The added costs of winter ocean warming for metabolism, arm regeneration and survival in the brittle star Ophionereis schayeri

Warming trends associated with climate change are generating increasing concern about the impact of temperature on animal physiology (Rohr et al., 2018; Bindoff et al., 2019) and energy budgets/allocation (Harianto et al., 2021; Rubalcaba et al., 2020; Yeruham et al., 2019; Sokolov, 2021). Environmental temperature affects most aspects of the biology of ectothermic organisms, especially metabolism, growth and reproduction (Pörtner and Farrell, 2008; Somero, 2012; Huey and Kingsolver, 2019). For marine invertebrates, recent studies have investigated the impact of increased temperature in the context of ocean warming effects on metabolic rate and other fundamental fitness parameters such as growth and reproduction (Wong and Hofmann, 2020; Reed et al., 2021; Gooding et al., 2009; Dworjanyn and Byrne, 2018; Balogh and Byrne, 2020; Torres and Giménez, 2020; Uthicke et al., 2020; Chirgwin et al., 2020).

Resting metabolism dictates the basic energy requirements of all organisms, with the remaining energy consumed or produced being available for activity, growth, reproduction and storage (Burton et al., 2011). As the energy requirements for metabolism increase, there are fewer resources to support other life processes, potentially impairing the survival of the individual and future generations by reducing reproductive fitness and success. Two variables that have a large influence on metabolism are temperature and body size (Gillooly et al., 2001). Increases in either lead to an increase in metabolic rate, especially for ectotherms with limited regulation of body temperature.

With respect to body size, metabolic rate (R) changes as organisms increase in size according to the power function R=aM^b, where M is the mass, a is a species-specific coefficient and b is the scaling factor (Krogh, 1916; Klieber, 1932; Hemmingsen, 1960). The value of b has been alternatively proposed as 0.67 and 0.75 and has been the subject of debate (Glazier, 2010, 2018). The underlying principles to explain this allometric phenomenon are also debated, with recent models predicting a range of b from 0.67 to 1.0 (White and Kearney, 2013; Glazier, 2010). This scaling factor can be affected by abiotic factors, including temperature and light, and biotic factors such as food availability (Verberk et al., 2016). With regard to climate change, this is important because the value of b decreases for many ectotherms as temperature increases (e.g. crustaceans and fish: Ivleva, 1980; Killen et al., 2010). A consequence of habitat warming is suggested to be decreasing body size in marine species, attributed to the temperature–size rule (Sheridan and Bickman, 2011; Forster et al., 2012). Thus, ocean warming can affect two aspects of the physiological responses to temperature: (1) influencing metabolic responses and (2) indirectly altering body growth dynamics. Metabolic scaling can place constraints on food webs and ecosystem functioning by restricting food supply and energy flow (Brown et al., 2004). Recent studies have examined metabolic scaling in the context of climate change across a range of marine taxa (Bruno et al., 2015), and specifically in echinoderms (Carey et al., 2014; Carey and Sigwart, 2014) and fish (Pauly and Cheung, 2018).

Most studies on the impacts of warming on the metabolism of marine invertebrates have focused on short-term effects (hours to a few weeks, e.g. Calosi et al., 2013; Christensen et al., 2011; Fang et al., 2015a,b; Carey et al., 2014; Minuti et al., 2021; Tepolt and Somero, 2014), although more recent studies have identified the need to focus on longer term, ecologically relevant exposure durations (months to years: Leung et al., 2021; Harianto et al., 2021; 2018; Suckling et al., 2015; Rohr et al., 2018; Godbold and Solan, 2013; Uthicke et al., 2020). These studies avail of the phenomenon of acclimation, in which organisms re-adjust their metabolism to cope with the new conditions (reviewed in Rohr et al., 2018). An organism's ability to acclimate to change, however, may be affected by the rate of change (acute versus slow), season, latitude and reproductive status (Peck et al., 2009a; Nguyen et al., 2011; Bates and Morley, 2020; Allen et al., 2016; Rohr et al., 2018; Donelson and Munday, 2012; Dupont et al., 2013; Kühnhold et al., 2019a; Byrne et al., 2020; Uthicke et al., 2020; Havrid et al., 2020). It is also often assumed that an organism is ‘acclimated’ after a certain period of time and that there are no further metabolic changes. However, this may not be the case for many species (Harianto et al., 2021; Morley et al., 2018; Bates and Morley, 2020), with organisms either unable to acclimate (Wilson and Franklin, 2000; Nilsson et al., 2010; Peck et al., 2009b; Peck et al., 2010) or unable to maintain the acclimation status (Sokolova 2021).

Extreme summer temperatures (heatwaves) are well known to have detrimental effects on shallow water marine invertebrates, including increased mortality (Garrabou et al., 2009; Caputi et al., 2019; Weitzman et al., 2021), especially when coincident with periods of low tide (see Glynn, 1968; Tsuchiya, 1983; Anthony and Kerswell, 2007). This is particularly concerning as climate change has increased the frequency and duration of summer heatwave events (Wernberg et al., 2012; Holbrook et al., 2019). In extreme cases, these thermal anomalies have caused mass mortality (Pearce et al., 2011; Wernberg et al., 2012; Hughes et al., 2017). Less attention has been given to the impacts of warming events in cooler seasons. Seasonality effects on metabolism and growth are well studied in planktonic communities and fish (Sampou and Kemp, 1994; Murrell et al., 2013; Wohlschlag and Juliano, 1959), but less so in marine invertebrates (Propp et al., 1983; Wittmann et al., 2008). Most ectotherms experience decreased metabolism, growth and feeding in winter as a consequence of colder temperatures and reduced food availability (Brockington and Clarke, 2001; Malanga et al., 2009; Percy, 1988). Thermotolerance may also change during cooler periods, making organisms more vulnerable to elevated temperatures (Richard et al., 2012; Pörtner, 2010; Chapple et al., 1998). This is a concern for regions that are experiencing warmer winters as a result of climate change (Hobday and Lough, 2011).

