Altering developmental oxygen exposure influences thermoregulation and flight performance of Manduca sexta Free

Flight evolved in insects only once, over 300 million years ago, with all winged insects stemming from that evolutionary event. In addition, many large flying insects have evolved the ability to regulate body temperature well above ambient temperatures. Endothermy most likely evolved in Protodonatan insects over 100 million years ago (May, 1982, 1979). This is presumed to be due to the advent of large highly active flight muscle and larger body size. In these species, elevated endothermic flight muscle temperatures are required to achieve fast muscle contractions necessary for flight and these are achieved and maintained by heat produced by flight muscle (Heinrich, 1974). Endothermic, flying insects are capable of some of the highest recorded metabolic rates (Harrison and Roberts, 2000). These metabolic rates are achieved through aerobic respiration and therefore require an adequate and rapid oxygen delivery system. This delivery is accomplished by a branching network of trachea and associated air sacs involved in convective movement that serve the insect tissues directly with oxygen. While this is an extremely high-capacity system, it does have limitations when ambient oxygen is reduced.

There are many studies that have examined the effect of oxygen exposure on the insect tracheal structure (Harrison et al., 2018; VandenBrooks et al., 2018), and body size and growth (Frazier et al., 2001; Harrison et al., 2015; Harrison and Haddad, 2011; Wilmsen and Dzialowski, 2023). At rest, most insects have an outstanding tolerance for hypoxia, with critical oxygen levels as low as 5–10 kPa (Greenlee and Harrison, 2005). However, safety margins for oxygen delivery are reduced during flight, whether that is tethered flight (Rascón and Harrison, 2005; Snelling et al., 2017) or free flight (Harrison and Lighton, 1998; Henry and Harrison, 2014; Joos et al., 1997). During flight, these critical oxygen levels rise to 15–20 kPa (Henry and Harrison, 2014; Snelling et al., 2017). Reduction in hypoxia tolerance during increased aerobic activity may be ameliorated by phenotypic plasticity when developing under altered oxygen levels. Most studies on adult insects use acute oxygen changes or temperature to test the limits of the insect tracheal system. There are limited studies looking at the effects of chronic larval developmental exposure on adult energetics and flight. Hypoxic and hyperoxic exposure during larval and pupal development of Manduca sexta result in changes in adult body size (Wilmsen and Dzialowski, 2023). Hypoxia results in smaller adults, while exposure to hyperoxia produces larger adults. Shiehzadegan et al. (2017) found that developing under hyperoxic and hypoxic conditions influenced the flight probability at low oxygen concentrations in Drosophila melanogaster. The potential role that developmental phenotypic plasticity during larval and pupal development may play in shaping the morphology and physiology of adult Manduca in response to these altered oxygen conditions is unknown.

In this study, we looked at the effects of three chronic developmental oxygen exposures (10%, 21% and 30% O₂) during larval and pupal development on the endothermic performance and flight critical oxygen levels in the adult hawk moth M. sexta. These moths are a well-established physiological model system for studying development in holometabolous insects. They are relatively large and are obligate endotherms during flight, necessitating an internal thorax temperature of up to 40°C, which is achieved by a 5-fold increase in metabolic rate from their resting levels (Heinrich, 1971a,b). We coupled chronic hypoxia and hyperoxia treatments with acute hypoxia and hyperoxia exposures to understand the developmental plasticity and limitation of oxygen delivery in this system. We expected animals reared in hypoxia to maintain increased metabolic rates and flight speeds as adults under acute hypoxic conditions when compared with those raised in hyperoxia and normoxia. This should be the case if hypoxic rearing induces the development of increased surface area in the tracheal system.

