Applications of a novel radiotelemetry method for the measurement of intrathoracic pressures and physiological rhythms in freely behaving mice

Abstract

Techniques to comprehensively evaluate pulmonary function carry a variety of limitations, including the ability to continuously record intrathoracic pressures (ITP), acutely and chronically, in a natural state of freely behaving animals. Measurement of ITP can be used to derive other respiratory parameters, which provide insight to lung health. Our aim was to develop a surgical approach for the placement of a telemetry pressure sensor to measure ITP, providing the ability to chronically measure peak pressure, breath frequency, and timing of the respiratory cycle to facilitate circadian analyses related to breathing patterns. Applications of this technique are shown using a moderate hypoxic challenge. Male C57Bl/6 mice were implanted with radiotelemetry devices to record heart rate, temperature, activity, and ITP during 24-h normoxia, 24-h hypoxia (${\text{F}\mathrm{I}}_{{\text{O}}_{\text{2}}}$ = 0.15), and return to 48-h normoxia. Radiotelemetry of ITP permitted the detection of hypoxia-induced increases in “the ITP equivalent” of ventilation, which were driven by increases in breathing frequency and ITP on a short-term time scale. Respiratory frequency, derived from pressure waveforms, was increased by a decrease in expiratory time without changes in inspiratory time. Chronically, telemetric recording allowed for circadian analyses of respiratory drive, as assessed by inspiratory pressure divided by inspiratory time, which was increased by hypoxia and remained elevated for 48 h of recovery. Furthermore, respiratory frequency demonstrated a circadian rhythm, which was disrupted through the recovery period. In conclusion, radiotelemetry of ITP is a viable, long-term, chronic methodology that extends traditional methods to evaluate respiratory function in mice.

NEW & NOTEWORTHY We have demonstrated for the first time in mice that radiotelemetry is an effective tool for the continuous and chronic recording of intrathoracic pressure (ITP) to facilitate circadian rhythm analyses. We show that continuous 24-h hypoxic stress alters the circadian rhythms of heart rate, body temperature, activity, and respiratory parameters, acutely and perpetually, through normoxic recovery. Radiotelemetry of ITP can complement traditional methods for evaluating respiratory function and better our understanding of respiratory pathophysiology.

INTRODUCTION

Respiratory function is an important measure for evaluating pulmonary and cardiac diseases in clinical populations as well as the effect of experimental interventions on respiratory health in preclinical animal models (40). An array of methods exists to measure respiratory function, including: phrenic recordings (21), invasive plethysmography (for review, see Ref. 12), whole body plethysmography (18, 38), the barometric method (23, 30), intrapleural pressure measurement (3), and in vitro lung and airway function. Although these techniques are biologically relevant, without telemetric monitoring they are limited in capturing the physiological changes to intrathoracic pressure (ITP; a surrogate for pleural pressure) over a 24-h period in the natural, undisturbed, unrestrained, freely behaving mouse. This limits the detectability of subtle changes in breathing cycles over long terms (i.e., weeks to months) or throughout sleep-wake cycles (i.e., circadian biology). Furthermore, some methods for measuring lung function are heavily invasive and introduce investigator interference, equipment that imposes animal stress, and confined spaces that restrict food and water availability (5, 11, 13). Thus, artificial conditions may not reflect natural ITP.

To account for time-of-day differences or changes that present during the progression of disease, radiotelemetry of ITP swings provides continuous, real-time in vivo monitoring (8). Pressure waves are used as surrogates for obtaining various respiratory parameters, including breathing frequency, inspiratory and expiratory timing, duty cycle, and respiratory drive. To accurately assess respiratory function, continuous and chronic pulmonary airflow measurements (acquired from telemetric recordings of ITP) can be combined with a number of methods (e.g., barometric, plethysmographic data) to permit the evaluation of changes in compliance and resistance (25).

Biological rhythms are vital for maintaining a healthy state, which, when disturbed, cause and/or aggravate disease states (19, 39). Circadian rhythm disruption in humans may be responsible for exacerbating the diurnal and nocturnal inflammatory responses in the lung (41). Hypoxia is a common feature of many disease states, including chronic obstructive pulmonary disease (COPD) and heart failure. In mice, moderate and severe hypoxia induce diverging pressor effects that are nonproportional, nonlinear, or synchronized with the time of day (1). Indeed, blood pressure is reduced during normoxic recovery from moderate hypoxia but increased following severe hypoxia (1). Furthermore, moderate hypoxia dampens the amplitude of circadian rhythm, while severe hypoxia abolishes this organized response (1). Thus, we used moderate hypoxia, a relevant cardiorespiratory challenge, to perturb these physiological rhythms and evaluate ITP, heart rate, body temperature, and activity in freely behaving mice.