Echinoderms have been a focal taxon in investigations of the impacts of warming on the metabolic responses of marine invertebrates. This ecologically important group appears to be particularly vulnerable to many of the stressors associated with climate change (Byrne, 2011; Byrne et al., 2013; Byrne and Fitzer, 2019). There have been many studies of the impact of temperature on the metabolic rate of sea urchins and sea stars, reflecting their ecological importance as grazers and predators, respectively. In general, metabolic rate increases with temperature up to tolerance limits, after which organisms may experience metabolic depression (Delorme and Sewell, 2016; Ulbricht and Pritchard, 1972; Harianto et al., 2021, 2018; Manríquez et al., 2019; Carey et al., 2016; McElroy et al., 2012; Uthicke et al., 2014; Détrée et al., 2020; Fly et al., 2012; Catarino et al., 2012; Siikavuopio et al., 2008; Sedova, 2000). Much less is known about the most diverse class of echinoderms, the Ophiuroidea (brittle stars, serpent stars and basket stars) with ∼2100 described species (Stöhr et al., 2012, 2016). Brittle stars play many key ecological roles as high-density suspension feeders (Allen, 1998; Davoult and Gounnin, 1995; Hily, 1991) and selective deposit feeders (Hoskins et al., 2003; Clements and Stancyk, 1984; Gunnarsson and Sköld, 1999), and as a food source of higher tropic levels in many marine ecosystems (Dahm, 1993; Packer et al., 1994; Ravelo et al., 2017). They are also important as ecosystem engineers, contributing to the flux of oxygen, nutrients and other compounds in marine sediments (Vopel et al., 2003; Davoult et al., 2009; Calder-Potts et al., 2018). Their capacity for fast regeneration is striking and allows them to regrow or replace damaged arms after sub-lethal predation and so serve as a ‘renewable’ food source for many other species, including fish and crustaceans (Bowmer and Keegan, 1983; Duineveld and van Noort, 1986; Aaronson, 1987, 1989, 1991, 1992; Munday, 1993; Sköld and Rosenberg, 1996; Sides, 1987). The multiple ecological roles that brittle stars play in marine ecosystems globally, together with their extensive regenerative abilities, make these animals prime model organisms to investigate a range of topics from cellular regenerative processes to trophic transfer in marine food chains (Mashanov et al., 2022; Zueva et al., 2018; Geraldi et al., 2017; Ravelo et al., 2017).

Efficient regeneration of lost or damaged tissues is characteristic of the Echinodermata, and for brittle stars the ability to autotomize and readily regrow lost body parts is key to their success (Wilkie, 1978; Emson and Wilkie, 1980). Regeneration processes are well documented for Amphiura filiformis, from gene expression to anatomy and histology and in response to a range of abiotic factors (Burns et al., 2011; Czarkwiani et al., 2013, 2016; Biressi et al., 2010; Wood et al., 2008; Hu et al., 2014; Dupont and Thorndyke, 2006; Nilsson and Sköld, 1996). Many burrowing brittle stars, including A. filiformis, have very long thin arms that are used for feeding and burrow irrigation (Woodley, 1975) and are quickly regenerated over several weeks (Clements et al., 1988; Stancyk et al., 1994; Clements et al., 1994; D'Andrea et al., 1996; Pape-Lindstrom et al., 1997; Dupont and Thorndyke, 2006; Hu et al., 2014). Regeneration in brittle stars with thicker, more robust arms which are used in prey capture, locomotion and escape takes longer (i.e. >10 months to years) and is assumed to incur greater costs (see Sides, 1987; Pomory and Lawrence, 1999, 2001; Biressi et al., 2010; Weber et al., 2013; Clark et al., 2007; Wood et al., 2010, 2011; Clark and Souster, 2012). In both groups, it has been shown that the resources used to replace lost tissue are physiologically costly (Pomory and Lawrence, 2001; Stancyk et al., 1994; Lawrence 2010). The effects on respiration are not as well known; however, burrowing brittle stars with thin arms either exhibit a spike in respiration at ∼1.5 weeks followed by a return to normal (Hu et al., 2014; Golde, 1991) or show no effect (Christensen et al., 2017). Ophiocoma echinata, a species with robust arms, experienced elevated respiration rate associated with regeneration persisting over the duration of the study (∼7 weeks) (Pomory and Lawrence, 1999).

The impact of unseasonal ocean warming on the metabolic cost and rate of regeneration in echinoderms is not known. We investigated the effects of increased sea temperature during winter on the metabolic rate and arm regeneration in Ophioneries schayeri, the most common shallow subtidal brittle star along the coast of SE Australia (Rowe and Gates, 1995). This species is an important scavenger and detritivore in the shallow subtidal rubble community (Byrne and O'Hara, 2017). It has robust arms that are extended out from rock crevices during feeding, making it susceptible to sublethal predation by fishes and other predators, with 69% of individuals in natural populations bearing evidence of recent and/or past regeneration (i.e. scars) (A.B.C., personal observation).

