Comparison of survival of preterm newborn rabbits at 25–28 days of gestation with perinatal therapies at birth transition

Abstract

Eligibility of ventilated preterm rabbit model to investigate extreme pulmonary immaturity at birth transition is unknown. By extending this model to early saccular stage of fetal lung development, we evaluated efficacy in survival, lung maturation, and underlying mechanisms of contemporary perinatal therapies. Pregnant New Zealand White rabbit does were given dexamethasone (DEX), or sham injection as control (NDEX), 48 and 24 h before delivery at gestational age (GA) of 25–28 days. At birth, newborn rabbits were anesthetized and randomly allocated to four groups receiving either surfactant or nonsurfactant for both DEX and NDEX, and mechanically ventilated within low tidal volumes. Ranges of time to maintain survival rate ≥ 50% in GA 25–28 days were 59–136, 138–259, 173–288, and 437 to ≥600 min, respectively, each across the four groups. The benefits of DEX and/or surfactant for survival were more obvious in GA 25–26 days, as judged by improved lung mechanics, lower lung injury scores, higher lung surfactant phospholipid pools, and surfactant protein mRNA expression, with DEX-surfactant combination being the most optimal for the outcome. In contrast, those of GA 27–28 days had variable but meaningful responses to the treatment. Cox regression analysis revealed GA, DEX, and surfactant being independently protective factors whereas pneumothorax was a risk factor. The extremely preterm rabbits at GA 25–26 days markedly responded to the perinatal therapies for longer survival, lung maturation and injury alleviation, and were relevant for study of preterm birth transition-associated morbidities and underlying mechanisms.

NEW & NOTEWORTHY An extremely preterm rabbit model with gestational age of 25–26 (term 31) days was established by mechanical ventilation with individually adjusted tidal volume at lower ranges. The administration of antenatal glucocorticoids and/or postnatal surfactant achieved significantly longer duration to maintain 50% survival and facilitated lung maturation and protection at early saccular stage. The usefulness of this model should be validated in future investigation of perinatal and neonatal morbidity and mortality at extremely preterm birth transition.

INTRODUCTION

Advances in perinatal care have led to increased survival of extremely preterm (EPT) infants (1, 2) whose pulmonary structural and functional immaturity is crucial in limiting the survival capacity and treatment options (3). Antenatal glucocorticoids, postnatal surfactant, and invasive and noninvasive ventilation strategies constitute the mainstay of respiratory support in preterm births (4–7). However, despite maximal noninvasive and invasive respiratory support, many EPT infants develop respiratory distress (8, 9), extubation failure (10, 11), pneumothorax (PTX) (9, 12), hospital-acquired infection (13), bronchopulmonary dysplasia (6, 9), and other major comorbidities associated with prematurity. Long-term neurodevelopmental impairment, growth failure, and frequent rehospitalizations (14–16) are still common in surviving EPT infants in early childhood. Therefore, approaches to optimize respiratory support for EPT infants at birth transition remained the focus of ongoing research (9, 17, 18). One hurdle is a lack of appropriate experimental animal models to study the pathophysiological, cellular, and molecular mechanisms underlying lung maturation and injury protection in EPT infants (19), especially those involving antenatal and peripartum adverse events and/or neonatal intensive care support following delivery. Due to limitations of the mouse, rat, or other large animal models in terms of sophisticated techniques and high costs, the ventilated preterm rabbit model at birth transition has been used for years (20–28). This model offers a promising alternative to simulate preterm birth and perinatal therapies and investigate pathophysiological mechanisms underlying critical illnesses, in which the lung development is as distinctive in stages as with advancing gestational days. However, the ventilated rabbit model of extreme prematurity, at the limit of viability, has not been reported.

The body-plethysmograph system for parallel ventilation of multiple neonate rabbits was originally designed for fixed peak inspiratory pressure (PIP) ventilation by Lachmann et al (29) in the 1980s where variable tidal volume (Vt) was passively obtained. It was further modified to offer customized, identical tidal volume in ranges under pressure control mode by individual adjustment of PIP (24, 26, 27). In this study, we adopted the original methodology and our previous experience from the ventilated moderate preterm and near-term [gestational age (GA): 27–30 (term 31) days] rabbit models (21, 25, 28, 30, 31). We aimed to explore the feasibility of this model at extreme prematurity (GA: 25–26 days). We hypothesized that a longer duration of survival may be achieved with the use of antenatal glucocorticoids and/or postnatal surfactant along with an individually adjustable, lower Vt ventilation strategy. We estimated benefits and risks as well as underlying mechanisms of the perinatal therapies in reference to that of the less immature newborn rabbits.

MATERIALS AND METHODS

Experimental Protocol and Animal Management

The study protocol was approved by the Ethics Committee of Children’s Hospital of Fudan University (No. 2019193), with experimental procedures complying with the requirement of Chinese national regulations for experimental animal care and protection. Healthy pregnant, GA date-mated New Zealand White rabbits were obtained from the Shanghai Songlian Experimental Animal Center and housed under standardized care conditions until delivery date. Does were allocated to either antenatal dexamethasone-treated (DEX) or nontreated (NDEX) group, receiving two intramuscular (im) doses of 0.1 mg/kg dexamethasone (Macklin, Shanghai, China) or sham injection of sterile normal saline (23, 25, 26) 48- and 24-h before scheduled delivery at 25, 26, 27, or 28 days.

