Fatty acid oxidation: a neglected factor in understanding the adjustment of mitochondrial function to cold temperatures

Temperature, through evolution, is a powerful selection factor for determining the adaptability and survivability of a species (Al-Fageeh and Smales, 2006; Blier et al., 2014; Pörtner et al., 2007). For ectothermic species, the internal body temperature depends directly on the environmental temperature and thus any thermal deviation can affect the efficiency of their integral biochemical processes. It is important to understand not only how these animals are responding to temperature changes in the short term but also how they adapt to longer periods of seasonal adjustments, especially at the level of their mitochondrial energy production system. Additionally, even in the context of climate change, studying cold acclimation is critical. In this context, between 1975 and 2015, the increase in the minimal winter temperature was half the increase in the maximal spring temperature (Vyse et al., 2019). The annual thermal range organisms are facing every year becomes more and more extended as the decades pass, thus creating larger extremes. As a result, it is critical to understand not only how species adjust over the week to an increase in temperature in the spring but also how they adjust to a decrease in temperature in the autumn.

Mitochondria are well recognized for their vital role in cellular energy production via a subset of processes collectively called oxidative phosphorylation (OXPHOS). This term aptly refers to the coupling of (1) the sequential electron oxidation via protein complexes of the electron transport system (ETS) to generate an electrochemical gradient across the inner mitochondrial membrane and (2) the utilization of this hydrogen ion gradient by the phosphorylation system to phosphorylate adenosine diphosphate (ADP) to adenosine triphosphate (ATP) for use in energy-demanding cellular activities (Fig. 1). OXPHOS, though once viewed as a linear chain of oxidation–reduction reactions, is in reality a merging point for several closely interacting biochemical pathways that support energy production from various substrates that donate electrons to the ETS in vivo (reviewed by Blier et al., 2014). The sites at which electrons enter the ETS vary by substrate; however, all electron pathways, including the NADH (electron entrance through complex I), succinate (electron entrance through complex II), electron transferring flavoprotein (ETF), glycerophosphate dehydrogenase, choline dehydrogenase, proline dehydrogenase, dihydroorotate dehydrogenase and sulfide:quinone oxidoreductase pathways (reviewed by Hidalgo-Gutiérrez et al., 2021; Lemieux and Blier, 2022), converge at the Q-cycle (ubiquinol/ubiquinone). This complexity regarding OXPHOS has remained largely unaddressed in scientific literature, with most studies focusing on canonical pathways of energy production through carbohydrate oxidation via the NADH and succinate pathways. As such, data regarding the effects of differing temperatures on various aspects of other pathways is severely limited. While the several aforementioned pathways function in concert in most eukaryotes, their contribution, role and control vary between species, organs or tissues of interest, and environmental conditions. Considering the simultaneously cooperative, competitive and altogether intricate relationship between these pathways, each pathway and its associated OXPHOS components may respond differently to thermal changes and thereby play a unique role in acute thermal sensitivity and/or long-term thermal acclimation. More information on this subject is required to understand the global adjustment of OXPHOS during an animal's adaptation to a new temperature, as such metabolic alterations may have far-reaching functional repercussions at all levels of biological organization. Specifically, different electron pathways are associated with varying OXPHOS efficiency. Pathway preference, control and regulation not only impact the amount of ATP produced per substrate molecule and/or the amount of oxygen required to fully oxidize a substrate molecule (thus altering the organism's capacity to sustain cellular functions in a specific environment) but also have implications for management of reactive oxygen species (ROS) production. As a natural by-product of well-regulated OXPHOS, ROS act as signaling molecules that link energy metabolism to various cellular activities and biological processes (Finkel, 2011). However, uncontrolled ROS production can then not only alter cell signaling but also damage cellular macromolecules (i.e. proteins, lipids and DNA), producing multilevel functional consequences and perpetuating a cycle of damage and dysfunction of cellular components, including the OXPHOS process itself (Abele et al., 2002; Jarmuszkiewicz et al., 2015). Consequently, both short- and long-term temperature-mediated modifications to OXPHOS function and regulation may play a role in determining organismal adaptability and survivability through changes in environmental conditions. Here, we focused on the fatty acid-activated ETF pathway and its associated steps (Fig. 1).

