Growing underground: Development of thermogenesis in pups of the fossorial rodent Ctenomys talarum

In mammals, during the pup’s development and adult life, integrated requirements of all activities of the individual must conform to a sustained rate of metabolism. Thus, partitioning the available energy according to short-term priorities at a specific moment allows animals to survive and optimize long-term reproductive success. In altricial rodents, thermal balance is a key factor for survival. When no exogenous source of heat is present, altricial pups rapidly lose heat, reaching ambient temperature (Ta). Fossorial rodents showed a strong dependence on burrows, where Ta remains relatively stable within narrow ranges. Pups of the fossorial rodent Ctenomys talarum are altricial, making them an excellent model to evaluate the development of thermogenic capacity. In this study, the ontogeny of the thermogenic capacity of pups of C. talarum was evaluated. Using respirometry techniques, non-shivering thermogenesis (NST), total thermogenic capacity (cold-induced maximum metabolic rate, MMR), and resting metabolic rate (RMR) in pups until post-weaning age (day 60) were analyzed. No NST was present in pups until day 60 despite the presence of molecular markers for NST in brown adipose tissue deposits, which became functional in adults. Although pups are altricial at birth, they maintain their thermal balance behaviorally during lactation. Total thermogenic capacity became fixed at an early age, indicating an improvement in shivering thermogenesis (ST) efficiency after day 10, which might be related to the development of musculature related to digging. Before the aboveground dispersal period (~day 60), pups gradually reached adult Tb by improving ST and thermal isolation, allowing them to confront climatic fluctuations on the surface.


Introduction
The energy budget, that is, a sum of processes associated with energy intake and expenditure, has enormous relevance to individual fitness since it integrates all aspects of the life of organisms. The changes in the energy budget during the lifespan are related to requirements and the use of time and space (Withers et al., 2016). Thus, partitioning the available energy according to short-term priorities at a specific moment allows animals to survive and optimizes long-term reproductive success (Wade and Schneider, 1992). In endotherms, the development of thermoregulatory mechanisms after birth is the determinant of the adult's fitness. The development of functional mechanisms that ensure thermal homeostasis is one of the main requirements that must be achieved before parental care independence (Thompson, 1992). Defined by their physiological maturity, different strategies for the development of endotherm pups have been proposed to exist between altriciality and precociality extremes (Hill, 1992). The distinction between altricial and precocial development is related to the time at which pups become furred, possess capable eyesight, coordinate their locomotor abilities, or, particularly, reach endothermy (McClure and Randolph, 1980). Precocial pups can use physiological thermoregulation at an early stage, whereas altricial pups show a low capacity to control their body temperature (T b ) due to the lack of functional thermogenic mechanisms, losing heat rapidly when they face an ambient temperature (T a ) below thermoneutral zone (TNZ; Hill, 1992). In the case of altricial pups, the requirement for warmth might be provided by the exogenous source of heat. Direct physical contact of pups with their parents might help to overcome the thermal imbalance. Thus, when physiological thermogenic capacity is limited, behavioral thermoregulation could restrict heat loss, for example, either by huddling with other members of the litter (Geiser and Kenagy, 1990), by behavioral thermotaxis (Farrell and Alberts, 2007), or by remaining in the nest (Antinuchi and Busch, 2001). Total thermogenic capacity is defined as the maximum amount of heat that can be produced to maintain a constant T b and can be approximated by cold-induced maximum metabolic rate (MMR). The components of total thermogenic capacity are shivering. Shivering thermogenesis involves rapid, involuntary contractions of skeletal muscles to produce heat, related to biochemical-mechanical processes at the cellular level, such as inefficiency of biochemical pathways or friction among myofibrils (refer to Hohtola, 2004). Unlike ST, NST has specialized tissues to produce heat without any muscle contraction. Brown adipose tissue (BAT) is the main source of NST (Janský, 1973;Foster and Frydman, 1979; but see Nowack et al., 2017 for a discussion about musclemediated NST). BAT is a highly vascularized tissue with a large number of mitochondria and respiratory enzymes and is found in discrete areas in the body, mostly surrounded by the main organs (Cannon and Nedergaard, 2004). The activation of BAT depends on the release of norepinephrine (NE) by the sympathetic nervous system, and its functioning is related to uncoupling the oxidative phosphorylation in BAT's mitochondria. The presence of uncoupling proteins (UCP), particularly UCP1, short-circuits the electron transport chain circuit, producing heat (Oelkrug et al., 2015;Chouchani et al., 2019).