On the scheduled delivery date, the does were sedated with intramuscular diazepam (5 mg/mL, Shanghai Xudong Haipu Pharmaceutical, Shanghai, China) followed by intravenous (iv) urethane infusion (5 mL/kg, 20%, ethyl carbamate, BBI Life Sciences, Shanghai, China) as anesthesia (30, 31). Then, sequential injections of lidocaine (20 mg/mL, Shandong Hualu Pharmaceutical Co. Ltd., Jinan, Shandong, China) were administered subcutaneously and intramuscularly to the abdominal wall followed by cesarean section (25, 27). Oxygen was provided by a face mask throughout the delivery procedure while the animal pulse was closely monitored, and a maintenance dose of urethane (1 mL/kg) was provided intravenously as needed. Following delivery, the does were euthanized by an overdose of potassium chloride. The fetuses were delivered sequentially from each uterus, dried, weighed, and were given intraperitoneally (ip) 0.05–0.1 mL of lidocaine and 10% glucose mixture (vol/vol 1:1) followed by tracheostomy and intratracheal intubation (30). Neonatal rabbits were randomly allocated to subgroups treated with either surfactant (S, administered intratracheally phospholipids 200 mg/kg) or control (C, administered sham air). Thus, in each GA day there consisted of four parallelly treated groups subjected to mechanical ventilation from birth: NDEX-C, sham-treated control; NDEX-S, surfactant only; DEX-C, DEX only; DEX-S, both DEX and surfactant. The surfactant preparation, was derived from porcine lungs by multiple saline bronchoalveolar lavage, centrifugation, methanol/chloroform extraction, and sterilization (31, 32). It contained 40 mg phospholipids/mL as sterile saline suspension with >45% disaturated phosphatidylcholine (DSPC) and >60% PC in total phospholipids (TPL) and 1% surfactant protein (SP)-B and -C. Nonventilated subgroups (time 0) were also obtained from selected does of DEX-treated and nonDEX-treated right after delivery, immediately euthanized, and their lungs were processed and biochemically analyzed for baseline surfactant contents.

Mechanical Ventilation, Lung Function, and Survival Time

The newborn rabbits were ventilated in a ventilator-plethysmograph system that enabled parallel ventilation of 12 animals (27, 30). A Servo ventilator (900 C, Siemens-Elema, Solna, Sweden) was set to pressure control mode and a fraction of oxygen ($\text{F}{\text{I}}_{{\text{O}}_2}$) of 1.0, positive end-expiratory pressure (PEEP) 2–3 cmH₂O, frequency 40/min, and an inspiration-to-expiration ratio 1:2. During the ventilation, the animals, kept at 37°C with a heating system, received hourly intraperitoneal injection of 50–100 µL of a mixed solution containing 2% lidocaine, 5% NaHCO₃, and 10% glucose (vol/vol/vol 1:3:6) (30). PIP was recorded with a pressure transducer (Shanghai Yangfan Electronic, Shanghai, China) and adjusted individually along with PEEP to obtain a target Vt (mL/kg birth weight, BW) of 4–6 mL/kg for newborn rabbits of GA 28 days (30, 31). To avoid early pneumothorax and death, during the initial 15–30 min of ventilation, PIP was limited to under 30 cmH₂O, and target Vt was set to 2–3, 3–4, and 4–6 mL/kg for those of GA 25, 26, and 27–28 days, respectively (based on pilot studies, in the NDEX-C group). Subsequent adjustment of PIP among the groups was based on the real-time Vt gain and was confirmed by the adequate lung expansion for gas exchange, oxygenation and peripheral perfusion based on visual inspection of skin color. Vt was measured by a pneumotachometer (RSS100-HR, Hans Rudolph Inc., Kansas City, KA), and simultaneously integrated with the recordings of PIP and PEEP on an individual basis by an automated physiologic monitoring system (PowerLab, ADInstruments Pty Ltd, NSW, Australia). These recordings were obtained at prespecified time points every 15–30 min until spontaneous death, or by prescheduled end-points which were set at 240, 360, 360, and 600 min for GA 25, 26, 27, and 28 days (Fig. 1, Supplemental Tables S2 and S4), respectively. The dynamic compliance of respiratory system (Cdyn, mL/kg/cmH₂O) was derived by dividing Vt (mL/kg) with the difference of PIP and PEEP (PIP-PEEP, cmH₂O) (27, 30).

Figure 1.

Survival curve in gestational age and intervention groups. Group definition: NDEX-C, neither antenatal dexamethasone nor postnatal surfactant treatment; NDEX-S, postnatal surfactant (200 mg/kg) only; DEX-C, antenatal dexamethasone (two doses of 0.1 mg/kg DEX at 24 h apart) only; DEX-S, both antenatal dexamethasone and postnatal surfactant. *P < 0.01, **P < 0.01 vs. NDEX-C; †P < 0.05, vs. NDEX-S (by log-rank test in Kaplan-Meier analysis). For group sample size, see Table 1. DEX, dexamethasone; DEX-C, dexamethasone-sham treated-control; DEX-S, dexamethasone-surfactant only; NDEX-C, control with no dexamethasone or surfactant; NDEX-S, control with surfactant only. The sample sizes of groups NDEX-C, NDEX-S, DEX-C and DEX-S in A [gestational age (GA) 25 days] were 23, 23, 21, 24; in B (GA 26 days) 61, 60, 36, 38; in C (GA 27 days) 51, 53, 30, 28; and in D (GA 28 days) 29, 30, 28, 26, respectively.