Therefore, the objective of this study was to investigate the use of radiotelemetry of ITP in freely moving mice to measure the effect of a moderately hypoxic insult on the magnitude and timing of respiratory parameters [e.g., inspiratory pressure (P_I), frequency of breathing (V_f), inspiratory and expiratory time (T_I and T_E, respectively), duty cycle (T_I/T_TOT), and respiratory drive (P_I/T_I)] and their respective circadian/physiological rhythms over an extended period. Using implantable radiotelemetry devices, we continuously evaluated ITP for 4 consecutive days and the changes elicited during the response to, and recovery from, a 24-h hypoxic challenge. We hypothesized that radiotelemetry of ITP would detect breath-by-breath and hour-by-hour changes in respiratory frequency, peak pressure, and the timing variables of the respiratory cycle (P_I, V_f, T_I, T_E, T_I/T_TOT, P_I/T_I) between normoxic and hypoxic states of mice over several days of experimental procedures. Our findings indicate that radiotelemetry of ITP is a sensitive methodology that detects normal changes in respiration over a 24-h cycle and that it was successful at detecting hypoxia-induced wake- and sleep-phase effects on respiratory parameters.

METHODS

Ethical approval.

Adult male C57Bl/6 mice were bred and aged beyond 8 wk before surgery. Animal housing was maintained at 24°C and 45% humidity and kept to a 12-h:12-h light-dark cycle (lights ON: 0800; lights OFF: 2000). Following telemetry implantation, animals were housed individually with food and water provided ad libitum. Housing and experimental procedures were approved by the Animal Care Committee at the University of Guelph and in conformity with the guidelines of the Canadian Council on Animal Care.

Telemetry.

Male mice (n = 5) were implanted with transmitters (HDX11; Data Science International, St. Paul, MN) equipped to measure ITP, heart rate, core body temperature, and activity. Local anesthetics lidocaine (3 mg/kg) and bupivacaine (1.5 mg/kg) were mixed 50:50 and administered subcutaneously at the incision site. Briefly, mice were anesthetized with isoflurane (2:98% isoflurane-oxygen), and, by use of a heating pad and lamp, body temperature was maintained at 37°C (confirmed by rectal thermometer). A small incision was made at the midline inferior to the xiphoid process through the skin and muscle layers of the abdomen. The esophageal hiatus was located on the inferior surface of the diaphragm, and a pressure catheter was inserted between the serosal and muscularis layers of the esophagus (25, 26) (Fig. 1). The catheter was advanced superiorly into the thoracic cavity while pressure was continuously monitored. Once the maximal pressure amplitude was obtained, the catheter was secured in place using VetBond (3M, London, ON, Canada) tissue adhesive. Two electrocardiogram leads were placed subcutaneously, one above the ribcage and the other above the abdomen, and secured to the underlying muscle layer. Postoperative analgesics meloxicam (5 mg/kg) and buprenorphine (0.1 mg/kg), were provided as required. Although circadian rhythms were restored within a few days postsurgery, mice were given 2 wk of recovery.

Fig. 1.

Schematic representation of the radiotelemetry device implantation (A) and the pressure catheter (blue) placement (B), which is secured between the muscularis and serosal layers of the esophagus.

In total, seven mice were instrumented, with a 100% survival rate postsurgery. There were no surgical complications. Mice (n = 2) were excluded from the experiment when the pressure tracings were insufficient or unstable postsurgery, likely due to misplacement of the pressure catheter.

Hypoxia study design.

Mice were transported in cages to a hypoxia chamber (830-ABB; Plas Laboratories, Lansing, MI), where they remained for the duration of the experimental protocol. Baseline normoxic recordings were obtained over a weekend, beginning on a Friday afternoon to ensure minimal disruption due to human interference. Weekly cage changes were done at least 2 days before start of telemetry recording to prevent environmental disturbance during the experimental protocol. Mice were made hypoxic at 0900 Monday morning, and exposure was sustained for 24 h, as previously described (1). Hypoxia was induced by progressively increasing N₂ levels (over 15 min) until desired O₂ levels (${\text{F}\mathrm{I}}_{{\text{O}}_{\text{2}}}$ = 0.15) were achieved as determined by an O₂ sensor (ProOx 110; BioSpherix, New York, NY). After 24 h, O₂ levels in the chamber were restored to ambient levels, and the animals were monitored for an additional 48 h. In the chamber, cages were placed on telemetry receiver bases (RPC-1; Data Science International, St. Paul, MN). Receiver signals were sent to a matrix (Data Exchange Matrix, Data Science International) located within the chamber, from which signals were relayed to a central matrix. Ambient temperature (C10T, Data Science International) and pressure (APR-1, Data Science International) were also recorded throughout the duration of the study. Signals were collected using computer acquisition software (Dataquest ART v.3.3, Data Science International).

Sampling and analysis.

Raw signals were measured for 30 s at 5-min intervals from implantable telemetry devices (at a sampling frequency of 600 Hz). To assess circadian rhythms, cosine wave characteristics were measured as previously described (1, 29). To determine the best-fit cosine wave using a nonlinear regression model, cosinor analysis was performed. This mathematical technique applied the best-fit cosine wave to data collected over 24-h periods and assigned wave characteristics. Mesor, an estimation of the mean statistic, acrophase, the Zeitgeber time at the cycle’s peak, and amplitude, the difference between the peak/trough and mean statistic, were recorded. Goodness-of-fit R² was calculated using GraphPad (Prism 6; GraphPad Software, Inc., La Jolla, CA). Differences in the best-fit values for mesor, amplitude, and acrophase between hypoxiaand 24 and 48 h of normoxic recovery to normoxic baseline used the extra sum of squares F test.

