Breathing Behaviors in Common Marmoset (Callithrix jacchus)

The respiratory system maintains homeostatic levels of oxygen (O2) and carbon dioxide (CO2) in the body through rapid and efficient regulation of frequency and dept (tidal volume) of breathing. The use of common marmoset (Callithrix jacchus), a New World non-human primate (NHP) model, in neuroscience is increasing, however, the data on their breathing is limited and their respiratory behaviors have yet to be characterized. Using Whole-body Plethysmography in room air as well as in hypoxic (low O2) and hypercapnic (high CO2) conditions, we sought to define breathing behaviors in an awake, freely behaving marmosets. Additionally, we instituted and optimized an analysis toolkit for unsupervised analysis of the respiratory activities in common marmoset. Our findings indicate that marmoset’s exposure to hypoxia decreased metabolic rate and increased sigh rate. However, the hypoxic condition did not augment the ventilatory response as reported in other animals. Hypercapnia, on the other hand, increased both the frequency and tidal volume as expected. In this study, we shed light on the breathing behaviors of common marmosets in a variety of O2 and CO2 conditions to further understand the breathing behaviors in NHPs.


Introduction
Mammals rely on fresh and continuous supply of oxygen (O 2 ) from the environment and efficient removal of carbon dioxide (CO 2 ) and other metabolic waste products from their body. The intricate respiratory system ensures the homeostatic state of the arterial partial pressure of O 2 (PO 2 ) and CO 2 (PCO 2 ) in the blood by executing rhythmic movement of the respiratory pump, which include the intercostals and the diaphragm muscles. The inception of this rhythm occurs within the preBötzinger complex (preBötC), a functionally specialized region in the ventrolateral medulla of the brainstem (1,2). Activities of the preBötC are modulated by specialized peripheral and central chemosensors that adjust the respiratory drive for homeostatic level of PO 2 and PCO 2 (3)(4)(5)(6)(7)(8)(9).
A large number of studies on homeostatic control of breathing have been done on rodent models, in which the experiments are mostly performed during the day i.e., rodent's normal inactive period. Since, in general, rodents have relatively reduced chemosensitivities compared with primates (10), there is little assurance on if these results can effectively be extrapolated to humans. Therefore, use of non-human primates (NHPs) has been proposed to fill this gap and translate rodent data to humans (11). The common marmoset (Callithrix jacchus) is a New World NHP with a small body size similar to a rat (250 -600 g). Ease of handling, high reproductive efficacy, and lack of zoonotic risks compared to Old World NHPs make marmoset an attractive and powerful NHP model for biomedical and neuroscience research (12).
Here, we used Whole-body Plethysmography to record breathing behaviors of awake common marmosets in room air, as well as during hypoxic (low inspired O 2 ) and hypercapnic (high inspired CO 2 ) conditions. We also helped developing an optimized analysis toolkit for unsupervised characterization of respiratory indices in laboratory animals. We found that exposure of marmosets to hypoxia decreased metabolic rate and increased sigh rate, while the hypoxic-induced augmentation of ventilatory response was diminished. On the other hand, hypercapnic conditions increased both frequency and depth of breathing similar to other mammals of similar size.

Animals
We used seven adult common marmosets (Callithrix jacchus) for measurement of breathing behavior (3 males, 4 females; 378 ± 12 g). All experiments were performed in accordance with National Institutes of Health Guide for the Care and Use of Laboratory Animals. Animals were housed in temperature-controlled facilities on a normal lightdark cycle (12h:12h, lights on at 7:00 AM). They lived in paired or family-grouped housing and were given tap water ad libitum.

Measurement of marmoset respiratory activity
Marmoset respiratory activity was measured using a Whole-body Plethysmography. Awake animals were placed in the Plexiglas chamber (~3 L) which was flushed with 21% O 2 , 79% N 2 , 22-24 °C, at a rate of 1.2 L/min during measurements of baseline respiratory behavior ( Figure 1). Concentrations of O 2 and CO 2 in the chamber were monitored using a fast-response O 2 /CO 2 analyzer (ML206, AD Instruments). All experiments were performed at the same time of day (between 1000 and 1400 hours) to account for circadian changes in base level physiology (13).
For measuring the respiratory behaviors during hypoxia, following a 40-minute baseline period, the chamber was flushed with 10% O 2 , 90% N 2 , 22-24 °C, at a rate of 1.2 l min -1 . After 10 minutes of exposure to hypoxic conditions, the chamber was then flushed with the room air for another 30 minutes ( Figure 2).
Marmoset respiratory activity was also measured during exposure to hypercapnic conditions. Following a 40-minute baseline period, the chamber was flushed with 6% CO 2 , 60% O 2 , 34% N 2 , 22-24 °C, at a rate of 1.2 l min -1 . After 10 minutes of exposure to hypercapnic conditions, the chamber was then flushed with room air for another 30 minutes.

