz et al. 2024). However, given that we used ELISA kits designed for human serum, some of the T4 values were beyond the stated measurement range. Such samples were excluded from the analysis. Finally, hormonal concentration was recalculated to the sample weight.

2.7.2. Ossification Progress

Post‐embryonic fish were fixed in 4% formaldehyde solution in PBS and then stained with 0.5% alizarin red S (Sigma) (Table 1). Soft tissues were bleached and cleared, and then whole‐mount samples were deposited in glycerol following Pichugin (2015). We analyzed neurocranial (supraethmoid, frontal, parasphenoid, and orbital series), dermal tooth‐bearing (premaxilla, maxilla, dentary, vomer), and primary tongue‐bite apparatus bones (palatine, lingual bone, and first gill arch). The ossification of each bone was ranked as follows (Supporting Information S1: Figure 1): 0—no ossification, 1—germ or primordium of the bone, 2—bone scute or fused plate, 3—bone reaching a definitive shape (more than three elements in the gill arch and orbital series). To compare the degree and rate of skull ossification among series, we used the individual sum of ranks.

2.7.3. Histological Preparations

The specimens were fixed in 4% formaldehyde solution in PBS. Then, fish were placed in standard histological cassettes and decalcified following the protocol: washing in running water for 24 h; decalcification in a 5% HNO₃ for 24 h; acid neutralization and tissue stabilization in a 5% KAl(SO₄)₂ for 24 h; washing in running water for 24 h. Decalcified samples were dehydrated in an ethanol series, cleared in the HistoChoice Clearing Agent (Sigma, H2779‐1L, #MKCF1386), and mounted using the Vitrogel synthetic mounting media (BioVitrum, 12‐005). Cross‐sections of the pharyngo‐branchial area (5–7 μm) were performed with a hand sliding microtome (Meditome M530). Ehrlich's hematoxylin and eosin counterstaining technique was used. The presence of active thyroid follicles with rectangular‐shaped thyrocytes and a colloid‐filled lumen (Eales and Brown 1993; Peter 2011) was checked using a light microscope (Keyence Biorevo BZ‐9000). At least 100 sections per fish were analyzed. The imaging of sections was performed with a microscope (Nexcope NCF950).

2.8. Data Processing

The data obtained on ossification (sum of ranks) and thyroid level (hormone concentration) were statistically tested. As a normality check was not always positive, the non‐parametric Kruskal−Wallis H‐test, followed by the van der Waerden post‐hoc test for multiple comparisons of small samples, was used (Conover 1999). The calculations were performed using the PMCMR module in R (Pohlert 2014).

To quantify the effect of the T3 content (factor #1) and temperature (#2) on the D(τ) (=variable), we evaluated a factorial MANOVA in SPSS v.22 (IBM Corp.; www.ibm.com/support/pages/downloading-ibm-spss-statistics-22). Pillai's trace was used as the test of significance. The partial effects of mean T3 at the time of 50% transition to the next stage and mean temperature during the previous developmental stage were tested in the model: “factor #1 + factor#2 + #1 × #2 + error.”

3. Results

3.1. Development of Different S. malma Morphs Under Different Temperature Regimes

Morphs had different initial T4 contents. It was significantly lower in the small‐sized insectivorous morphs GC and S (= dwarfs) than in the large‐sized piscivorous morphs SC and W (= predators) (Supporting Information S1: Table 1). The medium‐sized benthivorous morphs DV and N (= “normal” morphs) exhibited the intermediate values. By the onset of external feeding, the T4 level in all morphs was below the stated measurement range of the ELISA kit, but at the belly closure stage, it was detected in SC and W (Figure 3a).

Figure 3.