Manduca sexta (Linnaeus 1763) eggs were purchased from Carolina Biological Supply. Upon arrival, eggs were carefully separated and placed in containers with mixed hornworm media (Carolina Biological Supply) to hatch into larvae (see below). Once larvae were big enough to handle without harming them (about 11 days after arrival), they were weighed, an ID was assigned to each larva, and they were moved to individually labeled vials with the same hornworm media. Larvae were allowed to grow, and tubes were cleaned every other day to reduce mold growth. When larvae in their 5th instar had reached the wandering phase immediately prior to pupation, they were weighed again and placed in a dark vial containing wood shavings for a week to pupate. After 5 days in the dark, pupating larvae were checked daily to see whether they had completed pupation. Once they had pupated, they were removed from the dark vial in order to prevent them from entering diapause, weighed and placed into large plastic bins (57.2×40.6×32.4 cm). The pupae were checked daily to determine when they eclosed. Adult moths were marked with Testors™ acrylic paint, to keep track of the date they hatched, and maintained in the plastic bins. Adults were provided access to 25% sugar water ad libitum until they were used for experiments. The moths were weighed a final time 3–4 days after they had eclosed. Moths were maintained on an 18 h:6 h light:dark cycle, at 25°C, and a humidity of ∼38%.

In order to quantify the developmental effects of chronic hypoxia or hyperoxia exposure on the preflight warm-up metabolic rate of adult M. sexta, once eggs were received, they were immediately placed in one of three ­O₂ levels: 10%, 21% or 30% O₂. The animals were maintained chronically at these O₂ levels throughout their entire development through the adult stage. To maintain constant O₂, the vials containing the developing animals were place in tightly closed plastic bins and oxygen was regulated to within 0.2% O₂ of setpoint levels with a ROXY-4 gas regulator (Sable Systems International) and oxygen sensors (Maxtec, LLC). To maintain 10% hypoxia, the ROXY-4 controlled the influx of excess nitrogen gas into the chamber and for 30% hyperoxia, it controlled the influx of excess oxygen into the chamber. Small fans inside the bins pulled in some outside air to ensure humidity levels did not drop too low and circulated the air to ensure that the O₂ level was consistent throughout the container.

To look at the metabolic response as the pre-flight warm up metabolism to different chronic developmental oxygen levels, we measured metabolic rate in thermoregulating adult moths at various acute oxygen concentrations (5%, 10%, 21%, 25% and 30% O₂). All moths used were 3–4 days post-eclosion because this is the age when metabolic rates peak in adult moths (O'Brien, 1999). Adults were placed in a flow-through metabolic chamber (364 ml) that had a built-in IR window (32-806, Edmunds Optics) in the top to allow for simultaneous recording of thorax temperature and metabolic rates. The metabolic chamber was placed into a large custom-built temperature-controlled chamber and metabolic rates were measured at an air temperate of 25°C. During a trial, a moth was exposed to only three of the six different oxygen levels in a random order, either 5%, 10% and 21% or 21%, 25% and 30%. Two normoxia-reared moths, one hypoxia-reared moth and one hyperoxia-reared moth were exposed to all five levels of oxygen. The moth was allowed to acclimate to the given oxygen level for 5 min, then it was briefly agitated by shaking the metabolic chamber; there were beads on the floor of the metabolic chamber to aid this process without having to open the chamber. These moths were agitated to induce endothermic thermoregulation but were not flying. Once the moth started thermoregulating, a 5 min IR video was recorded using a FLIR A600 series IR camera (Fig. 1). The thermal images were corrected for the incomplete IR transmission of the thermal window. We measured the IR temperature through the IR window of a black body with known temperature ranging from 23 to 43°C. The calibration correction factor was determined by linear regression between the actual temperature (T_a) and the temperature measured through the IR window with the IR camera (T_ir), yielding the equation: T_a=1.3T_ir−6.9.