Depending on the pup's development strategy, the onset of functional total thermogenic capacity can vary (Brück and Hinckel, 1996). While thermogenic mechanisms in precocial pups are well developed right after birth, complete development takes days in altricial pups (Berthon et al., 1994;Farrell and Alberts, 2007;Canals et al., 2009;Lezama-García et al., 2022). For example, in rats, behavioral control of T b is important during the first days of life; by the second week, with the increase of body mass and thermal insulation, the appearance of ST enhances thermogenic capacity (Sant Anna and Mortola, 2003). In the white-footed mouse Peromyscus leucopus, NST appears by day 10, whereas ST was observed around 2 weeks after birth (Hill, 1976;. Deer mice Peromyscus maniculatus from a low-altitude population can shiver at day 14, but only NST is observed in individuals from a high-altitude population at this age . Given the variation between altricial species, it is difficult to establish a temporal pattern of postnatal development of different thermogenic mechanisms, but it is generally recognized that, in mammal pups, shivering is not functional at early stages, but NST tends to be active. Deposits of BAT are well developed and matured only a few days after birth, contrary to muscles, which show a slow development during ontogeny (Lagerspetz, 1966;. In this general framework, inhabiting a relatively thermally stable environment can diminish the cost of thermoregulation. Some mammalian species live in underground burrows that can buffer cold external T a s and its fluctuation. In particular, subterranean rodents occupy almost all available habitats, from cold to tropical, from mountain slopes to deserts, inhabiting tunnel systems parallel to the surface. Although subterranean burrows buffer T a and protect against predators, they are stressful in most cases because they are humid, dark, stagnant, and lack most clues available for orientation . Among rodents, different degrees of adaptation to the use of burrows are present , with some species adapted to digging and living almost exclusively underground (i.e., subterranean) and some other species also adapted to digging but with extensive use of surface (i.e., fossorial). Either subterranean or fossorial rodents, underground ecotope had determined the observed convergence in morpho-physiological features observed between both groups of rodents (e.g., low mass-specific minimal metabolic rate, a wide range of thermoneutrality, high thermal conductance, and low Tb; McNab, 1966;Vleck, 1979). Different studies have been carried out for evaluating the total thermogenic capacity, ST, and NST in adults of subterranean and fossorial species, finding significant variability in the type and proportion of the mechanisms on total thermogenic capacity. While in some South American Ctenomyid species, a combination of NST and ST is observed (e.g., C. porteusi), in other species, only ST is present (C. tuconax; Luna et al., 2019). Similarly, in African bathyergids, the cost of NST varies depending on the body mass and sociality (Luna et al., 2021). However, no studies have been performed to evaluate the ontogeny of total thermogenic capacity and the nature of thermogenic mechanisms involved in pups of subterranean or fossorial species.
