Evolution of Adaptive Non-Shivering Thermogenesis in Mammals

Abstract

Endothermy depends on the ability to generate heat beyond basal metabolic output. The transition from incidental metabolic heat production to regulated adaptive thermogenesis represents a critical, yet poorly understood milestone in vertebrate evolution. In vertebrates, adaptive thermogenesis includes shivering thermogenesis and non-shivering thermogenesis (NST), the latter involving molecular mechanisms that release energy in the form of heat independently of muscle contraction. This review synthesizes physiological, comparative, and evolutionary findings on adaptive NST, with a focus on adipose-based thermogenesis mediated by its unique mitochondrial uncoupling protein 1 (UCP1). We highlight a two-stage evolutionary model in which a pre-thermogenic adipose program governing UCP1 expression preceded the emergence of efficient UCP1-mediated thermogenesis in eutherian mammals. The source of NST and underlying mechanisms in other tissues and in non-eutherian endothermic vertebrates are less well understood. We discuss proposed UCP1-independent sources of cellular heat production in adipose tissue and muscle. Although these pathways can increase cellular energy expenditure, evidence supporting their role as dedicated and regulated thermogenic mechanisms remains limited to date. Collectively, these findings suggest that adaptive NST is not a singular conserved mechanism but a diverse and context-dependent evolutionary strategy with important implications for understanding human physiology and disease.

1. Thermogenesis

Heat production in animals can be broadly divided into basal metabolic heat, activity-associated heat, and adaptive thermogenesis (1). Basal metabolic heat arises unavoidably from cellular and organismal processes required to maintain homeostasis at rest and is reflected in the basal metabolic rate (2). Activity-associated heat is generated incidentally during muscular contraction and other forms of biological work. Because biochemical energy conversion to mechanical output is intrinsically inefficient, locomotion and other behaviors inevitably produce heat as a byproduct. Adaptive thermogenesis, by contrast, comprises inducible and reversible mechanisms that specifically increase heat production in response to cold exposure or other environmental stimuli, e.g., the photoperiod to anticipate seasonal changes (3, 4). These mechanisms are under neural and/or endocrine control and can be engaged independently of voluntary activity (5).

Adaptive thermogenesis is one of the most intriguing physiological innovations in nature. By enabling organisms to maintain elevated body temperatures relative to the environment, adaptive thermogenesis supports optimal enzyme kinetics, sustained locomotor performance, efficient nutrient assimilation, and energetically demanding processes such as brain function and reproduction. These benefits facilitate survival and ecological success across diverse environments, including cold climates where ectothermic competitors may be limited in activity (1). However, the evolution of thermogenesis causes considerable energetic costs. Heat production requires continuous substrate oxidation and therefore increased caloric intake to compensate for metabolic inefficiency (6). This complicates the ultimate question: why did thermogenesis evolve in ectotherms when heat production represents a form of energy dissipation? Selective pressures likely included advantages in foraging, reproductive success, predator avoidance, and the ability to remain active across variable environmental temperatures. Yet, despite extensive research, the precise evolutionary drivers of endothermy remain actively debated (1, 7). While endogenous heat production is rare outside the animal kingdom, occurring in only a few plant species, for instance, in thermogenic aroids that warm to attract pollinators (8), localized thermogenesis has evolved repeatedly across animals. Certain insects, such as bumblebees, warm their flight muscles to facilitate activity at low temperatures (9). These examples illustrate how transient, tissue-specific heat generation may have been an evolutionary precursor to sustained, whole-body endothermy.

2. From Ectothermy to Endothermy

To date, it remains unclear how and when thermogenic capacities evolved in vertebrates, with evidence supporting either a single origin or multiple independent evolutionary events in birds and mammals (1, 10, 11). The presence of localized endothermy in a small number of species that are traditionally classified as ectothermic complicates reconstructions of thermogenic evolution (11, 12). Such examples weaken the conventional separation between ectothermy and endothermy and suggest that thermogenic traits may have evolved incrementally through tissue- or organ-specific adaptations.

Among fishes, which are classically considered ectothermic, several shark and teleost species maintain tissue temperatures above that of the surrounding water, despite the high thermal conductance of water as compared with air (13, 14). In endothermic sharks and tunas, heat is generated by the high oxidative activity of internalized red muscle during continuous swimming and conserved via counter-current heat exchangers (15). In contrast, billfishes exhibit localized cranial endothermy via a specialized “heater organ” derived from extraocular muscle, enabling elevated brain and eye temperatures (16). These adaptations enhance neural processing and visual performance, giving them an advantage when hunting at high speeds in cold waters (17). Phylogenetic evidence suggests that these thermogenic strategies evolved independently multiple times, highlighting the repeated and context-dependent development of heat-producing systems in vertebrates (12, 14).