Metabolic rate was measured as CO₂ production (V̇_CO2; ml CO₂ h⁻¹). Dry CO₂-free air was pumped through the metabolic chamber at a rate of 200 ml min⁻¹. Gas went from the metabolic chamber outflow, through Nafion tubing surrounded by Drierite to remove water, and then to a calibrated Sable Systems International CA-10 carbon dioxide analyzer. A gas mixer (Brooks Instruments) was used to control the oxygen concentration throughout the experiment using a mix of oxygen balanced with nitrogen. The inflow gas to the metabolic chamber was subsampled and passed through a calibrated Sable Systems International SC-1B O₂ analyzer to verify the oxygen level. Carbon dioxide production was calculated as V̇_CO2=CO₂×f, where CO₂ is the excurrent fraction of carbon dioxide and f is the flow rate (ml h⁻¹) of gas through the chamber. Once the IR video was taken, the oxygen level in the chamber was changed, and the moth was allowed to acclimate to the new oxygen level for 5 min. This was repeated until IR videos had been taken at all three oxygen levels. The moth was weighed after the exposure was complete.

The critical oxygen level for flight was measured in adult moths 3–4 days post-eclosion using a flight mill (modified from https://tfrec.cahnrs.wsu.edu/vpjones/flight-mill-studies/#fmworks, accessed 1 February 2022). The day before the flight trial, moths were placed in the freezer for 5 min to reduce movement and then a small paintbrush was used to remove a patch of scales on the thorax. The cleared cuticle was gently roughened with fine grit sandpaper (P320) to allow the glue to stick, and cyanoacrylate glue was used to adhere a small 3 mm diameter×0.7 mm magnet (K&J Magnetics Inc.) to the thorax.

We ran flight trials at the beginning of the dark cycle with only yellow light to simulate dusk conditions, which is when these moths are most active. Moths were attached to the arm of the flight mill using the magnet glued to their thorax. Moths usually started flying immediately upon attachment to the flight mill and only small adjustments were needed to ensure they were flying in the correct orientation. The flight mill was placed in a large wooden box with a removable windowed lid that had a built-in fan to circulate air within the chamber. Each moth was flown for 5 min in normoxic air with the lid off to get baseline flight performance. The lid was then placed on the chamber and nitrogen was pumped into the wooden box at a rate of 3000 ml min⁻¹. This caused the oxygen level to decrease at a rate of 0.5% O₂ min⁻¹, such that it took 22 min for the chamber to reach 10 kPa (Fig. S1). The flight mill had a Hall effect sensor that recorded each rotation around the mill on a Powerlab 8_SP (ADInstruments). Based on the flight mill arm radius (19.0 cm) and the number of rotations per minute, flight speed (cm s⁻¹) was calculated. Oxygen level was continuously monitored using an oxygen probe (Maxtec, LLC). This allowed us to determine the partial pressure of O₂ (P_O2) when the moth ceased flight. We noted this P_O2 as the P_crit, or the critical limit of oxygen needed for flight. After the moth stopped flying, the lid was removed, and the moth typically began to fly again in normoxia. We allowed the moths to fly for another 5 min to ensure that they had not ceased flight due to exhaustion.

Average maximum thorax temperatures were calculated from the IR videos using FLIR Research Software. A region of interest (ROI) was positioned over the IR window of the metabolic chamber in each video. Maximum thorax temperature was taken to be the maximum temperature within that ROI for each frame. These maximum values were exported to an Excel file and analyzed with a custom-written Python script. A rolling average of 20 values of the maximum temperature within the ROI was calculated and the highest rolling average was used for each video; this value should correspond with the maximum metabolic rate measured for each trial.