The genus Ctenomys is distributed in several environments in the southern cone of South America, being one of the most speciose genera among subterranean and fossorial rodents, including species from ∼100 g (C. pundti) to more than 1,000 g (C. conoveri; Reig et al., 1990). Particularly, Ctenomys talarum is a solitary fossorial rodent that lives in the coastal grasslands of Argentina (Antenucci and Busch, 1992). Similar to other Ctenomys species, tunnel systems show a linear pattern with no differences in total burrow length between sexes, feeding mainly on aerial plant parts near the burrow openings (Antenucci and Busch, 1992;Comparatore et al., 1995). Males of C. talarum established a dominance hierarchy among them, monopolizing mating activity, and consequently, dominant male burrows occupied central distribution surrounded by several female burrows Antenucci and Busch, 1992). The reproductive season lasted ~9 months, beginning in June-July (austral Frontiers in Ecology and Evolution 03 frontiersin.org winter) and continuing until February (austral summer). Birth takes place from August through early December after a long gestation period (~100 days; Malizia and Busch, 1991;Fanjul et al., 2006). Females produced as many as two litters per reproductive season, with a mean litter size of 4 . For C. talarum, the cost of acceding to new food patches made through the construction of new tunnel systems is relatively high (Luna et al., 2002). During the ontogeny, C. talarum pups show an initial altricial phase, represented by a period of high thermal vulnerability. During lactation (~10 days), pups are not capable of maintaining T b on their own, depending, therefore, on their mother to thermoregulate. After this period, pups enhance their thermoregulatory capacity and start eating solid food. At approximately 30-40 days after birth, pups reach the T b of adults and become behaviorally and physiologically independent. At this age, the weaning period starts, and pups show active exploratory behavior. Around day 60, solitary pups disperse aboveground from their maternal burrow and construct their burrow (Malizia et al., 1995;Zenuto et al., 2002;Antenucci et al., 2003;Cutrera et al., 2003;Baldo et al., 2014). In this context, the aim of this study was to evaluate the thermogenic capacity development of pups of the fossorial rodent Ctenomys talarum. We also evaluated the presence of BAT during ontogeny and the timing of mitochondrial uncoupling proteins related to BAT-mediated NST. We hypothesized that no thermogenic capacity is attained during the altricial phase of their development (~day 10), increasing their thermogenic capacity after weaning, before the dispersal period. Furthermore, we expect that NST is completely functional after day 10, reaching the value of the adult during this period.

Animal capture and husbandry
Pregnant females were live-trapped in Mar de Cobo (37°45'S, 57°56'W, Buenos Aires Province, Argentina). In the laboratory, females were housed in individual cages (0.30 × 0.40 × 0.25 m), with wood shavings as nesting material. A total of 27 pregnant females were used in this study. The animal room was kept at 25 ± 1°C, and the photoperiod was LD 12:12 (lights on at 7.00 a.m.). Animals were fed with mixed native grasses, carrots, lettuce, corn, alfalfa, and sunflower seeds ad lib. As C. talarum does not drink free water, it was not provided. Pregnant females were monitored daily until pups were born. After birth, pups were kept in the same cage with their mother and siblings until day 60, after which the remaining pups were placed in individual boxes until they were released in the field. On days 2, 6, 10, 15, 30, and 60 after parturition, 1 pup from a randomly assigned mother was weighed, and metabolic variables and T b were measured. To avoid possible maternal effects during the experiment, only one pup of each litter on each day was used in NST max and MMR trials. Due to the difficulty in maintaining pregnant females and obtaining parturition in the laboratory, we used 4 pups from 4 different females that delivered more than 4 pups. In this case, 1 pup was used between days 2 and 6, and another between days 30 and 60. A total of 31 pups were used to estimate the pattern of NST max and MMR. The sex of the pups in the experimental design was not considered, using both males and females independently, because sex cannot be stablished externally in the early stages of development. Given that our primary focus was thermogenic capacity, we used another set of pups (7 pups from different litters) for the RMR experiment. This set of pups was maintained in the same husbandry conditions and was related to an undergoing study on thermoregulatory physiology.