In amphibians, which are semi-aquatic ectothermic vertebrates that rely predominantly on environmental heat sources for thermoregulation, no evidence of sustained endothermy has been detected so far (11, 18). In reptiles, thermogenesis is rare and circumstantial occurring primarily during offspring incubation or as a metabolic response to large meals (19). These transient thermal responses differ fundamentally from the persistent elevation of body temperature observed in birds and mammals, in which endothermy is believed to have evolved later and independently in the two lineages (6, 11). Whether the last common amniote ancestor of birds and mammals already possessed any form of proto-endothermy remains an open question (1, 10, 11). Fossil, physiological, and comparative evidence offer competing interpretations, and reconstruction is complicated by the absence of direct thermophysiological data from extinct species. As a result, the precise timing, sequence of events, and selective pressures underpinning the evolution of vertebrate thermogenesis remain subjects of active debate (11).

3. Fossil and Palaeophysiological Evidence for Thermogenic Evolution

Our understanding of when and how heat-producing mechanisms arose in terrestrial vertebrates remains limited, and relies primarily on indirect evidence from the fossil record (11). Recent advances in integrating macroevolutionary modelling with palaeoclimatic reconstructions have begun to provide more insights. For example, a recent study combined fossil evidence with macroevolutionary and palaeoclimatic models to investigate the thermophysiology of three major dinosaur lineages: ornithischians, theropods, and sauropodomorphs (20). The broad geographic distribution of ornithischians and theropods, spanning from tropical to polar regions, suggests the adoption of endothermy, which enabled these species to expand their range into cooler climates. In contrast, sauropodomorphs seemed to have been largely restricted to warmer habitats, consistent with an ectothermic physiology that relied more heavily on external heat to maintain body temperature. Such findings are particularly relevant for reconstructing the development of avian endothermy, given the direct lineage linking theropod dinosaurs to modern birds.

Additional insights into palaeophysiology are derived from morphological structures that are preserved in fossils. The interpretation of morphological data from fossils, however, is not straightforward, as seen for instance for the semicircular duct system of the inner ear. Araújo and colleagues demonstrated that semicircular canal thickness and endolymph characteristics differ systematically between extant ectotherms and endotherms (21). As body temperature rises, the viscosity of the endolymph decreases, requiring the physical shape of the canal to adapt. By applying these morphological criteria to extinct species spanning diverse lineages, from extant reptiles to extinct synapsids such as Lemurosaurus, the authors found that mammalian endothermy likely emerged during the Late Triassic. Earlier, however, semicircular canal dimensions were thought to be driven by complex movements and spatial sensitivity (22, 23). Thus, the observed morphological shifts may not reflect a response to endolymph viscosity rather a secondary consequence of the increased behavioral activity that endothermy permits. Others argue that it seems more likely that canal shapes are strongly influenced by spatial and developmental constraints, such as skull size (24, 25).

While fossils and palaeophysiological proxies therefore offer valuable, albeit sometimes inconclusive, evidence, alternative data sources remain limited. Molecular evidence is largely inaccessible in the fossil record, and ancient DNA sequencing efforts are currently limited to specimens no older than approximately one million years (26, 27). This temporal restriction precludes direct reconstruction of the molecular evolution of thermogenesis in mammals and birds, which diverged from their last common ancestor roughly 320 million years ago (28). Instead, comparative analyses of extant species can be paired with phylogenetic relationships to infer ancestral states and evolutionary transitions, thereby extending reconstructions of thermogenesis several hundred millions of years ago, deep into vertebrate history.

4. Bioenergetic Principles Underlying Cellular Heat Production

To understand how adaptive thermogenesis operates at the organismal level, it is essential to consider the underlying bioenergetic principles at the cellular level. Cellular heat is released during catabolism and oxidation of substrates at the mitochondrial respiratory chain, where chemical energy is conserved in the form of a proton motive force across the mitochondrial inner membrane (29). This gradient normally drives ATP synthesis, but it can also be dissipated as heat through proton leak pathways (30). Mitochondrial proton leak and increasing ATP turnover through for instance muscle shivering, cellular activity, or active ion transport, e.g. by transmembrane proteins, accelerates upstream catabolic reactions and elevates overall heat production. These bioenergetic processes form the fundamental mechanistic basis for cellular thermogenesis (Fig. 1).

Figure 1. Cellular sources of basal heat production.

Basal cellular heat arises from nutrient catabolism and subsequent substrate oxidation at the electron transport chain (ETC), ATP turnover, and the mitochondrial proton leak.

