Severe thermoregulatory deficiencies in mice with a deletion in the titin gene TTN

Defense of a constant core body temperature (T_b) by generating or dissipating heat is the defining feature of homeothermic endotherms. Thermoregulatory heat production and heat loss allows small animals to maintain thermal constancy even when challenged with temperatures outside of their thermal neutral zone. Large deviations in T_b can have severe consequences, such as reduced enzyme efficiency, altered diffusion capacity and changes in membrane fluidity. These critical cellular functions can result in loss of consciousness, inability to coordinate and execute motor functions, and even death (Morrison and Nakamura, 2011). T_b is maintained during cold exposure by vasoconstriction of peripheral vessels, piloerection, postural changes, shivering thermogenesis (ST) and regulatory non-shivering thermogenesis (NST; Hemingway, 1963). A previous study found that mice with a mutation in the gene for the muscle protein titin (TTN) exhibit decreased tremor frequency during ST and could not maintain euthermic T_b when exposed to ambient temperatures (T_a) <34°C (Taylor-Burt et al., 2015). In this study, we investigated the effect of the titin mutation on the other heat-generating mechanism, NST.

NST is a heat-generating mechanism that occurs through the uncoupling of oxidative phosphorylation and ATP production by uncoupling protein-1 (UCP1) in brown adipose tissue (BAT). UCP1 is expressed among many animals, including rodents (Cannon and Nedergaard, 2004; Depocas, 1960; Golozoubova, 2006; Lowell et al., 1993; Meyer et al., 2010; Nedergaard et al., 2001; Nicholls and Locke, 1984). Like basal metabolic rate (BMR), the contribution of NST to summit metabolism (summit oxygen consumption rate, V̇_O2sum) as well as V̇_O2sum itself, scales with body mass (Wunder and Gettinger, 1996). Thermogenic capacity, which can be approximated by V̇_O2sum, provides an estimate of an animal's ability to thermoregulate and refers to the total capacity for heat production during cold exposure. Thermogenic capacity can be measured when both ST and NST are maximally activated during acute cold exposure. The house mouse (Mus musculus), has a small body mass, which leads to relatively high thermal conductance and a greater contribution of NST to V_O2sum in order to offset heat loss in comparison with larger animals (Wunder and Gettinger, 1996).

The muscular dystrophy with myositis (mdm) mouse has a 779-bp deletion in the N2A region of the titin gene TTN (Garvey et al., 2002), and a previous study demonstrated severe thermoregulatory deficiencies (Taylor-Burt et al., 2015). Mice homozygous for the mdm mutation exhibit severe and progressive degeneration of skeletal muscles and lower body mass, stiffer gait and reduced lifespan compared with wild-type (WT) littermates (Garvey et al., 2002; Huebsch et al., 2005; Lopez et al., 2008). Muscle in mdm mice has higher passive stiffness than WT muscle (Hessel et al., 2017; Lopez et al., 2008; Monroy et al., 2017). Taylor-Burt et al. (2015) demonstrated in vitro that mdm mice have lower active muscle stiffness as well and a lower tremor frequency during ST. It is expected that if titin did not contribute to active muscle stiffness, then the titin mutation in mdm mice would not affect tremor frequency. However, tremor frequencies are dramatically reduced in mdm mice (Taylor-Burt et al., 2015). The deficiency in shivering frequency likely leads to the observed hypothermic state of mdm mice at temperatures below 34°C. The combination of increased thermal conductance and lower heat generation via ST results in an offset between heat production and heat loss, resulting in a failure to adequately thermoregulate even at T_a values modestly below the T_b set-point. Although no studies have previously investigated NST in mdm mice, the hypothermic state of mdm mice motivated us to determine whether an impairment in NST is also contributing to the decreased heat production.

We set out to determine whether mdm mice have impaired NST and whether this contributes to their observed hypothermia in addition to impaired ST and higher thermal conductance as compared with the WT mouse. We tested two hypotheses: (1) MR of mdm mice is greater at higher T_a than WT mice owing to a shifted thermoneutral zone (TNZ) of mdm mice; and (2) the inability to defend T_b cannot be explained solely by lower shivering frequencies; therefore, an impairment in the other heat-production mechanism, NST, is present.

