Heart rate response and bimodal gas exchange in three developmental stages of the bullfrog Lithobates catesbeianus (Anura: Ranidae)

ABSTRACT Measuring cardiorespiratory variables can be challenging in developing animals, especially when they use bimodal gas exchange to maintain metabolic activity. In tadpoles, gas exchange may occur through the integument and gills when breathing in the water and through the lungs when breathing air, with varying contributions of each respiratory structure during development. The interaction between aquatic and air breathing results in a complex physiological response that may affect the cardiac cycle. Measuring the heart rate (fH) together with aquatic and aerial gas exchange in anurans during their development can be challenging, since it may involve handling small animals and/or a certain degree of invasiveness (i.e., surgery to implant electrodes). Here, we evaluated concomitantly aquatic and aerial gas exchange, lung ventilation, and fH in three stages of development of the bullfrog Lithobates catesbeianus (Shaw, 1802). We built a novel, noninvasive, closed respirometry system capable of measuring fH, aerial and aquatic gas exchange simultaneously in animals of different sizes. Our integrative analysis revealed a decrease in the heart rate and an increase in oxygen consumption during the developmental stages of the bullfrog, but there was no adjustment of heart rate after or during air breathing. Moreover, tadpoles in metamorphosis showed higher oxygen consumption in air than in water, while aquatic breathing was responsible for releasing CO2. Our results are consistent with those found in the literature, yet our study represents the first non-invasive investigation to evaluate bimodal gas exchange and heart rate simultaneously. Moreover, our setup holds potential for further advancements that would allow for controlled water and air composition. This tool could greatly facilitate the investigation of how cardiorespiratory physiology responds to varying environmental conditions.


INTRODUCTION
The exchange of gases is essential for sustaining an organism's metabolic activities, involving the uptake of oxygen and the removal of carbon dioxide.Gas exchange between animals and the environment must occur cons tantly during all stages of development, even as distinct respiratory organs are being formed (Burggren and Doyle 1986a).During anuran development, these animals rely on a combination of respiratory organs, including the skin, gills (external or internal), and lungs, to extract oxygen from water and air (Burggren and West 1982).At the beginning of development, oxygen requirements for the embryo are met by diffusion through the skin (Burggren andInfantino 1994, Warkentin 2007), while larger larval stages utilize either aquatic respiration through gills and skin, and/or aerial respiration through lungs (if present) (Crowder et al. 1998).Bullfrog tadpoles use all three respiratory gas exchange surfaces during the premetamorphic and early metamorphic stages of development (Atkinson andJust 1975, Burggren andWest 1982).However, tadpoles of other species develop their lungs well before metamorphosis, displaying airbreathing behavior and lung inflation a few days after hatching (Schwenk and Phillips 2020).
Air breathing can be considered as an additional means to supply oxygen to meet metabolic demands, particularly in physiologically demanding environments.Various abiotic and biotic factors can influence the partial pressure of oxy gen (PO 2 ) in water, leading to diverse respiratory responses of the animal (Burggren and Infantino 1994).Chronic hypox ic conditions in water can increase respiratory surface area, respiratory function of the gills, and skin vascularization (Pinder and Burggren 1983), while acute hypoxia might increase gill and lung ventilation (Wassersug and Seibert 1975, Feder and Wassersug 1984, Burggren and Infantino 1994).The interplay between aquatic and air breathing is highly intricate and mediated by receptors associated with gills and lungs, as lung ventilation inhibits gill ventilation, possibly to reduce oxygen loss from the blood into the water (West andBurggren 1983, Feder andWassersug 1984).
Cardiorespiratory coupling consists of the interaction between the ventilatory cycle and the cardiac cycle, decreas ing (bradycardia) or increasing (tachycardia) heart rate in intermittent air breathers (Pinder and Burggren 1983).Adult frogs exhibit tachycardia shortly after lung ventilation and bradycardia in response to environmental hypoxia, although tadpoles apparently do not alter their heart rate in response to ventilation or aquatic hypoxia (West andBurggren 1982, Burggren andDoyle 1986b).Heart rate alterations during development, from egg to adult, may differ among species (Burggren and Pinder 1991).Burggren and Doyle (1986b) showed a decrease in resting heart rate during bullfrog development, with hatched larvae displaying the highest resting heart rate, which decreases throughout development.However, the decrease in heart rate during development could be linked to allometric scaling rather than other fac tors associated with ontogeny or developmental processes (Burggren and Pinder 1991).
Measuring heart rate simultaneously with aquatic and aerial gas exchange can be challenging depending on the developmental stage, as it might involve small animals and/or a certain degree of invasiveness (such as electrode implanting surgery).Longhini et al. (2017) established a noninvasive tech nique for measuring heart rate and buccal movements in larg er premetamorphic tadpoles, without accessing air breathing.
In the present study, we examined bimodal gas exchange in developing tadpoles using an innovative noninvasive system that concurrently measures aquatic and aerial gas exchange, lung ventilation, and heart rate in three developmental stages of the bullfrog Lithobates catesbeianus (Shaw, 1802).