Resting metabolic rate and non-shivering thermogenesis of pups Oxygen consumption was measured using a computerized positive pressure open-flow respirometry system (Sable System, United States). A cylindrical chamber (volume 500 mL) was used to estimate O 2 consumption. Chamber was placed inside an incubator (Simedix, model I-501, Ingelab, Argentina). Ambient temperature (T a ) was maintained with an accuracy of 0.2°C. The chamber received dry and CO 2 -free air ranging from 300 mL min −1 to 600 mL min −1 , depending on the body mass of the pups, from a mass-flowmeter (Side-Trak Sierra, model 830/840, Sierra Instruments, United States). To scrub CO 2 and water, the air was passed through a CO 2 absorbent (self-indicating soda lime, Laboratorios IQB, Argentina) and water scrubber (Drierite, W. A. Hammond Drierite Company, United States) before and after passing through the chamber. Excurrent air from the chamber was subsampled at 80 ± 10 mL min −1 , and oxygen consumption was obtained from an Oxygen Analyzer FC-1B every 1 s by an EXPEDATA data acquisition program (Sable System, United States). Rates of oxygen consumption were calculated using equation 4a of Withers (1977). All metabolic trials were performed between 9:00 and 17:00 h. The body mass of individuals (M) was measured before each metabolic trial using an electronic balance (model FX-3000, ± 0.01 g, A&D Company Limited, United States), and the pharyngeal temperature was recorded at the beginning and the end of each experiment with a YSI probe (model 93 k73545-402) connected to a Cole-Parmer thermistor meter (model 8,402-10, ± 0.1°C, Cole-Parmer Instrument Company, United States).
Resting metabolic rate was estimated as the lowest 5-8 min steady-state O 2 consumption of a total trial period of 35-40 min. A total trial period was chosen because C. talarum pups are highly dependent, during the early stage of their development, on mother and sibling contact (Zenuto et al., 2002), showing signs of hypothermia if pups are alone for more than 30 min . However, brief isolation did not affect the development of thermoregulatory physiology (Baldo et al., 2014). In addition, total chamber volume and air flow allowed a short equilibration period of the chamber (2-4 min to achieve 90% of equilibration, depending on the air flow; Lasiewski et al., 1966). After being placed in the metabolic chamber, pups showed an exploratory behavior, which was performed during the period in which the metabolic chamber was equilibrating (~5 min).
Maximal non-shivering thermogenesis (NST max ) was estimated using the same chamber and protocol used to estimate RMR, but after intramuscular injection of norephinefrine (NE). In eutherians, the O 2 consumption in response to NE occurs 10 min after the injection and lasts at least 5-10 min (Feist and Rosenmann, 1976). Doses of NE were estimated according to Wunder and Gettinger (1996) and described as NE (mg Kg −1 ) doses = 2.53 M -0.4 . Doses of NE and procedures were validated for adults of this species (Luna et al., 2012). The maximum 10-min steady-state oxygen consumption after the injection of NE was considered to be NST max , which includes both RMR and thermoregulatory NST (Wunder and Gettinger, 1996). To avoid Frontiers in Ecology and Evolution 04 frontiersin.org unnecessary long-time exposition in pups, we did not assess the effect of injection on pups. Values of O 2 consumption of adults between the control group and after saline solution injection were similar (Luna et al., 2012). For RMR and NST max estimations, the metabolic chamber was maintained at T a of 25°C, which is within the thermoneutral zone of adults for this species (Busch, 1989).

Cold-induced maximum metabolic rate
Maximum thermogenic metabolism was estimated in a HELOX atmosphere according to the procedure described by Rosenmann and Morrison (1974). A mixture of He (79%) and O 2 (21%) was passed through a mass flowmeter before entering the metabolic chamber (Side-Trak, Sierra model 830/840, Sierra Instruments, United States). Before MMR estimation, the flow rate was corrected for the He-O 2 gas mixture. As described before, the mixture was passed through a CO 2absorbent and water scrubber before and after passing through the chamber. Oxygen consumption of pups was recorded for a 30-40 min period at an ambient temperature of 19°C. This T a was chosen based on previous studies that determined the lowest T a within burrows during the reproductive period when pups are born (Cutrera and Antinuchi, 2004) and because pups fall into deep hypothermia even when there are huddling .

Shivering thermogenesis
Shivering thermogenesis (ST) was estimated for each pup according to the equation proposed for eutherian mammals (MMR = BMR + NST + ST; Wunder and Gettinger, 1996). Experimentally, ST was estimated as O 2 consumption after coldinduced minus NST max .