During ventilation, precordium pulsation, a sign of heartbeat, was closely observed and matched by color changes of lips, limbs, and trunks, to determine animal survival status, with the observer unblinded from the group. Survival time was recorded, and the duration to maintain ≥50% survival rate (ST₅₀) was obtained for each group. The animals were kept sedated during the entire ventilation period. Individual ventilation was terminated when spontaneous death occurred or preset ventilation time was reached. Euthanasia was performed by intracranial injection of 0.5 mL 2% lidocaine, which caused immediate cardiac arrest (27, 30).

Lung Sample Processing

At the termination of ventilation, PTX was inspected through abdominal incision and diaphragm, and was confirmed by the lung mechanic measurement (unanticipated Cdyn drop) when the pneumothorax was identified at postmortem examination. The appendix lobe of the right lung was removed for measurement of wet-to-dry weight ratio (W/D) for estimation of lung tissue fluid content (30). The remaining lungs were randomly allocated for either histopathological or biochemical measurements.

Lung Histopathology

The lungs allocated to histopathological measurement (n = 8–22/group) were fixed en bloc with 4% paraformaldehyde and embedded in paraffin, sectioned, and stained with hematoxylin-eosin. The lung sections of both lungs were examined under light microscope (Leica Microsystems, Wetzlar, Germany). To estimate lung expansion, a point-counting method based on semiquantitative geometry theorem was employed at ×200 magnification, to assess volume density (Vv) of aerated alveolar space in total parenchyma (27, 30, 31). Coefficient of variance [CV (Vv)] was estimated for field-to-field variation of alveolar aeration. A 4-score scale was used to estimate the magnitude of lung injury severity using the same semiquantitative methodology for items of injury pattern (27, 30, 31). A scale 0 was for none or very minor (<1%), 1 for mild-to-moderate (<25%), 2 for moderate-to-severe (25%–50%), 3 for very-to-severe (50%–75%), and 4 for widespread and most prominent (>75%) in the field. There were four items of lung injuries graded: edema, hemorrhage, inflammation (neutrophil infiltration), and small airway epithelial desquamation. A sum of each item score constituted the total score of lung injury (LIS_total) (31).

Biochemical Analysis

The lungs (8–25/group) of animals assigned to biochemical analysis were extracted intact bilaterally and weighed to obtain wet lung weight (WW), which was compared with the body weight (BW, WW/BW). Bronchoalveolar lavage (BAL) was obtained and performed on the left (or no PTX) side, which was repeated thrice using sterile saline at a volume equivalent to 15 mL/kg BW, and pooled BAL fluid (BALF), was immediately centrifuged at 2,000 rpm for 15 min at 4°C to remove cell debris and obtain supernatant for further use (30). The lobes after lavage were homogenized in saline (vol/vol = 1:9) with an automatic grinder (JXFSTPRP-24, Shanghai Jingxin Industrial development, Ltd, Shanghai, China) at 60 Hz for 1 min to obtain lung tissue homogenate (LH). The nonlavage side of the lungs was snap-frozen and kept at −80°C for further biochemical and mRNA measurements.

TPL were extracted by chloroform-methanol (vol/vol 2:1) from BALF (supernatant) and LH (33, 34), separately. DSPC was recovered by the neutral alumina column chromatography after exposing the isolated TPL to osmium tetroxide (26, 30, 35). TPL and DSPC were quantified with the molybdenum acid method for inorganic phosphorus determination (26, 30). Total protein (TP) content was quantified using BCA kit (Pierce BCA Assay Kit 23225, Thermo Scientific, Rockford, IL). These values are presented as mg/kg BW or ratio.

A part of the nonlavage side of the lung was ground and used with commercially available testing kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) to measure the contents of malondialdehyde (MDA) and reduced glutathione (GSH), and the activities of myeloperoxidase (MPO) and total superoxide dismutase (SOD).

Quantitative Polymerase Chain Reaction of mRNAs

The rest of nonlavaged lobe was used for qPCR analysis, which was performed with Roche LightCycler 480 II (Roche, Basel, Switzerland). The amplification reaction was carried out in 10-μL volume containing 5-μL SYBR Premix Ex Taq (Takara Bio Inc., Otsu, Shiga, Japan). Triplicate sample were employed to ensure quality control. Trizol reagent (Invitrogen, Carlsbad, CA) was used to extract total RNA from the lung tissue sample, and the concentration of RNA was measured by colorimetry. The PrimeScript RT reagent kit with genomic DNA Eraser (Takara Bio Inc., Otsu, Shiga, Japan) was utilized to synthesize complementary DNA (31). Target gene sequence information was obtained through nucleotide databases (www.ncbi.nlm.nih.gov/gene/). The target mRNAs included: 1) surfactant synthesis related mRNAs: surfactant protein (SP)-A, -B, -C, and phosphocholine cytidylyltransferase (CCT); 2) growth factors (GFs): vascular endothelial GF (VEGF), insulin-like GF-1 (IGF-1), and IGF-2, platelet-derived GF (PDGF)-B; and 3) inflammatory mediators: interleukin-6 (IL-6) and IL-8. The mRNA expression of each target gene was normalized to β-actin, using the ΔΔCT method for mean fold changes calculation (36). All the primer sequences are summarized in the electronic Supplemental Table S1 (all Supplemental Tables are available at https://doi.org/10.6084/m9.figshare.13341215.v2).