Developmental profiles of whole‐body thyroid hormones in different Salvelinus malma morphs (SC—Stone charr, DV—Dolly Varden, GC—Goldfish charr; W—Widehead, N—nosed; S—Smallmouth morphs) reared under temperature regimes that imitated the morph‐specific temperature dynamics on the spawning grounds (see Figure 2). The mean ± 95% CI and min–max values of T4 (a), and the mean values of T3 (b) are shown for the six morphs at the moments of 50% stage transition; * and ** indicate the stages at which significant differences were observed, as referenced in Supporting Information S1: Tables 1 and 2, respectively. The mean ± SE and min‐max values of T3 are shown for the riverine (c) and lacustrine (d) morphs from the stage of somitogenesis. D(τ)—amount of developmental temperature. Table 1 shows the number of specimens used.

T3 developmental profiles (Figure 3b) were similar in different morphs, but the T3 content varied (Figure 3c,d, Supporting Information S1: Table 2). All morphs displayed a peak in T3 at hatching, followed by a decline during post‐embryonic development. Both dwarfs had the lowest T3 levels throughout development, whereas predators were characterized by the highest T3 content. Statistical tests revealed a significant excess of T3 levels in predators over dwarfs at most developmental stages (Supporting Information S1: Table 2).

At the beginning of external feeding, all morphs possessed single active follicles located along the dorsal elevation of the ventral aorta in the branchial region (Figure 4a,b). Measuring 15–35 (mean 25) μm in diameter, they were filled with eosin‐stained colloid, and the surrounding follicular cells (thyrocytes) had large central nuclei stained with haematoxylin. One month after the belly closure, we found minor changes in the state of the “gland” in S and GC, with only 1–3 enlarged (30–50 μm) follicles detected on separate sections. At the same stage, W and especially SC demonstrated an advanced state of thyroid “gland”. They had typical of salmonids diffuse disseminated islets of follicles (up to 5–6 per section) (Figure 4c,d) (Chanet and Meunier 2014), and demonstrated an increase of thyroid hormone content (Figure 3b).

Figure 4.

Histological structure of the branchial region of Salvelinus malma morphs. Single active thyroid follicles in juveniles of the predatory and dwarf morphs switching to external feeding (a, b), diffuse islets of enlarged follicles in predatory morphs at the end of the experiment, at the belly closure stage (c, d). More details in the text. CA, cartilage; CT, connective tissue; TF, thyroid follicles; VE, vessels. Scale bar is 50 μm. Serial 5 μm sections were prepared and stained with hematoxylin and eosin. Images were digitally acquired using a ×20 objective.

Since the somitogenesis, the morphs differed in the developmental rate, and these differences progressively increased throughout the development (Table 2). In terms of D(τ), the dwarfs showed stunted, but predators showed accelerated development compared to the “normal” morphs. As a result, by the belly closure stage, the differences between the predatory W and dwarf GC reached 0.30 D(τ), which is 17% of the amount of developmental temperature required for the “normal” morphs (Table 2).

Table 2.

The values of D(τ) at the moments of 50% stage transition for different experimental series of Salvelinus malma morphs.