Animal experiments were approved by the Institutional Animal Care and Use Committee of Northern Arizona University, under protocol 05-014-R3, which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, International.

Breeding pairs of B6C3Fe a-/a-mdm mice (Mus musculus Linnaeus 1758) from the Jackson Laboratory (Bar Harbor, ME, USA) were housed on a 14 h:10 h light:dark cycle at 23–24°C at Northern Arizona University. In this study, all animals were housed in an environmental chamber set to 34°C upon weaning, and experiments were conducted during the dark phase of the light cycle, when mice are active. WT mice were fed LabDiet 5001 Laboratory Rodent Diet and mdm mice were fed LabDiet 5LJ5 PicoLab High Energy Mouse Diet (LabDiet, St Louis, MO, USA). mdm mice were provided the high-fat-diet chow owing to high mortality rates when fed standard laboratory chow. WT mice were housed singly or with siblings and mdm mice were housed with the mother and other mdm siblings owing to mortality if separated. Each cage was equipped with bedding and enrichment.

Sample sizes for WT and mdm mice at each experimental temperature were as follows: 20°C (n=7 WT, n=5 mdm), 24°C (n=6 WT, n=7 mdm), 29°C (n=6 WT, n=8 mdm) and 34°C (n=7 WT, n=8 mdm), unless otherwise noted. Body mass differed significantly between genotypes (Welch's test, P<0.001), with mdm mice (7.2±0.21 g; n=9) at approximately one-third of the mass of the WT mice (26.0±1.82 g; n=7).

To measure core T_b, recording devices were surgically placed in each mouse's peritoneal cavity at the age of 25–30 days. Different temperature recording devices were used because of the smaller size of the mdm mice compared with WT siblings. mdm mice were implanted with a Biothermo13 passive integrated transponder (PIT) tag (0.109±0.030 g; Biomark, Boise, ID, USA). The PIT tag measured and recorded T_b at a resolution of ±0.5°C. WT mice were implanted with a PhysioTel TA-F10 telemetry device (1.6 g; Data Sciences International, St Paul, MN, USA), which measured T_b at a resolution of ±0.05°C. The surgical protocol was modified from Richter et al. (2017). Briefly, mice were anesthetized with isoflurane USP (MWI Veterinary Supply, Boise, ID, USA) using a Forane vaporizer (Ohio Medical Corporation, Gurnee, IL, USA). Anesthesia was induced by placing the mouse in a small chamber supplied with a 1–2% isoflurane–oxygen mixture. Upon unconsciousness, the mouse was transferred to a heating pad, where anesthesia was administered using a nosecone that delivered a 1–2% oxygen–isoflurane mixture. Once pedal reflex was absent, the abdominal fur was shaved and the exposed skin was scrubbed with povidone iodine and 70% ethanol. The skin and muscle were incised and bluntly dissected to expose the peritoneal cavity. The device was then placed freely inside the peritoneal cavity and the muscle and skin were closed with 3-0 Maxon absorbable monofilament suture. Buprenorphine, an analgesic agent, was administered subcutaneously (0.10 mg kg⁻¹) for 72 h following surgery. Post-surgery recovery lasted for 7 days at 34°C before further experiments were conducted.

Respirometry was conducted as in Richter et al. (2015) and modified for mice. Briefly, mice were placed in 9 liter metabolic cages during the experiments. Oxygen consumption rate (V̇_O2) and carbon dioxide (CO₂) production were measured using an open-flow respirometry system (Promethion, Sable Systems International, Las Vegas, NV, USA). Each cage had its own gas analysis chain for respiratory gases comprising a flow controller, capacitive water vapor partial pressure analyzer, spectrophotometric CO₂ analyzer and fuel cell O₂ analyzer. Water vapor was continuously measured and its dilution effect on O₂ and CO₂ was compensated mathematically. The flow rate was maintained at 2 l min⁻¹ and subsampled at 250 ml min⁻¹, resulting in ∼12 complete air exchanges per hour. The gas analyzers were calibrated using 100% N₂ and a span gas with a known concentration of CO₂ mixed with N₂ prior to each set of experiments for an animal. The cages were housed inside an environmental chamber (Percival Scientific, Perry, IA, USA) so that T_a could be controlled. Bedding was removed from each cage prior to metabolic trials. Cages sat on top of either an antennae or telemetry receiver (Biomark; Data Sciences International) to collect T_b data throughout the experiment. T_b was recorded once per minute using either PhysioTel RPC-1 Receivers for the PhysioTel TA-F10 telemetry devices or HPR Plus PIT Tag Reader and Antennae for the BioTherm13 PIT tags (Data Sciences International; Biomark). T_a was recorded once per minute using a datalogger (Onset HOBO Data Logger, Bourne, MA, USA) with a resolution of ±0.14°C.