Data Analysis

Quantitative data are expressed as means ± standard deviation (SD). SPSS 23.0 (IBM, Armonk, NY) was applied for data presentation and statistical analysis. Kaplan–Meier analysis and multivariate Cox regression model were used for survival analysis. The validity of the proportionality assumption attached to the Cox model was examined by Nelson Aalen cumulative hazard plots and by testing Schoenfield residuals (37). One-way analysis of variance (ANOVA) followed by Bonferroni post hoc test was performed for comparing quantitative data among GA or treatment groups. Correlations of continuous and categorical parameters with survival time were tested by Pearson and Spearman correlation analysis, respectively. Pearson Chi-square test was performed on proportional data. A P value < 0.05 was considered statistically significant for the difference.

RESULTS

Five-hundred and sixty newborn rabbits (male:female 15:14) were enrolled in this study. Table 1 shows groups, BW, occurrence of PTX, and LIS_total. Supplemental Table S2 in the supplemental material illustrates the numbers of animals in each group and those underwent lung processing, as well as the time course of early death or PTX occurrence. BW and WW increased in parallel with GA in both DEX and NDEX subgroups and were generally 10%–40% lower in DEX subgroups [Table 1 and Supplemental Table S3]. The average values of WW/BW were generally above 2% (ranging from 2.3% to 2.9%) in all groups, but significantly lower in DEX than in the NDEX counterparts at GA 26 and 27 days (Supplemental Table S3). Values of W/D were lower in DEX of GA 26 and 27 days (Supplemental Table S3). During the ventilation, PEEP fluctuated passively as the peak inspiratory pressure (PIP) varied individually (to achieve adequate tidal volume), and the number of animals in the parallel plethysmograph-ventilator system decreased as time elapsed. The measured value of PEEP varied and average PEEP over the time was 1–3 cmH₂O. The measured Vt values, based on both PIP and PEEP, or PIP-PEEP, was around 1–3 and 3–6 mL/kg for those of GA 25–26 and 27–28 days, respectively, which modestly deviated from the intended ranges. The Vt ranges (1–3 mL/kg) used for the GA 25–26 days were associated with high prevalence of PTX (Table 1 and Supplemental Table S2), whereas a relatively lower PTX (30%) was identified in the lower portion of Vt ranges. In contrast, there were fewer to none pneumothoraces in the GA 27 and 28 day GA groups (Table 1 and Supplemental Table S2). Notably, by combining the GA 25–26 day groups, DEX and/or surfactant reduced PTX by ∼15%–30% (P < 0.05).

Table 1.

Survival and major morbidities of ventilated preterm rabbits in gestational age and intervention groups

GA, days	Group	n	BW, g	ST50	PTX, n%	LIStotala
25	NDEX-C	23	17.7 ± 2.5	59 (42, 106)	18 (78)	10.5 ± 2.9
	NDEX-S	23	17.6 ± 2.9	136 (97, 200)	7 (30)*	5.0 ± 2.1*
	DEX-C	21	14.5 ± 1.2**‡	122 (50, 144)	17 (81)†	7.7 ± 2.0
	DEX-S	24	15.5 ± 1.6**‡	130 (101, 180)	10 (42)§	4.3 ± 1.9**
26	NDEX-C	61	25.2 ± 5.0	138 (57, 220)	33 (54)	8.9 ± 2.3
	NDEX-S	60	24.3 ± 4.5	156 (84, 305)	23 (38)	6.3 ± 2.1**
	DEX-C	36	16.3 ± 2.7**‡	259 (134, 330)	9 (25)*	8.2 ± 1.6
	DEX-S	38	15.7 ± 2.8**‡	231 (180, 289)	8 (21)*	7.8 ± 2.0
27	NDEX-C	51	28.3 ± 3.9	173 (90, 321)	15 (29)	9.0 ± 2.4
	NDEX-S	53	27.8 ± 4.3	285 (167, 360)	6 (11)	4.6 ± 1.8**
	DEX-C	30	22.4 ± 3.5**‡	281 (198, 423)	7 (23)	8.3 ± 3.4†
	DEX-S	28	22.2 ± 3.3**‡	288 (146, 400)	3 (11)	7.5 ± 2.9
28	NDEX-C	29	35.5 ± 5.5	572 (325, >600)	0	4.1 ± 3.2
	NDEX-S	30	34.4 ± 6.2	437 (241, >600)	0	3.2 ± 1.9
	DEX-C	28	28.1 ± 3.3**‡	>600	0	3.6 ± 1.3
	DEX-S	26	27.3 ± 4.0**‡	>600	0	4.3 ± 1.4

Values are presented as means ± SD or median (interquartile range). Group definition: DEX-C, antenatal dexamethasone (two doses of 0.1 mg/kg DEX at 24 h apart) only; DEX-S, both antenatal dexamethasone and surfactant. NDEX-C, control with no dexamethasone or surfactant; NDEX-S, surfactant (200 mg/kg) only. ^a8–22 for LIS_total in each group. *P < 0.05, **P < 0.01 vs. NDEX-C; †P < 0.05, ‡P < 0.01, vs. NDEX-S; §P < 0.05, §§P < 0.01 vs. DEX-C by one-way analysis of variance followed by Bonferroni post hoc test or Pearson Chi-square test, whichever applicable. BW, birth weight; GA, gestational age; LIS_total, total lung injury score; PTX, pneumothorax; ST₅₀, survival time to maintain 50% survival rate.