Animals used and their maintenance
Measurements of heart rate and gas exchange in air and water were conducted on three groups representing different developmental stages of L. catesbeianus, following the classi fication of Gosner (1960).The stages, number of animals and mean body mass ± SD were as follows: Larval stages 28-36 (N = 5, 1.56 ± 0.19 g), characterized by the beginning of hind limb development; premetamorphosis, stages 39-41 (N = 15, 4.32 ± 0.18 g), with fully formed hindlegs; and metamorphosis, stages 42-43 (N = 5, 3.14 ± 0.24 g), with front limbs externalized.The animals were acquired from the bullfrog farm (Centro de Aquicultura da Unesp -CAUNESP) at the College of Agri cultural and Veterinary Sciences, São Paulo State University (FCAVUNESP) in Jaboticabal, São Paulo State, Brazil.All tadpoles were maintained under natural photoperiod (12:12 h light:dark) at 25 ± 1°C in a tank (50 x 50 x 40 cm) filled with dechlorinated tap water continuously aerated by an air pump.The tadpoles were fed commercial fish food daily, except for 24 hours prior to measurements.The experimental approach was approved by the Ethics Committee on the use of Animals of University of São Paulo, campus Ribeirão Preto (CEUAFF CLRPUSP, Protocol 17.5.119.59.3).

Heart rate, aerial and aquatic breathing
We constructed a unique apparatus capable of simul taneously measuring heart rate (f H ), aerial and aquatic gas exchange in animals of varying sizes.We used a plastic syringe (60 mL) in a horizontal position as a respirometer chamber, since closed respirometry can be performed with small animals using plastic syringes (Stevens 1992, Lighton 2008).Bimodal respiration was possible by filling the syringe partly with dechlorinated water and the other part with atmospheric air.The chamber volume was adjusted to a tadpoles' size, ranging from 10 to 15 mL air and from 15 to 20 mL water.
Heart rate was measured through noninvasive elec trocardiogram (ECG) recordings following Longhini et al. (2017) and Altimiras and Larsen (2000).We used two pieces of wire (steam Ø 1.0 mm) placed in parallel (about 2 cm apart) and fixed perpendicularly at the bottom of the syringe (horizontally positioned).Both electrodes were connected to a differential AC amplifier (AM Systems, model 1700, Se quim, WA, USA) by a cable (3' with 5pin).The amplifier was configured to record with a gain of 10k, high cutoff 5Hz and lowcutoff 0.1-300 Hz (according to animal size).The signal was recorded (1 kHz sample rate) by a PowerLab acquisition system (ADInstruments, Sydney, Australia), using digital filters (bandpass 100 Hz low and 5 Khz high) of Labchart software (version 8, ADInstruments) (Fig. 1).We used two systems to continuously measure gas exchange in water and air during the experiment (Fig. 1).Aquatic gas exchange: on the underside of the syringe, two aluminum tubes were glued (Ø 4.0 mm) to each end of the syringe to insert probe sensors for aquatic O 2 and CO 2 .These sensors were connected to O 2 (Firesting) and CO 2 (Presens) analyzers which were connected to a computer and data recorded using the Pyron Oxygen logger (O 2 ) and the Presens measurement studio 2 (CO 2 ) software (Fig. 1).Air system: two aluminum connec tors (Ø 3.0 mm) for both air inlet and outlet were glued to the top of the syringe.The outlet connector was connected by a hose to a desiccant box upstream to a gas analyzer (ADInstruments) and at the outlet of the gas analyzer was connected to the input of the respirometer, achieving close system respirometry.The gas analyzer was connected to the PowerLab acquisition system (Fig. 1).