Although limited, current literature addressing changes in general lipid metabolism implicates the transportation and β-oxidation of fatty acids as an important determinant of an animal's adaptability to temperature fluctuations (Blier et al., 2014). For example, compared with their warm-adapted counterparts, cold-adapted carp (Cyprinus carpio) liver and brain showed greater expression of genes involved in fatty acid metabolism (i.e. FABP1, GPD1, ME1, MTP, ACBP) and long-chain unsaturated fatty acid synthesis (i.e. FACL1, SCD1, ELOVL2) (Gracey et al., 2004). Further, both laboratory-induced and naturally occurring cold acclimation stimulate enzymatic changes in ectotherm lipid catabolism. Specifically, carnitine palmitoyl transferase I (CPT-I) (Rodnick and Sidell, 1994), total CPT (Hunter-Manseau et al., 2019) and hydroxyacyl-CoA dehydrogenase (HOAD) activity (Guderley and Gawlicka, 1992; Kyprianou et al., 2010) increased after cold acclimation in various ectotherm species. In the outer oblique muscle of cold- versus warm-acclimated striped marsh frogs (Limnodynastes peronii), only HOAD activity was increased (Rogers et al., 2007), but both CPT and HOAD activity were increased in muscles of fish species from polar versus temperate regions (Crockett and Sidell, 1990) and in horse mussel (Modiolus modiolus) muscles collected in winter versus summer (Lesser and Kruse, 2004). Finally, a recent study in zebrafish showed that fasting-induced modification of lipid catabolism not only improved but was also necessary for cold resistance (Lu et al., 2019). Although these studies together emphasize the importance of lipid metabolism in ectothermic thermal adaptation, further work is required to fully elucidate how the multiple elements comprising fatty acid β-oxidation (FAO) in mitochondria readjust in tandem to accommodate temperature change in these animals. One study of intact red muscle mitochondria from the shorthorn sculpin (Myoxocephalus scorpius) showed that FAO of palmitoylcarnitine, a long-chain fatty acid, was decreased by half after cold acclimation (Guderley and Johnston, 1996) but this study failed to consider that fatty acids of different chain lengths follow distinct β-oxidation pathways and that changes in the oxidation of shorter fatty acids could have occurred simultaneously. Thus, questions remain about how thermal variation impacts oxidation of specific chain lengths of fatty acids in addition to the questions concerning the adjustment of global functional pathways in intact mitochondria, rather than specific genes or enzymes.

Occupying freshwater in temperate regions, the flatworm planarian Dugesia tigrina is regularly exposed to large daily and seasonal water temperature variations; in fact, its worldwide distribution and abundance are seemingly controlled by the water temperature (Chandler, 1966; Claussen and Walters, 1982; Hay and Ball, 1979; Kawakatsu, 1965; Lascombe et al., 1975; Reynoldson et al., 1965). Further, compared with other planarian species, D. tigrina is known for its well-developed response to temperature changes from 5 to 25°C and vice versa (Claussen and Walters, 1982), indicating a pronounced capacity to activate a thermal adaptation response. Studying this eurythermal species provides the opportunity to understand thermal adjustment of OXPHOS in a physiological, rather than pathological, context, potentially illuminating the range of metabolic adaptations that promote survival through thermal stress and across wide temperature variations. Studying D. tigrina also offers several practical advantages for researchers (Gentile et al., 2011). As an invertebrate, this inexpensive organism is easily purchased and simply housed in the laboratory, requiring no regulatory oversight and little maintenance to create a thriving population, and can remain accessible for experiments for extended periods. The ability to use whole organisms immediately ensures fresh samples and minimizes preparation time, which can be an obstacle for tissue collections by dissection, for example, where there is an increased risk of poor mitochondrial preservation. Studying whole-organism metabolism also offers an alternative perspective alongside studies focusing on specific organs and tissues. The aquatic habitat of this ectotherm allows for better control of temperature and environmental conditions over time while its small body size is conducive to rapid and equal distribution of the desired exposure. Both factors facilitate acquisition of clearer and more consistent data when addressing specific questions about long- and short-term thermal adaptation. For the above reasons, we opted to study D. tigrina and contend that it is an excellent model for initial investigations into poorly explored OXPHOS pathways, whose results may form the basis for species comparison as the body of work concerning the role of these pathways in the thermal adaptation of ectotherms grows.