V̇_O2 and CO₂ were recorded once per second using SableScreen v3.3.11 acquisition software, with 60 samples averaged per minute for final analysis (Sable Systems International). V̇_O2 (ml min⁻¹) is expressed as ml O₂ g⁻¹ body mass h⁻¹ for each animal and is referred to hereafter as MR. Raw data were processed using ExpeData v1.8.4 (Sable Systems International). A subset of 10 min of each metabolic trial was selected for analysis based upon the following criteria: (1) recordings occurred after the fasting period; (2) T_a did not exceed the specified temperature range for that experiment (20, 24, 29 and 34°C); and (3) the 10 data points for oxygen consumption (ml O₂ g⁻¹ h⁻¹) occurred when animal movement was minimal. Movement was assessed by BXYZ-R beam arrays (Sable Systems International).

All mice underwent four 4-h (mdm) or 5-h (WT) experiments where T_a was manipulated once per day for four consecutive days. To minimize the effect of diet-induced thermogenesis, WT mice were fasted for 2 h at the onset of the experiment and mdm mice were fasted for only 1 h due to their fragility. All mice were weighed before and after each experiment. When calculating mass-specific MR, mass was assumed to be lost linearly across the experimental period. The experimental temperatures (20, 24, 29 and 34°C) were chosen based on a previous study which found that 34°C was the lower critical limit of the TNZ in mdm mice (Taylor-Burt et al., 2015). The T_a=34°C was used to establish resting MR for this study.

To compare the capacity for NST between genotypes, we first fasted mice either for 1 or 2 h (mdm and WT, respectively; Speakman, 2013) at 34°C, measured MR (as above) and then briefly removed them from the metabolic cages to subcutaneously administer 1.2 mg kg⁻¹ of noradrenaline (Wunder and Gettinger, 1996). Mice were then placed back into the metabolic cages and measured for MR and T_b over the next 2–3 h. The duration of the metabolic curve, referred to as total effect time, as well as time to peak MR and peak MR were measured to test for differences in noradrenaline response between genotypes. In addition, peak T_b and average T_b were compared between genotypes after noradrenaline injection to test whether T_b followed the same trend as MR. Peak T_b was defined as the highest 60 s average, and average T_b was the average T_b during the test.

Statistical analysis was performed using JMP v12 (SAS Institute, Inc., Cary, NC, USA). Values are reported as means±s.e.m. and an alpha level of 0.05 was used for all tests. For statistical analysis, body mass was normally distributed but variance of residuals was unequal between groups, so a Welch's test, where the means are weighted by the reciprocal of the group mean variances, was used to identify significant differences. Throughout the study, the mdm group at 20°C had a large within-group variance with an outlier, leading to unequal variances when compared with WT mice. Therefore, we used two statistical methods to analyze V̇_O2 with and without the outlier. For our primary and more conservative method, data were ranked using temperature as a blocking variable (20, 24, 29 and 34°C). A mixed-model ANOVA with subject nested within genotype and fixed effects of temperature and genotype was conducted for T_b, MR and Q₁₀ effects data. Steel–Dwass multiple comparisons were used to identify significant differences between genotypes at all temperatures. In our secondary method with the outlier removed, residuals of the V̇_O2 data were normally distributed with equal variance; therefore, data were analyzed using two-way ANOVA with subject nested within genotype and fixed effects of temperature and genotype. Tukey’s HSD test was conducted post hoc. Removing the outlier for the E–O MR data did not change statistical significance, so this was not reported. For the noradrenaline-stimulated thermogenesis data, Welch's test for unequal variances was used to identify significant differences in peak T_b, average T_b and time to peak. t-tests were used for peak MR, total effect time and NST capacity. To identify significant differences in absolute contributions to thermogenic capacity, t-tests were used to compare ST and NST between genotypes.