Measurement of UCP1 and COXII content and COX activity
After NST max and MMR, pups were euthanized, and interscapular brown adipose tissue (iBAT) was carefully dissected, frozen in liquid nitrogen, and stored at −80°C. All other main organs were also frozen and stored at −80°C for further analysis. We adhered to the 2012 Revised International Guiding Principles for Biomedical Research Involving Animals developed by the Council for International Organizations of Medical Sciences (CIOMS) and the International Council for Laboratory Animal Science (ICLAS). Pups used in RMR experiments were released at the capture site after day 120. Interscapular BAT was homogenized with a manual homogenizer in STE buffer (250 mM sucrose, 5 mM Tris-HCl, 2 mM EDTA, pH 7.4) with protease and phosphatase inhibitors (10 μM Leupeptin, 10 μM Pepstatin, 0.2 mM PMSF, and 0.2 mM Orthovanadate) in a proportion of 10 mL of buffer per gram of tissue and filtered through a layer of inert gauze. Total protein content was determined using the Bradford method (Bradford, 1976). Samples of iBAT were denatured, and 40 μg of proteins per line were loaded and run in an SDS-PAGE (3% stacking gel and 12% running gel), according to Laemmli (1970), and electrotransferred onto a nitrocellulose filter, as described by Puigserver et al. (1991), using a Trans-Blot ® Turbo™ Transfer System (Bio-Rad, United States). After that, membranes were blocked in 5% nonfat powdered milk in TBS-Tween and incubated with the corresponding antibodies (UCP1, COXII antibodies, Alpha Diagnostics, United States). Protein bands on the nitrocellulose filters were visualized by Immun-Star ® Western C ® Kit reagent (Bio-Rad, United States) using the Western blotting detection systems. The chemiluminiscence signal was captured using a Chemidoc XRS densitometer (Bio-Rad, United States) and analyzed with Quantity One Software (Bio-Rad, United States). COX activity was measured using a spectrophotometric method (Wharton and Tzagoloff, 1967;Chrzanowska-Lightowlers et al., 1993). Briefly, aliquots of the iBAT homogenates were incubated in 0.1 M NaPO 4 H 2 , pH 7.0, in the presence of 2 μg/mL Catalase and 5 mM substrate DAB (3, 3′ diaminebenzidine-tetrachloride). After 30 s, 100 μM reduced cytochrome c was added to start the reaction, and the absorbance variation was recorded for 20 min at 450 nm.

Statistics
Physiological variables (NST max , MMR, and RMR) were analyzed by linear mixed-effects model with the R package "nlme" (Pinheiro et al., 2022) in R (R Core Team, 2021) inside RStudio (R Studio Team, 2020). To account for repeated measurements, individual IDs were included as a random factor when comparing NST max and MMR. Body mass was included as a covariate in the model. Assumptions were verified using residual plots and tested for normality of residuals by the Shapiro-Wilk test. To evaluate each fixed factor (treatment and day), we used hypothesis tests with a significance level set to α = 0.05 for all analyses. To explore changes in metabolic rate, the variance correction function varIdent was used. We performed a general linear mixed model with heterogeneous variance for the absolute T b obtained before metabolism measurement (cold-induced or NE injection) with the same fixed and random effects as the models detailed above, without using body mass as a covariate. To compute p-values for multiple comparisons, the R package "emmeans" (Lenth, 2022) was used when we found significant differences between means. For ΔT b , we used linear models to test for differences between days in each experimental condition (cold-induced or after NE injection). In addition, linear models were used to evaluate differences in the enzymatic activity of COX between days. Finally, we used the Kruskal-Wallis test for differences in protein content (UCP1 and COXII) in iBAT. Linear models and the Kruskal-Wallis tests were performed in the R package "R stats" in R (R Core Team, 2021) inside R Studio (R Studio Team, 2020).