Flatworms are known to use fatty acids to support regeneration (Nery da Matta et al., 1994) and reproduction (Angerer et al., 2019; Huang et al., 2012), while new roles, and even new fatty acids, are still being explored and identified in these animals (Angerer et al., 2019). Despite their essential nature to planarian life, lipid metabolism and fatty acid processing in the flatworm are poorly characterized and not well understood. Although planaria were once thought to take on the fatty acid profile of their diet, two studies in particular have strongly refuted this idea and concluded that, while more work is required to fully understand what is happening, they are in fact capable of de novo lipogenesis (Angerer et al., 2019; Makhutova et al., 2009). The study of Angerer et al. (2019) specifically identified a major discrepancy between the fatty acid composition of the flatworm's lab diet, beef liver, and the fatty acid composition of the flatworm species itself. This finding begs the question of how dietary fatty acids are being used, as the flatworms' diet in the wild and in the lab are still considered significant sources of fatty acids. While it is possible they are being transformed into specific lipid species intended for specific functions in the organisms, consumed fatty acids may also be oxidized as fuel. As eukaryotes, flatworms depend on mitochondria for energy production but, as yet, their substrate preferences are unknown. Here, our intention was to provide information on the capacity of D. tigrina to oxidize specific fatty acids for energy production, but also broach how these specific FAO pathways participate in thermal adaptation.

In this study, we explored thermal adaptation of FAO in the freshwater planarian Dugesia tigrina. We studied (1) the long-term thermal adaptation of OXPHOS by subjecting D. tigrina to a 4 week acclimation period at 20°C (normothermic conditions) or 10°C (cold acclimated) and (2) the acute thermal sensitivity of OXPHOS by measuring oxygen consumption of these previously acclimated D. tigrina at assay temperature of 20 and 10°C. We utilized high-resolution respirometry to evaluate quantitative changes in various FAO routes fueling the ETF pathway in intact, functional mitochondria from the whole organism. Using three unique fatty acid substrates for the ETF pathway (i.e. palmitoylcarnitine, octanoylcarnitine and acetylcarnitine), we assessed whether temperature effects were specific to β-oxidation of certain fatty acid chain lengths (long- or medium-chain fatty acids, using palmitoylcarnitine and octanoylcarnitine, respectively) and/or localized to specific enzymes involved in the ETF pathway, including carnitine palmitoyltransferase 2 (CPT2; needed for palmitoylcarnitine and octanoylcarnitine oxidation only), carnitine-acylcarnitine translocase (CACT; needed for oxidation of each of the three fatty acids) and carnitine acyltransferase (CAT; needed for acetylcarnitine oxidation only) (Fig. 1).

Dugesia tigrina (Girard 1850) were obtained from Ward's Natural Science Canada (St Catharines, ON, Canada) and maintained in an aerated medium of dechlorinated water mixed with preconditioned water from a goldfish tank (100 ml l⁻¹). The tanks were kept inside low-temperature incubators (Fisherbrand Isotemp, Fisher Scientific, Ottawa, ON, Canada) set to keep the water temperature in the tank at either 20°C (normothermic group) or 10°C (cold-acclimated group). These temperatures were selected from the optimal temperatures of D. tigrina, as it can maintain larger body size between 10 and 20°C (Folsom, 1976; Russier-Delolme, 1972). The water temperature was recorded twice a day and the mean±s.d. for each group was 20.2±0.7°C (n=42 time points) and 10.0±0.5°C (n=117 time points), respectively. The planarians were acclimated for 4 weeks at the assigned temperature before proceeding with experiments. They were fed beef liver twice a week. After each feeding, the water was replaced with water pre-adjusted to the appropriate temperature. Planarians used for experiments were starved for a minimum of 1 day prior to being used for experiments.