For calculations of noradrenaline-stimulated thermogenesis, subsequent rise in MR after noradrenaline injection yielded a metabolic curve from which thermogenic capacity due to NST was calculated by taking the integral of the increase in MR after noradrenaline injection, computed numerically using the trapezoidal rule. The average of 10 consecutive minutes of MR measures that met the criteria (as above) before the animal was removed from the cage was considered their baseline MR. This was subtracted from the calculation to account for animals that had higher or lower V̇_O2 before injection. The area calculation began with the onset of injection and ended when V̇_O2 reached previously calculated baseline values.

mdm mice exhibited hypothermic characteristics, as shown by their lower T_b values in comparison with WT mice at lower T_a. There was a significant effect of genotype and T_a on T_b (P<0.001) but no interaction between T_a and genotype. We found a significant effect of subject (P<0.05). Post hoc tests revealed significant differences between genotypes at all temperatures (P<0.05; Fig. 1). At 20, 24, 29 and 34°C, we observed the following differences in T_b between WT and mdm mice, respectively: 8.8, 4.2, 1.9 and 1.4°C.

At 24°C, thermoregulatory mechanisms in mdm mice are insufficient to maintain constant and euthermic T_b, resulting in significantly lower MR (Fig. 2) and T_b (Fig. 1) as compared with WT mice. In addition, mdm mice activate ST and NST more than WT mice at 29°C, as demonstrated by their higher MR. Despite their higher MR, their T_b remained lower than that of WT mice.

There was a significant interaction between genotype and temperature on MR (P<0.05). The MRs of WT and mdm mice did not significantly differ at 34°C (P>0.05). At 29°C, the MR of mdm mice (2.4±0.12 ml O₂ g⁻¹ h⁻¹) was significantly higher than that of WT mice (1.9±0.09 ml O₂ g⁻¹ h⁻¹; P<0.05). At 24°C, the MR of WT mice (3.9±0.19 ml O₂ g⁻¹ h⁻¹) was significantly higher than that of mdm mice (3.1±0.34 ml O₂ g⁻¹ h⁻¹). WT mice (4.8±0.29 ml O₂ g⁻¹ h⁻¹) did not have significantly different MRs from mdm mice (3.5±1.51 ml O₂ g⁻¹ h⁻¹) at 20°C (P=0.1939); this was likely due to the large variability in mdm MR at 20°C compared with WT mice (Fig. 2). When the mdm outlier at 20°C was removed, MRs of WT mice at 20°C were significantly higher than those of mdm mice. Additionally, the mdm mean MR dropped to 2.9±0.89 ml O₂ g⁻¹ h⁻¹ as opposed to 3.5±1.51 ml O₂ g⁻¹ h⁻¹.

In general, mdm mice had MR_observed values far lower than MR_expected values, with the largest differences in E–O MR at the lower temperatures (Fig. 3). WT mice tended to remain close to MR_expected, except at lower temperatures, where MR_observed was higher than predicted. The higher E–O MR values for WT mice at lower temperatures are likely a result of certain subjects having higher activity, and thus higher MR at lower temperatures (observational).

A significant effect of genotype was found on E–O MR: mdm mice had significantly higher E–O MR compared with WT mice at all temperatures (P<0.001). The largest differences in E–O MR were seen at the two lowest temperatures. At 24°C, mdm mice (1.0±0.08 ml O₂ g⁻¹ h⁻¹) had significantly higher E–O MR than WT mice (−0.4±0.10 ml O₂ g⁻¹ h⁻¹; P<0.05), which had higher observed values than predicted, making their E–O MR negative. At the lowest temperature of 20°C, mdm mice (3.0±0.46 ml O₂ g⁻¹ h⁻¹) failed to attain their expected MR. WT mice, in contrast, had higher observed than expected values (0.4±0.32 ml O₂ g⁻¹ h⁻¹; P<0.05).