Experimental protocol and data analysis
Each animal was individually measured at constant water (24 ± 0.3 °C) and air (25 ± 0.3 °C) temperature.All experiments were conducted on unanesthetized and un restrained animals, which were kept in the respirometer chamber for at least 40 minutes before the measurements began.During the acclimatization period, the aerial phase was kept open to the surrounding air in the room.After the acclimatization period, all the water within the respirome ter was gently replaced by airsaturated water at the same temperature, with care taken not to disturb the tadpole.Following this step, the respirometer was sealed (both air and water) to conduct measurements for 30-40 minutes.Rodgers et al. (2016) showed that background respiration caused by microbial growth was not significant for several hours of measurement.In our setup, measurements of background respiration for one hour have not yielded sig nificant microbial oxygen consumption, suggesting that two 40minute periods of freshly inserted water samples into the syringe did not lead to significant background respiration.
Occasionally during experiments, a tadpole would start floating just above the electrodes, which required adjust ments to the amplifier's lowcut setting in order to enhance the signal quality.Additionally, when a tadpole ascended to the surface to breathe air, noise was generated and the sig nal was lost, but the signal was quickly reestablished when the tadpole submerged itself again onto the electrodes.As a result, the animals maintained contact with the ECG elec trodes for a significant portion of the experiment, providing nearly continuous f H recordings throughout the procedure (Fig. 2A).Heart rate data analysis was carried out using Labchart software by applying the cyclic measurement tool with cyclic detection by ECG mode, adjusting the QRS width and filtering high pass with a 0.16Hz cutoff.We analyzed samples (at least one minute of continuous recordings) at  intervals 5, 10, 20, 30 and 40 minutes.We considered the movements performed by a tadpole within a chamber to be sufficient to allow for mixing of water, but we did not test for the existence of an oxygen gradient within the aquatic phase (Rodgers et al. 2016).
Air and aquatic gases were continuously measured throughout all experiments.We used the rate of PO 2 decline and PCO 2 increase in air and water to calculate MO 2 and MCO 2 in air and water for each individual.Massspecific MO 2 and MCO 2 (μmol gas.g 1 .h 1 ) in the aerial and aquatic phases were calculated.Following Lefevre et al. (2011), the fall in PO 2 or the increase in PCO 2 during time (h), the volume of air or water (L: for the air phase, the volume of air in the closed system; for the aquatic water phase, the volume of water in the res pirometer minus the tadpole's mass), the capacitance of O 2 and CO 2 (μmol mmHg −1 L −1 ) in air and water at experimental temperatures, and body mass (g) were used to calculate gas exchange.By constantly measuring the aerial phase, we were able to tally the occurrences of ventilatory events for each individual throughout an experiment (Fig. 2B).
To minimize visual disturbances to the animals, the respirometer was covered with an opaque material.After completing the experiment, the tadpoles were staged, gently dried using paper towels, and weighed to the nearest 0.001 g.