Dugesia tigrina were rinsed twice with MilliQ water. Whole animals were immediately weighed, and had an average wet body mass of 4.32±0.93 mg (n=242) and 3.91±0.80 mg (n=287) for the normothermic and cold-acclimated conditions, respectively. Two to 11 animals (depending on the mass) were then transferred into 1 ml of ice-cold respiration medium (MiR05, containing 110 mmol l⁻¹d-sucrose, 60 mmol l⁻¹ potassium lactobionate, 0.5 mmol l⁻¹ EGTA, 1 g l⁻¹ fatty acid-free BSA, 3 mmol l⁻¹ MgCl₂·6H₂O, 20 mmol l⁻¹ taurine, 10 mmol l⁻¹ KH₂PO₄ and 20 mmol l⁻¹ potassium-Hepes, at pH 7.1; Gnaiger et al., 2000). The animals were gently homogenized on ice by hand in a Potter–Elvehjem homogenizer with one pass, and 300 μl of homogenate, containing 23.0±6.9 mg of D. tigrina wet mass, was immediately transferred into each respiration chamber (Oxygraph 2k, OROBOROS Instruments Inc., Innsbruck, Austria) containing 1700 μl of MiR05, for a final volume of 2 ml. Based on preliminary assays that optimized intra-sample reproducibility, rate of respiration, preservation of coupling (i.e. low LEAK state) and membrane integrity (i.e. low effect of exogenous cytochrome c), three protocols were used to evaluate mitochondrial function and respiration at 20 or 10°C. Datlab7 software (OROBOROS Instruments Inc.) was used for data acquisition and analysis.

Each protocol included three coupling states: (1) LEAK (before the addition of ADP, where the phosphorylation system is inactive), (2) OXPHOS (coupled respiration in the presence of saturating ADP) and (3) ET (non-coupled electron transfer after uncoupler addition to reach maximal respiration without inhibition). By considering the presence or absence of respiratory changes in the last two coupling states, it is possible to determine OXPHOS limitation by the phosphorylation system; where a change between control and experimental groups is observed in the OXPHOS state but not in the ET state, the difference between groups can be attributed to changes in phosphorylation system capacity rather than changes in the capacity of oxidative pathways. All three protocols started with the addition of a distinct fatty acid substrate, i.e. palmitoylcarnitine (0.04 mmol l⁻¹), octanoylcarnitine (0.2 mmol l⁻¹) or acetylcarnitine (5 mmol l⁻¹), alongside malate (2 mmol l⁻¹), inducing the LEAK state of the ETF pathway. Malate was provided with each of the fatty acids as a metabolite of the citric acid cycle to prevent accumulation of acetyl-CoA and the concurrent inhibition of FAO. In most tissues, malate alone supports no or very low respiration (Gnaiger, 2020), and here, as it was provided with each fatty acid substrate, the differences in fatty acid chain length specifically would be responsible for differences in temperature effects between protocols. The subsequent sequence of substrate, uncoupler and inhibitor additions was identical for each protocol: ADP (2.5 mmol l⁻¹; OXPHOS state with the specific fatty acid), cytochrome c (10 μmol l⁻¹; test for outer mitochondrial membrane integrity), uncoupler titration (carbonyl cyanide p-trifluoro-methoxyphenyl hydrazone, FCCP, up to the maximal rate between 0.125 and 0.5 μmol l⁻¹; ET state), glutamate (10 mmol l⁻¹), pyruvate (5 mmol l⁻¹), succinate (10 mmol l⁻¹), uncoupler titration [FCCP, up to an additional 0.05–0.725 μmol l⁻¹; ET state for the combined NADH and succinate (NS) pathway, through complexes I and II, and FAO] and azide (100 mmol l⁻¹; residual oxygen consumption after inhibition of complex IV; ROX). Glutamate, pyruvate and succinate were added in order to measure the maximal ET capacity and express the respiratory flux not only in pmol O₂ s⁻¹ mg⁻¹ wet mass, but also as the flux control ratio (FCR) normalized to the maximal ET capacity in the presence of substrates feeding electrons into the NS pathway and FAO. Mitochondrial respiration was corrected for oxygen flux due to instrumental background and for ROX. Cytochrome c control efficiency was calculated as follows: [(OXPHOS rate with cytochrome c−OXPHOS rate without cytochrome c)÷OXPHOS rate with cytochrome c]. The thermal sensitivity of each pathway and state was also compared on the same scale using the flux relative to 10°C.

Statistical analyses were performed using SigmaPlot 13 (Aspire Software International, Ashburn, VA, USA) and figures were made using GraphPad Prism 7.01 (GraphPad Software Inc., San Diego, CA, USA). The criteria of normality and homogeneity of variance for analysis of variance (ANOVA) were tested for each variable with Shapiro–Wilk and Brown–Forsythe tests, respectively. To meet the criteria for the ANOVA, a square root transformation was used for five datasets (i.e. ET state in flux per mass with palmitoylcarnitine+malate, octanoylcarnitine+malate and acetylcarnitine+malate as well as ET state in FCR with acetylcarnitine+malate, and OXPHOS state for the flux relative to 10°C), and a cube root transformation was used for one variable (i.e. ET state for the flux relative to 10°C). P<0.05 was considered significant.