Increased temperature has the expected influence on respiration rate in summer-collected O. schayeri with a Q₁₀ of 2.6. (19–25°C; Christensen et al., 2011). Q₁₀ is the factor by which chemical reaction rates change with a 10°C change in temperature and has often been used to describe the dependence of metabolism on temperature in biological systems (Mundim et al., 2020). The factor is reported to be 2–3 for most echinoderm species (Lawrence, 1987; Peck, 2016). Ophioneries schayeri experiences decreased survivorship at elevated temperatures (+4°C for 5 weeks), with 25°C being near its upper thermal limit in experiments conducted in summer (Christensen et al., 2011). We examined the impact of a +3°C warming in O. schayeri in winter, when the ambient sea surface temperature (SST) was 15.5°C, slowly introducing a temperature increase to 18.5°C to avoid thermal shock. The animals were acclimated for 8 weeks, during which time we compared metabolic rate and arm regeneration of the +3°C and ambient group. As a species common to shallow water temperate conditions where it experiences a seasonal range ∼10°C in its habitat (Wolfe et al., 2020), we hypothesized that O. schayeri should have a broad thermal tolerance and be able to withstand increased temperature, even in winter. This is an important consideration as eastern Australia is experiencing a disproportionate increase in winter sea temperatures as well as thermal anomalies (marine heatwaves) (Hobday and Lough, 2011; Babcock et al., 2019). The +3°C warming in winter reflects the winter marine heatwave conditions experienced recently in SE Australia for weeks to months (Oliver et al., 2017, 2018).

We hypothesized that exposure to 18.5°C would not have deleterious effects on O. schayeri as Christensen et al. (2011) kept animals at 19–22°C (summer temperature range) for weeks with no mortality. We expected to see a 2- to 3-fold increase in metabolism and faster regeneration in the warm treatment, allowing the brittle star to more quickly replace a lost arm. Regenerating brittle stars were also expected to have elevated respiration rates as a reflection of the increased energy needed to replace the lost tissues. As O. schayeri is an active and highly mobile species with robust arms, we hypothesized that it would place more emphasis on regaining function, as indicated by the presence of fully developed muscles and tube feet, than on increasing the length of the regenerating arm.

Ophioneries schayeri Müller and Troschel 1844 were collected during low tide (1.0 m depth) from Little Bay (33°58′S, 151°14′E), Sydney, Australia, in August 2018 and transported in ambient sea water to the Sydney Institute of Marine Science (SIMS, Mosman, NSW, Australia). They were placed in flow-through aquaria for 4–5 days to adjust to the new conditions and to check for post-collection health. During this time, they were fed with commercial flake fish food every 2–3 days. For the regeneration experiment, specimens with a disc diameter (dd) >10 mm with intact arms (mean±s.d. dd 14.8±2.8 mm, range 10.1–19.6 mm, n=122) were selected. Smaller specimens (<10 mm dd) were only used for respirometry. The animals that autotomized an arm during collection were placed into aquaria (80 l) supplied with ambient (13.5–19.9°C) filtered (70 µm) flow-through sea water (FSW; 1.5 l min⁻¹) and fed on the same schedule as the other animals.

Temperatures were automatically controlled using a thermocouple–solenoid feedback system where cold and warm FSW were mixed in a header tank, one for each temperature treatment. Header tanks were also aerated to ensure >90% O₂ saturation. The header tank continuously supplied water to the containers in each temperature treatment at a rate of 0.1 l min⁻¹ so that the water in each container completely turned over approximately every 20 min. The experiments were conducted in a temperature-controlled room with a 12 h:12 h light:dark cycle.

After the 4–5 day post-collection period, 6–8 animals were haphazardly placed into flow-through food grade plastic containers (2 l) with sand, cleaned rocks and an individual air stone connected to a common air pump. The ambient temperature treatment (15.5°C) was the SST at the time of collection (early August 2018) and the warm (+3°C, 18.5°C) treatment reflects the mean SST predicted for the region with greater warming in winter than summer (Hobday et al., 2016) as well as the temperature increase experienced in a recent heatwave (Oliver et al., 2017). The 18.5°C treatment is also the temperature experienced in late spring, which was the end time of the experiment. The temperature was slowly adjusted +1°C every 3 days until the target temperature was achieved. In a pilot study, an increase to 20°C (SST in late spring/early summer) was too stressful, causing arm autotomy and disc lesions.

At the end of a 1 week adjustment to the target temperature treatment, animals were anesthetized in 1:1 sea water: 7% magnesium chloride solution and photographed with a digital camera. These images were used to identify individuals by their unique/characteristic disc pigment patterns. The animals were allowed to recover in experimental water and did not exhibit any deleterious effects of anesthesia. Approximately half (ambient treatment) to two-thirds (warm treatment) of the animals were induced to autotomize the arm just to the left of the madreporite at a position 20–25 segments from the disc by gently tugging on the arm with forceps at the desired region (Fig. S1). They were then haphazardly placed back into treatment containers with separate ones for intact and automized animals. For the warm treatment, there were 24 whole animals distributed over four containers (controls) and 38 autotomized animals distributed over five containers. For the ambient treatment, 30 whole or autotomized animals were distributed over five containers each (n=6 per container). The whole and autotomized O. schayeri were maintained in these conditions, at a constant temperature, for an additional 10 weeks. They were fed every other day with commercial flake fish food. The tanks were monitored daily to document mortality. Dead animals and autotomized arms were removed promptly.