To obtain respiratory waveforms from the telemetry data, raw time and pressure signals were exported as text files from a computer acquisition software (Dataquest ART v.3.3, Data Science International). Waveform text files were then imported into Spike2 analysis software (Cambridge Electronic Design, Cambridge, UK) to obtain waveform data files. With a custom-designed computer script, data files were analyzed with Spike2 to provide output parameters; inspiratory (P_I) and expiratory pressure, inspiratory and expiratory timing (T_I and T_E, respectively), duty cycle (fraction of a full respiration spent in inspiration, T_I/T_TOT), frequency of breathing (V_f), and an index of respiratory drive (inspiratory pressure/inspiratory time, P_I/T_I). T_I begins when the pressure becomes more negative than the end-expiratory value (Fig. 2A). Data were analyzed only during quiet eupneic breathing (Fig. 2C). Hourly averages were created beginning on Sunday at 0900 and ending on Friday at 0900. Graphic and statistical analyses were completed using GraphPad (Prism 6). Statistical analyses were performed on 12-h averages during both lights ON and lights OFF. Paired Student’s t tests were used for comparison between two groups. For comparison among three or more groups, a one-way ANOVA was used to detect a significant difference (P < 0.05), followed by a protected least significant differences (LSD) test. Normoxic during lights ON was compared with hypoxic and recovery during lights ON, whereas normoxic during lights OFF was compared with hypoxic and recovery during lights OFF. Differences were considered significant at P < 0.05.

Fig. 2.

Representative eupneic respiratory waveform (A) with corresponding points of measurement (B) used to calculate indexes of respiratory parameter. Representative ineligible waveforms produced by radiotelemetry. Eupneic breathing (C), signal loss (D), nonrhythmic breathing with unstable baseline (E), and abdominal positive pressure deflections (F–H). P_I, inspiratory pressure; P_I/T_I, respiratory drive; T_I, inspiratory time; T_E, expiratory time; T_I/T_TOT, duty cycle.

RESULTS

Pressure and time are the primary measures of radiotelemetry that are used to generate respiratory waveforms (Fig. 2A). From these waveforms, individual indexes (Fig. 2B and Table 1) were derived to evaluate V_f, T_I, T_E, and T_I/T_TOT, as well as P_I and P_I/T_I. Since these indexes are inherently dependent on accurate pressure and time measurements, any factor that affects pressure or time will subsequently affect respiratory data. Therefore, it was critical that we remove any confounding waveforms to ensure the data was reflective of eupneic respiration (Fig. 2C). Ineligible waveforms (<5% of data) included those created by sneezes, movement artifacts, as well as catheter slippage. Such instances are depicted in Fig. 2, D–H: signal drop (Fig. 2D), nonrhythmic breathing with unstable baseline (Fig. 2E), and abdominal pressure (positive pressure deflections; Fig. 2, F–H). Once ineligible waveforms had been excluded, individual indexes of respiration were derived from eupneic breathing and plotted to depict the temporal trends (Fig. 3).

Table 1.

Daily averages for respiratory parameters during lights ON and lights OFF

	Normoxia	Hypoxia	24-h Recovery	48-h Recovery
Frequency, breaths/min
Lights ON	142 ± 2	151 ± 2*	139 ± 2	135 ± 3*
Lights OFF	154 ± 2	177 ± 3*	150 ± 2	146 ± 3*
TI, s
Lights ON	0.111 ± 0.001	0.103 ± 0.003*	0.102 ± 0.003*	0.104 ± 0.003*
Lights OFF	0.107 ± 0.002	0.104 ± 0.002	0.103 ± 0.003	0.104 ± 0.004
TE, s
Lights ON	0.350 ± 0.007	0.318 ± 0.010*	0.372 ± 0.011	0.384 ± 0.009*
Lights OFF	0.323 ± 0.007	0.262 ± 0.010*	0.336 ± 0.010	0.354 ± 0.011*
TI/TTOT
Lights ON	0.254 ± 0.005	0.245 ± 0.008	0.214 ± 0.008*	0.220 ± 0.006*
Lights OFF	0.234 ± 0.006	0.292 ± 0.007*	0.241 ± 0.009*	0.234 ± 0.007*
PI, cmH2O
Lights ON	11.1 ± 0.3	13.4 ± 0.9*	13.6 ± 0.8*	13.4 ± 0.6*
Lights OFF	12.1 ± 0.4	14.2 ± 0.4*	12.5 ± 0.6	12.3 ± 0.5
PI/TI, cmH2O/s
Lights ON	112.2 ± 4.3	138.0 ± 7.5*	131.2 ± 4.6*	127.1 ± 4.9*
Lights OFF	124.7 ± 5.0	149.7 ± 5.7*	119.5 ± 3.5	131.4 ± 4.7

Values are means ± SE. P_I, inspiratory pressure; P_I/T_I, respiratory drive; T_I, inspiratory time; T_E, expiratory time; T_I/T_TOT, duty cycle.

^*P < 0.05 vs. normoxic baseline as determined by a protected least significant difference test; n = 5.

Fig. 3.