5. Adaptive Thermogenesis

Potent thermogenic mechanisms must be tightly regulated to prevent excessive energy wasting while allowing rapid adjustments to environmental temperature changes (5). Constitutive metabolic inefficiency alone cannot provide this flexibility; instead, thermoregulation requires inducible mechanisms that can be recruited and disengaged as needed. In endothermic vertebrates, adaptive thermogenesis is typically divided into two categories: shivering thermogenesis and non-shivering thermogenesis (NST) (5). NST, the focus of this review, includes molecular and cellular processes capable of dissipating chemical energy as heat without significant muscle contraction (31).

5.1. Shivering

Skeletal muscle is a major contributor to whole-body heat production in vertebrates due to its large mass and high oxidative capacity (32). Muscle thermogenesis is a ubiquitous and evolutionarily ancient source of heat (6, 9), with both ectothermic and endothermic vertebrates generating substantial heat as a byproduct of ATP turnover during muscular activity. In some taxa, such as endothermic fishes, sustained muscle activity supports regional or whole-body warming (14, 16), while in birds and mammals, shivering thermogenesis provides a regulated version of this mechanism. Through rapid, involuntary oscillatory contractions, shivering increases ATP turnover and heat production during acute cold exposure (33, 34). Although effective, shivering is energetically costly and incompatible with sustained locomotion (35–37), underscoring the importance of mechanisms that can take over during chronic cold adaptation.

5.2. Non-shivering thermogenesis (NST)

NST is a form of adaptive heat production that does not involve muscle contractions but instead relies on chemical energy dissipation as heat. At the molecular level, NST involves pathways that increase mitochondrial respiration, which can occur through increased mitochondrial proton leak across the mitochondrial inner membrane or increased ATP turnover through ATP-consuming processes, including futile metabolic cycles. Such mechanisms accelerate substrate oxidation and other exergonic catabolic reactions to enhance heat production. Adaptive NST harnesses these mechanisms in a controlled, inducible manner, enabling endothermic vertebrates to maintain thermal homeostasis in challenging environments.

Among vertebrates, brown adipose tissue (BAT) with its unique mitochondrial uncoupling protein 1 (UCP1) represents the best-characterized heater organ for regulated NST. Proposed UCP1-independent mechanisms, including candidate pathways in adipocytes and skeletal muscle, are discussed in later sections.

5.2.1. Regulation of BAT and UCP1-mediated adaptive NST

In mammals, BAT is a highly specialized thermogenic organ characterized by dense mitochondrial content, rich vascularization, multilocular lipid droplets, and extensive sympathetic innervation (31). Adaptive NST in BAT is mediated primarily by UCP1, a mitochondrial transmembrane protein that is permeable for protons when activated (38). By dissipating the proton motive force without ATP synthesis, UCP1 can increase respiration independent of ATP homeostasis, thereby maximizing heat evolution from mitochondria. Together, these features enable rapid, inducible, and reversible increases in heat production, clearly differentiating BAT-mediated thermogenesis from both basal metabolic heat and activity-associated heat generation.

Thermogenic activation is initiated by cold-induced sympathetic nervous system signaling, resulting in the release of noradrenaline from postganglionic neurons innervating brown adipocytes (Fig. 2). Noradrenaline activates adrenergic receptors (e.g., β₃; ADRB3 in many rodents) on the adipocyte plasma membrane, triggering a cAMP–protein kinase A (PKA) signaling cascade that stimulates lipolysis through hormone-sensitive lipase (HSL) and adipose triglyceride lipase (ATGL). The resulting free fatty acids serve as both oxidative substrates and direct activators of UCP1, overcoming nucleotide inhibition and enabling UCP1-mediated proton leak across the mitochondrial inner membrane (31). Despite recent cryogenic electron microscopy structures of UCP1 (39, 40), key aspects of the molecular mechanism underlying proton translocation and activator binding remain unresolved (41).

Figure 2. Regulation of adipose-based UCP1-depedent adaptive NST.

Cold-induced sympathetic noradrenaline release activates adrenergic receptors (β₃; ADRB3) in the plasma membrane of brown adipocytes, triggering a cAMP-protein kinase A (PKA) signaling cascade that stimulates hormone-sensitive lipase (HSL) and adipose-tissue triglyceride lipase (ATGL) to promote lipolysis. Free fatty acids activate UCP1, thereby increasing uncoupled mitochondrial respiration and heat output. In parallel, transcriptional regulators such as peroxisome proliferator-activated receptor γ (PPAR γ) and PPAR γ coactivator 1 α (PGC1α) are recruited supporting UCP1 expression and inducing the thermogenic gene programs in brown adipocytes.