Discussion
To attain thermal independence before the dispersal period, pups must fully acquire thermoregulatory capacity (Hill, 1992). The main objective of this study was to determine the thermogenic development of pups of the fossorial C. talarum, stablishing the onset of thermogenic capacity and its components. To the best of our knowledge, this is the first study to establish the postnatal development of non-shivering thermogenesis (NST) and cold-induced maximum metabolic rate (MMR) in a fossorial species. Pups of C. talarum showed a biphasic pattern of mass-specific RMR during their development. Mass-specific RMR increases until day 10, with a subsequent decline to day 60, confirming the pattern found by Zenuto et al. (2002). In this species, acquisition of thermoregulatory capacity takes days, since the onset of homeothermy is observed around day 30, a period in which pups showed T b of adults (Zenuto et al., 2002;Cutrera et al., 2003;Baldo et al., 2014). Notably, no thermogenic mechanism is observed until day 10, and NST max does not differ from RMR up to day 60.
In altricial species, it has been described that the first thermogenic mechanism to appear is NST Lezama-García et al., 2022), involving the maturation of BAT a few days after birth . In pups of the golden hamster  Mesocricetus auratus, the onset of NST occurs around 14 days after birth, whereas in the Djungarian hamster Phodopus campbelli, the onset of NST occurs after day 9 (Newkirk et al., 1995). Despite the high metabolic activity of BAT, the common rat showed a weak NST on day 5 after birth (Spiers and Adair, 1986). In the golden Syrian hamster Mesocricetus auratus, thermogenic capacity develops within a week after birth and is determined by qualitative changes in mitochondrial function related to the initiation of UCP synthesis (Houštěk et al., 1990). In newborn rats, NST capacity appears after weaning, but UCP1 is detectable in iBAT just after birth, reaching its maximum value on day 10 (Porras et al., 1990). Interestingly, C. talarum pups showed a complete absence of NST until day 60, but this does not imply the complete absence of BAT. Interscapular BAT patch is directly recognizable, increasing in size with age but never exceeding 1% of the body mass. Neonatal exposure to cold can establish modifications in NST capacity, induced by changes in neuronal stimulation and content of UCP in BAT (Morrison et al., 2000;Sant Anna and Mortola, 2003). However, even though pups of C. talarum can be exposed to cold when isolated from the mother and siblings, UCP1 content does not vary between days. In the same way, different molecular markers of the oxidative capacity of BAT ensure NST (Klingenspor, 2003) does not vary between days. For example, the activity of cytochrome c oxidase (COX), a marker enzyme for the mitochondrial membrane, and the content of COXII, commonly used to estimate respiratory capacity in BAT mitochondria (Klaus et al., 1988;Klingenspor et al., 1996), remain similar between days. Although unlikely, given the relative thermal stability of burrows, it would be worth establishing whether acclimatization to low T a s by mothers might affect the magnitude of pup's thermogenesis, as has been observed in the surface-dwelling altricial leaf-eared mouse Phyllotis darwini during the first days of life (Canals et al., 2009). Day 60 corresponds to the beginning of the period of dispersion of this species, which occurs on the surface (Malizia et al., 1995). Interestingly, the size of iBAT and the amount of UCP1, as well as markers of the oxidative capacity of BAT in adults, are similar to that in juveniles (Luna et al., 2012). Due to the complete maturation of endocrine and neural pathways (Symonds et al., 2012;Symonds, 2013;Lezama-García et al., 2022), changes in the efficiency of BAT-related NST might establish the functional NST after day 60. Complete maturation of neuro-endocrine pathways would allow an increased capacity for fatty acid oxidation and tricarboxylic acid cycle activity to uncouple oxidative phosphorylation to produce heat (Chouchani et al., 2019). For example, in P. maniculatus, a delay in the onset of NST in pups reared at high altitude was associated with a delay in the maturation of nervous and O 2 supply systems Velotta et al., 2020).