Statistical analysis
We adjusted a generalized linear model (GLM) to analyze all the measured variables.ANOVA test was per formed to evaluate each GLMs and its components, and Tukey's post hoc test to obtain the pairwise comparisons.Specifically, we compared heart rate during experimental time and among the three groups, and gas exchange (MO 2 and MCO 2 ) in air, water and among tadpoles` stages.To examine the influence of body mass (g) on wholebody gas exchanges (μmol gas h 1 ), we performed a linear regression between log 10 of the MO 2 and MCO 2 in the air and water by log 10 of body mass, adjusting two lines for aerial and aquatic gas exchanges.All models met the assumption of homo geneity of variance (Levene) and normality distribution (ShapiroWilk).The level of significance for all analyses was 0.05.Statistical analyses were conducted in Jamovi software, version 2.3 and GraphPad Prism, version 6 for Windows from GraphPad Software (San Diego, California USA), was used to plot the graphics.

RESULTS
Variability in heart rate (f H ) was observed throughout the experimental duration (Fig. 3A) across all stage groups, but without significant difference (F 4, 115 = 0.62, p = 0.65).Nev ertheless, the f H varied significantly among different develop mental stages (F 4, 115 = 12.6, p < 0.001) (Fig. 3B), since the larval group showed a higher f H than the premetamorphic (Tukey p = 0.004) and metamorphic (Tukey p < 0.001) groups.There was no significant difference between premetamorphic and metamorphic stages (Tukey p = 0.07) (Fig. 3B).While animals remained submerged on top of the ECG electrodes for most of the time, tadpoles periodically surfaced to breathe air.This continuous behavior allowed for the simultaneous measurement of aerial and aquatic gas exchange.Ventilation events (mean ± SE) pertaining to air intake were observed in premetamorphic (10.6 ± 0.82) and metamorphosed tadpoles (15 ± 3.83).While larvalstage tadpoles also engaged in aerial gas exchange, the system sensitivity was inadequate for detecting ventilatory events in this stage.Furthermore, no discernible heart rate adjust ments were noted after or during air ventilation.
There were significant differences in massspecific oxygen consumption among stage groups and between aerial and aquatic phases (Table 1).In the aerial phase, tadpoles in metamorphosis consumed more oxygen than those in premetamorphosis (Tukey p < 0.001) and larval (Tukey p < 0.001) stages (Fig. 4A, Table 2).In the aquatic phase, premeta morphic tadpoles showed higher oxygen consumption only than larvae (Tukey p = 0.004).Moreover, only metamorphic tadpoles did take up significantly more oxygen in air than in water (Tukey p < 0.001) (Fig. 4A).Tadpoles in all stages released more CO 2 in water than air (Tables 1, 2) (Fig. 4B), with premetamorphic (Tukey p < 0.001) and metamorphic (Tukey p < 0.001) animals secreting more CO 2 in water than the larval stage.On the other hand, in the aerial phase there were no significant differences among stage groups (Table 2) (Fig. 4B).