Both WT and mdm mice exhibited similar metabolic profiles following noradrenaline injection, but WT mice had a significantly higher capacity for NST, found using area under the curve (143.9±20.01 ml O₂ g⁻¹ h⁻¹), than mdm mice (62.3±22.78 ml O₂ g⁻¹ h⁻¹; P<0.05; Fig. 4). There were no significant differences between genotypes in either latency to reach peak MR (P=0.60) or the total effect time (P=0.08; Fig. 4). WT mice took 17.3±2.20 min to reach peak MR and mdm mice took 22.8±9.65 min. Total effect time for WT and mdm mice was 66.3±4.92 min and 48.3±10.44 min, respectively. Interestingly, genotypes did not differ in peak MR reached after injection (P=0.21), with WT mice reaching 7.2 ml O₂ g⁻¹ h⁻¹ and mdm mice reaching 6.4 ml O₂ g⁻¹ h⁻¹. However, WT mice (39.7±0.08°C) reached a higher peak T_b (P<0.05) than mdm mice (38.5±0.35°C) during the trial. In addition, WT mice had a significantly higher average T_b (38.7±0.24°C) than mdm mice (37.3±0.33°C; P<0.05).

In this study, we examined the capacity for NST in mdm mice to investigate whether this component of heat production is impaired. Our results demonstrate that a deletion in the titin gene TTN not only results in a deficiency in ST, but it impacts NST as well, likely contributing to lower T_b, MR and thermogenic capacity. Previously, it was shown that the capacity for ST of mdm mice is compromised (Taylor-Burt et al., 2015). We confirmed the inability of mdm mice to maintain homeothermy by measuring their T_b and MR across a range of T_a values. Compared with WT mice, mdm mice had lower T_b values at all T_a values and lower MRs at 24°C. Because mdm mice could not defend core T_b, we observed significant Q₁₀ effects on MR. E–O MR was significantly higher for mdm mice. Further, we found that the capacity for NST was significantly lower in mdm mice compared with WT mice, indicating that this component of heat generation is also impaired. The relative contributions of basal MR, ST or NST to V̇_O2sum did not differ between WT and mdm mice.

mdm mice had significantly lower T_b values than WT mice, even at 34°C (Fig. 1). These findings suggest that in order for mdm mice to have a euthermic T_b set-point similar to that of WT mice, T_a needs to be even higher than 34–35°C, as previously reported by Taylor-Burt et al. (2015) for mdm mice to maintain euthermy.

We do not think that mdm mice are entering torpor at the temperatures they were tested. A reduction in T_b set-point due to torpor is unlikely owing to the pattern of MR and T_b observed. Torpor is characterized by a ‘decided’ reduction in T_b set-point that leads to a precipitous decline in MR and T_b (Geiser et al., 2014). During hypothermia, T_b set-point is at euthermic levels, but the animal cannot support the metabolic heat production needed to maintain euthermia; therefore, both T_b and MR decrease slowly as thermoregulatory mechanisms fail to meet demand and then decline rapidly once the body cools further. mdm mice exhibited increases in MR upon cold exposure followed by a gradual reduction, which was likely hypothermia. This indicates that thermoregulatory mechanisms of mdm mice are insufficient to maintain T_b. Hudson and Scott (1979) measured T_b and MR in M. musculus and found that torpid mice with T_b of 32°C had a metabolism 50% of what was observed at euthermic T_b levels. The mice in our study had similar T_b values of ∼29.3°C at 20°C but had higher MRs of ∼3.5 and ∼2.0 ml O₂ g⁻¹ h⁻¹ at 20 and 34°C, respectively, suggesting that mdm mice lack sufficient thermoregulatory capacity at T_a values <34°C and do not adopt the behavioral strategy of hypometabolism through torpor.