Temporal respiratory responses and recovery to hypoxic exposure. Three-hour binned averages of respiratory frequency (A), inspiratory time (B), expiratory time (C), duty cycle (D), inspiratory pressure (E), and respiratory drive (F). Black bars indicate the sleep phase of mice; n = 5. P_I, inspiratory pressure; P_I/T_I, respiratory drive; T_I, inspiratory time; T_E, expiratory time; T_I/T_TOT, duty cycle.

To visualize the effect of a 24-h hypoxic challenge on respiratory trends, we plotted 3-h binned averages acquired over five experimental days. Although these averages were sufficient for our study, temporal plots can be adjusted to depict more detailed changes in respiratory parameters, such as those that occur on an hourly or even minute-to-minute basis. This offers researchers a unique ability with radiotelemetry to customize their data visualization and illustrate changes in respiratory parameters as they occur in real-time. Although useful in identifying data trends, temporal plots were not used for statistical analyses in our study. Rather, we were primarily interested in delineating circadian cycle differences (during lights ON vs. lights OFF) in the pattern of breathing. Figure 4 represents the cumulative 48-h normoxic baseline average compared with 24-h averages of hypoxic exposure and normoxic recovery. Subsequently, 24-h averages were analyzed during their respective sleep and wake cycles to determine circadian differences (Table 1 and Fig. 5).

Fig. 4.

Daily averages of respiratory patterns in response to and recovery from a hypoxic challenge: respiratory frequency (A), inspiratory time (B), expiratory time (C), duty cycle (D), inspiratory pressure (E), and respiratory drive (F). Scatter plots composed of individual hourly averages with representative mean values. P_I, inspiratory pressure; P_I/T_I, respiratory drive; T_I, inspiratory time; T_E, expiratory time; T_I/T_TOT, duty cycle. *Significance (P < 0.05) vs. normoxic baseline as determined by a protected least significant distance test; n = 5.

Fig. 5.

Circadian averages of respiratory patterns in response to and recovery from a hypoxic challenge: respiratory frequency (A), inspiratory time (B), expiratory time (C), duty cycle (D), inspiratory pressure (E), and respiratory drive (F). P_I, inspiratory pressure; P_I/T_I, respiratory drive; T_I, inspiratory time; T_E, expiratory time; T_I/T_TOT, duty cycle. Values expressed as means ± SE. *Significance (P < 0.05) vs. normoxic baseline as determined by a protected least significant distance test; n = 5.

Hypoxia increases respiratory rate by reducing expiratory time.

In response to hypoxia, respiratory rate was significantly increased during the 24-h hypoxic exposure period (Fig. 4A) during both lights ON and OFF (Fig. 5A). Following the return to normoxia, respiratory frequency progressively decreased and was significantly reduced following 48 h of normoxic recovery during both sleep and wake cycles (Fig. 5A).

To further deduce the cause of reduced frequency of breathing, we evaluated the individual components of total respiratory time, which includes T_I and T_E. Consistent with the increased respiratory frequency, T_I was reduced during the 24-h hypoxic exposure (Fig. 4B). This reduction persisted throughout the 48 h of normoxic recovery and was primarily driven by decreases that occurred during lights ON (Fig. 5B), as inspiratory time was unaltered during lights OFF. Although statistical significance was reached, the changes observed in T_I (~7%) were not sufficient to reconcile the changes observed with respiratory frequency (~15%). This is likely because T_I accounts for ~25% of the duty cycle, whereas T_E accounts for ~75% (Table 1 and Fig. 4D). Consequently, T_E had a more profound effect on total respiratory time than T_I and therefore a greater effect on respiratory frequency. As expected, changes in T_E closely reflected those of respiratory frequency.

As a primary determinant of the increase in respiratory rate (Fig. 4A), T_E was significantly reduced during the 24-h hypoxic exposure (Fig. 5C) during both lights ON and OFF (Fig. 5C), similar to that reported for hypoxic challenges in mice (38) and rats (33). Following the return to normoxia, T_E progressively increased and was significantly elevated following 48 h of normoxic recovery during both lights ON and OFF (Fig. 5C). Consistent with the temporal trends (Fig. 3, A–C), 24-h circadian waveform analyses indicated that changes in respiratory frequency were primarily driven by alterations in T_E, with minor contributions from T_I.

Hypoxia increases inspiratory pressure and respiratory drive.

In response to hypoxia, P_I was significantly elevated during the 24-h exposure period (Fig. 4E) during both lights ON and OFF (Fig. 5E). Following the return to normoxia, P_I remained elevated throughout the 48 h of recovery (Fig. 4E); the increase in P_I was only due to increases during lights ON (Fig. 5E). As expected, changes in P_I were similar to those of respiratory drive (Fig. 3, E and F). Respiratory drive, as assessed by P_I/T_I, was significantly elevated during the 24-h exposure period during lights ON and OFF. Upon return to normoxia, respiratory drive remained significantly elevated for 48 h and was primarily driven by increases during the sleep cycle, as wake cycle respiratory drive was similar between baseline and recovery (Fig. 5F).

Applications of telemetry of ITP, heart, temperature, and activity demonstrate the effects of moderate hypoxia on physiological rhythms.