Calculation of metabolic rate
For measuring metabolic rate (MR), we calculated CO 2 production using the following equation MR = ∆CO 2 x flow rate / body mass where ∆CO 2 is the percent change in the [CO 2 ], flow rate is the flow rate through the plethysmography chamber (1.2 l min -1 ), and body mass is marmoset body mass (g).

Data analysis
Data were tested with Shapiro-Wilk test for normality and statistically compared in Prism 8 (Graphpad, Inc; RRID: SCR_002798). Plethysmography data were imported to Python using Neo Python package (14,15). We wrote a custom Python script using methods from Neurokit2, NumPy, and Pandas software packages (16)(17)(18). Neurokit2 methods were used for signal cleaning and extraction of instantaneous frequency, T TOT (total time of breath), and amplitude (i.e., tidal volume) from trough to peak of the signals ( Figure 3).
During hypoxia and hypercapnia challenges, we analyzed the respiratory signals in 1minute epochs to consider local changes in respiration parameters.
Frequencies above 200 cycles per minute (~3.3 Hz) and amplitude above .5 a. u. were excluded from analysis, as they were likely artifact resulting from movement inside the chamber. The calculated V T (tidal volume) was normalized to the body mass of each animal. High frequency breathing behavior (i.e., sniffing) was defined as any breathing frequencies above 150 cycles per minute (2.5 Hz). Apneas were defined by breathing cycles with T TOT greater than 3 seconds. Augmented breaths (i.e., sighs) were readily identifiable by using the criteria described in rats (6) and measured during the baseline and experimental conditions.
Two measures of rate variability were also calculated as described elsewhere (19). SD1 is a measure of dispersion of T TOT perpendicular to the line of identity in the Poincaré plots, therefore demonstrating short term variability. SD2 is a measure of dispersion of T TOT along the line of identity in the Poincaré plots, demonstrating long term variability in respiratory rate.

Data Availability
All the data and codes will be available on the NGSC GitHub.

Validation of the Neurokit2 as an analysis toolkit in experimental animal models
To analyze resting rate of breathing (ƒ R ), tidal volume (V T ), and minute ventilation (V E ), we first benchmarked the Neurokit2 analysis toolkit against the conventional method of analyzing respiratory data (6,9) in conscious mammals. We did not identify any differences in values of ƒ R , V T , and V E .

Hypoxic Ventilatory Response
We measured changes of ƒ R , V T , and V E during systemic hypoxic challenges (10% O 2 in the inspired air) with respect to the baseline (expressed as a percentage of changes from the baseline, e.g. % Dƒ R ). The magnitude of the % Dƒ R , % DV T , and % DV E were not different in female and male during hypoxia, therefore we combined all the data from both sexes. Overall, changing the inspired O 2 from 21% (room air) to 10% did not elicit ventilatory response in adult marmosets ( Figure 5). However, during hypoxia, the metabolic rate (MR) was decreased by ~ 35%.

Hypercapnic Ventilatory Response
We then measured changes in ƒ R , V T , and V E before and during hypercapnic challenge (6% CO 2 in the inspired air). Similar to hypoxia, the magnitude of the % Dƒ R , % DV T , and % DV E in female and male during systemic hypercapnia were similar. Increasing CO 2 inside the chamber augmented frequency (% Dƒ R , by 40 ± 4%), V T (by 191 ± 20%), and V E (by 306 ± 22%; Figure 6).

Regularity of breathing
The cycle-to-cycle dispersion of T TOT in female and male marmosets were shown in Poincaré plots (Figure 7). We also quantified the regularity of breathing (9) by SD1 and SD2 (see Methods and (19)). The baseline SD1 was similar between female and male marmosets (608 ± 165 vs. 490 ± 185 respectively, Figure 7). The baseline SD2 was also comparable in female vs. male marmosets (846 ± 253 vs. 619 ± 222, respectively, Figure 7).

