Sex differences in impacts of early gestational and peri-adolescent ozone exposure on lung development in rats: Implications for later life disease in humans

Abstract

Air pollution exposure during pregnancy may affect fetal growth, and fetal growth restriction (FGR) is associated with reduced lung function in children that can persist into adulthood. Using an established model of asymmetrical FGR in Long-Evans rats, this study investigated sex differences in effects of early life ozone exposure on lung development and maturation. Adverse health effects for (1) gestational exposure (with impacts on primary alveolarization), (2) peri-adolescent exposure (with impacts on secondary alveolarization), and (3) cumulative exposure across both periods were evaluated. Notably, female offspring were most affected by gestational ozone exposure, likely due to impaired angiogenesis and corresponding decreases in primary alveolarization. Females had diminished lung capacity, fewer mature alveoli, and medial hypertrophy of small and large pulmonary arteries. Males, especially FGR-prone offspring, were more affected by peri-adolescent ozone exposure. Males had increased ductal areas likely due to disrupted secondary alveolarization. Altered lung development may increase risk of developing diseases such as pulmonary arterial hypertension (PAH) or chronic obstructive pulmonary disease (COPD). PAH disproportionately affects women. In the U.S., COPD prevalence is increasing, especially in women; and prevalence for both men and women is highest in urbanized areas. This investigation underscores the importance of evaluating results separately by sex, and provides biologic plausibility for later consequences of early life exposure to ozone, a ubiquitous urban air pollutant.

Introduction

Epidemiologic studies provide evidence that associations between air pollution and respiratory health outcomes are significantly modified by sex. In women, for example, exposure to ozone and particulate matter (PM) was associated with a higher risk of all-cause and cardio-respiratory mortality (approximately twice that of men).¹ Similarly, Shin and colleagues examined mortality attributable to air pollution exposure in Canadian populations and showed that women had higher mortality risk for ozone exposure,² higher mortality risk for circulatory disease associated with warm season ozone exposure,³ and overall, women exhibited higher mortality risk compared to men.⁴

There are also clear sex-based differences in lung disease incidence and clinical phenotype.⁵ Furthermore, the prevalence, morbidity, and mortality of lung diseases such as asthma, chronic bronchitis, and chronic obstructive pulmonary disease (COPD) appear to be increasing in women.^{6, 7} The global COPD burden attributable to air pollution (i.e., PM, ozone) and high temperature exposure also show increasing trends, especially in lower socioeconomic regions.⁸ Although COPD affects both men and women, on a global basis COPD prevalence is increasing more rapidly in women, particularly younger women.⁹ As of 2019 and prior to the COVID pandemic, worldwide COPD was estimated by the WHO to be the 3^rd leading cause of death despite significant underreporting.¹⁰ In the U.S., COPD-related hospitalizations and deaths in women now surpass those in men.⁹ Now affecting > 7 million women in the U.S., the prevalence of COPD for U.S. men and women is highest in urban areas.^{5, 7}

Concordant evidence that air pollutant exposure increases the prevalence and severity of existing pulmonary conditions is provided by epidemiologic, controlled human clinical, and laboratory animal investigations.^{11, 12, 13, 14, 15} However, more research will be useful to define the role that poor air quality may play in causing detrimental effects to health and thus increasing lung disease incidence, especially in vulnerable populations. Two critical windows of susceptibility for exposure to environmental agents are pregnancy¹⁶ and early (post-natal) life.¹⁷ Consistent with the Developmental Origins of Health and Disease (DOHaD) concept, the early life environment is a key period for fetal epigenetic programming.¹⁸ Therefore, exposure to environmental agents (e.g., air pollution) or other maternal stressors during pregnancy may contribute disproportionately to risk of disease later in life.¹⁹

The number of epidemiological investigations on air pollutant exposure during pregnancy has been steadily increasing. Within pregnancy, the beginning of gestation was identified as a potentially sensitive exposure period based on associations with epigenetic change in children (e.g., global and locus-specific DNA methylation, shortened telomere length) (systematic review of 32 publications).²⁰ In other reports, air pollution and environmental exposures were associated with a range of adverse pregnancy outcomes including increased uterine vascular resistance, altered placental vascularization, preterm birth, and intrauterine or fetal growth restriction (FGR).^{21, 22}

In turn, FGR is often associated with impaired lung function in children²³ that can persist into adulthood.²⁴ Among multiple antenatal factors, FGR was the only maternal or prenatal characteristic that was highly predictive of chronic lung disease risk.²⁵ In another report, FGR was found to be worse than extreme prematurity for effects on the developing lung.²⁶ Although underlying mechanisms are not fully understood, data in humans^{27, 28} and rodents²⁹ show that FGR is also associated with increased risk of developing asthma or asthma-like lung changes, respectively. Other lung diseases associated with FGR include bronchopulmonary dysplasia (BPD), pulmonary arterial hypertension (PAH), and COPD and moreover, preterm infants affected by FGR have a higher incidence of both BPD³⁰ and PAH.^{31, 32}

In the seminal report by Barker et al,²⁴ lower body weight at birth and the first year of life was associated with worse lung function in adulthood; and later in life, death from COPD. Barker et al postulated that the same intrauterine influences which served to limit fetal weight gain, may have irrecoverably constrained lung and airway growth. More recent evidence similarly suggests that COPD is not simply a disease of older heavy smokers but may instead be associated with insults occurring during fetal life and postnatally when lung growth and development are still changing rapidly.^{33, 34} As reviewed by Wang et al,³⁵ new data further indicate that the trajectories of lung function impairment and airway remodeling in asthma often precede symptom onset and may be present at birth.