Crucially, UCP1-mediated uncoupling fulfills the defining criteria of adaptive NST. It is inducible by environmental cues, rapid, reversible, quantitatively sufficient to affect whole-body heat balance, and integrated into neural and endocrine thermoregulatory networks. Regulation occurs at multiple levels, including sympathetic input, substrate supply, UCP1 activation and inhibition, and transcriptional programs that support mitochondrial biogenesis and oxidative capacity. During chronic cold exposure, transcriptional regulators such as PGC-1α orchestrate the maintenance of the thermogenic phenotype (42).

Besides brown adipocytes, beige adipocytes are thermogenic adipocytes that emerge within white adipose depots in response to stimuli such as cold exposure and β-adrenergic signaling (43). They share elements of the thermogenic gene program similar to classical brown adipocytes, including the capacity to express UCP1, but differ in developmental origin and regulatory context (44).

5.2.2. Evolution of BAT and UCP1

BAT-mediated NST has traditionally been viewed as a generic mammalian innovation, a perspective shaped by early comparative studies focused primarily on a handful of rodents (31). Comparative and phylogenetically informed analyses have since revised this perspective, revealing that the key molecular components underlying BAT thermogenesis have a deeper evolutionary history than previously appreciated (45, 46). Notably, UCP1 orthologs are present in non-mammalian vertebrates, including fish (47) and amphibians (48), demonstrating its origin before the divergence of ray- and lobe-finned vertebrates about 420 million years ago. UCP1’s ancestral function, however, is unknown and was likely unrelated to regulated heat production.

In ectothermic fish and amphibians, UCP1 appears to be dominant in non-adipose tissues (47, 48), possibly representing an ancestral stage (‘Stage 0’). Current evidence supports a two-stage evolutionary model for the emergence of adipose-based adaptive NST (49) (Fig. 3). In the first pre-thermogenic stage (‘Stage 1’), UCP1 expression and elements of a thermogenic regulatory architecture became established in adipose tissue, generating a “proto-BAT” state without efficient thermogenic uncoupling activity but a “beige-like” gene signature. In the critical stage (‘Stage 2’), which appears to have occurred in the common ancestor of eutherian mammals, specific amino acid substitutions rendered UCP1 competent to catalyze regulated proton leak, enabling uncoupled respiration for physiologically effective adaptive NST.

Figure 3. Two-stage evolution of adipose-based adaptive NST.

In an ancestral stage (‘Stage 0’), a UCP1 version with an unknown function is present in fish and amphibians, mainly in non-adipose tissues (liver, kidney and brain). The pre-thermogenic stage in mammals (‘Stage 1’) is characterized by the establishment of UCP1 expression and elements of a thermogenic regulatory architecture in adipose tissue, followed by a eutherian mammal-specific stage (‘Stage 2’) enabling efficient UCP1-mediated mitochondrial uncoupling and physiologically effective adaptive NST. For monotremes, UCP1 tissue expression is scarcely characterized, and the reptilian/avian lineage has lost the UCP1 gene.

Consistent with this model, thermogenic BAT and UCP1 have also been identified in afrotherian mammals, suggesting that adipose-based adaptive NST evolved in warm climates, prior to the radiation of modern eutherian mammals into colder Holarctic environments (50). In the afrotherian tenrec, rated as “proto-endothermic” mammal due to body temperatures tracking close to ambient temperatures, BAT localizes near the gonads, raising the possibility that early BAT thermogenesis facilitated reproductive success and offspring incubation, aligning with hypotheses that parental care and reproductive benefits drove the evolution of endothermy (51). Consistent with this, reproductive tenrecs exhibit increased homeothermy relative to nonreproductive states (52).

Importantly, BAT-mediated adaptive NST is not universally conserved among extant mammals. Comparative genomic analyses reveal multiple independent losses of functional UCP1 across diverse eutherian lineages (45). These losses highlight the evolutionary plasticity of mammalian thermoregulation and demonstrate that endothermy can be maintained through alternative strategies including enhanced insulation, behavioral thermoregulation, or increased metabolic heat production in other tissues.

Together, these findings establish adipose-based adaptive NST mediated by UCP1 as the most thoroughly validated example of adaptive NST in vertebrates. At the same time, the evolutionary history of BAT underscores that it represents only one of several possible solutions to the thermoregulatory challenges faced by endothermic animals. This framework is essential for interpreting proposed UCP1-independent thermogenic pathways and their physiological or biomedical relevance.