Circadian rhythms are endogenous oscillations with a period of ~24 h that include both physiological and behavioral rhythms, which can also be influenced by external stimuli (e.g., temperature, hypoxia). For example, circadian rhythmicity is present in the blood pressure patterns of mammals, but exercise or hypoxia can augment or supress the rhythm amplitude. Circadian clock refers to the molecular mechanism driving a rhythm. At the molecular level, clock genes intrinsically cycle providing the molecular mechanism driving rhythmicity, which can both drive behavior and be influenced by behavior. In our study, we defined circadian rhythms as oscillations in physiological variables, as determined by cosinor analysis (1, 29), in response to a normoxic baseline, hypoxia, and normoxic recovery. Circadian rhythms were characterized by their mesor, amplitude and phaseshift with comparisons further based on R², a goodness-of-fit value (as shown in Tables 2 and 3). Any changes to the mean values during lights ON and OFF simply reflects changes to the biological parameter, independent of 24 h rhythm. A unique benefit of radiotelemetry is that measurements of heart rate, activity level, and body temperature are measured simultaneously with ITP (Fig. 6). These supplemental measurements provide additional physiological details, which improve our understanding of the effects of hypoxia. To show that telemetry is a sensitive method capable of detecting changes in the physiological rhythms of respiratory parameters (e.g., V_f, T_I, T_E, and P_I), we evaluated data over the course of 4 days (24-h baseline normoxia, 24-h hypoxia, and 48 h of normoxic recovery) using cosinor analysis. Certainly, telemetry of ITP (e.g., V_f and T_E) and biological parameters (e.g., body temperature and heart rate) revealed changes to cosine wave parameters (e.g., mesor, amplitude, phase shift) in response to hypoxia (Tables 2 and 3 and Figs. 7–9). For instance, heart rate data provide insight as to whether a specific insult, such as hypoxia, elicits respiratory changes independent of cardiovascular changes. Here, we demonstrate that, similarly to respiratory frequency, heart rate was increased during the 24-h hypoxic exposure period and was reduced during lights ON following 48 h of normoxic recovery (Fig. 6, A and B). Cosinor analysis revealed that the goodness-of-fit value R² remained moderate throughout the entirety of the experimental procedures, ranging from 0.39 to 0.55 (Table 2 and Fig. 7).

Table 2.

Cosinor analysis of physiological parameters: temperature, HR, and activity

	Mesor	Amplitude	Acrophase, h	R2
Temperature, °C
Normoxia (baseline)	36.0 ± 0.1	0.4 ± 0.1	14.4 ± 0.2	0.63
Hypoxia	35.3 ± 0.1*	1.0 ± 0.1*	20.5 ± 0.1	0.77
24-h Recovery	36.3 ± 0.1*	0.5 ± 0.1	20.5 ± 0.2	0.62
48-h Recovery	36.1 ± 0.1	0.7 ± 0.1*	14.2 ± 0.1	0.74
HR, beats/min
Normoxia (baseline)	519 ± 6	30.6 ± 8.4	20.9 ± 0.3	0.39
Hypoxia	503 ± 6	31.1 ± 8.2	13.9 ± 0.3	0.41
24-h Recovery	494 ± 5*	36.2 ± 7.2	14.2 ± 0.2	0.55
48-h Recovery	482 ± 6	40.2 ± 9.0	14.4 ± 0.2	0.49
Activity, AU
Normoxia (baseline)	4.4 ± 0.4	1.8 ± 0.6	20.4 ± 0.3	0.30
Hypoxia	3.7 ± 0.3	2.0 ± 0.5	13.7 ± 0.2	0.47
24-h Recovery	5.2 ± 0.4	2.2 ± 0.6	20.3 ± 0.3	0.39
48-h Recovery	4.2 ± 0.5	2.0 ± 0.7	14.2 ± 0.4	0.26

Values are means ± SE. HR, heart rate; mesor, midline estimating statistic of rhythm; amplitude, half the extent of predictable variation within a cycle; acrophase, the time of overall high values recurring in each cycle; R², degree of curve fit.

^*P < 0.05; n = 5.

Table 3.

Cosinor analysis of respiratory parameters: frequency of breathing, inspiratory pressure, respiratory drive, inspiratory time, expiratory time, and duty cycle