Series	Somito‐ genesis	Eye pigmen‐tation	Hatching	Start of feeding	Belly closure
1. Development of six S. malma morphs under specific temperature regimes
GC	0.26–0.28 (0.27)	0.44–0.48 (0.46)	1.08–1.13 (1.10)	1.49–1.54 (1.51)	1.87–1.92 (1.90)
S	0.25–0.27 (0.26)	0.43–0.47 (0.44)	1.05–1.10 (1.08)	1.45–1.50 (1.47)	1.82–1.87 (1.85)
N	0.24–0.25 (0.25)	0.40–0.44 (0.42)	0.99–1.05 (1.02)	1.42–1.48 (1.45)	1.73–1.79 (1.76)
DV	0.24–0.26 (0.25)	0.40–0.44 (0.41)	0.99–1.04 (1.01)	1.39–1.44 (1.41)	1.70–1.75 (1.72)
SC	0.23–0.26 (0.24)	0.37–0.40 (0.39)	0.97–1.02 (0.99)	1.34–1.39 (1.36)	1.64–1.69 (1.67)
W	0.23–0.26 (0.24)	0.35–0.39 (0.37)	0.93–0.98 (0.95)	1.27–1.34 (1.31)	1.58–1.63 (1.60)
2. Development of four S. malma morphs at a standard temperature
Sst	0.25–0.27 (0.26)	0.43–0.45 (0.44)	1.03–1.10 (1.06)	1.42–1.51 (1.45)	1.80–1.88 (1.83)
Nst	0.24–0.26 (0.25)	0.40–0.46 (0.43)	0.99–1.09 (1.03)	1.41–1.49 (1.46)	1.73–1.83 (1.78)
DVst	0.24–0.27 (0.25)	0.40–0.44 (0.42)	1.00–1.05 (1.03)	1.39–1.44 (1.42)	1.72–1.76 (1.74)
Wst	0.23–0.26 (0.25)	0.36–0.40 (0.38)	0.95–0.99 (0.97)	1.31–1.38 (1.34)	1.60–1.65 (1.63)
3. Development of the Dolly Varden morph at six different temperature regimes
TGC	0.24–0.26 (0.25)	0.39–0.43 (0.41)	0.99–1.03 (1.01)	1.38–1.42 (1.40)	1.69–1.73 (1.71)
TS	0.24–0.26 (0.25)	0.39–0.43 (0.41)	0.99–1.03 (1.02)	1.38–1.43 (1.41)	1.70–1.74 (1.72)
TN	0.24–0.26 (0.25)	0.39–0.43 (0.41)	0.99–1.03 (1.01)	1.38–1.42 (1.40)	1.69–1.73 (1.71)
TDV (own)	0.24–0.27 (0.25)	0.39–0.43 (0.41)	0.99–1.03 (1.01)	1.38–1.42 (1.40)	1.69–1.73 (1.71)
TSC	0.24–0.26 (0.25)	0.39–0.43 (0.41)	0.98–1.01 (0.99)	1.36–1.40 (1.38)	1.67–1.72 (1.69)
TW	0.24–0.26 (0.25)	0.39–0.42 (0.40)	0.98–1.01 (0.99)	1.36–1.40 (1.38)	1.66–1.69 (1.67)
4. Development of the Dolly Varden morph under altered thyroid status at a standard temperature
Thyo	—	0.40–0.43 (0.42)	1.02–1.07 (1.03)	1.41–1.50 (1.43)	1.80–1.90 (1.83)
Control	—	0.40–0.44 (0.42)	1.00–1.05 (1.02)	1.40–1.44 (1.42)	1.71–1.75 (1.73)
T3	—	0.40–0.44 (0.42)	0.97–1.03 (0.99)	1.36–1.44 (1.39)	1.68–1.74 (1.69)

Note: Min–Max (Mean) values are shown.

Analysis of skull bones revealed the same trend: dwarfs have delayed development, while predators are ahead of the “normal” morphs (Figure 5a,b). By the time of foraging onset, SC had significantly outperformed GC in ossification progress; by the stage of belly closure, both predators had significantly outperformed both dwarfs; 1 month later, SC also significantly outperformed N in ossification (Supporting Information S1: Table 3). The differences in ossification progress were more pronounced in the riverine morphs, with SC being the most ossified and GC the least ossified at the end of the experiment (Figure 5c). The most significant disparities in ossification rates between predators and dwarfs were observed in the bones of the primary tongue‐bite apparatus.

Figure 5.

The sum of the ossification ranks of nine cranial elements in different Salvelinus malma morphs reared under temperature regimes that imitated the morph‐specific temperature dynamics on the spawning grounds (see Figure 2). The mean ± SE and min‐max values are shown for the riverine (a) and lacustrine (b) morphs. D(τ)—amount of developmental temperature. The mean values at the moments of 50% stage transition are shown for the six morphs simultaneously (c,* indicates the stages at which significant differences were observed, as referred in Supporting Information S1: Tables 3). The experimental series are labeled as in Figure 1; Table 1 shows the number of specimens used.