6. UCP1-independent heat production and proposed NST mechanisms

Although UCP1-mediated mitochondrial uncoupling in brown adipocytes represents the canonical NST mechanism in mammals, multiple studies have proposed that heat production can also be supported by UCP1-independent pathways (Fig. 4). Interest in these mechanisms increased following observations that UCP1-deficient mice can survive prolonged cold exposure after acclimation, maintain near-normal resting energy expenditure and are resistant to diet-induced obesity when housed below thermoneutrality (53–57). These findings demonstrate that whole-body heat balance can be preserved in the absence of UCP1, suggesting compensatory mechanisms, such as increased shivering, altered behavior, enhanced insulation, or metabolically inefficient pathways across multiple tissues, contributing to maintaining body temperature in UCP1-deficient animals.

Figure 4. Scheme of proposed ATP-consuming futile cycles contributing to UCP1-independent heat production in adipocytes and skeletal muscle.

Four ATP-consuming futile cycles and their key molecular components are depicted. Lipid futile cycling may occur during simultaneous triglyceride lipolysis and re-esterification, where ATP is consumed during fatty-acid activation and glycerol metabolism, for example, when glycerol is phosphorylated by glycerol kinase (GYK). Futile creatine cycling involves ATP-dependent phosphocreatine formation by creatine kinase B (CKB) and its hydrolysis by tissue-nonspecific alkaline phosphatase (TNAP). Calcium cycling in adipocytes and muscle is driven by ATP-dependent sarco/endoplasmic-reticulum Ca²⁺-ATPase (SERCA1 and 2b) activity, balanced by Ca²⁺ release/leak through ryanodine receptors. Current evidence suggests that RyR2 is the principal isoform in adipocytes, whereas RYR1 is the only isoform in mammalian skeletal muscle. Potential modulators are another-regulin (ALN) in adipocytes and sarcolipin (SLN) in muscle. These pathways associate with elevated energy expenditure, whereas their regulation, magnitude, and contribution to whole-body thermoregulation require further evaluation.

Several ATP-consuming futile substrate cycles have been described as candidate sources of UCP1-independent heat production, potentially residing in both brown and beige adipocytes, as well as in skeletal muscle. These pathways share a common principle: energy is dissipated through repeated biochemical cycling without net substrate synthesis, thereby increasing ATP demand, stimulating upstream oxidative metabolism, and ultimately generating heat.

6.1. Lipid futile cycling

An intensively studied UCP1-independent pathways involves futile cycling between triglyceride lipolysis and re-esterification. In adipocytes, specifically beige adipocytes, cold exposure has been shown to enhance fatty acid turnover through coordinated activation of lipolytic enzymes, acyl-CoA synthetases, glycerolipid synthesis enzymes, and glycerol kinase (GYK). This cycle consumes ATP while regenerating triglycerides, thereby driving mitochondrial respiration and releasing heat. Experimental work in mice indicates that lipid cycling increases during cold exposure and can partially compensate for the loss of UCP1-mediated uncoupling during mild or chronic cold adaptation (55, 56, 58).

From a bioenergetic perspective, lipid futile cycling is a plausible source of heat production, as it is energetically costly and tightly linked to mitochondrial ATP synthesis. However, this pathway also serves broader roles in cellular lipid remodeling, signaling, and metabolic flexibility (59–63). Thus, whether lipid futile cycling in adipocytes evolved specifically as adaptive NST, or primarily reflects a compensatory metabolic response revealed under experimental conditions, remains unresolved. Moreover, quantitative assessments of its contribution to whole-body thermogenesis in non-rodent species are currently lacking.

6.2. Futile creatine cycling

Another proposed mechanism is a futile cycle of creatine phosphorylation and dephosphorylation in brown and beige adipocytes (64, 65). In this model, mitochondrial creatine kinases transfer phosphate groups from ATP to creatine, generating phosphocreatine, which is then hydrolyzed by a proposed phosphatase, TNAP (Tissue-Nonspecific Alkaline Phosphatase), resulting in increased ATP turnover and mitochondrial respiration. Experimental manipulation of components of this pathway in mice alters energy expenditure and cold tolerance, suggesting that creatine cycling can modulate adipocyte metabolism (64–66).

The physiological relevance of this mechanism as a dedicated thermogenic system remains debated (67). The creatine kinase system is ubiquitous in tissues with fluctuating energy demands, where it primarily functions as an energy buffer rather than a heat-producing pathway (68). Evidence that adipocytes uniquely repurpose this cycle for heat production relies heavily on genetically modified mouse models, and independent confirmation of its thermogenic magnitude, regulation, and integration into thermoregulatory circuits remains limited (67).