For whole-animal metabolic rates and thorax temperatures at different acute oxygen levels, the data were analyzed separately as either acute hypoxia exposures 21%, 10% and 5% O₂ or acute hyperoxia exposures 21%, 25% and 30% O₂. Each dataset was analyzed using linear mixed models with chronic developmental oxygen level and acute oxygen level as independent factors. To account for repeated measures, the subject was used as a random factor in the model. Mass was initially used as a covariate in the analysis of the metabolic data and surface temperature data. Holm post hoc test was run to examine differences between chronic oxygen groups when appropriate. Mass was not a significant covariate factor in the initial analysis of surface temperature data, so was removed as a covariate in the analysis. We also fitted non-linear exponential plateau curves between metabolic rate and thorax temperature in each group to look at their relationship, to determine whether independent relationships for each chronic oxygen exposure or a single relationship of all oxygen exposures best described the data. The best-fit equation was in the form T_Th=T_M−(T_M−T₀)^−(k×MR) where T_Th is the thorax temperature, T_M is the maximum plateau temperature, T₀ is the temperature at a metabolic rate (MR) of 0, and k is a rate constant. We used linear regression to examine the relationship between developmental oxygen exposure and critical flight P_O2. One-way ANOVA was used to examine mass, critical flight P_O2 and flight speed, with developmental oxygen as the factor. Data are presented as means±s.e.m., except for metabolic rate data, which are the estimated means from the mixed model with mass as a covariate. Statistical analysis was conducted in GraphPad Prism or JAMOVI version 2.3.

There were no significant differences between chronic rearing oxygen treatments in the body mass of adults used in the metabolic experiments (Table 1).

Chronic oxygen exposure during development influenced the pre-flight warm up endothermic response of M. sexta. There was no effect of chronic developmental rearing O₂ level on metabolic rate during endothermic thermoregulation with acute oxygen changes (Fig. 2, Table 2). The CO₂ production from hypoxia-raised M. sexta trended lower at all acute oxygen exposures but was not significantly lower compared with that of moths raised in hyperoxia or normoxia. There were no differences between the metabolic response between normoxia- and hyperoxia-reared adults (Fig. 2). There was a significant main effect of chronic rearing condition on thorax temperature (Table 2, Fig. 3). Adults raised in 10% oxygen had a lower maximum thorax temperature than normoxia- and hyperoxia-reared adults in response to acute hyperoxia (Fig. 3B). There was also a significant effect of acute oxygen level on thermoregulation and metabolic rate within each group. All rearing treatments had significantly lower metabolic rates and thorax temperatures in acute 5% oxygen compared with acute normoxia (Figs 2 and 3). Chronic hyperoxia exposure did not have any significant effects on metabolic rate or thorax temperature during thermoregulation compared with normoxia. An exponential plateau curve was fitted for each developmental oxygen treatment: T_Th=T_M−(T_M−T₀)^−(k×MR). The data were best fitted with individual curves for each chronic treatment rather than a single curve representing all the data (F_6,108=6.89, P<0.0001; Fig. 4): hypoxia T_M=35.6°C, T₀=26.9°C, k=0.13; hyperoxia T_M=38.0°C, T₀=25.5°C, k=0.15; and control T_M=38.9°C, T₀=27.4°C, k=0.09. The hypoxia-reared animal's thorax temperature plateaued at a lower thorax temperature (T_M) than that of normoxia- or hyperoxia-reared animals.

When comparing critical oxygen levels for tethered flight, chronic rearing oxygen level had a significant effect (Fig. 5A, Table 3). There was a significant linear relationship between developmental O₂ level and critical flight oxygen (P_crit=0.28×P_O2+8.5; P=0.01). Hypoxia-reared adults had significantly lower flight P_crit compared with hyperoxi­a-reared animals (Fig. 5A). Hyperoxia-reared adults tended to stop flying at 17 kPa, while hypoxia-reared adults flew at O₂ levels as low as 11.5 kPa. We also examined differences in flight speed prior to decreasing oxygen and compared those across the different treatments. There were no significant differences in tethered flight speed based on chronic developmental O₂ exposure (Table 3, Fig. 5B).