3.2. Development of Different S. Malma Morphs at a Standard Temperature

At a standard temperature, the morphs exhibited the same T3 profile as under a natural temperature regime, peaking at hatching and declining at the belly closure stage (Figure 6). However, the T3 concentration values changed (Supporting Information S1: Figure 2). At standard temperature, the T3 content in the DV and N morphs at the eye pigmentation and hatching stages, and in the predatory W during embryogenesis, was lower than under the natural regime. In contrast, the T3 level in the dwarf S morph developed at a standard temperature was slightly increased than under the natural regime. As a result, differences in T3 dynamics among morphs were less pronounced at a standard temperature. Only the predatory W differed significantly from the other morphs by a higher hormone level during embryogenesis and hatching (Supporting Information S1: Table 4). The correlation (R²) between the mean T3 level and D(τ) for stage transition in the combined dataset from the two experiments was 0.830.

Figure 6.

Whole‐body T₃ developmental profiles in different Salvelinus malma morphs reared at a standard temperature. The mean values of T3 are shown for the morphs in the experiment at the moments of stage transition (a, * indicates the stages at which significant differences were observed, as referenced in Supporting Information S1: Table 4). The mean ± SE and min‐max values of T3 are shown for lacustrine morphs (b). D(τ)—amount of developmental temperature. The morphs are labeled as in Figure 1; Table 1 shows the number of specimens used.

At standard temperature, the predator W demonstrated accelerated development, while the dwarf S lagged behind the “normal” morphs: DW and N (Table 2). Meanwhile, at standard temperature, predator and “normal” morphs, which experienced a higher winter temperature, exhibited retarded development than at the natural temperature regime. Conversely, dwarf, which experienced a lower temperature at standard conditions, exhibited an accelerated development compared to dwarf developed at natural temperature. The same trend was noted for the skull bone ossification. However, by the end of the experiment, the W_st series significantly exceeded the N _st and S _st series in the degree of ossification (Figure 7a and Supporting Information S1: Table 5).

Figure 7.

The sum of the ossification ranks of nine cranial elements in the Salvelinus malma morphs from different experiments. The mean ± SE and min‐max values are shown for the morphs reared at a standard temperature (a, *** indicates the stage at which significant differences were observed, as referenced in Supporting Information S1: Tables 5); the morphs are labeled as in Figure 1. The mean ± SE and min‐max values are shown for the Dolly Varden (DV) morph reared at different temperature regimes (b, ** indicates significant differences referenced in Supporting Information S1: Tables 7); t_SC, t_DV, t_GC, t_W, t_N, t_S—series reared at Stone charr, Dolly Varden, Goldfish charr, Widehead, Nosed, and Smallmouth morphs temperature regimes. The mean ± SE and min‐max values are shown for the Dolly Varden (DV) morph reared under modified thyroid status at a standard temperature (c, * indicate significant differences between series, p ≤ 0.042); T3—hyper‐, Control—eu‐, and Tio—hypothyroid series. D(τ)—amount of developmental temperature. Table 1 shows the number of specimens used.

3.3. Development of the Dolly Varden Morph Under Different Temperature Regimes

The experimental series of DV did not differ in the initial T4 supply (H_{3; 12} = 0.6 p = 0.730), ranging from 4.5 to 6.0 μg g⁻¹ (mean 5.2 ± SE 0.13 μg g⁻¹). Herewith, DV reared at different morph‐specific thermal regimes demonstrated different T3 developmental dynamics (Figure 8a). From four to six of the 15 comparison pairs of the experimental DV series differed significantly in T3 levels at different developmental stages (Supporting Information S1: Table 6). The DV series reared under typical predator (SC and W) temperature regimes (cold) consistently displayed higher T3 levels than the DV series reared under typical dwarf (GC and S) thermal regimes (warm) (Figure 8b,c). Maximum differences in the between‐series hormone concentration were observed at the eye pigmentation stage, which coincides with the appearance of pronounced temperature differences at the natural spawning grounds of the morphs (Figure 2).

Figure 8.