6.3. Calcium futile cycling

Calcium cycling across the endoplasmic reticulum (ER) has also been proposed as a source of UCP1-independent heat production in adipocytes. Two related mechanisms have been suggested: partial uncoupling of SERCA (Sarcoplasmic/Endoplasmic Reticulum Calcium ATPase)-dependent ATP hydrolysis from Ca²⁺ transport through binding of regulatory peptides (69), and regulated release of Ca²⁺ from the ER, mediated by ryanodine receptor RyR2 in adipocytes, followed by the ATP-dependent reuptake, thereby generating heat through futile cycling (70).

Manipulation of these calcium-dependent pathways in mouse models can influence energy expenditure and glucose homeostasis (69, 70). However, several challenges complicate the interpretation. Importantly, Ca²⁺ signaling is essential for cellular viability and homeostasis, and alterations in ER Ca²⁺ handling may have broader implications beyond thermogenesis, as it is tightly linked to protein folding, ER stress responses, and lipid biosynthesis. Consequently, demonstrating that Ca²⁺ cycling can be selectively and reversibly recruited for heat production without disrupting other essential cellular processes remains a major experimental challenge. Additionally, while direct imaging of Ca²⁺ transients in BAT during cold exposure is still limited, recent evidence indicates that adrenergic stimulation can modulate intracellular Ca²⁺ dynamics via ER remodeling in thermogenic adipocytes (71). Quantitative measurements of the thermogenic contribution of Ca²⁺ cycling in adipocytes, however, are scarce, and its evolutionary trajectory is unresolved.

6.4. Muscle-based NST

Beyond shivering, skeletal muscle has been proposed to contribute to adaptive NST through molecular mechanisms that increase ATP turnover without producing mechanical work (72–74). Central to this hypothesis is again calcium futile cycling, in which leakage of Ca²⁺ through the ryanodine receptor (RyR) is matched by ATP dependent Ca²⁺ uptake into the sarcoplasmic reticulum (SR) by SERCA. This cycle increases ATP turnover without generating mechanical work, promotes substrate flux into the mitochondria and activates oxidative phosphorylation to replenish ATP, resulting in increased heat production both directly through ATP hydrolysis by SERCA and indirectly by supporting enhanced mitochondrial substrate oxidation. A well-known example of this principle is the heater organ of billfishes, a highly specialized structure derived from extraocular muscle and adapted for rapid Ca²⁺ cycling without contraction (16). However, this organ is unique to a narrow evolutionary lineage and does not constitute broad evidence for a generalized muscle-based NST mechanism across vertebrates.

A proposed regulator of muscle-based NST is sarcolipin (SLN), a small peptide that can reduce the coupling efficiency of SERCA, thereby increasing ATP consumption and heat production at a given calcium transport rate. Mouse models with altered SLN expression exhibit changes in cold tolerance and energy expenditure, suggesting that SLN-mediated modulation of Ca²⁺ cycling can influence muscle thermogenesis (72). While these findings demonstrate that SERCA efficiency can affect muscle energy turnover, they do not in themselves establish SLN-mediated Ca²⁺ cycling as a physiologically relevant contributor to adaptive NST. Notably, recent work indicates that SERCA can generate detectable heat in resting skeletal muscle even when the Ca²⁺ leak through RyR1 is experimentally minimized, consistent with an intrinsic thermogenic component of SERCA activity that is not necessarily dependent on SLN (75). SLN is expressed widely across vertebrate taxa, including ectotherms that do not exhibit endothermic thermoregulation (76). This broad phylogenetic distribution argues against a role specifically associated with the evolution of endothermy. To date, clearly defined regulatory pathways for SLN-mediated thermogenesis are lacking, and it remains unclear how this would be coordinated with contractile function during physical demands. For RyR, however, differences in the sensitivity for Ca²⁺-induced Ca²⁺ release between the only isoform in mammalian skeletal muscle, RyR1 (77), and the two isoforms αRyR and βRyR in the muscles of ectothermic vertebrates (e.g., amphibians) have been shown (78). This indicates that RyR1 in the mammalian skeletal muscle can have a wider, tunable range of Ca²⁺-leakage from the SR that changes ATP-hydrolysis rates of SERCA, without the initiation of contraction (74). Higher sympathetic input coincides with RyR1-phosphorylation in the body core muscles, indicating that heat generation in resting muscle could support core temperature (79). Quantitative bioenergetic analyses assessing how much heat such a mechanism would produce are limited (76, 80).

Pathophysiological evidence highlights the risks associated with altered muscle Ca²⁺ fluxes. Mutations in the ryanodine receptor RYR1 can lead to malignant hyperthermia in humans, a condition characterized by uncontrolled Ca²⁺ release in response to specific triggers, excessive ATP consumption, and life-threatening hyperthermia (81). While this pathology does not preclude physiological regulation, it highlights the need for Ca²⁺ cycling to be tightly regulated.