Oxygen uptake rates (used as a proxy for metabolic rate) for whole and regenerating animals at 18.5°C and regenerating animals at 15.5°C were examined at four time points: 2, 4, 6 and 8 weeks post-autotomy. The 15.5°C whole treatment was only examined at the first three time points because a temperature failure occurred before respiration measurements could be made during week 8. Mass-specific oxygen uptake (mg O₂ g⁻¹ h⁻¹) was measured as a decline in oxygen using closed-chamber respirometry. At week 2, 3–4 animals were selected from each container for respirometry measurements to achieve a broad size range for the metabolic scaling study (see below). During subsequent weeks, only two random animals were selected from each container for measurement. Animals were placed in stirred 1 l respirometry chambers containing aerated sea water of the appropriate temperature and scrubbed rocks to provide shelter, and allowed to settle for 15 min prior to readings being taken. The chambers were made from watertight transparent domestic food containers made of polypropylene plastic that has a low oxygen permeability (Gilroy and Bessman, 1978). The decline in dissolved oxygen was monitored continuously for 75 min using an optical fluorescence quenching probe, either a Vernier Optical DO probe connected directly to a LabQuest2 interface, or a GoDirect Optical DO probe connected via Bluetooth to a computer running Vernier Graphical Analysis software (Vernier Science Education). The probes were calibrated daily using aerated sea water and one point calibration. After measurements, animals were gently removed, blotted dry and weighed to the nearest 0.01 g with an electronic balance. Water volume (ml) of the chamber was also measured with a graduated cylinder. Blanks containing sea water and scrubbed rocks were routinely run (n=6 per temperature at each measurement period) to control for probe drift and background oxygen changes.

To increase the size range to investigate the influence of size on respiration, small animals (<10 mm dd) were also acclimated to the experimental temperatures as above (but not used for autotomy induction). After being at temperature for 2 weeks, their metabolism was measured using closed-chamber respirometry as above, but using 50 ml chambers. These data were combined with the respiration measurements for the larger whole animals at the appropriate temperature (see above), also taken at week 2.

After respiration measurements were performed at each time point, the animals that had been induced to autotomize were anesthetized (as above) and the regenerating arm was photographed. An additional arm photograph was taken at week 10. Animals were viewed with a Leica M125 stereomicroscope fitted with an IC80 HD camera. Two measurements on the regenerating arms were made using ImageJ software (NIH, Bethesda, MD, USA): length of the regenerating portion from the plane of autotomy to the arm tip, and width of the arm on either side of the autotomy site (Fig. S2). Photomicrographs of the regenerating arms were examined to follow regrowth of the arm skeleton, spines and tube feet and the appearance of arm pigment.

After photography at week 10, and while still under anesthesia, the regenerating arms of 8 individuals (n=4 each treatment) were removed and placed in 10% buffered formalin in sea water for histological examination. The tissue was decalcified in 0.5 mol l⁻¹ EDTA, for 4 days followed by three, 10 min consecutive washes in 70% ethanol and then through a graded ethanol series to 100% ethanol. They were then cleared in xylene for 2 h and embedded in paraffin wax and sectioned (5 μm thick) beyond the position of the original break in order to compare with the intact basal region. The sections were mounted on glass slides and stained using Mallory's trichrome (aniline blue, acid fuchsin, orange G), which stains the connective tissue and muscle (Ross and Pawlina, 2020; M.B., personal observation), prominent features of functional arms. The sections were photographed using an Olympus BX51 compound microscope fitted with an Olympus XC50 color view digital camera.

Animals that experienced autotomy during collection were maintained separately in ambient flow-through conditions (see above), during which time they experienced daily (±1–1.5°C) and seasonal (late winter/early spring to early summer, 13.5–19.9°C) changes in water temperature. These brittle stars were used as a proxy for changes in metabolism and regeneration under ‘natural’ conditions during the 11 weeks that the animals were in captivity. Fifteen animals were haphazardly sampled to measure respiration (as above) at 11 weeks post-collection. On the day of measurement, the ambient temperature of the incoming water was 18.8°C and this was the temperature used in the respirometry. After respirometry, animals were anesthetized, and the regenerating arm was photographed and measured as above.

After a visual inspection of the respirometry data, Excel was used to fit a linear regression to the decline in oxygen over time. Mass-specific oxygen uptake (Ṁ_O2) was then calculated by dividing the amount of oxygen consumed in a given amount of time (e.g. 60 min) by the wet mass of the animal (mg O₂ g⁻¹ h⁻¹).

To determine the effects of size on metabolism, the oxygen uptake rate (mg O₂ h⁻¹) and mass (g) were log transformed and plotted for whole animals in both temperature treatments. A linear regression was fitted to the data to determine the scaling coefficient, b (slope of the line). The slopes of the linear regressions were compared using ANCOVA in R (http://www.R-project.org/) to determine whether temperature had any effect.