Whole‐body T3 developmental profiles in the experimental series of the Dolly Varden (DV) morph of S. malma. (a–c) DV reared at temperature regimes that imitated the natural temperature dynamics on the spawning grounds of the six morphs (see Figure 2); the mean values are shown at the moments of stage transition (a, ** indicates the stages at which significant differences were observed, as referenced in Supplementary Tables S6) and the mean ± SE and min–max values are shown for the temperature regimes of riverine (b) and lacustrine (c) morphs. (d) the rearing under altered thyroid status at a standard temperature, compared with the control series; the mean ± SE and min–max values of the T₃ content are shown at the moments of stage transition (*indicates significant differences between series, p ≤ 0.035). D(τ)—amount of developmental temperature. The experimental series are labeled as in Figure 7; Table 1 shows the number of specimens used.

In the DV series reared under different temperature regimes, developmental rate differed: it was accelerated under the predatory (SC and W) and retarded under the dwarf (GC and S) regimes (Table 2). The rates of cranial bone ossification differed too. By the end of the experiment, the DV series reared under predatory and dwarf conditions exhibited significant differences in the sum of ossification ranks (Figure 7b): the t_sc series outperformed the t_s series (Supporting Information S1: Table 7).

Quantifying the effects of temperature and T3 content revealed that the time to reach the next developmental stage correlated with the average temperature at the previous stage (mean R ² = 0.975), as well as with the T3 level at the stage‐transition (=0.858). MANOVA confirmed the significant effect of both factors on the developmental rate (in the D(τ) values), but with a dominant effect of the hormone level (model F = 5.13 p = 0.0063 R = 0.610; Supporting Information S1: Table 8). At the same time, the T3 level at the stage‐transition was inversely correlated with the average temperature of the previous stage (mean R ² = 0.894), with the exception of the last developmental stage.

3.4. Development of the Dolly Varden Morph under Altered Thyroid Status at a Standard Temperature

Thyroid therapy of DV (the “T3” series) resulted in the elevated T3 level compared to control group both during therapy (>10 ng g⁻¹) and after 3 days of exposure to pure water (H_2;9 = 5.6 p = 0.003) (Figure 8d). The “Thio” series showed a reduced hormone level at the end of the experiment (p = 0.035).

Analysis of the developmental rate revealed a minimal advance of the “T3” series compared to the control (Table 2). The retardation of the “Thyo” series development was noted at the belly closure stage. Herewith, by the time of external feeding onset, the “T3” series significantly outperformed the control series in the state of cranial bone ossification (H_2;24 = 5.1 p = 0.042), approaching the W morph in this indicator (27.8 ± 1.48 vs. 26.1 ± 1.84 sum of ranks, respectively; p = 0.410). By the end of the experiment, the difference in ossification between the “T3” and the control series increased (Figure 7c; H_2;24 = 17.3 p = 0.022). The “Thio” series delayed in ossification with growth; at the end of the experiment, this parameter was significantly lower than in the control (p = 0.023) and “T3” (p = 0.002) series.

4. Discussion

Thyroid hormones facilitate acclimation to the temperature, and both temperature and thyroid axis activity are considered to be crucial drivers of adaptive radiation in fish (Davis 1997; Eales 2019; Clarke and Johnston 1996; Harada et al. 2003; Кitano et al. 2010; Little et al. 2013; Little and Seebacher 2014; Kahilainen et al. 2014; Lorgen et al. 2015; Zak et al. 2017; Bolotovskiy et al. 2018; Iwen et al. 2018; Little 2021; Salis et al. 2021; Karagic et al. 2022). However, the interplay between temperature and thyroid hormone level, which has also been suggested as a potential factor contributing to diversification, has not received much attention. Our experimental study demonstrates that the dynamics of ambient temperature and thyroid hormone level, both individually and in interaction, can shape the development of the ecologically plastic salmonid S. malma, resulting in the emergence of adaptive phenotypes. Of the hypotheses proposed in the introduction, the third integrated scenario, which suggests an ecological function of thyroid hormones (temperature influences the thyroid axis, which, in turn, affects development and adjusts it to the temperature), appears to be the most plausible.