Collectively, lipid cycling, creatine cycling, and calcium cycling illustrate how metabolically inefficient biochemical processes can increase ATP turnover and heat release in adipocytes and muscle (Fig. 4). They are plausible sources of heat production and can elevate cellular energy expenditure. However, current evidence has not yet established them as dedicated mechanisms that fulfil the criteria of regulated adaptive NST, nor can they be fully refuted. Most supporting data derive from genetically modified mouse models under laboratory conditions, and comparative data across species are largely missing. In addition, adipose tissue, particularly beige adipose tissue, is cellularly heterogeneous, and adipocyte subpopulations with differing UCP1 expression show variable expression of components of these futile cycles (70, 82, 83), complicating inference about their contribution to regulated thermogenesis. From an evolutionary perspective, these pathways may reflect ancestral metabolic programs that predate the emergence of UCP1-mediated uncoupling and can be co-opted or amplified under certain physiological or experimental contexts. Alternatively, they may act as compensatory mechanisms in species or individuals with reduced BAT activity or non-functional UCP1.

7. Biomedical implications

Insights into the evolution and diversification of both UCP1-mediated and UCP1-independent NST have important biomedical implications, given their significant impact on whole-body metabolism and energy homeostasis.

7.1. Obesity and metabolic diseases

NST is an appealing target to support weight loss and mitigate obesity and its associated metabolic disorders by increasing energy expenditure (46). Following the rediscovery of functional BAT in adult humans (84), more than a century after its initial description (85), and the identification of beige adipocytes within human white adipose tissue (WAT) (86, 87), there has been intense interest in strategies that activate thermogenic pathways in adipose tissue.

Activation of BAT improves lipid and glucose homeostasis, ameliorating obesity-related metabolic dysfunction (88). This has been demonstrated in experimental mouse models and supported in humans, where BAT activity correlates with healthier lipid and glucose profiles (89–91). However, targeting UCP1-dependent pathways may have limited therapeutic potential in obese individuals, who often have very low or undetectable BAT mass and reduced UCP1 expression (92–94). These limitations have driven interest in UCP1-independent pathways, including potential thermogenic mechanisms in skeletal muscle. Because muscle constitutes a large fraction of body mass, even modest increases in ATP-consuming futile cycles or mitochondrial inefficiency could meaningfully elevate energy expenditure. As a result, the proposed muscle-based thermogenic mechanisms may represent promising complementary or alternative therapeutic strategies (95).

Pharmacological approaches aimed at controlled mitochondrial uncoupling have also regained attention. Although the chemical uncoupler 2,4-dinitrophenol (DNP) was banned by the FDA due to a narrow therapeutic window and fatality (96), recent efforts aim to achieve controlled and tissue-specific uncoupling. Over the past decade, several compounds have been investigated (97), these include BAM15 (98), endogenous metabolites like N-acyl amino acids (99, 100), and repurposed drugs such as niclosamide ethanolamine (100, 101). Controlled-release formulations of protonophores are also being explored to enhance hepatic fat oxidation and reduce steatosis without causing systemic toxicity (102).

Recent research has further refined the understanding of the central mechanisms that govern energy balance and thermogenesis. In particular, a thermogenic regulatory node in the central nervous system that modulates energy expenditure and body weight in mice was recently described (103), highlighting the potential of neural targets for metabolic control. Concurrently, thermogenic capacity and energy homeostasis can be regulated not only by changes in sympathetic neural activity, but also by regulating the architecture of the sympathetic nervous structure in adipose tissue, as it was recently shown for a leptin–BDNF signaling pathway (104). These emerging insights expand the conceptual framework for the development of novel therapeutic strategies for obesity from a focus on peripheral thermogenic mechanisms alone toward an integrated, circuit-level regulation of energy balance.

7.2. Cancer and cancer cachexia

Cold exposure, via activation of BAT and UCP1, has been reported to reduce tumor growth, potentially by diverting circulating substrates (including glucose) toward thermogenic tissues (105). Similarly, mitochondrial uncouplers are also emerging as potential therapeutic agents in cancer because they can interfere with tumor metabolism (106), although their clinical applicability remains uncertain. However, extensive research is still needed to determine which cancer types and patient populations might benefit from such interventions and to evaluate long-term safety (107).

BAT and browning of WAT have long been indicated in cancer cachexia (108–110). Cachexia is a multifactorial wasting syndrome characterized by involuntary weight loss, adipose tissue depletion, muscle atrophy, and elevated energy expenditure. Both, experimental models and clinical observations, suggest that tumor-derived and host inflammatory signals can induce browning of WAT and aberrant activation of thermogenic programs, including increased UCP1 expression potentially contributing to increased metabolisms and wasting (108–110). Recently, however, the association between cachexia and BAT was challenged by retrospective analyses reporting no consistent link between the two (111), highlighting the need for further investigations.