The respiration rate data were analyzed by two-way ANOVA with temperature and animal treatment (whole versus regenerating animals) as fixed factors for each measurement time. One-way ANOVA was used to analyse the respiration rate among weeks (time) for each of the animal treatment–temperature groups. Data on respiration rates of regenerating animals held at 15.5°C and 18.5°C, and those placed in ambient flow-through (fluctuating) conditions (13.5–19.9°C) were analyzed by one-way ANOVA, and significant differences were assessed by Tukey's post hoc test. For all these analyses, the required assumptions for data normality and homogeneity of variance were confirmed using the Kolmogorov–Smirnov test and Levene's test, respectively (Sokal and Rohlf, 1987).

The effect of temperature on respiration, Q₁₀, was calculated using the equation ⁠, where R₁ is the respiration rate measured at temperature 1 (T₁) and R₂ is the respiration rate measured at temperature 2 (T₂). Only respiration rates from whole animals were used, while respiration rates for summer animals at 19°C and 25°C were taken from Christensen et al. (2011) (19°C: 0.025±0.003 mg O₂ g⁻¹ h⁻¹ and 25°C: 0.045±0.005 mg O₂ g⁻¹ h⁻¹).

The extent of arm regeneration was measured as the increase in arm length and the return to original arm width (proportional regrowth). Length was calculated by taking the difference between the average regrowth lengths between successive measurements (e.g. average week 4−average week 2) and then dividing by 2 to give a weekly growth rate (mm week⁻¹). The proportional regrowth was calculated by dividing the width of the regenerating portion at the widest point (base) by the width of the arm at the plane of autotomy (Fig. S2). The weekly change in proportional regrowth was calculated by taking the difference in the average proportional regrowth in successive weeks and dividing by 2 to provide a weekly change. It was assumed that all growth at week 2 was new growth. The effect of temperature (15.5 and 18.5°C) on arm regrowth data over time was analysed by a repeated measures (RM) ANOVA of ln(x+1) transformed arm length measurements over 10 weeks. RM ANOVA was also used to test for the effect of temperature (15.5 and 18.5°C) over time, on the proportional regrowth (with proportions being log transformed).

Respiration rates for regenerating animals at 15.5°C and whole and regenerating animals at 18.5°C were examined at four time points (Fig. 1), while the 15.5°C whole treatment was only examined at three time points. Respiration rates were generally higher in the warm treatment, but there was no consistent difference in respiration rates between animal categories (i.e. whole versus regenerating). Two-way ANOVA (temperature×animal treatment) indicated that respiration differed, with the animals at 18.5°C (whole and regenerating) having higher respiration rates at week 2 (F_1,1=17.88, P=0.0001) and week 4 (F_1,1=12.70, P=0.0009). At week 8, one-way ANOVA indicated significantly higher respiration rates in 18.5°C regenerating animals (F_2,25=7.96, P=0.0023) than for animals in the other treatments.

While the respiration rates for all treatment–temperature groups varied (Fig. 1), one-way ANOVA did not reveal any significant differences among the weeks (15.5°C: whole F_2,35=2.52, P=0.094 and regenerating F_3,39=0.41, P=0.745; 18.5°C: whole F_3,34=2.04, P=0.126 and regenerating F_3,50=1.53, P=0.217).

Respiration differed in the regenerating animals held at constant (15.5 or 18.5°C) and fluctuating ambient temperature (13.5–19.9°C) at 11 weeks post-collection (ANOVA: F_2,33=5.96, P=0.006). Tukey's test indicated that the difference was between the constant temperature treatments (15.5°C=fluctuating<18.5°C; Fig. S3).

Respiration rates were measured on a broad range of sizes (mass) of whole animals at each of the constant temperature treatments (15.5°C: 0.1–7.68 g; 18.5°C: 0.19–7.94 g). The metabolic scaling exponent, b, estimated by the slope of the linear regression of the log-transformed data, was 0.746 at 18.5°C and 0.635 at 15.5°C (Fig. 2). While temperature elevated the respiration rate, the slopes of the lines were not significantly different (F_1,78=0.7365, P=0.393), indicating that temperature did not affect metabolic scaling.

Mortality of O. schayeri occurred in both temperature treatments but was highest at the warmer temperature (18.5°C, 40% overall mortality) (Fig. 3). Regenerating brittle stars held at 18.5°C experienced the greatest mortality (55.2% survival). Deaths occurred over the duration of the study, with the greatest mortality in the first 4–5 weeks. In the ambient temperature treatment (15.5°C), mortality was only observed in the regenerating animals (87% survivorship) and occurred during week 6.

Ophioneries schayeri exhibited a high capacity for regeneration (Fig. 4A), with average arm length regrowth after 10 weeks reaching 7.39±3.0 mm (mean±s.d., n=24) and 15.2±7.0 mm (n=18) at 15.5 and 18.5°C, respectively. Arm length initially increased rapidly during weeks 4–8 at both temperatures but slowed by the end of the experiment, particularly in the ambient temperature treatment (Table 1). There was a significant increase in regenerating arm length over 10 weeks (RM ANOVA: F_4,33=93.537, P<0.001) and this was greater in the warmer treatment (RM ANOVA: F_1,36=10.564, P<0.0025). Arm regrowth was not limited only to length; increases were also observed in proportional regrowth (gain in width) over time, with the greatest increases in the warmer treatments (Fig. 4B). This proportional arm recovery was significant over the 10 weeks (RM ANOVA: F_4,28=17.816, P<0.001) and was greater in the warmer treatments (RM ANOVA: F_1,31=9.809, P<0.0038). The trends in increasing arm width were similar in the two treatments, with the greatest rate of increase observed in the first 6 weeks (Table 1).