DISCUSSION
Our integrative analysis revealed noteworthy variations in heart rate and oxygen consumption among different de velopment stages of L. catesbeianus.The findings showed a decrease in heart rate and an increase in air breathing as de  A B velopment progressed.Specifically, the larval group exhibited a higher heart rate compared to both the premetamorphic and metamorphic groups, likely attributed to alterations in the intrinsic frequency of the cardiac pacemaker and allome tric scaling effects (Burggren andDoyle 1986b, Burggren andPinder 1991).Moreover, the process of metamorphosis entails significant body reorganization in tadpoles, including reshap ing and repositioning of the heart to accommodate the new body plan (Sandoval et al. 2022).These transformations within the cardiovascular system also contribute to the reduction in heart rate as the tadpole undergoes its transition into an adult form.While fluctuations in f H were observed across all groups during the experiment, these variations lacked significance, suggesting that the animals were not in a stressed state.
Tadpoles of some species inflate their lungs very early in development (Crowder et al. 1998, Phillips et al. 2020, Schwenk and Phillips 2020) and employ strategies like bub blesucking or breachbreathing to take in air.Bubblesucking is often utilized by smaller tadpoles, while breachbreathing is favored by larger ones, involving breaking the water's surface tension (Schwenk and Phillips 2020).We observed ventilatory events in intermediatestage tadpoles (larval group) using bubblesucking, while the system only detected such events in larger stages (premetamorphic and metamorphic groups).Smaller tadpoles employing bubblesucking might move smaller gas volumes, potentially falling below the system's detection threshold.
During the larval stage, tadpoles exhibited similar ox ygen consumption rates in both air and water, indicating ef fective bimodal respiration catering to their metabolic needs.Interestingly, this trend persisted in the premetamorphic stage, with a slight upturn in aquatic oxygen consumption.It is likely that at this stage, the skin plays a prominent role in gas exchange due to ongoing limb formation in the cranial part of the abdominal cavity, possible hindering pulmonary ventilation and air breathing.However, as limbs become ex ternalized in the metamorphic stage, air breathing becomes dominant, enabling a transition from aquatic to terrestrial environments (McDiarmid and Altig 1999).
The results showed that the release of CO 2 was consis tently higher in water across all developmental stages.This suggests a significant contribution of cutaneous gas exchange in water during this stage, concomitantly to gill involution during metamorphosis.Furthermore, as air breathing becomes more effective (metamorphic stage), respiratory organs exhibit specialization: lungs primarily handle oxygen uptake, while skin primarily facilitates CO 2 release into water (Burggren and West 1982).
The observed variation in gas exchange can be attribut ed to changes in animal mass (see Material and Methods).Our results demonstrated a positive relationship between animal body mass and gas exchange (MO 2 and MCO 2 ) in water, a relationship welldocumented in the literature (Lighton 2008, Kozlowski et al. 2020).However, the absence of correlation bet ween body mass and air breathing was due to animals in the premetamorphic stage.These animals presented higher body mass coupled with lower metabolic rates, likely influenced by tissue reorganization processes and the potential interference of developing limbs on pulmonary ventilation.Conversely, metamorphic stage tadpoles exhibit lower mass due to tail absorption, resulting in a positive correlation between body mass and metabolic rate when the premetamorphic stage data is excluded from the analysis.
Studies investigating heart rate in conjunction with aquatic and aerial breathing have been conducted in certain fish species (Burggren 1979, Sacca and Burggren 1982, Bar rionuevo and Burggren 1999, Altimiras and Larsen 2000).While increased heart rate linked to aerial respiration has been noted in air breathing fish (Singh and Hughes 1971), tadpoles, on the other hand, do not seem to display venti latory tachycardia associated with air breathing (West and Burggren 1982).In our study, we similarly did not observe an adjustment in f H linked to ventilatory events.This could be attributed to the tadpoles' less developed nervous system for cardiorespiratory regulation, immature peripheral reflexes governing heart rate, or the dominance of skinmediated gas exchange, which might keep breathing rates constant in water (Buggren and Doyle 1986b).Longhini et al. (2017) were the first to use a noninvasive approach to measure f H and gill ventilation in premetamor phic tadpoles of L. catesbeianus.Our study builds on this by demonstrating the feasibility of simultaneously measuring aerial and aquatic gas exchange alongside heart rate in con scious tadpoles.Although some data on cardiorespiratory physiology in frog larvae exist, comparing our results with those of prior studies is challenging due to methodological disparities.These differences involve analyzed larval stages, temperature conditions, experimental setups, invasiveness levels, and acclimatization times.For example, Burggren and Doyle (1986b) recorded f H in various bullfrog developmental stages at 20 °C, obtaining 40 -50 bpm in larval, < 40 bpm in premetamorphic, > 40 bpm in metamorphic tadpoles, and 30 bpm in adults.Longhini et al. (2017) measured a f H of < 20 bpm in premetamorphic stages at 15 °C, < 60 bpm at 25 °C and < 80 bpm at 30 °C.West and Burggren (1982) showed that premetamorphic bullfrogs increased lung ventilation from 12 to 51 events per hour during increasing hypoxia (O 2 decreasing from 82 to 21 mmHg), without altering f H (⁓50 bpm at 20 °C).Burggren and West (1982) also quantified MO 2 and MCO 2 across skin, gills, and lungs in different developmental stages.They did not find lung ventilation in larval stages, but skin and gill MO 2 (3.4 and 2.3 μmol O 2 .g 1 .h 1 , respectively) and MCO 2 (2.6 and 1.7 μmol CO 2 .g 1 .h 1 , respectively) gas exchange, while premetamorphic gas exchange through skin, gills, and lungs were, respectively, MO 2 4.6, 1.2 and 1.2 μmol O 2 .g 1 .h 1 , and MCO 2 4.2, 3.0 and 0.2 μmol CO 2 .g 1 .h 1 .In postmetamorphic animals without gills, cutaneous and pulmonary gas exchange contributed (MO 2 0.7 and 4.3 μmol O 2 .g 1 .h 1 ; MCO 2 4.9 and 0.6 μmol CO 2 .g 1 .h 1 , respectively).Despite methodological disparities, our findings align with existing literature in response patterns.This suggests that our experimental setup is dependable for assessing the cardiore spiratory system, even while concurrently measuring oxygen consumption in aquatic and aerial phases.
While there is some literature on L. catesbeianus tadpole's gas exchange and heart rate, our study stands out as the first noninvasive investigation to concurrently eval uate bimodal gas exchange and heart rate across different developmental stages.Furthermore, our approach holds the potential for further enhancements, enabling controlled modifications in water composition (such as pH, tempera ture, or pollutant levels) and air composition (such as gas concentrations), ultimately reducing extraneous factors that affect the capture of electromyogram signals associated with gill ventilation or to f H in smaller animals.Through these enhancements, larval model organisms could serve as valuable tools for exploring cardiorespiratory physiology under varying environmental conditions.

