All methods and procedures were approved by Seattle Children’s Research Institute’s IACUC and done in accordance with NIH guidelines. C57BL/6 mice were group housed on a 12:12 light-dark cycle with access to chow and water ad libitum. Female and male mice were group housed until breeding. In house, timed pregnant mice were obtained by placing a male mouse in a harem at 5 pm and separating the male mouse from the female mice by 10 am the next morning. Female mice were checked for sperm plugs. If a plug was present, the female mouse would be housed separately.

Four treatment groups were studied: (1) term mice that were born naturally on gestational day (G) 19.5; preterm mice born via cesarean section on either (2) G18.5 or (3) G17.5, 1 or 2 days preterm, respectively; (4) LPS-induced preterm mice born on G18.5.

To generate the lipopolysaccharide (LPS) 18.5 group, LPS (EMD Millipore: Calbiochem, lipopolysaccharide E. coli O111:B4, lot # D00144477) was injected i.p. (0.14 mg/kg, dissolved in 0.1% sterile saline for a final concentration of 1 mg/ml) to pregnant female mice on gestational day (G) 17.5 at 5 pm. This dose of LPS was chosen based on previous reports using LPS to induce premature birth (Guo et al., 2013). Once in solution, LPS was sonicated and then stored as a frozen stock. Dams for Term 19.5 and C-S 18.5 groups received control injections of saline (5 μl) on G17.5 at 5 pm.

Pregnant females at G17.5 or 18.5 were euthanized via cervical dislocation without anesthesia. A horizontal cut was made across the abdomen. The uterine horns were removed, and the pups were quickly dissected out of the uterus and embryonic sac. A paper wipe was used to dry the pup and remove excess fluid from the mouth. Six pups were removed from each female. The first 2–3 pups were used for subsequent breathing measures during the first 10 min of life.

All pups were removed from the mother immediately after birth and placed on a circulating water pad set to 32°C. A cardboard box top with holes cut into the side was placed over the pups. Survival was assessed during the first 2 h after birth. Death was defined by cessation of breathing and failure to respond to tactile stimulation.

Breathing was recorded within the first 10 min after birth. Each recording lasted for 3–5 min. The C-S 17.5 group was not included due to difficulty in getting the pups to breathe repeatedly. Breathing patterns were recorded from unrestrained mice by whole-body, flow, barometric plethysmograph (flow: 60 ml/min; compressed air 21% O₂, 0.1% CO₂). Changes in pressure were measured between the experimental and reference chamber (both 60 ml) using a BUXCO pressure transducer and preamplifier. The signal was digitized at 1,000 Hz (Digidata 1322A, Molecular Devices, Sunnyvale, CA) and captured using Axoscope 10.3 data acquisition software (Molecular Devices). Data were analyzed off-line using Clampfit 10.3 software (Molecular Devices). The ambient chamber temperatures were maintained at 32°C (Physitemp TCAT-SDF Controller, Clifton, NJ). However, other factors that influence gas volume measurements such as body temperature, barometric pressure, and relative humidity were not accounted for in calculations of tidal volume. Thus, breath volumes were normalized to calibrations made during each recording using a 50-μl Hamilton syringe attached to the experimental chamber. The amplitude ratio of breaths compared to calibrations was used to determine a value in microliter for each breath. To account for differences in body size, breath volumes were also normalized to body weight (μl/g). Although we acknowledge the limitations of plethysmography for precisely measuring tidal volume, our conclusions are primarily based on the relative size of breaths between experimental groups, not precise volumes. As a result, we do not expect these limitations to have a significant impact on the interpretation of our data.

Data were analyzed and graphed using GraphPad Prism 6 software. Data are presented as means±SE and statistical significance was determined at p < 0.05. Significance is indicated by: ****p < 0.0001, ***p < 0.001, **p < 0.01, and *p < 0.05. Appropriate t-tests (with or without Welch’s correction) or one- or two-way ANOVA with Bonferroni corrected post hoc tests were used to compare groups. Repeated measures ANOVA were used where appropriate. Details of each comparison and factors used for each two-way ANOVA are provided in figure legends.