Survival Time

In general, the survival time increased as GA advanced in 25–28 days (Table 1 and Supplemental Table S2, and Fig. 1), i.e., ranges of ST₅₀ being 59–136, 156–259, 173–288, and 437 to >600 min, respectively, each across the four treatment groups (Fig. 1, Supplemental Table S2). The ST₅₀ for those in NDEX-S, DEX-C, and DEX-S of GA 26 days was prolonged approximately equivalent to that of the NDEX-C of 27 days, but this effect was variable in those of GA 27 days compared with the corresponding counterparts of GA 28 days.

Measurement of Lung Mechanics

Overall, Cdyn increased as GA advanced, with surfactant use resulting in more prominent effect in GA 25–27 days (Fig. 2). In GA 25 and 26 days, surfactant improved Cdyn, with added effect discernible to that of DEX (Fig. 2, Supplemental Table S4). In the GA 27 days, both DEX and surfactant improved Cdyn, but the additive effect of surfactant on DEX was modest (Fig. 2, Supplemental Table S4). In the GA 28 days, changes of Cdyn over time were relatively stable, with Cdyn in NDEX-S being relatively lower than that in NDEX-C (Fig. 2, Supplemental Table S4).

Figure 2.

Dynamic compliance of the respiratory system in gestational age and intervention groups. Data are presented as means ± SD, derived from the values at the specified time point of ventilation (see Supplemental Table S4 for original data). For group definition, see Fig. 1 legends. For group survival time, see Supplemental Table S2. The sample sizes of groups NDEX-C, NDEX-S, DEX-C, and DEX-S in A [gestational age (GA) 25 days] were 23, 23, 21, 24; in B (GA 26 days) 61, 60, 36, 38; in C (GA 27 days) 51, 53, 30, 28; and in D (GA 28 days) 29, 30, 28, 26, respectively.

Lung Histopathology and Morphometry

Values of Vv were the lowest, and those of CV (Vv) the highest in NDEX-C at GA 25 and 26 days. Vv increased as GA advanced and it improved with surfactant at GA 25–27 days and by DEX at GA 25–26 days (Supplemental Table S5). Values of LIS by injury items are shown in Supplemental Table S5. Surfactant-treated subgroups had mild hemorrhage and inflammation in those of GA 25 and 27 days. More prominent mitigation of small airway impairment was found in the DEX-S group of GA 25 days. LIS_total decreased in respective NDEX-C groups with the advancing GA (Table 1). At 25 day GA, both NDEX-S and DEX-S groups had lower LIS_total, whereas only NDEX-S had significantly lower LIS_total in the GA 26–27 day groups. Representative photomicrographs of the lung samples are shown in Supplemental Fig. S1 (all Supplemental Figures are available at https://doi.org/10.6084/m9.figshare.13341221.v1).

Biochemical Analysis of BALF and Lung Tissue Homogenate

In the nonventilated subgroups, TPL and DSPC in BALF and LH were ∼1–5-folds higher in DEX than that in NDEX, whereas DSPC/TPL in the BALF was ∼40%–50%. Variable but modest increment of TPL and DSPC was seen in LH and BALF + LH with advanced GA (Supplemental Table S6).

TPL_BALF and DSPC_BALF increased by 1–4 folds in the consecutive GA days in NDEX-C or NDEX-C0 (Table 2 and Supplemental Table S6). With ventilation, in DEX-C groups, none or onefold increment for both TPL_BALF and DSCP_BALF were found compared with the DEX-C0 of the same GA. (Table 2, Supplemental Tables S6, and S7). Similarly, TPL and DSPC of LH + BALF in NDEX-C groups increased by 0–3 folds and 1–7 folds compared with the corresponding NDEX-C0 groups of the same GA, respectively, with relatively more changes in GA 25–26 day groups. Such changes in the DEX groups were less substantial (Supplemental Tables S6 and S7). In the ventilated groups, as GA advanced, values of TPL and DSPC increased in BALF (Table 2) and remained almost unchanged in LH (Supplemental Table S7). In general, groups treated with surfactant had improved TPL_BALF, DSPC_BALF, and DSPC_LH, mainly in the GA 25–27 days (Table 2 and Supplemental Table S7). The average level of DSPC_LH was the highest in DEX-S (Supplemental Table S7), and that of DSPC/TPL in BALF around 50% across all GA groups (Table 2), and in LH around 20%–30% and increased in NDEX-S and DEX-S groups (Supplemental Table S7). The DSPC/TPL of BALF + LH as total pool was higher in DEX-S in almost all GAs (Supplemental Table S7). Groups with longer ST₅₀ and higher Cdyn usually had higher DSPC_BALF, DSPC_LH, and total DSPC (Supplemental Table S7). The TP levels in BALF were the lowest in GA 25 days compared with 26–27 days (Table 2), but not significantly different in various treatment groups. DSPC/TP increased in the NDEX-S groups of GA 25–27 days, but not in other groups.

Table 2.