	Mesor	Amplitude	Acrophase, h	R2
Frequency, breaths/min
Normoxia (baseline)	166 ± 2	14.7 ± 2.2	14.3 ± 0.1	0.69
Hypoxia	174 ± 2*	28.1 ± 3.3*	14.3 ± 0.1	0.78
24-h Recovery	150 ± 3*	2.9 ± 3.5*	20.8 ± 0.9	0.03
48-h Recovery	139 ± 1*	7.7 ± 2.0*	20.4 ± 0.3	0.42
PI, cmH2O
Normoxia (baseline)	13.3 ± 0.3	0.4 ± 0.5	19.1 ± 1.2	0.03
Hypoxia	16.1 ± 0.4*	0.8 ± 0.6	8.3 ± 0.8	0.08
24-h Recovery	12.7 ± 0.3	1.2 ± 0.4	21.5 ± 0.3	0.29
48-h Recovery	12.7 ± 0.4	0.2 ± 0.5	11.7 ± 2.1	0.01
PI/TI, cmH2O/s
Normoxia (baseline)	168.8 ± 4.7	11.4 ± 6.6	20.3 ± 0.6	0.12
Hypoxia	168.5 ± 4.2	22.0 ± 6.0	8.3 ± 0.3	0.39
24-h Recovery	136.3 ± 3.0*	13.1 ± 4.6	3.12 ± 0.27	0.28
48-h Recovery	129.4 ± 4.3*	7.7 ± 6.1	18.1 ± 0.8	0.07
TI, s
Normoxia (baseline)	0.102 ± 0.001	0.004 ± 0.001	4.6 ± 0.5	0.18
Hypoxia	0.103 ± 0.001	0.004 ± 0.002	5.2 ± 0.5	0.19
24-h Recovery	0.102 ± 0.001	0.006 ± 0.002	13.9 ± 0.3*	0.41
48-h Recovery	0.106 ± 0.002	0.002 ± 0.002	2.8 ± 1.1	0.04
TE, s
Normoxia (baseline)	0.319 ± 0.004	0.026 ± 0.006	11.0 ± 0.2	0.51
Hypoxia	0.285 ± 0.006*	0.047 ± 0.009	4.8 ± 0.2	0.57
24-h Recovery	0.346 ± 0.007*	0.010 ± 0.011	9.4 ± 1.1	0.04
48-h Recovery	0.391 ± 0.008*	0.024 ± 0.011	17.5 ± 0.5	0.18
TI/TTOT
Normoxia (baseline)	0.42 ± 0.0	0.030 ± 0.007	4.7 ± 0.2	0.49
Hypoxia	0.40 ± 0.0*	0.060 ± 0.009*	11.1 ± 0.16	0.67
24-h Recovery	0.45 ± 0.0*	0.011 ± 0.010	8.8 ± 1.0	0.05
48-h Recovery	0.50 ± 0.0*	0.023 ± 0.012	17.1 ± 0.54	0.14

Values are means ± SE. P_I, inspiratory pressure; P_I/T_I, respiratory drive; T_I, inspiratory time; T_E, expiratory time; T_I/T_TOT, duty cycle; mesor, midline estimating statistic of rhythm; amplitude, half the extent of predictable variation within a cycle; acrophase, the time of overall high values recurring in each cycle; R², degree of curve fit.

^*P < 0.05; n = 5.

Fig. 6.

Temporal and circadian measurements of heart rate (A and B), body temperature (C and D), and activity (E and F) during the response to and recovery from a hypoxic exposure. Temporal values expressed as 3-h binned averages (A, C, and E), whereas sleep and wake cycles are expressed as means ± SE. *Significance (P < 0.05) vs. normoxic baseline during the respective sleep or wake cycle as determined by a protected least significant distance test; n = 5.

Fig. 7.

Cosinor analysis of biological parameters during baseline (normoxia), hypoxia, and 24 h and 48 h of recovery. Graphic representation of cosinor analysis of body temperature (A), heart rate (B), and activity (C); n = 5.

Fig. 9.

Cosinor analysis of respiratory parameters during baseline (normoxia), hypoxia, and 24 h and 48 h of recovery. Graphic representation of cosinor analysis of inspiratory time (A), expiratory time (B), and duty cycle (C); n = 5. P_I, inspiratory pressure; P_I/T_I, respiratory drive; T_I, inspiratory time; T_E, expiratory time; T_I/T_TOT, duty cycle.

Fig. 8.

Cosinor analysis of respiratory parameters during baseline (normoxia), hypoxia, and 24 h and 48 h of recovery. Graphic representation of cosinor analysis of respiratory frequency (A), respiratory drive (B), and inspiratory pressure (C); n = 5. P_I, inspiratory pressure; P_I/T_I, respiratory drive; T_I, inspiratory time; T_E, expiratory time; T_I/T_TOT, duty cycle.

Simultaneous measurements of body temperature and activity level were made. Whereas body temperature remained unchanged throughout our study (Fig. 6, C and D), we observed a marked increase in activity levels during lights OFF with hypoxia and during the initial 24 h of recovery (Fig. 6, E and F). A strong R² was calculated for body temperature, confirming the maintenance of circadian rhythm, despite a hypoxic insult (Table 2). Meanwhile, the R² for activity level was relatively low (0.26 to 0.47) and aligned more closely with values calculated for a number of respiratory parameters.

To evaluate circadian rhythmicity of the respiratory system, cosinor analysis was performed. Cosine waves were easily applied to frequency of breathing and expiratory time data sets, as they demonstrated clear rhythmicity through baseline, hypoxia stress, and recovery (i.e., goodness-of-fit value remained relatively strong; refer to Table 3). Interestingly, some, but not all, respiratory parameters have circadian rhythm.

In some instances, the fit of the cosine wave to the 24-h data was extremely poor. Normoxic baseline and 48-h recovery for P_I produced R² values of 0.03 and 0.01, respectively, which led us to conclude that it does not have specific daily changes. Furthermore, T_I and P_I lacked pronounced periodicity, with maximum R² values of 0.41 and 0.29, respectively. Although minor changes were seen in T_I, alterations in the duty cycle were driven primarily by changes in T_E, as the mesor was significantly reduced during hypoxia exposure and increased during normoxic recovery compared with baseline (Table 3). These findings indicate that radiotelemetry of ITP is a rigorous surgical technique for characterizing the peaks and nadirs of intrathoracic pressures and breathing patterns over several days, which may have significant implications for chronotherapy use in preclinical studies (9).