7.3. Human mutations and polymorphisms

Genetic variation in thermogenic pathways also has clinical relevance. Several single nucleotide polymorphisms (SNPs) have been identified in the UCP1 promoter and codon region, although their effects on body mass, adiposity, and metabolic disease risk remain inconsistent across studies (112). Some promoter variants have been associated with NST capacity (113) and resting metabolic rate (114), yet UCP1 polymorphisms alone do not appear to exert strong, uniform effects on thermogenic phenotype.

Beyond UCP1, other genes influence thermogenic potential. One example is the R577X polymorphism in ACTN3, which results in α-actinin-3 deficiency of homozygous individuals and is predicted to occur in over one billion humans worldwide (115, 116). While this variant is known for its associations with athletic performance and muscle fiber composition, recent evidence proposed that the polymorphism impacts cold tolerance by a shift towards low-twitch type I muscle fibers and an increase in muscle tone (116). Although this may highlight that genetic variation could fine-tune thermogenic capacity in human populations, no direct thermogenic measurements of the muscles were performed, and the thermogenic potential of the two main fiber types (I and IIa) in humans is unknown.

Other genetic conditions illustrate the risks associated with unconstrained heat production. Malignant hyperthermia (MH) is a severe reaction to certain volatile anesthetic gases and muscle relaxants such as succinylcholine which can be fatal if not immediately treated. Typical symptoms are a rapid increase in body temperature and muscle rigidity. Several genetic mutations are known to increase the susceptibility to MH, with the most common ones located in the gene for the ryanodine receptor type 1 (RyR1) Ca²⁺-release channels (117). MH-associated RyR1 mutations represent gain-of-function alterations that destabilize channel gating in the presence of specific pharmacological triggers (118). Although MH is a pathological condition and is mechanistically distinct from physiological Ca²⁺-mediated heat production in skeletal muscle, it underscores the importance for NST mechanisms to be tightly regulated, avoiding catastrophic metabolic outcomes.

Taken together, modulation of NST represents a promising therapeutic strategy for metabolic disease and hypermetabolic disorders. Importantly, from an evolutionary perspective, thermogenic pathways are constrained by trade-offs between energetic efficiency, thermal benefit, and tissue integrity, such that mechanisms optimized for survival in cold environments may become maladaptive when chronically or ectopically activated in modern metabolic or disease contexts. A deeper understanding of the evolution, regulation, and plasticity of NST will be essential for developing interventions that safely harness thermogenic pathways. Such insights could lead to novel treatments for obesity, metabolic dysfunction, cachexia, malignant hyperthermia, and other conditions in which thermogenic balance is disrupted.

8. Conclusion

NST is a vital part of mammalian cold adaptation and a key innovation in the evolution of endothermy. Comparative, physiological, and genomic evidence indicates that UCP1-mediated thermogenesis in BAT emerged early in mammalian evolution and became deeply integrated into neural and endocrine thermoregulatory networks. At the same time, the repeated loss of UCP1 across multiple mammalian lineages, together with emerging evidence for UCP1-independent thermogenic mechanisms, demonstrates that NST is not a uniform or indispensable feature of endothermy.

Rather than representing a single conserved mechanism, adaptive NST appears to include a spectrum of partially redundant and context-dependent strategies shaped by evolutionary history, ecological niche, and organismal constraints. Classical BAT provides an efficient and rapidly deployable heat source in small-bodied and cold-exposed mammals, particularly during early development and arousal from torpor/hibernation. In contrast, alternative thermogenic mechanisms in adipocytes and skeletal muscle may provide flexibility when the efficient UCP1-mediated thermogenesis in BAT is reduced or absent, highlighting the evolutionary plasticity of mammalian heat-producing systems.

Despite considerable advances, major questions remain unresolved. These include the relative timing of NST evolution in relation to elevated basal metabolic rates, the identity of the primary thermogenic tissues of early mammals, and the evolutionary significance and functional magnitude of UCP1-independent pathways. Addressing these questions will require the fusion of phylogenetics, comparative genomics and evolutionary physiology, paired with modern molecular and bioenergetic technologies capable of quantifying thermogenic protein function across species.

Ultimately, understanding the evolutionary diversification of NST not only illuminates the origins of mammalian endothermy but also offers valuable insights into metabolic flexibility, adaptation to thermal environments, and the constraints that shape energy balance in vertebrates.