Biochemical analysis of bronchoalveolar lavage fluid samples

GA, days	Group	TPL, mg/kg	DSPC, mg/kg	TP, mg/kg	DSPC/TPL, %	DSPC/TP, mg/mg
25	NDEX-C	2.72 ± 1.27	1.10 ± 0.53	11.1 ± 5.50	46.7 ± 21.7	0.09 ± 0.03
	NDEX-S	19.4 ± 6.96**	6.33 ± 2.49**	17.7 ± 6.33	35.4 ± 15.9	0.40 ± 0.21*
	DEX-C	3.35 ± 1.13†	1.30 ± 0.51†	13.7 ± 6.54	47.7 ± 4.19	0.08 ± 0.06‡
	DEX-S	12.9 ± 6.73*	7.23 ± 5.68*	22.0 ± 11.2*	48.1 ± 18.1	0.34 ± 0.26
26	NDEX-C	6.98 ± 3.91	3.60 ± 2.05	29.4 ± 17.1	51.3 ± 15.5	0.14 ± 0.07
	NDEX-S	17.5 ± 6.12**	7.71 ± 3.04**	35.2 ± 14.2	46.4 ± 14.9	0.23 ± 0.09**
	DEX-C	13.1 ± 6.41*	5.53 ± 3.17	27.7 ± 15.3	46.7 ± 16.5	0.25 ± 0.16
	DEX-S	26.0 ± 10.3**§§	11.8 ± 4.90*‡§§	40.0 ± 16.3	45.2 ± 11.8	0.35 ± 0.21**
27	NDEX-C	14.0 ± 5.82	6.01 ± 2.77	47.0 ± 23.7	43.4 ± 10.1	0.14 ± 0.04
	NDEX-S	38.0 ± 19.1**	18.4 ± 11.7**	38.7 ± 12.3	45.9 ± 11.8	0.44 ± 0.27**
	DEX-C	7.85 ± 4.98‡	2.82 ± 2.57‡	22.9 ± 11.9*†	39.8 ± 20.9	0.13 ± 0.06‡
	DEX-S	16.5 ± 5.01†	8.03 ± 3.76§	34.5 ± 22.5	47.4 ± 15.6	0.27 ± 0.15
28	NDEX-C	30.0 ± 14.9	14.2 ± 7.40	27.8 ± 6.9	48.9 ± 12.9	0.51 ± 0.24
	NDEX-S	37.7 ± 25.7	23.1 ± 13.2	34.5 ± 15.3	54.1 ± 10.2	0.72 ± 0.41
	DEX-C	15.8 ± 4.10	5.83 ± 3.92‡	41.7 ± 15.2*	36.5 ± 18.0	0.15 ± 0.07**‡
	DEX-S	30.1 ± 11.4	11.0 ± 6.85	43.6 ± 14.7*	36.4 ± 14.0	0.25 ± 0.14*‡

Values are presented as means ± SD. For group definition, see Table 1. *P < 0.05, **P < 0.01 vs. NDEX-C; †P < 0.05, ‡P < 0.01, vs. NDEX-S; §P < 0.05, §§P < 0.01 vs. DEX-C by one-way analysis of variance followed by Bonferroni post hoc test. The sample sizes were 10–39/group. For sample size in each group and the time point of sample collection, see Supplemental Table S2.

DEX-C, dexamethasone-sham treated-control; DSPC, disaturated phosphatidylcholine; NDEX-C, control with no dexamethasone or surfactant; NDEX-S, control with surfactant only; TPL, total phospholipid; TP, total proteins.

There were no significant changes in values of MDA among the GA and intervention groups. MPO was associated with GA (r = 0.24, P < 0.01), but was not influenced by the interventions in each GA stratum. GSH was also associated with GA (r = −0.35, P < 0.01) and it decreased by DEX in those of GA 26 days. SOD decreased as GA advanced (r = −0.56, P < 0.01) and it increased by DEX in those of GA 25–26 days (Supplemental Table S8).

Measurement of mRNA

Accompanying longer ST₅₀, significantly increased expression of SP-A and SP-B mRNAs was found in DEX-C and DEX-S groups of GA 25 and 26 days (Supplemental Table S8). GA was closely correlated with the expression of mRNA of IGF-1 (r = −0.72, P < 0.01), IGF-2 (r = −0.43, P < 0.01), PDGF-B (r = −0.15, P < 0.05), VEGF (r = 0.48, P < 0.01), and IL-6 (r = 0.34, P < 0.01), but not affected by the interventions in any GA subset, whereas IL-8 was affected by neither GA nor interventions (Supplemental Table S8). In the nonventilated subgroups, modestly enhanced expression was found in SP-A, SP-B, and SP-C mRNA as GA advanced whereas that of CCT was not (Supplemental Table S9).

Multivariate Cox Regression and Correlation Analysis for Survival

The survival analysis by multivariate Cox regression revealed that GA, DEX, surfactant, and Cdyn were protective factors (Table 3); all being closely and positively correlated with ST₅₀ (r = 0.617, 0.162, 0.250, respectively, P < 0.05). PTX and W/D were risk factors and W/D correlated negatively with ST₅₀ (r = −0.452, P < 0.05). SP-A and SP-B mRNA expressions, and LIS_total were modestly correlated with ST₅₀ (r = 0.184, 0.156, −0.345 respectively, P < 0.05) but not to be of predictive value by Cox regression model. DSPC_BALF was a modest, though statistically significant, risk factor, but it was not correlated with the survival time.