We compared the survival rates among neonatal mice that were born naturally on G19.5, mice born 1 or 2 days prematurely via cesarean section, or mice born vaginally 1 day prematurely due to maternal inflammation. Survival rates during the first 2 h after birth were different across manipulations (Figure 1A). Most mice born naturally at term on G19.5 (Term 19.5; n = 52) survived (94%). Mice delivered prematurely by cesarean section (C-S) on G18.5 (C-S 18.5; n = 48) also exhibited a high survival rate (81%), close to those of term mice (p > 0.05), indicating that 1 day of prematurity alone is not a major contributor to mortality in these animals. However, survival of mice born prematurely on G18.5 in response to maternal administration of low-dose LPS on G17.5 (LPS 18.5; n = 99) was 50%, which was lower than both the C-S 18.5 and Term 19.5 groups (p < 0.001) suggesting that the inflammatory response contributes significantly to the greater rate of mortality in these pups. A fourth group of mice delivered via C-S on G17.5 (C-S 17.5; n = 25) exhibited survival of 0%, which was lower than the LPS 18.5, C-S 18.5, and Term 19.5 groups (p < 0.0001). Survival was not different between male and female mice in any group (two-way ANOVA with Bonferonni post hoc tests, p > 0.05). In the Term 19.5 group, 92% of males and 96% of females survived. In C-S 18.5 mice, 78% of males and 84% of females survived. And in LPS 18.5 mice, 62% of males and 48% of females survived. Thus, sex was not predictive of survival in these premature mice.

In general, younger pups weighed less than older pups (Figure 1B). C-S 17.5 pups (n = 25) weighed 0.83 ± 0.02 g, which was less than LPS (n = 126) and C-S 18.5 pups (n = 69) (1.02 ± 0.01 g and 1.09 ± 0.01 g, respectively), as well as Term 19.5 pups (n = 83; 1.26 ± 0.01 g; p < 0.0001). The weights of LPS 18.5 and C-S 18.5 pups were both different from Term 19.5 pups (p < 0.0001). The weights of LPS 18.5 pups were also lower than C-S18.5 (p < 0.01), despite being the same gestational age. Within each treatment group, pups that died less than 2 h after birth weighed less than those that survived (Figure 1C). LPS 18.5 pups that died (n = 16) weighed 0.92 ± 0.03 g and those that survived (n = 12) weighed 1.04 ± 0.02 g (p < 0.01). C-S 18.5 pups that died (n = 6) weighed 0.86 ± 0.06 g and those that survived (n = 14) weighed 1.14 ± 0.02 g (p < 0.0001). Term 19.5 pups that died (n = 1) weighed 1.01 g, whereas survivors (n = 21) weighed 1.30 ± 0.03 g. Overall, newborns that failed to survive (n = 48; 0.87 ± 0.01 g) for 2 h weighed less than did those that survived (n = 46; 1.18 ± 0.02 g; p < 0.0001) (Figure 1D).

Breathing activities were measured using whole-body, barometric plethysmography, and representative recordings for each experimental group are shown in Figure 2A. Minute volumes were different across treatment groups. Minute volumes in LPS 18.5 mice (0.16 ± 0.03 ml/min/g), but not C-S 18.5 mice (0.31 ± 0.10 ml/min/g), were smaller than Term 19.5 mice (0.51 ± 0.06 ml/min/g; p < 0.001) (Figure 2B). This effect was primarily associated with slower breathing frequencies in LPS 18.5 (16.21 ± 4.79 breaths/min) and C-S 18.5 mice (37.03 ± 8.08 breaths/min) compared to Term 19.5 mice (124.7 ± 8.64 breaths/min; p < 0.0001) (Figure 2C). However, tidal volumes normalized to body weights in LPS 18.5 pups (12.67 ± 1.08 μl/g), but not in C-S 18.5 pups (9.37 ± 0.95 μl/g), were larger than in Term 19.5 pups (7.09 ± 1.00 μl/g; p < 0.001) (Figure 2D). The durations of individual breaths, measured as the peak half-widths (peak widths at half-height), were larger in LPS 18.5 pups (524.0 ± 39.0 ms) than in C-S 18.5 (328.2 ± 48.8 ms) and Term 19.5 pups (151.0 ± 19.6 ms; p < 0.001) (Figure 2E). Mice in the LPS 18.5 group (n = 40) therefore had very low breathing rates with large amplitude and long-duration breaths, whereas Term 19.5 mice (n = 47) had more frequent and smaller breaths. C-S 18.5 mice (n = 28) were generally intermediate, suggesting that premature termination of intrauterine fetal development and the presumed maternal inflammation responses invoked by administration of LPS combine to influence breathing patterns observed in the newborn pups. Breathing patterns also varied with body weights. Lower weights were related to slower breathing rates in LPS 18.5 (p < 0.01) and C-S 18.5 mice (p < 0.01), but not Term 19.5 mice (p > 0.05) (Figure 2F). Breath half-widths were also significantly related to weight in LPS 18.5 (p < 0.05) and C-S 18.5 mice (p < 0.05), but not Term 19.5 mice (p > 0.05) (Figure 2G). Thus, breathing patterns were related to birth weights only when pups were born prematurely.