In this study, acute and chronic exposure to hypoxia impacted thermoregulation and flight in the moth M. sexta. Chronic developmental hypoxia at 10% O₂ resulted in significantly lower thorax temperatures in adult moths. Metabolic rates and thorax temperatures were significantly lower in moths tested in acute 5% oxygen. However, the relationship between metabolic rate and thorax temperature was affected by chronic hypoxia with the exponential curve fit at a lower plateau thorax temperature, indicating that hypoxia-raised moths were warming less efficiently than the other two groups. Additionally, acute oxygen exposure of 10% or less decreased metabolic rate and thorax temperature in all the groups. Flight P_crit was affected by developmental oxygen in a linear fashion, with hypoxia-raised moths displaying a significantly lower critical oxygen level than hyperoxia-raised moths. These differences did not significantly impact flight speed; however, when taken together, flight was influenced by the rearing oxygen levels.

Acute changes in environmental oxygen levels are known to affect metabolic rates in insects. Few studies have examined the effect of chronic developmental oxygen levels on aerobic activity in insects. Here, we found a trend toward lower thermoregulating metabolic rate in moths raised in hypoxia that was consistent across all oxygen levels tested. Similarly, hypoxia- and hyperoxia-reared animals had significantly different thorax temperatures. Changes in metabolic rate and thorax temperature in response to acute oxygen treatment were similar across developmental oxygen groups. When looking within groups, hypoxia-reared insects had significantly lower thorax temperatures at 5% O₂ while the thorax temperature of hyperoxia-reared and control insects was significantly lower at 10% and 5% O₂ when compared with normoxia (21% O₂). There are mixed results in the few studies that looked at the question of how developmental hypoxia affects adult phenotype. One study (Shiehzadegan et al., 2017) found that development in hypoxia had little to no effect on the flight performance in D. melanogaster in response to acute hypoxia. Leaf-cutter bees (Megachile rotunda) showed increased viability when developing under chronic hypoxia during their pre-pupal diapause state (Abdelrahman et al., 2014) even though there were no changes in body mass or size. Multiple studies suggest that although body size is reduced in hypoxia-reared insects (Harrison and Haddad, 2011; Loudon, 1988; Peck and Maddrell, 2005), there tends to be an increase in tracheole branching that compensates for the lack of oxygen during development (Harrison et al., 2018; Henry and Harrison, 2004; VandenBrooks et al., 2018). It is possible that the changes in tracheal branching do not adequately compensate for the deficit of oxygen in M. sexta (Wilmsen and Dzialowski, 2024) and instead the compensatory mechanism is to decrease metabolic rate and therefore thorax temperature during energetically costly activities. The moths in this study were thermoregulating at elevated thorax temperatures and were active with values slightly below typical flight metabolic rates. Heinrich (1971a,b) measured tethered flight metabolic rates around 40 ml O₂ g⁻¹ h⁻¹ or 32 ml CO₂ g⁻¹ h⁻¹ when assuming a respiratory quotient (RQ) of 0.8 (Fernández et al., 2017) while Fernández et al. (2017) found slightly higher rates of 41 ml CO₂ g⁻¹ h⁻¹. The thermoregulating values for normoxia-reared animals in the current study were close to 30 ml CO₂ g⁻¹ h⁻¹. It is possible that we would see more stark effects of hypoxic and hyperoxic development if we had measured flight metabolic rates. However, in bumblebees, there are no differences in maximal flight metabolic rates in response to acute hypoxia between queens and workers (Walter et al., 2021). When we looked a thermoregulating metabolic rate in the same bumblebee species, we found differences between castes (Wilmsen, 2022). Moths are known to fly at lower thorax temperatures when exposed to low environmental temperatures (Heinrich, 1971a) and are capable of regulating thorax temperature during activity (Heinrich, 1971b). This may indicate that there are other physiological responses to chronic hypoxia that we did not measure such as changes in ventilation rate of the tracheal system, resulting in changes to convective oxygen delivery. Late 5th instar M. sexta have been shown to ventilate their tracheal system only in extremely low oxygen levels (1–2 kPa O₂) and decrease ventilation in 0 kPa (Greenlee et al., 2013). Adult Schistocerca americana raised in hypoxia have lower ventilation rates at 21 kPa when compared with the normoxia-reared adults (Harrison et al., 2006), indicating that exposure to chronic hypoxia induces changes in ventilation rates into adulthood in some insects. It remains to be seen whether ventilation rates of active M. sexta are influenced by the levels of developmental oxygen. We did not find any effect of chronic hyperoxia on thermoregulating metabolic rate, supporting the idea that oxygen is not limiting this endothermic insect in normoxia.