Sigh frequency, sniffing, and apnea index
Since incidences of sighs, apneas, and sniffing could contribute to the irregularity of respiration, we measured frequencies of these essential features of breathing behavior. It has been shown that sigh can be generated within the inspiratory rhythm-generating circuits of the preBötC (6,(20)(21)(22)(23)(24), and may be modulated by excitatory signals from central chemocenters (6,9,20,25,26). In adult marmosets, sigh frequencies were not different when compared to those in male animals during the baseline (12 ± 2 vs. 12 ± 3 hr -1 in male). It has also been shown that hypoxia increases frequency of sighs in rodents (6,20). Consistent with those results, hypoxia increased sigh events by average of 3.5-folds (12 ± 2 vs. 36 ± 10 hr -1 in room air; Figure 8).
Lastly, we analyzed high frequency breathing (sniffing) in marmosets. On average, marmosets spent 139 ± 46 sec hr -1 sniffing during the baseline recording and there were no differences in sniffing time between female and male marmosets (133 ± 110 vs. 144 ± 39 sec hr -1 in male). Neither hypoxia nor hypercapnia changed the sniffing time (data not shown).
The common marmoset (Callithrix jacchus) is a small New World primate (41). Recently marmosets have been proposed as a powerful animal model in neuroscience research (42)(43)(44)(45), especially to study vocal communication (46). Compared to rodents, marmoset's central nervous system more closely resemble humans' in terms of physiological function and anatomy of the brain (47). It was proposed recently that using marmosets in physiological research can fill the gap between rodents' and humans' studies (11). In addition, considering the similarity of the brain structure and circuit connectivity between primates, marmosets provide an attractive opportunity to study cortical (i.e., voluntary) control of respiratory motor activity (48) as well as coordination of complex respiratory functions during vocalization. Accordingly, in this study, we used Wholebody Plethysmography to characterize respiratory behaviors in awake, freely behaving adult marmosets.
Using the whole-body respiratory measurement in conscious animals requires sophisticated algorithms to distinguish the respiratory signals from noises (i.e., movements). To avoid this problem, respiratory activities are often recorded when the animal is asleep or anesthetized. There are absolute advantages to studying the homeostatic control of breathing physiology in awake animals, despite the increased variability. Therefore, to overcome this challenging task, we validated and used a new Python package, Neurokit2 (see Methods section), for unsupervised analysis of respiratory signals obtained from experimental animals. We then tested the ventilatory response to hypercapnia (increased inspired CO 2 to 6%). Currently, it has been postulated that distributed chemosensitive regions in the rodent's medulla (49)(50)(51)(52)(53) act as central respiratory chemosensors and are responsible for mounting of about 70% of the hypercapnic respiratory response (the mechanism that adjusts breathing in accordance with increase in PCO 2 ). Specialized peripheral chemoreceptors located in the carotid bodies (and aortic bodies in some species) are responsible to the remaining 30% of hypercapnia-induced augmentation of breathing.
In our experiments, to minimize the input from peripheral chemosensors, we applied hyperoxic hypercapnia (60% O 2 /6% CO 2 balanced with N 2 ) as it has been shown that hyperoxia (> 50% O 2 in inspired air) inhibit the drive from the carotid body chemoreceptors (54,55). Since marmosets lack aortic bodies (56), the hypercapnic ventilatory response reported here is driven by the central CO 2 respiratory chemocenters. It is proposed that neurons and astrocytes in the retrotrapezoid nucleus (RTN), medullary raphé, and preBötC are primarily responsible for central CO 2 sensing in rodents (6,(57)(58)(59)(60)(61). Recently the RTN was mapped in rhesus macaque's brainstem, a species of Old-World monkeys (62), however, the location of RTN, raphé, preBötC, and other respiratory centers have not yet been mapped in marmosets.
In awake, freely behaving marmosets, hyperoxic hypercapnia increased both ƒ R and V T ( Figure 6). However, the augmentation of ventilation (V E ) was mainly due to increase in V T (by ~ 250%) rather than ƒ R . These preliminary data are comparable to data from human (63) as well as recent data obtained from conscious rodents (11,64).
The hypoxic ventilatory response (HVR) in common marmoset was absolutely interesting as there was little or no increase in ƒ R and V T during hypoxic exposure ( Figure  6). Other than HVR, mammals can reduce oxygen demand by optimizing and decreasing the rate of their metabolism (65). During hypoxia, adult marmosets, decreased their metabolic rates by ~ 35%, which is consistent with previous data from cats (66), rodents (67), pygmy marmosets (68) and human (69,70).
Although hypoxic conditions (low inspired O 2 ) in marmoset's habitat (sea-level forests of the Amazon) are rare, hypoxia might occur during sleep (i.e., sleep apnea) or as a result of a disease state. It is commonly believed that the HVR is biphasic in adult mammals. During acute hypoxia, ventilation is depicted by an initial increase followed by a subsequent decline to a value above the baseline. This biphasic hypoxic response has been reported in humans, rats, and other mammals (71)(72)(73)(74)(75). It is believed that the rapid initial hypoxic-induced increase in V E is due to activation of peripheral chemosensors (i.e., carotid bodies). The mechanism of hypoxic ventilatory decline (HVD) is not fully understood. It is proposed that desensitization of peripheral chemoreceptors might have a role (76), though significant evidence suggest that, at least in rodents, astrocytes (the numerous star-shaped glia cells) in preBötC (6,8,77) and RTN (78,79) are capable of acting as central respiratory oxygen chemosensors and contribute to the HVD possibly via vesicular release of adenosine triphosphate (ATP). In addition to preBötC and RTN, rostral ventrolateral medulla (rVLM) and the nucleus of the solitary tract (NTS) in the brainstem are proposed to have oxygen sensing capabilities (80,81). However, more research is required to understand if the 'distributed central oxygen chemosensors' hypothesis (11) can be generalized to primates. The brain is highly susceptible to low oxygen levels. Supply of oxygen may be decreased in clinical conditions (such as sleep apnea or stroke) or environmental settings (such as exposure to carbon monoxide). Therefore, an understanding of how the brain maintains homeostatic levels of oxygen and responds to hypoxic events is of longstanding interest. Studies in common marmosets will shed some light on this problem and might fill the gap between rodent and human research to better understand the homeostatic control of breathing and its disorders.