Breathing parameters were also compared between mice that died (n = 23) within the first 2 h post-delivery and those that survived (n = 47) (Figure 3). Overall, mice that died had lower minute volumes (0.04 ± 0.01 vs. 0.37 ± 0.03 ml/min/g; p < 0.0001) (Figure 3A) primarily associated with slower breathing rates (2.93 ± 0.57 vs. 75.10 ± 8.98; p < 0.0001) (Figure 3B). Mice that died also had larger tidal volumes (12.90 ± 0.74 vs. 8.19 ± 0.89 μl/g; p < 0.001) (Figure 3C) and longer breath half-widths (628.8 vs. 254.4 ms; p < 0.0001) (Figure 3D) than those that survived. These trends in breathing parameters between survivors and non-survivors were apparent in all treatment groups.

Because mice breathe faster than humans, apneas are typically defined as the cessation of two or more respiratory cycles (Nakamura et al., 2003). This duration is considered to be equivalent to a pause of >10 s in humans (Davis and O’Donnell, 2013). However, the different breathing patterns reported here were characterized by strikingly different respiratory cycle lengths. For example, the “slow, large breathing pattern” often observed in LPS 18.5 mice was characterized by very long baseline cycle lengths, and therefore a 10-s interval would not be considered an apnea in these mice. Yet, this 10-s interval will certainly cause more hypoxia to the brain than an apnea of >two cycle lengths in a mouse breathing normally at 2 Hz. Thus, from this perspective, it seems that under baseline conditions, a typical LPS 18.5 mouse experiences an “apnea” almost every respiratory cycle. However, using different definitions of apnea for each distinct breathing pattern would seem arbitrary and confusing. Thus, in the present study, we present a careful quantification of breathing rate and other parameters without attempting to score the breaths or use different apnea definitions.

The quantitative differences in breathing parameters described in the previous paragraph were associated with qualitatively different breath types. We observed four distinct patterns, which we refer to as “eupneic activity,” “gasping,” “sawtooth breaths,” and “large breaths” (Figure 4A). “Eupneic” or normal breathing activity (n = 50) was characterized by relatively high respiratory frequencies and inspiratory efforts with small half-widths (109.7 ± 5.9 ms) and small tidal volumes (4.05 ± .027 μl/g) (Figure 4B). Occasionally, we observed “gasps” (n = 9). This respiratory activity exhibited half-widths similar to those observed with eupneic breathing (153.0 ± 16.0 ms; p > 0.05), but gasps are differentiated from “eupneic activity” by the much larger tidal volumes (13.92 ± 1.50 μl/g; p < 0.0001) (Figure 4B). Gasps and eupneic breathing are also observed in mature mice (Baertsch et al., 2019) and are not specific to perinatal respiratory activities.