Development of the regenerating arms with respect to morphological features (e.g. ossicles, spines, tube feet) was also faster in the warmer treatment (Fig. 5A,B). At 2 weeks, most of the brittle stars at 18.5°C had evidence of ossicle formation in the regenerating arm (Fig. 6A). By week 4, much of the regenerating arm had fully formed, with tube feet and arm spines evident and the pigment-banding pattern was also established (Fig. 5A). By week 6, the dorsal arm plates were present, and the arm had darker pigmentation. At week 8, the pigmentation on the aboral surface of the regenerating arm was fully present and the regenerating portion of the arm matched the non-regenerating part in everything except width (Fig. 5A). The animals regenerating at 15.5°C did not exhibit evidence of segmentation or ossicle formation at 2 weeks (Fig. 6B), while those at 4 weeks were similar anatomically to those seen in the warmer treatment at 2 weeks (Fig. 5B).

Histology of the regenerating arms at 10 weeks showed that they were similar for animals that had regrown in control and warm treatments, with well-developed intervertebral ligaments and muscles, both of which are required for locomotory function (Fig. 7). They were a functional miniature of the non-regenerating portion of the arm (Fig. 7C,F).

Brittle stars held under naturally varying temperature regrew a mean (±s.d.) arm length of 13.2±3.6 mm (n=12) over this period. The morphology and pigmentation of the regenerating portion was similar to what was observed in the warm animals at weeks 6 and 8 (Fig. 5C).

As expected, body mass had a significant effect on the respiration rate of O. schayeri. The values of b seen here (0.635 at 15.5°C and 0.746 at 18.5°C; Fig. 2) agree with the range predicted for other organisms (Glazier, 2010) and also fall into the range reported for echinoderms (b=0.6–0.8: Lawrence and Lane, 1982), including other ophiuroids (b=0.68: Hughes et al., 2011). Temperature, at least with respect to the size range of the O. schayeri examined, does not appear to affect the scaling coefficient, as also reported for other echinoderms (Carey et al., 2014, 2016). This lack of dependence on temperature means that animals of all sizes are similarly affected by temperature change. This finding for echinoderms is important as it contrasts with the trend for other marine invertebrates, such as crustaceans (Strong and Daborn, 1980) and bivalve mollusks (Clark et al., 2013), which show increased sensitivity to temperature extremes as size increases (Pörtner and Knust, 2007), leading to overall decreases in body size and decreased reproductive fitness with warming (Somero, 2010; Peck et al., 2009a).

Temperature plays a major role in the metabolic rate of O. schayeri as evidenced by the Q₁₀ values (Table 2). At week 2, the Q₁₀ for the present study was much higher than expected (∼23) (Table 2). The expected Q₁₀ of 2–3 assumes that all the changes are due to the kinetic effects of temperature on the biochemical reactions and that there are no additional metabolic pathways affected. The higher respiration rates at 18.5°C may be due to upregulation of coping mechanisms to deal with the out of season temperature increase (Minuti et al., 2021). As an example of this, the sea cucumber Holothuria scabra exhibited a Q₁₀ of 13.2 over the temperature range 22–29°C during which its levels of HSP70 increased (Kühnhold et al., 2019b). The respiration rates in the warm treatment (18.5°C) are more similar to the respiration rates measured on summer animals at 25°C as evidenced by the Q₁₀ of 1 between these temperatures. As 25°C is thought to be near to the upper limit of thermotolerance for O. schayeri as determined in summer (+4°C) (Christensen et al., 2011), 18.5°C may be an upper limit for this species in winter. Specimens collected around the same time were able to acclimate to 17.5°C (+2°C) over 10 weeks without obvious deleterious effects for a separate study (A.B.C., unpublished observation).

Large increases in metabolism with increased temperature in relation to season have also been reported for other echinoderms. In the brittle star Ophiura ophiura, summer respiration measured at 15.5°C was 7 times higher than that measured for animals collected in early spring at 10.5°C (Wood et al., 2010). This lower spring metabolic rate recorded was attributed to low activity levels and reproductive dormancy. Increased food availability is thought to contribute to the large Q₁₀ (30) observed in the Antarctic sea urchin, Sterechinus neumayeri, with a +3°C increase in temperature during the summer months (Brockington and Clarke, 2001). Starved urchins exposed to the same temperature increase exhibited a more normal Q₁₀ (2–3). The temperature treatment groups of O. schayeri were fed the same amount of food, but the warmer treatment may have increased foraging time, allowing animals in the 18.5°C treatment to gather more food. The greater activity and increased specific dynamic action (the elevation in metabolism associated with digestive processes; Secor, 2009) due to the increased food levels could have contributed to the larger than expected Q₁₀ levels observed. Controlled studies measuring actual consumption and assimilation efficiencies are needed to examine this further.