The morphs of S. malma differ in their thermal preference and, like many other salmonids, actively select sites with certain temperature regimes for reproduction (Elliott 1975; Ohlberger et al. 2013; Esin et al. 2021b; Markevich et al. 2022). The fertilized eggs fall to specific temperature conditions at the spawning sites due to homing and differential migratory ability of morphs (Markevich et al. 2021; Esin et al. 2021c). As a result, the early development of the different morphs occurs under different temperature conditions, with variation in thyroid hormone dynamics, developmental rates, and heterochronies in skull bone ossification. Morphs adapted to cold‐water conditions demonstrate elevated T3 levels and accelerated (in terms of accumulated thermal energy, D(τ)) rate of the life stage transition and cranial bone ossification, while morphs that prefer warm conditions exhibit lower T3 levels, as well as slower life stage transition and skull ossification rates. Such physiological and developmental disparities ensure phenotypic and life history variation, leading to the occupation of specific foraging and habitat niches by the morphs (Esin et al. 2021b). This phenomenon is common in many fish taxa, including salmonids (Murray and McPhail 1988; McCullough 1999; Pankhurst and King 2010).

Changing the temperature regime in the experiments led to changes in the physiology and developmental timing of S. malma morphs. All experimental series maintained a characteristic profile of T3 levels with a peak at hatching and a decline by the transition to external feeding. Meanwhile, the generalized morph (DV) displayed an increase in T3 content, accelerated development, and cranial bone ossification in response to lower temperatures (the predator‐like temperature regime). In contrast, under warm conditions (the standard and dwarf‐like temperature regimes), the DV demonstrated a decline in all indicators. In both cases, the temperature‐induced changes led to phenotypic consequences. At the belly closure stage, the DV reared at the predator‐like temperatures resembled the predator, while those reared at the dwarf‐like regime were closer to the dwarf. Similar, predictable, temperature‐induced physiological and developmental effects were observed in the predator (W) and the dwarf (S) reared at the standard temperature, which is warm for the predator and cold for the dwarf. Consequently, ambient temperature drives the heterokairic shifts (Lofeu and Ahi 2026), underlying developmental plasticity, which in turn, provides an important source for the diversification and adaptive radiation of S. malma.

Experimental outcomes indicate that temperature influences the T3 dynamics, the developmental rate, and the rate of cranial bone ossification. However, manipulations with the thyroid status in DV revealed that the level of T3 itself affects developmental and cranial bone ossification rates, and its effect, especially on the latter, is more pronounced than thermal influence. These findings are consistent with the role of thyroid hormones in the development of charrs and many other teleosts (Eales and Brown 1993; Kitano et al. 2010; Esin et al. 2021a; Borisov and Shkil 2025), and additionally emphasize the role of the maternal thyroid supply in influencing the developmental trajectory of fish.

Changes in thyroid hormone dynamics are considered to be a part of the complex mechanism underlying thermal compensation and acclimatization to changing ambient temperatures in vertebrates. In mammals and birds, thyroid signaling is considered as an essential in adaptive thermogenesis, the regulation of basal metabolism and energy expenditure (Silva et al. 2017; Zekri et al. 2021). In teleosts, however, evidence of the role of the thyroid axis in adaptive tuning of physiology and development to temperature changes is fragmentary and often conflicting (Cheverie and Lynn 1963; Umminger 1978; Little 2021). In some cases, a short‐term decrease in temperature can lead to a reduction in the thyroid axis activity, and vice versa. At the same time, chronic exposure to low or high temperatures may reduce metabolic changes due to the physiological overcompensation and/or the adaptation to new temperature regimes (Seebacher et al. 2015). For fish with wide latitude distribution, a reduction in T3 level in response to water warming towards the south has been observed (Levin and Bolotovskiy 2015). Considering that changes in thyroid hormone level mediate metabolic acclimation and ensure thermal plasticity in fish (Cheverie and Lynn 1963; Umminger 1978), we suggest that adaptive changes in the thyroid axis activity are an important mechanism underpinning the shift in thermal optima in Salvelinus morphs.