The 4 week acclimation to cold temperature did not affect the oxygen consumption normalized to the wet mass of D. tigrina in either the OXPHOS or ET states, regardless of the fatty acid substrate provided (Fig. 2). Oxygen flux per mass (Fig. 2) was higher at the warm assay temperature of 20°C than at 10°C with each of the fatty acid substrates: palmitoylcarnitine (P=0.002 and P=0.003, for OXPHOS and ET, respectively; Fig. 2A,B), octanoylcarnitine (P=0.010 and P=0.001, for OXPHOS and ET, respectively; Fig. 2C,D) and acetylcarnitine (P<0.001 and P<0.001, for OXPHOS and ET, respectively; Fig. 2E,F).

In contrast to the data presented as flux per mass (Fig. 2), FCR (Fig. 3) was changed by either the 4 week acclimation or the assay temperature, depending on the respiratory state measured and the fatty acid provided. Fatty acid β-oxidation with the medium-chain fatty acid octanoylcarnitine (P=0.003 and P=0.021 for OXPHOS and ET states, respectively; Fig. 3C,D), and the short-chain fatty acid acetylcarnitine (in the OXPHOS state only, P=0.015; Fig. 3E), increased after cold acclimation, whereas there was no change when the long-chain fatty acid palmitoylcarnitine was the available fatty acid substrate (Fig. 3A,B). The increase in respiration following cold acclimation when respiration was measured with acetylcarnitine as the substrate (versus octanoylcarnitine) was less pronounced and only reached significance in the OXPHOS state (P=0.015; Fig. 3E), but not the ET state (P=0.129; Fig. 3F). A warmer assay temperature (i.e. 20 versus 10°C) decreased the FCR in the palmitoylcarnitine-fueled OXPHOS state (P=0.011; Fig. 3A), but increased the FCR in the acetylcarnitine-fueled ET state (P=0.021; Fig. 3E).

Thermal sensitivity of the three FAO pathways fueled by the specific chain length fatty acids was compared on the same scale by normalizing the respiratory flux per mass at 20°C relative to that at 10°C (Fig. 4). Thermal sensitivity of the FAO pathways did not differ in either the OXPHOS (Fig. 4A) or the ET state (Fig. 4B) and was not altered by prior acclimation.

LEAK respiration was measured in the presence of each fatty acid (+malate) before the addition of ADP and was expressed as FCR (Fig. 5). FCR of LEAK respiration is also referred to as the coupling-control ratio where values of 0 and 1 indicate a fully coupled system or a fully uncoupled system, respectively (Gnaiger, 2020). The FCR for LEAK below 0.1 in most samples, regardless of the substrate combination, indicated good coupling of D. tigrina mitochondria. With palmitoylcarnitine as the substrate, the FCR for LEAK was lower at the 20°C assay temperature than at 10°C (P=0.012; Fig. 5A), but was unaffected by the acclimation temperature (P=0.413; Fig. 5A). With either octanoylcarnitine or acetylcarnitine, neither the assay nor the acclimation temperature affected the FCR for LEAK (Fig. 5B,C).

The cytochrome c control efficiency, indicating the integrity of the outer mitochondrial membrane (Gnaiger, 2020), was not affected by the acclimation temperature (data not shown; P=0.373 with palmitoylcarnitine; P=0.953 with octanoylcarnitine; P=0.242 with acetylcarnitine). The cytochrome c control efficiency was lower at the cold assay temperature, reaching significance with palmitoylcarnitine (P<0.001, without significant interaction, P=0.443) or acetylcarnitine as the substrate (P<0.001, without significant interaction, P=0.448), but not with octanoylcarnitine as the substrate (P=0.062, without significant interaction, P=0.436). The cytochrome c control efficiency obtained here at 20°C (mean±s.d. 0.25±0.13) was comparable to the effect observed in permeabilized D. tigrina (0.23±0.08, n=21).