DISCUSSION

Herein, we demonstrate a surgical method for continuous measurement of respiratory ITP, which allows the user to determine respiratory patterns (e.g., breathing frequency, timing, and respiratory drive). Using a moderately hypoxic respiratory challenge, we demonstrate that radiotelemetry can detect subtle changes in ITP, respiratory timing, and drive, as well as their respective circadian rhythms, for extended periods using cosinor analysis. In some cases, hypoxia had nonuniform effects on the physiological rhythms of mice. Combining intrapleural pressure data with respiratory flow measurements (e.g., inspiratory and expiratory flow, timing) and volumes (e.g., tidal volume) at the same points during the respiratory cycle would permit the assessment of dynamic lung mechanics to provide a comprehensive overview of respiratory function (3; for review, see Ref. 5).

One key advantage of radiotelemetry of ITP is that measurement of associated respiratory variables can be evaluated with minimal experimental intervention. Following device implantation, respiratory parameters are evaluated without the use of anesthesia, immobilizing restraints, or confining chambers. By minimizing researcher interference, animal stress, and food/water restriction, radiotelemetry permits the continuous assessment of respiration in a natural physiological state over days to weeks. Ventilatory plasticity has been well investigated using circadian rhythm of respiratory patterns in rats by radiotelemetry of diaphragm electromyography (EMG) (37), electroencephalography and neck muscle EMG (4), and plethysmography (4, 27). In 2000, Seifert et al. (30) used the barometric method and found that ventilation, tidal volume, and frequency exhibit daily rhythms in rats, oscillating in unison with body temperature, activity, and metabolic rate. Similarly, in rats, continuous hypobaric hypoxia (10.5% O₂) for 7 days depresses the amplitudes of the circadian patterns for body temperature and metabolism without affecting the period (22; for review, see Ref. 24). Here, we provide a much shorter hypoxic exposure time (24 h), in normobaric conditions, to demonstrate an altered circadian physiology, lasting through recovery. This suggests that, even in recovery, episodes of hypoxia can have lasting impacts, which may be important to clinical conditions of pneumonia, as an example. Our approach for the placement of a telemetry device to effectively record ITP, a surrogate for pleural pressure, over the long term and to detect differences in respiratory timing and drive in mice lends itself well to future work in genetically altered mechanistic models.

The reliability of radiotelemetry to produce consistent and accurate data has previously been demonstrated in rats, dogs, and primates (9, 25, 28). Regardless of the species, the reliability of this technique is dependent upon proper probe placement and successful recovery from the initial surgery. The pressure probe must be sufficiently advanced within the serosal layer of the esophagus to avoid slippage (Fig. 1). During inspiration, the probe can become lodged in the esophageal hiatus and/or abdominal cavity, resulting in positive pressure deflections at the onset of inspiration (Fig. 2, F–H). These types of artifacts must be excluded from analysis to ensure that pulmonary ITP is taken only during eupneic breathing (Fig. 2C). Indeed, manually selecting only eupneic respiratory signals is time intensive; however, it reduces data variability and improves the power to detect subtle changes while limiting animal usage. Therefore, although radiotelemetry may require a surgical procedure and extensive data processing, the resultant findings have strong statistical power and new physiological relevance, which is accompanied by more conventional telemetry variables (e.g., heart rate, temperature, activity).

Radiotelemetry can be implemented to monitor changes in respiratory timing and ITP, which can be further evaluated using cosinor analysis to characterize the circadian profiles of biological mechanisms associated with the lungs. In response to a stress (in our case, moderate hypoxia), radiotelemetry had the capacity to detect small differences in the mesor, amplitude, and acrophase of frequency of breathing, inspiratory and expiratory time, and inspiratory pressure. Further investigation allowed us to measure subtle changes in respiratory drive (assessed by P_I/T_I and breathing frequency) and duty cycle (calculated by T_I/T_TOT). Respiratory drive is defined as the frequency and intensity output of the respiratory centers, which cannot be quantified directly or captured by a single variable (15). The former is directly measured from breathing frequency, while intensity output is more challenging to capture, with many different indexes used in publications. Quantifying the neural output (i.e., phrenic nerve) from the respiratory centers as a function of time provides a better reflection of drive; however, the invasive nature and difficulty in obtaining a stable recording limit its use. Because of these limitations, rate of change of tidal volume (V_t/T_I) or inspiratory pressure (P_I/T_I) are alternative approaches (14, 16). For the latter, one can use either transdiaphragmatic pressure (P_di) or pleural pressure (P_pl) to assess drive specific to the diaphragm or all inspiratory muscles, respectively. In our study, ITP is a surrogate for P_pl; thus, we assessed respiratory drive by two indexes: the rate of rise of ITP during inspiration (P_I/T_I) and breathing frequency. Caution should be taken in using P_I/T_I (or V_t/T_I) to reflect respiratory drive when neural muscular transmission and diaphragm function are not intact. In disease states where this may be a suspected, the biopotential channels could be adapted to record either the phrenic nerve or, more easily, and when neural muscular transmission failure is not a concern, diaphragm EMG. Thus, the rate of rise of the neural signal would serve in the assessment of respiratory drive.