Studies have shown thermoregulatory deficiencies specific to NST in UCP1-null mice, mice with induced obesity and Type II diabetes due to leptin alterations (ob/ob and db/db), and mice with inherited defects of fatty acid oxidation. UCP1-null mice were able to acclimate and tolerate 18°C with a well-defended T_b that was not significantly different from that of WT mice (Golozoubova, 2006). Similarly, administration of leptin to ground squirrels failed to alter their thermogenic capacity (Boyer et al., 1997). However, adaptive NST was significantly altered in UCP1-null mice (details below). Mice with induced rapid early-onset obesity (ob/ob) and diabetes (db/db) have markedly reduced capacity for NST (Yen, et al., 1974). T_b of ob/ob and db/db mice was 30±1.4°C and 26.8±8°C, respectively, after 90 min of exposure to 4°C. This is comparable to mdm mice at 20°C, with T_b values of 29.3±1.2°C. In addition, mice homozygous for the inactivated allele BALB/cByJ, which encodes the short-chain acyl CoA dehydrogenase, have abnormal NST (Guerra et al., 1998). In BALB/cByJ mice, T_b decreased by 10°C in less than 4 h at 4°C, which is similar to mdm mice at 20°C.

The impact of ST on maintenance of T_b can be assessed by blocking muscular activity via curare-like drugs, which competitively block the binding of acetylcholine to the motor endplates of striated muscles (Bowman, 2006; Kashimura et al., 1992). In a study that reduced shivering by 50% via curare, WT mice were still able to maintain a T_b of 35.4±0.4°C at 4°C (Bal et al., 2012), likely because of their ability to compensate via NST. It is difficult to compare thermoregulatory deficits of mdm mice with those of genetically altered mice in other studies owing to 4°C being the standard cold temperature versus 20°C, which was used in our study as this is the low temperature limit of mdm mice. Nonetheless, the mdm mice in our study expressed more severe thermoregulatory defects than mice with deficits in either ST or NST discussed above. This is likely due to the combined defects of both ST and NST in mdm mice.

mdm and WT mice maintained similar RMRs at a T_a of 34°C (Fig. 2). These values are slightly higher but comparable to resting MRs reported for M. musculus that were fasted between 5 and 30 h (Hudson and Scott, 1979). The mice in our study were fasted for only 2 h (WT) or 1 h (mdm), which could account for the slightly higher MRs we found.

mdm mice reach thermoregulatory failure at T_a values below 29°C and are more cold-stressed than WT mice at this T_a. This is demonstrated by significantly higher MRs of mdm mice compared with WT mice (Fig. 2). WT mice had comparable MRs at 34°C and 29°C likely because these T_a values are near or within their TNZ (31–35°C; Hudson and Scott, 1979). Interestingly, although mdm mice have higher MRs than WT mice at 29°C, they maintained a significantly lower T_b. These results indicate that at this T_a, mdm mice work harder than WT mice to thermoregulate, yet they still cannot maintain or attain euthermic T_b and are thus hypothermic. The failure to maintain a higher T_b despite higher MRs is likely due to the combination of increased thermal conductance of the much smaller mdm mice and their compromised thermoregulatory capacity.

At 24°C, there was a large difference in MR between mdm and WT mice, with WT mice having significantly higher MRs (Fig. 2). There were no significant differences in MRs between WT and mdm mice at 20°C, which was interesting considering T_b differences were even more different between genotypes. This is likely due to the effect of a single outlier in the mdm group that greatly increased the standard error of this group as compared with the WT group. When we analyzed the data with the outlier removed, WT mice had significantly higher MRs than mdm mice.

As previously stated, MR scales with body size and it is necessary to account for the much smaller size (7 g) of mdm mice and its impact on MR. A similarly sized mouse, Perognathus longimembris (little pocket mouse), is highly abundant and one of the smallest rodents in North America (Bartholomew and Cade, 1957). Previous groups have studied this species because of its ability to maintain homeothermy despite its small size. A previous study investigating MR of the little pocket mouse reported regressions across T_a values (Chew et al., 1967) to investigate allometric scaling of MR with T_a (Kleiber, 1932). This allowed us to compare MR of the little pocket mouse with MR of mdm mice (using Q₁₀-corrected MR) at our desired T_a. For each of the temperature ranges (20, 24, 29 and 34°C), differences in MR_expected between the little pocket mouse and mdm mice were as follows: 0.54, 1.69, 1.65 and 0.85 ml O₂ g⁻¹ h⁻¹. Surprisingly, MR_expected of mdm mice was less than predicted MRs for the size-matched little pocket mouse at all of the test temperatures.