The constant temperature of the experiment may have also contributed to the elevated metabolic rate respiration that O. schayeri was experiencing. In nature, over the 3 months of the study, seasonal change resulted in an increase in ambient temperature in the subtidal habitat of this species (Wolfe et al., 2020). To some extent, there were also some daily temperature fluctuations depending on the tidal cycle (Wolfe et al., 2020). The decreased temperature pulses associated with night or high tide are believed to provide some organisms with a ‘recovery period’, thus increasing their thermal maxima (Klein et al., 2019; Lugo et al., 2020; Vajedsmiei et al., 2021). Studies incorporating an offset to reflect the seasonal temperature regime are needed to determine whether constant temperature itself is a stressor for O. schayeri.

Ophioneries schayeri has considerable capacity for arm regeneration, which is enhanced by warmer temperatures, as typical for brittle stars (Weber et al., 2013; Wood et al., 2010, 2011; Christensen et al., 2017). The sigmoidal pattern of arm regrowth that we observed is similar to that described for other species (Donachy and Watabe, 1986; Stewart, 1996; Sides, 1987; Biressi et al., 2010; Weber et al., 2013). As many estimates of the time required to complete arm regeneration in ophiuroids are based on linear growth patterns (Sullivan, 1988; Stancyk et al., 1994; Pomory and Lawrence, 1999; Clark et al., 2007; Clark and Souster, 2012), there is a need to reassess this approach, especially for robust armed species.

Regeneration is thought to be a costly process in many echinoderms as it requires a large proportion of energy and resources to be allocated to replacing lost body parts (Fielman et al., 1991; Pomory and Lawrence, 2001; Dupont and Thorndyke, 2006). With the exception of Hemipholis cordifera (Christensen et al., 2017), previous studies on the impact of regeneration on metabolic rate in brittle stars (A. filiformis, Microphiopholis gracilima, O. echinata) were undertaken at control temperatures (Hu et al., 2014; Golde, 1991; Pomory and Lawrence, 1999). Our study is the first to consider the impact of warming on the costs of arm regeneration in ophiuroids.

For O. schayeri, the similarity of the metabolic rate between intact and regenerating animals in the two temperature treatments in the early part of the experiment (week 2) is most likely influenced by the low amount of regeneration that was taking place at that time. While there was measurable regeneration at that point, it represented only a very small increase in mass and may correspond to the lag phase in arm growth observed in Ophioderma longicauda (Weber et al., 2013). Other brittle stars exhibit increases in the respiration rate 1–3 weeks post-autotomy, and O. schayeri appears to follow a similar time line, with regenerating animals having higher respiration rates than whole ones at week 4, albeit not significant (Fig. 1). The lack of significant differences between the whole and regenerating animals at 15.5°C over the remainder of the experiment indicates that the metabolic cost of regeneration is minimal under low/seasonal temperatures. Based on the respiration profile of regenerating O. echinata (at 22°C) (Pomory and Lawrence, 1999), a brittle star with robust arms and similar biology and ecological niche to O. schayeri, we had expected to observe elevated respiration rates.

While the 18.5°C animals (both treatments) appear to experience a metabolic depression at week 6 (Fig. 1), the difference between weeks was not significantly different. The apparent slowing of respiration may be an attempt to alleviate the stress of the higher temperatures (Harianto et al., 2018, 2021; Minuti et al., 2021) or an inability to maintain the elevated rate over prolonged exposure (Leung et al., 2021). Ophioneries schayeri may have been reallocating resources during this quiescent period to further support coping mechanisms as the rates then increased in the following period. This decrease in respiration is also interesting in that it coincides with the highest estimated growth rate. Brittle star respiration rates are reported to be highly variable, both between individuals and over time (Pomory and Lawrence, 1999; Christensen et al., 2017). As the respiration rates measured here only represent what was happening within those animals on that given day, it is impossible to know what they were over the 2 weeks that the growth was occurring. It may be that the day the measurements were taken was a ‘recovery’ period after a period of more intense metabolic activity.

The amount of arm that we removed, approximately two-thirds of the total length, is similar to the extent of sublethal loss seen in the natural population, likely in response to predation (A.B.C., personal observation). This extent of arm section removal may have stimulated the high regeneration rate observed. Rapid replacement of lost arms is crucial to the survival of brittle stars as the arms are used not only in locomotion but also in food gathering/capture and respiration (Donachy and Watabe, 1986; Woodley, 1975; Fielman et al., 1991; Beardsley and Colacino, 1998). This replacement comes with the dilemma of regaining length or regaining locomotory ability (segmentation, muscles, ligaments and tube feet) first (Dupont and Thorndyke, 2006). This tradeoff appears to be correlated most to the length lost (Dupont and Thorndyke, 2006). In burrowing species with thin arms, if the arm loss is close to the disc, there is a need to re-establish the length first to reach the sediment surface for food gathering and ventilatory purposes. Filter-feeding epifaunal brittle stars may also fall into this category, as they experience rapid regeneration rates (Sides, 1987). However, in those species that are more reliant on their arms for locomotion, the need to grow muscles and tube feet is more important to re-establish functional recovery.