Table 3.

Multivariate Cox regression analysis for survival time in ventilated newborn rabbits

Factor	aHR	95% CI	P
GA, days
25	1 (ref)
26	0.770	0.523–1.135	0.186
27	0.585	0.368–0.928	0.023
28	0.333	0.188–0.851	0.017
Group
NDEX-C	1 (ref)
NDEX-S	0.863	0.636–1.170	0.343
DEX-C	0.548	0.367–0.818	0.003
DEX-S	0.545	0.362–0.820	0.004
PTX	3.287	2.511–4.307	<0.001
Mean Cdyn	0.204	0.064–0.653	0.007
W/D	1.144	1.029–1.272	0.013
LIStotal	0.974	0.911–1.051	0.469
DSPCBALF	1.020	1.000–1.041	0.046

P values represent statistical significance by multivariate Cox regression analysis. aHR, adjusted hazard ratio; CI, confidence interval; DEX-C, dexamethasone-sham treated-control; DEX-S, dexamethasone-surfactant only; DSPC_BALF, disaturated phosphatidylcholine in bronchoalveolar lavage fluid; GA, gestational age; LIS_total, total score of lung injury; mean Cdyn, the average Cdyn during the whole period of ventilation, for original data see Supplemental Table S4; NDEX-C, control with no dexamethasone or surfactant; NDEX-S, control with surfactant only; PTX, pneumothorax; W/D, wet-to-dry lung weight ratio.

DISCUSSION

This study revealed the feasibility of ventilated preterm rabbit model by simulating perinatal therapies for longer survival as determined by ST₅₀, accelerated lung maturation and lung protective effects against injury from immaturity. The lung development in GA 25–26 days should be at the early saccular stage as verified by BW, WW, Vv, lung histology, TPL, DSPC, SP and CCT mRNA expression, and Cdyn in reference to that of the newborn rabbits from GA 27–28 days. The GA-associated lung immaturity was also corroborated by a high prevalence of PTX and early death in the GA 25 days. In contrast, fetuses delivered at GA 27–28 days had a much longer ST₅₀, i.e., from 3 to ≥10 h, with fewer or no PTX, or early deaths, suggesting moderate-to-late preterm for maturation by the same data file variables. For comparison among treatment groups, the ST₅₀ of NDEX-C and the phospholipid analysis, in each GA category including the NDEX-C0 subgroup, served as the baseline for fetal lung maturation.

The ventilated animal model was used in animal studies testing therapeutic efficacy of antenatal glucocorticoids and/or postnatal surfactant for understanding various underlying mechanism of perinatal and neonatal respiratory therapies at birth transition. These studies were mainly performed in moderate-to-late preterm and near-term rabbits (GA 27–30 days) corresponding to late saccular and early alveolar stages of lung development (20, 22, 24–26, 30). By tailoring the Vt to GA and comparing the four treatment modalities as perinatal therapies, we managed to have reduced prevalence of pneumothorax and early death due to immaturity. The response to antenatal DEX in EPT fetuses appears to be complex as evidenced by decreases in BW and WW, but improved lung mechanics and endogenous phospholipid pools as favorable responses to the perinatal therapies. Although both BW and WW of fetal lungs declined markedly after DEX, WW/BW remained >2% and alveolar aeration (Vv) appropriate for GA maturation and morphometry. Notably, Cdyn, W/D, Vv, and LIS_total did not show any deterioration in the DEX exposed groups (Supplemental Tables S3, S4, and S5), suggesting that DEX did not exert adverse effects on fetal lung structure and function despite its intrauterine growth-limit potential. We used dexamethasone rather than betamethasone as dexamethasone is widely used in countries where betamethasone is not available (38). Notably, the differences in clinical outcomes between betamethasone and DEX are only minor (39). The DEX and surfactant combination was most effective at GA 25–26 days, which is in concordance with the clinical findings in human EPT newborns (12). Further effort is needed to standardize the DEX dosing and timing to minimize the observed effect on growth restriction. As the overall effects by antenatal glucocorticoids and surfactant were marginal in GA 28 days, unless for specific purpose, the ventilated preterm rabbit model at this GA may not require these therapies to enhance lung maturation and survival.

The measurements of surfactant phospholipid pools in both alveolar and lung tissue compartments provided relevant information on changes in both exogenous and endogenous surfactant phospholipids over the GA and the ventilation time course, as predicted by the previous studies (23, 26). In general, the content of TPL (BALF + LH) and DSPC/TPL was higher in DEX-C0 subgroups compared with NDEX-C0 subgroups of the same GA, which was also true for the advancing GA. This implied the effects of DEX on accelerated synthesis of endogenous phospholipids including surfactant although the magnitude of phospholipid increment by DEX diminished with the advancing GA (Supplemental Table S6). In the ventilated groups, two DEX-exposed groups (DEX-C and DEX-S) had higher DSPC content in LH than NDEX groups of the same GA, particularly in GA 25–26 days (Supplemental Table S7). In contrast, the levels of TPL and DSPC in BALF in the GA 27–28 day groups were lower in DEX-treated subgroups than in the NDEX counterparts (Table 2). This discrepancy may involve complex mechanisms of exogenous surfactant phospholipid reutilization by the alveolar epithelial cells, as well as the variation in the length of ventilation between different experimental groups (40). A higher amount of DSPC_BALF was consistently associated with improvement of Cdyn and ST₅₀ across the GA studied (Figs. 1 and 2), which is known for surfactant synthesis enhancement by antenatal glucocorticoids in preterm fetal lungs (25, 41).