Our data indicate that even if mdm mice could obtain euthermic T_b, they would still have a deficit in MR. E–O MR of WT and mdm mice differed significantly at all T_a values. In general, WT mice had MR_expected close to MR_observed, leading us to conclude that their Q₁₀ approximated 2.4, which is typical of euthermic M. musculus (Hudson and Scott, 1979). The T_b data also confirm that WT mice were euthermic for all trials (Fig. 1). mdm mice, in contrast, had much higher MR_expected than MR_observed, leading to high E–O MR.

Maximal V̇_O2 following noradrenaline injection correlates well with body mass and is affected by acclimation temperature, with cold-acclimated animals having higher V̇_O2 responses. Allometric relationships for maximal metabolic response to noradrenaline in small mammals have been described by Wunder and Gettinger (1996) and include a BMR component that can be used for animals acclimated to 23°C. Using this equation, the V̇_O2 response of WT and mdm mice to noradrenaline was predicted to be 6 and 11 ml O₂ g⁻¹ h⁻¹, respectively. Observed maximal V̇_O2 responses to noradrenaline values for WT and mdm mice were 7.2 and 6.4 ml O₂ g⁻¹ h⁻¹, which did not significantly differ. WT mice did not differ from the Wunder model output and mdm mice failed to increase MR after noradrenaline injection accordingly. These conclusions are also reflected in the NST contributions to thermogenic capacity (Fig. 5). These comparisons provide further evidence that mdm mice have a lower V̇_O2 response to noradrenaline than WT mice.

Although mdm mice have a lower capacity for NST, the latency from noradrenaline injection to peak MR, total effect time of noradrenaline, and peak MR were similar to WT mice (Fig. 4). Peak T_b and average T_b of WT mice were significantly higher after noradrenaline injection than in mdm mice (Fig. 4). These results are consistent with the finding that mdm mice have a reduced capacity for NST. Similarly, mdm mice were unable to maintain euthermic T_b. This could have been due to their increased thermal conductance or to a depletion of metabolic fuels needed for NST, primarily free fatty acids and glucose (Townsend and Tseng, 2014).

Based on the severe phenotype of the mdm mice compared with UCP1-null mice, the null hypothesis is that the lower capacity of NST in mdm mice is due to the absence of UCP1. In WT and UCP1-null mice that were warm-acclimated to 30°C, noradrenaline evoked large increases in MR for WT mice and small increases for UCP1-null mice (Golozoubova, 2006). The small increase seen in UCP1-null mice was attributed to either UCP1-independent adrenergic thermogenesis or a general activation of adrenergic receptors. The mice in our study did not have a significantly different peak V̇_O2 response to noradrenaline compared with WT mice (Fig. 4); therefore, a solely adrenergic effect causing MR increase in mdm is not likely. In this study, we did not quantify expression of UCP1 or excise putative BAT from mice in the study; however, an interesting line of thought is the possibility for decreased UCP1 expression or decreased sensitivity in BAT to noradrenaline in mdm mice.

Although only MR and T_b were measured, mdm mice appear to be compromised in heat production at the BAT level. It is unlikely they are affected at the sympathetic outflow level based on the results from noradrenaline injection. The NST response was inferred from indirect calorimetry by mimicking a maximal sympathetic response through administration of noradrenaline, which is known to stimulate UCP1 activity in BAT. Because noradrenaline injection resulted in a lower than expected metabolic response and animals could not maintain euthermic T_b, we hypothesize that mdm mice are impacted at the level of BAT.