Figure 1 .
Figure 1.Scheme of non-invasive apparatus to measure gas exchange in water (A) and air (B), and heart rate (C).

Figure 2 .
Figure2.Representative data recording of electrocardiogram (A) and aerial ventilation (B) in a premetamorphic Lithobates catesbeianus at 25°C.In B the signals show a ventilatory event where the tadpole renewed the air in its lungs, resulting in a marked drop in PO 2 and an increase in PCO 2 .Following the ventilatory event, the expired air was mixed with the remaining air within the closed respirometry system, resulting in a PO 2 slightly lower, and a PCO 2 slightly greater, than before ventilation.

Figure 3 .
Figure 3. Heart rate (f H ) variation during the experiment (A) and individual f H (B) in larval (blue line and points), premetamorphosis (orange line and points) and metamorphed individuals (green line and points) of Lithobates catesbeianus.All bars represent mean ± s.e.m.

Figure 4 .
Figure 4. Mass-specific oxygen consumption (A) and carbon dioxide released (B) for aerial (red lines and points) and aquatic (blue lines and points) gas exchange during development of Lithobates catesbeianus.

Figure 5 .
Figure 5. Relationship between Log 10 whole-body oxygen consumption (A) and carbon dioxide release (B) (µmol h -1 ) in air (filled symbols) and water (open symbols), and Log 10 body mass (g) in larval (blue triangles), premetamorphic (orange squares) and metamorphic (green circles) stages of Lithobates catesbeianus.Each point represents a measurement from a single animal.The regression lines correspond to aerial (red) and aquatic (blue) gas exchange.Dotted lines represent no significant correlation.

Table 1 .
GLM analysis of gas exchange (µmol gas g -1 h -1 ) in air and water among stage groups.

Table 2 .
Results of Tukey post-hoc comparison tests regarding gas exchange (µmol gas g -1 h -1 ) in air and water among stage groups.