However, we also observed two types of activity patterns that were limited to the first hour after birth and are not observed in mature mice. Inspiratory efforts we refer to as “large breaths” (n = 40) were characterized by large tidal volumes (13.01 ± 0.64 μl/g), as well as very long half-widths (624.0 ± 30.3 ms) (Figure 4, red bars and trace). The long durations of these respiratory events clearly differentiate this pattern from gasps (p < 0.0001). Large breaths were typically associated with long pauses between breaths, resulting in very slow respiratory frequencies. Somewhat similar to large breaths were respiratory events that we refer to as “saw tooth pattern” (n = 8) (Figure 4, orange bars and trace). The tidal volume of saw tooth breaths (16.23 ± 2.29 μl/g) was similar to the large breaths (p > 0.05), and half-widths (363.3 ± 56.0 ms) were intermediate between gasps (p < 0.01) and large breaths (p > 0.05). However, sawtooth breaths have two characteristics that differentiate them from the large breaths. Most obvious is the superimposition of eupneic-like breaths that create the saw tooth pattern in the pressure curves. Secondly, the onset of these breaths is different from the large breaths. It appears that the onset may reflect a prolonged gasp in the saw tooth pattern, while in the large breaths, the onset is much rounder (Figure 4A, compare red and orange traces). Quantified tidal volumes and half-widths for each breath pattern are shown in Figure 4B.

All four breathing patterns were observed during the first hour after birth in both term and preterm mice. However, differences were observed in the prevalence of these patterns among groups (Figures 4C,D). During the first 10 min post birth, the probability of eupneic breathing was significantly greater in Term 19.5 mice (0.81 ± 0.06) than in LPS 18.5 (0.08 ± 0.05) and C-S 18.5 (0.33 ± 0.10) mice (p < 0.001). In contrast, the probability of large breaths was higher in LPS 18.5 (0.72 ± 0.08) and C-S 18.5 (0.42 ± 0.10) mice than Term 19.5 mice (0.083 ± 0.04; p < 0.05). Gasping and saw tooth breathing patterns were comparatively rare and their prevalence did not differ between term and preterm mice (p > 0.05).

These breathing patterns were predictive for survival (Figure 5). Pups that exhibited large breaths shortly after birth did not transition to normal breathing and did not sustain life. An example recording of an LPS 18.5 mouse with a large breathing pattern that fails to transition to normal breathing is shown in Figure 5A. Indeed, 100% of mice, term or preterm, that died within the first hour of life exhibited large breaths during the first 10 min after birth. Moreover, 67% of mice that exhibited large breaths (n = 30) during the first 10 min of life did not survive (Figure 5C). In contrast, preterm or term mice that exhibited either eupneic activity (n = 32), gasping (n = 2), or the saw tooth (n = 6) patterns survived the first hour of life after birth (p < 0.01) (Figure 5C). An example recording of a C-S 18.5 mouse with a saw tooth breathing pattern that successfully transitioned to eupneic breathing is shown in Figure 5B. Among pups that survived, breathing pattern transitions between the first 10 min of life and 1 h post birth are quantified in Figure 5D, demonstrating that saw tooth and large breath patterns can transition to eupnea to enable survival.

For pups that survived the first hour of life (n = 30), transitions in quantitative breathing parameters between the first 10 min of life and 1 h post birth are shown in Figure 6. Although overall changes in minute volume were small during the transition to normal breathing (0.34 ± 0.04 to 0.42 ± 0.03 ml/min/g; p > 0.05) (Figure 6A), breathing rates increased (56.6 ± 10.27 to 97.9 ± 8.8 breaths/min; p < 0.01) (Figure 6B), whereas tidal volumes (9.69 ± 1.18 to 5.51 ± 0.78 μl/g; p < 0.01) and breath half-widths (321.5 ± 41.6 to 121.0 ± 14.5 ms; p < 0.0001) decreased (Figures 6C,D). Thus, faster breathing with smaller breaths developed in pups that survived the first hour of life. Among LPS 18.5 pups that survived (n = 9), changes in minute volumes (0.31 ± 0.08 to 0.56 ± 0.07 ml/min/g; p < 0.01), breathing rates (32.09 ± 15.08 to 111.67 ± 17.96 breaths/min; p < 0.0001), tidal volumes (12.07 ± 1.55 to 6.55 ± 1.87 μl/g; p > 0.05), and breath half-widths (402.9 ± 71.3 to 95.0 ± 14.7 ms; p < 0.001) during this period were the most pronounced, suggesting that breathing must undergo the most dramatic changes during the first hour of life in LPS 18.5 pups in order to ensure survival.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.