We found that P_crit during flight on a flight mill was significantly lower in hypoxia-reared animals but there were no differences in normoxic flight speed. The ability to fly at lower oxygen levels in the hypoxia-reared adults may be driven by decreased thorax temperatures and metabolic rates seen during thermoregulation. This could indicate differences in metabolic rate during flight between chronic developmental groups; however, we did not directly measure metabolic rate during the P_crit flight trials. In fact, we saw the opposite trend in our data (Fig. 5B), with hypoxia-reared animals tending to fly faster than the other two groups. This is most likely not due to size differences as it has been shown that insect critical oxygen level is not affected by size (Harrison et al., 2014; Henry and Harrison, 2014; Lease et al., 2012). It is also possible that rearing in hypoxia may not have had a significant effect because the critical oxygen concentration for M. sexta larvae is around 5 kPa (Greenlee and Harrison, 2005). Developing in 10% oxygen was limiting in our study during pupation, as seen by the increased pupation duration of hypoxia-reared animals (Wilmsen and Dzialowski, 2023).

Despite the paucity of data looking at activity after developing in chronic hypoxia and hyperoxia environments, there are many studies that have looked at the P_crit of acute oxygen exposures during various activities. Overall, there is a broad range of critical oxygen levels during flight, as high as 20 kPa in locusts (Rascón and Harrison, 2005) and as low as 5 kPa in bees (Joos et al., 1997) and 10 kPa in dragonflies (Henry and Harrison, 2004). Here, we found that critical oxygen level for flight in moths was around 15 kPa for normoxia-reared moths and dropped to about 12 kPa in hypoxia-reared moths. These values are very similar to dragonfly P_crit during flight and higher than those of bees. There are many factors that potentially play a role here. One may be the difference in function of asynchronous and synchronous flight muscles. Bees, flies and several other insects have evolved a system that requires less nervous input, resulting in a >1:1 or 1:>1 nerve to muscle contraction that allows bees to increase muscle contraction frequency (reviewed in Hickey et al., 2022; Josephson et al., 2000). This is known as asynchronous flight muscle control. However, dragonflies (Odonata) and moths (Lepidoptera) rely on synchronous control or 1:1 nerve to muscle contraction. Within this system, synchronous insects are able to achieve similar maximal metabolic rates to asynchronous insects (Harrison and Roberts, 2000), but wingbeat frequency is quadrupled in asynchronous species (Ha et al., 2013), with M. sexta wingbeat frequencies around 25 Hz (Willmott and Ellington, 1997) and Bombyx species with frequencies as high as 111 Hz. How wing beat frequency plays a role in safety margins and how that would be affected by chronic development in hypoxia or hyperoxia is unknown. Clearly, this is a gap in our understanding of how insects with high maximal metabolic rates respond to chronic oxygen changes during development and should be addressed in future studies.

In summary, we found that exposure to chronic developmental hypoxia and hyperoxia has effects on metabolic rate and flight in adult M. sexta. Hypoxia-reared moths had lower thorax temperatures overall and especially in acute normoxia and hyperoxia. There was also a relationship between critical flight oxygen levels, with hypoxia-reared moths being most tolerant to hypoxia during flight. Influences of hypoxia rearing on flight metabolism remain to be explored.

The Results and Discussion in this paper are reproduced from the PhD thesis of S.M.W. (Wilmsen, 2022).