From anatomical analysis, the morphology of the regenerated arms (10 weeks) of animals in both temperature treatments appeared fully formed with ossicles, spines and pigment similar to those of adjacent non-regenerating arms and the histology indicated that the regenerated arms also had well-developed muscles and overall normal tissue morphology. This shows the importance of a return to a functional arm to facilitate prey capture and escape. In contrast, the lower muscle density in regenerated arms of O. ophiura was hypothesized to indicate a strategy to recover sensory functioning before movement (Wood et al., 2010). The low growth rate observed in the deep-water, heavily calcified species Ophiocten sericeum may be due to the short time period of the study (20 days at 5 and 8.5°C), with the animals potentially still being in the regeneration lag phase as a result of the energy needed for calcification. The polar brittle stars Ophionotus victoriae and O. ophiura, exhibit a very long lag phase (5–7 months) before any measurable growth is observed (Clark et al., 2007; Clark and Souster, 2012). In these polar species, regeneration occurs very slowly, and calcification levels are low as a result of the low calcite saturation (Watson et al., 2012), while food availability and assimilation rates are low and seasonal (Peck, 2016).

Ophioneries schayeri in the warmer treatment not only regrew their arms faster but also regained arm functionality more rapidly. While arm growth at +3°C slowed at the end of the experiment, lengths were still increasing. In contrast, arm length regrowth of the individuals in the colder treatment appeared to have reached a plateau, although the arms continued to increase in width, further reinforcing the idea that there is an emphasis on achieving a fully functional arm over length.

The marked difference in arm regeneration rates due to temperature came at a high cost, as evidenced by the significantly higher metabolic rate of O. schayeri in the warm temperature treatment at week 8 (Fig. 1). Not only were the brittle stars dealing with regeneration but they also had the additional stress of elevated temperature. This is interesting in that increased respiration leads to a decreased scope of growth when upper thermal limits have been exceeded (Jutfelt et al., 2021). This supports the idea that brittle stars prioritize regeneration over all other processes (Pomory and Lawrence, 2001). While regeneration itself does not appear to be particularly stressful at the lower temperature (15.5°C, lack of increased respiration rate and low mortality rates), at the elevated temperature (+3°C), the cost increases. Regeneration may pose additional risks to the brittle stars as regenerating animals in both temperature treatments experienced mortality. Increased mortality among regenerating O. longicauda in response to time and temperature would support this hypothesis (Weber et al., 2013).

The greater mortality of O. schayeri in the warm (+3°C, 18.5°C) treatment was surprising, not only because this temperature is normally experienced in late spring but also given the results of Christensen et al. (2011) where this species tolerated 20°C. The cumulative effects of exposure to warm conditions may have been detrimental to O. schayeri. While organisms may tolerate thermal changes over the short term, thermotolerance decreases with exposure time (Pörtner, 2010). Some organisms are unable to fully acclimate to the change in regimes, particularly when those changes occur out of season (Harianto et al., 2021) and with prolonged exposure. The elevated respiration rate and Q₁₀ suggest that the animals had upregulated additional metabolic pathways, with the renewal of gametogenesis, which occurs at this time (Selvakumaraswamy and Byrne, 1995), potentially being one of these.

While not always leading to instantaneous mortality, heatwaves can reduce survivorship and increase susceptibility to disease. The sea urchin Heliocidaris erythrogramma survived a 10 day ‘summer’ heatwave (+2°C medium and +3°C strong), but experienced increased mortality in the days following a return to control temperatures (Minuti et al., 2021). This species also experienced elevated respiration rates that persisted at least 10 days after the return to control temperature. The temperature regime applied in the present study emulated a winter heatwave, resulting in elevated respiration rates and high mortality rates in O. schayeri. It is a concern that the respiration rates remained elevated for so long (at least 4 weeks), indicating that O. schayeri may be particularly vulnerable to ocean warming, especially marine heatwaves. As weather patterns continue to change, the extremes of summer are not the only concerns as shown here in the sensitivity of O. schayeri to warming in winter. Given the high levels of sublethal predation and incidence of regenerating O. schayeri in nature, the higher mortality of the regenerating brittle stars in the elevated temperature treatment is also a concern. Reduced survivorship of regenerating animals during heatwaves may lead to higher incidences of mortality and local extirpation events.

Seasonal acclimation in winter is a well-documented phenomenon in fish and terrestrial inhabitants, when enzyme function and performance levels are decreased, presumably as a result of colder temperatures. Windows of thermotolerance also shift in cooler months (Richard et al., 2012), potentially making exposure to a ‘normal’ summer temperature harmful, if not lethal. Climate change increases the rate of ocean warming and the frequency of elevated temperature at heatwave levels, as now reported to occur across seasons. An early warming trend for the Austral winter (Hobday and Lough, 2011) has already led to phenological shifts in marine taxa, resulting in ecological mismatches such as with reproduction, feeding and seasonal productivity (Poloczanska et al., 2013, 2016). In the present study, the increased temperatures out of season caused a large increase in respiration in O. schayeri and would likely place growth and reproductive success at risk as more resources would be needed to support basic metabolism and coping mechanisms. The energy needed to support increased metabolism may also lead to an increase in foraging behavior to gather food, placing the brittle stars at a higher risk of predation. While the ability to regrow lost arms quickly at the higher temperature comes with an increase in energy demands (positive feedback cycle), it also comes with a higher mortality rate.

We thank the following people for their support during this research: Sergio Torres Gabarda (gathering animals, tank set up and maintenance, encouragement), Matt Clements (arm measurements), Hamish Campbell (arm measurements), Andrew Niccum (aquarium system maintenance) and Dr Matthew Pyne (Lamar University, USA; statistics). A special thanks to Mathew Downs (Zoology Department, University of Otago, New Zealand) for his expert assistance with histological examination. This is SIMS contribution number 303.