Titin has been implicated as a regulator in mitochondrial respiration as well as bioenergetics. The sallimus (sls) gene in Drosophila, whose product is homologous to the NH2-terminal half of titin in vertebrates, has been identified as a transcription regulator of mitochondrial respiration (Jumbo-Lucioni et al., 2012). Jumbo-Lucioni et al. (2012) observed the natural variation between state 3 and state 4 mitochondrial respiration and found a direct effect of sls on mitochondrial function. Homozygous sls^d00134 flies, which have a missing allele from the sls gene, had 17% lower mitochondrial state 3 and 18% higher state 4 rates than controls. These results reveal that sls is a novel gene hub for regulation of mitochondrial respiration and that specific alleles of this gene can control naturally occurring variation in mitochondrial function. In addition, truncating titin variants (TTNtv) that cause genetic dilated cardiomyopathy (DCM) have been shown to alter mitochondrial bioenergetics (Verdonschot et al., 2018). In patients with TTNtv, increased expression of genes across all electron transport chain complexes as well as ATP synthase was found. These findings suggest a compensatory response of increased oxidative phosphorylation components in order to counteract limited contractile ability of cardiac tissue as an indirect effect of a titin mutation.

Defects in oxidative phosphorylation can affect BAT thermogenic activity (Kajimura and Saito, 2014). During BAT-mediated thermogenesis, UCP1 uncouples heat production from ATP synthase; therefore, it is likely that defects in oxidative phosphorylation could affect this output. As mentioned above, titin can modulate mitochondrial bioenergetics and transcription of oxidative phosphorylation components, and therefore could be a potential explanation for the reduced NST capacity of mdm mice. Taylor-Burt et al. (2015) demonstrated a direct effect of titin stiffness on the rate of ST, and our results indicate that an NST deficiency could reflect a regulatory effect of titin on metabolic processes. Whether these effects are due to titin signaling or other pathways is a question that could be explored in future work.

Thermogenic capacity is significantly reduced in mdm mice compared with expected values for their body size (Fig. 5) (Taylor-Burt et al., 2015). Many small mammals, including deer mice, increase their thermogenic capacity solely by increasing NST after cold acclimation (Van Sant and Hammond, 2008). Although NST is regarded as the most plastic component of V̇_O2sum, animals can also increase their capacity for ST after cold acclimation (Nespolo, et al., 1999). This leads to the hypothesis that animals with deficient ST would increase NST capacity to compensate. It is interesting that mdm mice have a reduced thermogenic capacity for both ST and NST.

Our results show significant differences in MR as well as the capacity for NST even after accounting for body mass differences (WT mice are approximately four times larger than mdm mice). Because of the allometric relationship between body mass, MR and capacity for NST, our comparisons allowed us to conclude that the severe defects in thermoregulation are not just due to body size. Comparisons between mdm mice and the little pocket mouse demonstrate that mdm mice do not exhibit similar MRs to size-matched animals, as shown by their smaller MRs, even at thermoneutrality (Chew et al., 1967). Additionally, as animals become smaller, the contribution of NST to thermogenic capacity increases to offset the balance in increased thermal conductance due to increased surface area (Wunder and Gettinger, 1996). V̇_O2 responses to NE in mdm mice were less than half the predicted value (11 ml O₂ g⁻¹ h⁻¹; Wunder and Gettinger, 1996). Our findings demonstrate that mdm mice do not activate more heat-generating mechanisms like other small animals to offset the increased heat loss owing to small body size, and that their thermoregulatory defects are either due to direct or indirect effects of the N2A deletion in titin.

A previous study showed an unexpected link between a small titin mutation, ST and thermoregulation (Taylor-Burt et al., 2015). The present study expands on the previous study through continuous monitoring of T_b and investigation of NST and MR. A link between titin mutations and metabolic deficits might have implications for understanding inherited diseases due to titin mutations, which are increasingly found in humans. We conclude that a deletion in the N2A titin gene TTN reduces not only the rate of ST in mdm mice (Taylor-Burt et al., 2015), but also NST. Collectively, mdm mice exhibit a significantly compromised thermogenic capacity. When comparisons are made to account for the small body size of mdm mice, these deficiencies are exacerbated. It is not clear how titin modulates both NST and body size, but it could be through an indirect role in regulating oxidative phosphorylation or other pathways. Future studies should investigate these possible links.

We thank Anthony Hessel, Johhn Lighton, Uzma Tahir, Tinna Traustadóttir and our reviewers for their helpful comments on this manuscript. We would also like to give a special thanks to Danielle Dillon, Destiny Simpson and the team at Sable Systems International for their technical help with this project.