In favor of this suggestion, our experimental results demonstrated that the standard temperature mitigated, but did not eliminate, the disparities in T3 levels between the Salvelinus morphs. Predators maintained a significantly higher level of hormone and accelerated development compared to generalists and dwarfs. These data suggest that morphs are adaptive hormonal phenotypes with specific thermal preferences and tolerance. It is in good agreement with previously discovered selective signatures in genes regulating thyroid signaling, craniofacial development, and metabolism in Salvelinus morphs (Woronowicz et al. 2024). It is reasonable to suppose that the polymorphism of the Kamchatkan S. malma is genetically structured, and that the maternal thyroid supply in different morphs is to some extent predetermined. Hypothetically, predatory morphs follow an “energy‐maximizing” strategy (rapid development and ossification in cold water with high thyroid consumption), while dwarf morphs follow a “cost‐minimizing” strategy (slow development in warmer water with reduced thyroid supply). Such thyroid‐dependent physiological and developmental decisions ensure harmonization of the developmental timing with seasonal environmental changes in the appropriate habitat (Esin et al. 2021b), and lead to the phenotypic consequences underlying the morphological diversification of morphs.

Noteworthy, morph‐specific thyroid dynamics and developmental divergence occurred prior to the establishment of a fully functional hypothalamic‐pituitary‐thyroid axis, which occurs after the development of thyroid follicles, immediately before the larva‐juvenile transition (Vergauwen et al. 2018). Based on ELISA and histological study, a more intensive consumption of hormones was detected in predators. This resulted in earlier depletion of the maternal supply, more intensive development of thyroid follicles, and earlier onset of endogenous hormone synthesis compared to generalized morphs and dwarfs. Considering that these occurred at the end of the experiment, at the belly closure stage, we conclude that divergence in the S. malma early development is largely due to differences in consumption of the maternal thyroid hormones, which is known to be a vital factor in fish development and early specialization (Ruuskanen and Hsu 2018; Vergauwen et al. 2018; da Silva Rodrigues et al. 2025).

Our data suggest that the interplay between ambient developmental temperature and maternal thyroid hormone consumption, together with other ecological factors (Schluter 2000), contributes to adaptive radiation in Salvelinus. Thus, lacustrine and riverine morphs adapted to similar feeding strategies (generalized benthivorous/sedentary insectivorous/migratory piscivorous) demonstrate similar thermal preferences and have undergone parallel physiological and developmental specializations. The generalized morph occupies spawning grounds with moderate temperatures and passes a medium hormone supply to its offspring. The insectivorous dwarf prefers warmer reproduction sites and passes a low hormone reserve. The predators choose colder spawning grounds and pass a maximum thyroid reserve.

The main findings of the study can be summarized as follows. 1. Ambient temperature affects thyroid hormone dynamics, which, in turn, influences developmental and cranial ossification rates, that is, provides heterokairic changes. 2. Lacustrine and riverine morphs with similar feeding strategies demonstrate similar thermal preferences, parallel physiological adaptations, and developmental specializations. 3. The adaptive specialization of early ontogeny is realized under predetermined temperature conditions through the tuning of developmental programs constrained by maternal thyroid supply. 4. The interplay between maternal thyroid supply and developmental temperature facilitates the adaptive radiation of Salvelinus morphs.

Ethics Statement

The experiments were done on embryonic and early post‐embryonic Salvelinus malma (Salmonidae), a least concern (IUCN; https://www.iucnredlist.org/fr/species/202647/18232551) widespread species of charr in north‐east Asia. All applicable international, national, and institutional guidelines for the care and use of this species were strictly followed. The procedures with fish were carried out according to the guidelines and following the laws and ethics of the Russian Federation, and approved by the ethics committee of the A.N. Severtsov Institute of Ecology and Evolution, Russian Academy of Sciences (Approval ID N 88 issued 22/03/2024).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File

Data Availability Statement

The data that supports the findings of this study are available in the supplementary material of this article.

All data including database of primary measurements are included in the Supplementary Information.