In mice, we found that the increase in frequency of breathing was modulated by reductions in T_E, not T_I, similarly reported in response to hypoxia in mice (38), rats (33), and goats (7). Conversely, in rat, respiratory load-induced hypoxia decreases frequency of breathing by increases in T_E, not T_I (31, 32). Furthermore, we observed that increases in respiratory drive were reflected by elevations in inspiratory pressure. Taken together, respiratory frequency is modulated through changes in T_E rather than T_I.

Importantly, this technique allows users to subdivide and investigate how respiratory drive (i.e., respiratory intensity and frequency) is disrupted in disease and stress (e.g., hypoxia). Furthermore, investigating time-of-day therapy for respiratory diseases may provide additional therapeutic advantages by increasing drug effectiveness and/or decreasing side effects. Monitoring the physiological rhythms of respiration provides a preclinical framework for assessing optimal timing of pharmacotherapy delivery. Indeed, the advantages of radiotelemetry and circadian analysis discussed here provide a clear justification for their use in evaluating diurnal and nocturnal changes in lung physiology.

Circadian rhythms are present in almost every organ (42), highlighting their importance for maintaining homeostasis in the body. In fact, rhythm disturbances adversely affect organs and their function (10, 19). Similar to our findings, Allwood et al. (1) showed that hypoxia disrupts the circadian patterns of body temperature and heart rate. In another study, the effects of acute hypoxia on the circadian patterns of body temperature and metabolism were investigated in humans (6). They found that oxygen consumption and pulmonary ventilation lost their circadian oscillations during hypoxia compared with the clear circadian patterns they maintained in normoxia (6). In 2007, a review concluded that hypoxia alters the circadian patterns of critical variables (including body temperature and metabolism); thus, it is plausible that the physiological rhythms of other parameters would also be disrupted by hypoxia (24). Mortola (24) speculated that perturbations to the healthy circadian oscillations of many variables might be contributing factors to symptoms associated with altitude sickness, including insomnia/disrupted sleep, fatigue, nausea, and appetite loss. With the lungs comprising the third most circadian genes of all organs (42), rhythm regulation may be a key factor in respiratory disease pathogenesis. To address this, radiotelemetry is useful for quantifying the occurrence of frequency of breathing irregularities, such as apneas and Cheyne–Stokes respiratory patterns. Human conditions, including COPD, asthma, sleep apnea, and heart failure, frequently show time-of-day variation in symptom severity and pulmonary function (34, 35). This ability is relevant in the context of evaluating pharmacological safety and respiratory pathologies, which are known to exhibit circadian rhythms. Since pharmaceuticals target specific genes, which often have circadian oscillations in their expression (10, 20), treatment during the peak phase of gene activity (or when the severity of symptoms is at its maximum), may improve drug efficacy; this strategy is known as “chronotherapy” (17).

Conclusion.

Using radiotelemetry, a powerful tool for the continuous evaluation of ITP (and associated respiratory variables/patterns of breathing), we demonstrate that moderate hypoxia disrupts respiratory circadian rhythm in the absence of overt pathology in freely moving, nonsedated mice. Whether these disruptions lead to chronic deficits remains to be elucidated. Radiotelemetry effectively assesses the simultaneous effects of hypoxia on the body and circadian rhythms over the long term. Our findings may be useful in guiding chronotherapies in a preclinical setting. Future studies using telemetric recording of ITP could also perform this technique in parallel to methods that record volumes and capacities to continuously evaluate resistance and compliance on a breath-by-breath time scale. These data would significantly improve our understanding of the pathophysiology of respiratory conditions, including COPD, emphysema, and asthma.

GRANTS

This work was supported by Canadian Institutes of Health Research (CIHR) Grants MOP-111159 and PJO-162481, the Natural Sciences and Engineering Research Council of Canada (NSERC) (J. A. Simpson), Heart and Stroke Foundation of Canada Grant (J. A. Simpson), and NSERC (K. R. Brunt). A. J. Foster was supported by the Canadian Lung Association. J. P. Marrow was supported by an Alexander Graham Bell Canada Graduate Scholarship-Doctoral (CGS D) NSERC. J. A. Simpson is a New Investigator with the Heart and Stroke Foundation of Canada.

DISCLAIMERS

The funding sources had no influence on study design, sample collection, data analysis and interpretation, or preparation of this manuscript.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.J.F., M.A.A., and J.A.S. conceived and designed research; A.J.F. and M.A.A. performed experiments; A.J.F. and J.P.M. analyzed data; A.J.F. and J.P.M. interpreted results of experiments; A.J.F. and J.P.M. prepared figures; A.J.F. and J.P.M. drafted manuscript; A.J.F., J.P.M., M.A.A., K.R.B., and J.A.S. approved final version of manuscript; J.P.M., K.R.B., and J.A.S. edited and revised manuscript.