Frontiers in Ecology and Evolution 07 frontiersin.org As proposed, ST matures later than NST because of the slow development of muscles Lezama-García et al., 2022). In C. talarum pups, shivering was characteristically observed over the first days (Baldo et al., 2014) but appeared to be fully functional after day 15. Aerobic-resistant muscles that generate repeated contractions could be mainly recruited to produce heat; thus, the proportion of this type of muscle can establish ST potential (Hohtola, 2004). Muscles of altricial rodents are relatively immature at birth, and they mature due to changes in muscle fiber structure and hypertrophy (White et al., 2010). From day 10, offspring of C. talarum start to show exploratory behaviors, wandering the tunnel systems and beginning to scratch and push the soil (Vassallo, 2006;Echeverría et al., 2016). The progression of digging behavior has particular energetic requirements to loosen, remove, and break down the soil to construct a new burrow during dispersal. Similar to ST, aerobic-resistant muscles are used during digging activities (Alvarez et al., 2004). Thus, muscular work performed during digging could also be exploited as a source of heat via ST (Luna and Antinuchi, 2007). The onset of functional ST overlaps with the beginning of the exploration of tunnels near the thermally stable nest, tunnels that usually show T a s below TNZ (Cutrera and Antinuchi, 2004). In this context, even though it is not clear how muscle maturation occurs during the first days of life, the capacity for ST can be associated with an increase in digging capacity. Furthermore, cold acclimation in adult rodents was proposed to lead to muscle training and an increase in ST capacity (Shefer and Talan, 1998). As digging becomes more efficient, ST might also do so. Interestingly, from day 15 until dispersal, total thermogenic capacity relies entirely on ST, increasing its specific energetic cost due to the progressive decrease in mass-specific RMR and the constancy of total thermogenic capacity.
The total thermogenic capacity of pups from day 10 is similar to those observed in adults (Luna et al., 2012). The consistency of massspecific MMR over the lifespan of this species addresses important questions about the mechanistic meaning of physiological adjustments within upper and lower metabolic limits. A fixed metabolic ceiling determines that any energetic modification must be within the aerobic scope. The net aerobic scope represents the capacity of aerobic processes above minimal metabolism (see Nespolo et al., 2017 for a discussion). For example, unlike juveniles, adults of C. talarum show a functional NST that represents ~40% of the total thermogenic capacity (Luna et al., 2012). Thus, from day 10, when the metabolic ceiling is established, there is an increase in the net aerobic scope given the progressive decrease in mass-specific RMR. Any energy input intended for other processes, in addition to growth, must be within metabolic limits. In fact, the part of the energy cost of NST in adults could be related to the difference between mass-specific RMR at day 60 (1.75 mL O 2 /gh; this study) and in adults (0.89 mL O 2 /gh; Luna et al., 2012). Such metabolic trade-offs between different energyconsuming mechanisms might ensure a balanced daily energy expenditure, hence survival. Maximal metabolism is an ecologically relevant trait (Bozinovic and Rosenmann, 1989), and although it appears to be infrequently reached during life, maximal metabolism can be affected by natural selection (Rezende et al., 2004). It is not the focus of this study to evaluate which factor determines mass-specific MMR constancy, but it has been proposed that there would have been an evolutionary limitation on reaching high MMR values (Luna et al., 2012(Luna et al., , 2019 because of limitations in respiratory and cardiovascular systems imposed by burrow's atmosphere (i.e., low O 2 and high CO 2 ). Therefore, the value of maximal metabolism might be independent of the physiological mechanism that elicits it. In adults of C. talarum, cold-induced MMR also does not vary after acclimation to cold, being similar to exercise-induced MMR (see Luna et al., 2012). As proposed, further studies are needed to assess the consequences of fixed ceilings and net aerobic scope in subterranean and fossorial species when compared to surface-dwelling species (Luna et al., 2015(Luna et al., , 2019. Heat balance in endotherms results from the integration of a coordinated system of heat production and loss (Naya et al., 2013). One of the main characteristics of underground rodents is their high Relationship between the mass of interscapular brown adipose tissue patch (iBAT) and body mass during postnatal development in the fossorial Ctenomys talarum. The solid line represents the regression line. Frontiers in Ecology and Evolution 08 frontiersin.org