For a eurythermal species, such as D. tigrina, it is critical to preserve mitochondrial metabolism during temperature variations. It is the only way to maintain an appropriate energetic balance, a suitable level of activity and an acceptable level of ROS to ensure proper cell signaling without causing damage to cellular components (reviewed by Blier et al., 2017). Our study looked at the ETF pathway, which is fed by fatty acid substrates providing electrons to the ETS in order to produce ATP (Fig. 1). Our data showed that the short-term, i.e. immediate, thermal sensitivity was similar for oxidation of short-, medium- and long-chain fatty acid substrates. However, after a 4 week acclimation to low temperature, while there was no adjustment in the long-chain FAO in D. tigrina, there was a positive adjustment in the medium-chain FAO. This specific adjustment of various chain length β-oxidation has been entirely understudied and could be of considerable importance in understanding the overall adjustment to temperature variations.

Our results have shown that cold acclimation specifically increased octanoylcarnitine oxidation in D. tigrina. Because the resulting increase was measured in both the OXPHOS and ET states, the change was not an outcome of an adjustment of the phosphorylation system capacity; that would have been observed if there was a change only in the state dependent on that system, i.e. the OXPHOS state (Lemieux et al., 2011; Lemieux and Warren, 2012). Furthermore, as the ETF and NADH pathways intersect at complex I, it is necessary to consider the potential involvement of the NADH pathway as it relates to the results herein. Increased NADH pathway capacity following cold acclimation in D. tigrina has been shown previously in the context of carbohydrate oxidation (Scott et al., 2019). However, in the presence of the appropriate fatty acid+malate substrate combination supplementing FAO and activating the ETF pathway, the contribution of the NADH pathway is minor and, according to Gnaiger (2020), would not limit oxygen consumption through the ETF pathway. Conversely, an increase in NADH pathway capacity would only account for our results if (1) cold acclimation increased respiration similarly in each set of substrates feeding FAO regardless of fatty acid chain length, or (2) cold acclimation increased respiration only in the substrate combination producing higher respiration rates in relation to the other fatty acids, as a higher rate of oxygen consumption would require more support from the NADH pathway. Here, oxygen flux per mass was similar between the three fatty acid substrate protocols, except under certain conditions where respiration with acetylcarnitine was higher. Even so, the FCR for acetylcarnitine under the ET state was not adjusted following cold acclimation; with this substrate, the occurrence of a change only in the OXPHOS state suggests an adjustment in the phosphorylation system, as shown previously in Scott et al., 2019. The higher oxygen consumption in flux per mass when acetylcarnitine is the available substrate could cause the phosphorylation system capacity to become a limiting factor as a result of the demand for an overall higher working capacity. Therefore, the adjustment of FAO following cold acclimation observed here is exclusive to medium-chain FAO through the ETF pathway.

Finally, palmitoylcarnitine oxidation was not modified by the 4 week cold acclimation in either the OXPHOS or ET states, indicating that D. tigrina did not alter its capacity to oxidize long-chain fatty acids after long-term cold exposure. In contrast to our data, one study detected an increase in palmitoylcarnitine oxidation in the red muscle of the shorthorn sculpin following cold acclimation (Guderley and Johnston, 1996). However, because this study only examined long-chain FAO, it is impossible to know whether, in the red muscle of the sculpin, responses of convergent FAO steps fueled by other substrates, such as octanoylcarnitine, would be more affected, less affected, or present at all. In future studies, we strongly recommend that researchers use an array of fatty acid substrates to distinguish which specific steps of FAO are modified by temperature. Given the differences in our results compared with those for sculpin, it is possible that thermal responses of FAO could be as diverse as ectotherm species themselves. In the planarian flatworm, medium-chain fatty acids may be preferred to shorter or longer chain fatty acids for structural components in contrast with higher organisms.

Our study showed that acute immediate thermal sensitivity was similar for short-, medium- and long-chain fatty acid substrates, but long-term acclimation to cold temperature showed distinct responses of these chain-length-specific FAO steps alongside each other. Our data suggest that, in D. tigrina, 4 weeks of cold acclimation does not affect CACT (needed for palmitoylcarnitine oxidation only), CAT (needed for acetylcarnitine oxidation only, but in both OXPHOS and ET states) or CPT2 (needed for oxidation of both palmitoylcarnitine and octanoylcarnitine), but prompts a specific adjustment in the capacity to oxidize medium-chain fatty acids. This emphasizes the need to study various steps and pathways of FAO with multiple protocols in order to understand the global changes in functional pathways following thermal acclimation. Although we identified changes in medium-chain fatty acid oxidation, it is possible that other steps and pathways are also modified with cold acclimation. For example, pathways that involve CPT1b, via the knockout of CPT1b, have been shown to reduce survival to cold stress in the zebrafish (Lu et al., 2019).