Although we did not directly measure the synthesis and/or the clearance of surfactant phospholipids, there was a trend toward a decline in the mRNA expression of CCT, a time-limit enzyme in PC synthesis, in those of GA 25–27 days, exposed to DEX in the ventilated groups (Table S8). The differences in mRNA expression of SP-A, SP-B, SP-C, and CCT between the nonventilated and ventilated groups, with or without DEX, (Table 2, Supplemental Tables S6, S7, S8, and S9), implied enhanced impacts of DEX and ventilation on the endogenous surfactant system. This is in accordance with the findings in very preterm rat and lamb models, in which antenatal glucocorticoids enhance SP mRNA expression in the animal lungs (42, 43). As a result, the responses to DEX and surfactant, either alone or in combination, as well as the variable SP and CCT mRNA expressions, may also be extrapolated in the GA 27 and 28 days (Figs. 1, 2, and Supplemental Fig. S1) respectively, for estimation of relative lung maturation over the GA range in this model. It suggests GA-specific differential effects of the perinatal therapies for the lung maturation and protection against injury.

The measurement of antioxidants in the lung tissue did not reveal any substantial changes across the GA and interventions studied. Neither did we identify any substantial changes in mRNA expression of IL-6, IL-8, and growth factors among different groups. These data imply that $\text{F}{\text{I}}_{{\text{O}}_2}$ (1.0) used with the intention to minimize hypoxemic stress over the relatively short ventilation periods employed in this study did not result in any noticeable oxygen exposure-related adverse effects. However, adverse effect of long-term hyperoxic exposure on survival and lung injury may not be ruled out.

In the multivariate Cox regression analysis, GA, DEX, surfactant, and the mean Cdyn over time were protective for survival, whereas PTX was a risk factor (Table 3). Likewise, since the survival length, i.e., ST₅₀, is GA-dependent, it was hard to adjust for the effects of physiological, histopathological, and biochemical parameters across different gestational ages and intervention categories simultaneously. The BW was also a protective factor; however, since it also correlated with GA (r = 0.691, P < 0.001), it was not included in the multivariate Cox regression model for survival.

There are several limitations in this study. Although we tailored Vt on GA, and on an individual basis, the range of survival time and the prevalence of pneumothorax in GA 25–26 days remained highly variable. A deviation to a smaller Vt from the target in this population was also observed. Despite smaller Vt was provided, some EPT rabbits were intolerant to the ventilation, presumably as a result of impaired gas exchange with hypoxemia and hypercapnia. Therefore, for specific study purpose using this model with EPT animals, surviving status needs to be well defined and measured. Other efforts are needed to optimize ventilation strategies by alternative ventilator settings, fluid balance and nutrition. Besides, the small body size limited blood gas measurement, the intraperitoneal fluid administration to maintain homeostasis was empirical, which leave room for further improvement. Moreover, an electrocardiogram may be applied in future to monitor cardiovascular function and validate survival status. Considering the frequent PTX occurrence in those of GA 25–27 days, no extended pressure was applied to the airways during lung fixation, which exerted modest impact on lung morphometry among the groups (27, 30). Due to a relatively short pregnancy period in rabbits, the two-day DEX injection interval before delivery might have magnified the side-effects of DEX (19) such as lower BW and WW. It is difficult to directly correlate the developmental stages of fetal lung maturity by GA in preterm rabbits with that of the gestational week in human fetuses. Last, the preterm birth in this experiment was not inducted but arbitrarily terminated according to the protocol, which was different from the real clinical situation. Our results, nevertheless, provided a comprehensive profile in the assessment of preterm lung maturation and response to the perinatal therapeutic modalities against extremely immaturity-associated lung injury. Being a progression over the previous ones, the usefulness of this model needs to be validated in future investigation of pathogenesis, pathophysiology, and mechanism of other pulmonary and nonpulmonary morbidities, and therapies, at birth transition.

In conclusion, the ventilated newborn rabbit model at GA 25–26 days is feasible to investigate extreme and very preterm lung maturation and injury severity. It facilitated a relatively longer survival time using the perinatal therapeutic modalities, though technical issues remain to be solved, such as optimal tidal volume and adequate blood gas analysis adjusted on the basis of GA and treatment modalities. It is warranted to use this model for study of perinatal and neonatal morbidity and mortality at preterm birth transition.

GRANTS

This study is supported by grants from the National Natural Science Foundation (No. 81501288, to Y. Dong), Shanghai Municipal Commission of Health (to Y. Dong), and international travel grants from the Laboratory of Neonatal Diseases, National Commission of Health, and Children’s Hospital of Fudan University (to V. K. Rehan).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

S.L. and B.S. conceived and designed research; S.L., X.G., and Y.X. performed experiments; S.L., X.G., and B.S. analyzed data; S.L., X.G., Y.X., Y.D., and B.S. interpreted results of experiments; S.L. prepared figures; S.L. and X.G. drafted manuscript; S.L., X.G., Y.D., V.K.R., and B.S. edited and revised manuscript; S.L., X.G., Y.X., Y.D., V.K.R., and B.S. approved final version of manuscript.