thermal conductance (see McNab, 1966). Thus, not only changes in thermogenic capacity but also thermal insulation are important to maintaining thermal homeostasis (Withers et al., 2016). Early development in altricial pups represents a stage of dependence on external thermal sources (MacArthur and Humphries, 1999;Antinuchi and Busch, 2001;Zepeda et al., 2018), which mostly depends on thermal resources provided by the mother and behavioral thermoregulation, such as huddling with their siblings (Zenuto et al., 2002). The effectiveness of huddling depends not only on the reduction of the individual's exposed surface but also on the augmented heat transfer between littermates and the mother (Webb et al., 1990;Gilbert et al., 2007). In the altricial P. darwini, a strong maternal effect on the thermoregulatory development of pups was observed until day 7 (Canals et al., 2009). McClure and Randolph (1980) found in the eastern wood rat Neotoma floridana pups that poor insulation after birth is related to the maximization of heat transfer. In C. talarum, as no thermogenic mechanisms are observed until day 10, behavior can alleviate the need for these energetically costly mechanisms. Until day 10, hairless pups spend most of the time in contact with their mother and siblings Baldo et al., 2014). Zenuto et al. (2002) established that the increase in mass-specific RMR until day 10 was related to the increase in fur coverage and the increase in the intake of solid food after day 7. After day 10, pups remained huddled for shorter periods of time due to the increment of agonistic aggressive behavior between them Baldo et al., 2014). Thus, the onset of ST also coincides with the time that pups start to spend less time huddled and become furrier, increasing their insulation. Therefore, thermoregulatory benefits associated with the underground environment might allow a short period where behavioral strategies are important to avoid thermogenic costs. During lactation, pups might invest energy to accelerate the production of new tissue and its maintenance (Hill, 1992), without compromising their thermal balance. The thermally stable environment of the nest and behavior might lead to a delay in the onset of ST, hence independent thermoregulation, until the period in which pups find new thermal environments by exploring and digging nearby tunnels.
In conclusion, although pups of C. talarum were altricial at birth (Zenuto et al., 2002;Cutrera et al., 2003;Baldo et al., 2014), they maintained behaviorally their thermal balance during lactation. Until day 10, pups are dependent on their mother to obtain heat, remaining in the nest where they commonly huddle with their littermates Baldo et al., 2014). During this period, resting metabolism increases, but no thermogenic mechanisms were observed. As most of the energy is destined for growth, pups at this age showed impossibility to maintain T b when confronted with T a s below adults TNZ. From day 10, mass-specific RMR showed a progressive decrease and MMR became fixed, establishing a specific increment of ST until day 60. Postnatal development of musculature related to digging could be employed as a heat source when pups start to explore and dig in the surrounding of the nest, but it appears to be inefficient until day 30 when pups achieved adult T b . No functional NST was observed in pups until day 60 besides the presence of UCP1 and other molecular markers (e.g., COX activity) of NST in BAT. Attaining complete thermal independence is critical after day 60 since juveniles disperse at this age (Malizia et al., 1995). During their aboveground dispersal, juveniles must dig a new burrow or eventually find an empty burrow, so they are exposed to aboveground climatic fluctuations. Interestingly, in adults, NST complies with around 40% of the total thermogenic capacity (Luna et al., 2012), so, at some point during or after dispersal, NST becomes functional.

Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics statement
The animal study was reviewed and approved by the Institutional Committee for the Care and Use of Laboratory Animals (number 2555-06-14 RD141/15; CICUAL), Facultad de Ciencias Exactas y Naturales (FCEYN), and Universidad Nacional de Mar del Plata (UNMDP).

Author contributions
FL designed the study. CDA supported the respirometric experimental procedures. JS-S and JO supported the molecular experimental procedures. FL and CDA analyzed the data, prepared the manuscript, and created figures and tables. All authors contributed to the article and approved the submitted version.

Funding
This project was supported by the Consejo Nacional de Investigaciones Científicas y Técnicas of Argentina (CONICET, PIP3164) and the Agencia Nacional de Promoción de la Investigación, el Desarrollo Tecnológico y la Innovación (Agencia I+D+i, PICT 3785).
Frontiers in Ecology and Evolution 09 frontiersin.org