The regulation of chain-length-specific FAO pathways in proportion to each other and our results herein have possible implications for ROS management, according to a hypothesis developed by Speijer (2011, 2014, 2019). This hypothesis postulates that the ratio of electrons coming from FADH₂ versus electrons coming from NADH (i.e. the FADH₂/NADH ratio for a given substrate) is correlated with ROS production originating from mitochondrial respiration. Within this framework, FAO has a much larger FADH₂/NADH ratio than glucose breakdown. Further, the longer the fatty acid, the higher the FADH₂/NADH ratio, and the higher the level of ROS production, as complex I, known as a major source of ROS (Brand, 2010; Murphy, 2009), is confronted with a scarcity of its acceptor Q (Speijer, 2011, 2014). We observed no change in long-chain FAO in the cold-acclimated planarian, which, if increased, would increase the mitochondrial radical burden. Rather, cold acclimation stimulated medium-chain FAO. Preferential induction of medium- versus long-chain FAO after a cold temperature challenge could be a mechanism by which D. tigrina maintains adequate mitochondrial function while preventing an undesirable increase in the FADH₂/NADH ratio. The ability to limit the mitochondrial burden of oxygen radicals in this way may contribute to this planarian's impressive thermal adaptability and purported immortality (Sahu et al., 2017).

Another important aspect in our study was the effect of cold acclimation on feeding in D. tigrina, and how this could also be linked with changes in mitochondrial function. In our experiment, the D. tigrina exposed to 10°C strongly reduced their feeding compared with those exposed to a temperature of 20°C, in agreement with previous studies in D. tigrina (Chandler, 1966; Dahm, 1958; Pickavance, 1971), as well as in various species of fish (Lopez-Olmeda and Sanchez-Vazquez, 2011; Sala-Rabanal et al., 2003). The temperature threshold at which D. tigrina stops feeding varies depending on the population, but temperatures between 3 and 11°C have been reported (Chandler, 1966; Dahm, 1958; Pickavance, 1971). In our study, after the 4 week period of cold acclimation, even with the reduction of feeding, no significant reduction in body mass was observed. Interestingly, fasting has been linked with the capacity to adjust to cold temperature in various species. In zebrafish (Lu et al., 2019) and in Drosophila melanogaster (Le Bourg, 2013), fasting has been shown to increase acute cold resistance. The increase in lipid catabolism was even essential to increase cold resistance in zebrafish, as inhibition of FAO by suppressing the mTOR pathway weakened the fasting-induced cold resistance (Lu et al., 2019). Additionally, suppression of FAO using mildronate reduced survival rates following cold stress, while the FAO activator fenofibrate caused the opposite effect (Lu et al., 2019). These studies indicate the importance of changes in FAO in helping ectothermic animals adapt to cold temperatures, and reinforce the importance of studies of changes in the various steps and pathways of FAO following thermal acclimation.

Our study looked at the ETF pathway, fed by various fatty acid substrates providing electrons into the ETS in order to produce ATP (Fig. 1), and emphasizes the need for understanding the specific changes in fatty acid β-oxidation to increase resistance to thermal acclimation. To our knowledge, this is the only study to look at thermal sensitivity and the impact of thermal acclimation on multiple steps and pathways of fatty acid β-oxidation in intact, functional mitochondria. Together, the current study and the previous one from Scott et al. (2019) identify the important steps and pathways of OXPHOS adjustment to cold acclimation in D. tigrina, i.e. the medium-chain fatty acid β-oxidation, the complex I and the phosphorylation system capacity. These two studies are building a global picture of several neglected OXPHOS changes following thermal acclimation in a feasible animal model, and highlight the need to study these processes, as well as many other pathways and steps (Lemieux and Blier, 2022), in other ectothermic species. Comparing the changes observed in eurythermal species such as D. tigrina and in animals experiencing less tolerance to variation in temperature could illuminate what allows thermal plasticity in ectothermic animals. More studies are needed to understand the specific changes in the various OXPHOS pathways and steps, in order to delineate the adjustment following long- and short-term acclimation to both cold and warm temperatures in more species and tissues.

We are grateful to Dr Roshani Payoe for comments on the manuscript.