There is, therefore, an urgent need to better understand and elucidate mechanisms by which exposure to air pollution early in life may alter lung development. To this end, we have established a rat model of FGR induced by ozone exposure early in gestation.^{36, 37, 38} This model used a 2-day ozone exposure during implantation [gestation day (GD) 5-6] to induce late-term (GD21) growth-restricted fetuses. Male fetuses exhibited FGR when dams were exposed to either 0.4 or 0.8 ppm ozone, whereas female fetuses were more consistently affected by dam exposure to 0.8 ppm ozone.³⁶ Despite having reduced body weight, affected fetuses had normal body length, indicative of asymmetrical FGR and late pregnancy utero-placental insufficiency. Asymmetrical FGR is the most common form of FGR in humans.³⁹ Additionally, fetal sex-based differences in placental efficiency and metabolism were observed in this model. Furthermore, female and male growth-restricted fetuses exhibited very different patterns of brain and hepatic gene expression change.⁴⁰

Using this model, in the present study, offspring from dams exposed to air or to ozone in early gestation were then exposed to air or ozone for three (3) additional times during peri-adolescence. Offspring from the ozone-exposed dams are referred to herein as FGR-prone offspring. With this approach, effects of early life ozone exposure during two key periods of lung development and maturation in rats could be assessed: (1) gestation and neonatal life with potential impacts on primary or bulk alveolarization and micro-vascularization processes and (2) peri-adolescence with possible effects on secondary alveolarization processes.^{41, 42}

We hypothesized that FGR-prone offspring would be at increased risk for reduced lung growth and maturation. Secondly, owing to the sex differences in placental and fetal effects observed, we hypothesized that female and male FGR-prone offspring would exhibit unique lung developmental pathologies related to FGR. Thirdly, because factors contributing to FGR can modify lung development making it more vulnerable to postnatal insults,⁴³ we predicted that FGR-prone offspring would incur greater pulmonary effects due to repeated ozone exposure during peri-adolescence. After the last peri-adolescent exposure, offspring were euthanized and differences in lung capacity and maturation were evaluated based on lung volume and area assessments, airspace and vascular morphometry, and vascular gene expression.

Materials and Methods

Rat model of fetal growth restriction and study approval.

Three cohorts of timed-pregnant Long-Evans rats (11-weeks-of-age) from Charles River Laboratories Inc. (Raleigh, NC) were transferred to the U.S. EPA’s AAALAC-approved facility on GD1 (plug-positive day). Dams were individually housed in polycarbonate cages and maintained on a 12h light/dark cycle at ≅ 22°C and 50% relative humidity. Dams were randomized by weight into two groups (air or ozone; n = 12/exposure). Because previous studies showed that 0.8 ppm ozone consistently induced FGR in both female and male GD21 fetuses, this concentration was also used for the present study. All dams received a modified, phytoestrogen-free American Institute of Nutrition growth and lactation diet (D15092401; Research Diets) and water ad libitum throughout pregnancy and lactation. This diet was used because standard rat chow may include plant-based protein sources containing phytoestrogens (natural compounds that behave like estrogen). Such compounds have been shown to impact pregnancy outcomes. The Institutional Animal Care and Use Committee of the U.S. EPA approved all procedures involving animals.

Exposure groups and timeline of gestational and peri-adolescent ozone exposures.

Based on this 2 x 2 factor study design (with one factor for the “dam” peri-implantation ozone exposure and the other for “pup” peri-adolescent ozone exposure), there were 4 groups of female offspring and four groups of males. Female (F) groups were abbreviated according to their sex, dam, and pup exposure to air (A) or ozone (O₃) as: F A:Ax3, F O₃:Ax3, F A:O₃x3, and F O₃:O₃x3. Male (M) groups were similarly described. As shown in Figure 1, pregnant rats were exposed on GD5-6 to filtered air or 0.8 ppm ozone for 4h (7:00–11:00 am) in Rochester-style Hinners chambers controlled for temperature (≅ 22.5°C) and humidity (≅ 50%). On postnatal day (PND) 4, litter size was standardized to 10 pups/litter inclusive of 5 females and 5 males whenever possible. On PND 19, offspring were weaned, pair-housed and fed a standard growth diet (Purina 5001, Brentwood, MO) ad libitum. To minimize offspring size differences across this prolonged investigation, the two smallest females and two smallest males in each litter were selected for this study. One female and male per litter were randomly assigned to the peri-adolescent air group and the other to the ozone group. Then, offspring were exposed either to air or to ozone for 4h (7:00–11:00 am) once/week, across PND33-47, for a total of 3 exposures (Fig. 1). By the first peri-adolescent exposure, offspring had been weaned for only 2 weeks, and thus they were initially exposed to 0.4 ppm ozone. When the 0.4 ppm exposures appeared to be sufficiently tolerated (based on visual examination during and after the exposures), 0.8 ppm ozone concentrations were used for their second and third exposures. Offspring body weights were obtained weekly just prior to exposure. No food or water were available during the exposures. Ozone was generated by a silent arc discharge generator (OREC^™) delivered to the chambers using a mass flow controller. Throughout exposures, ozone concentrations were monitored in real-time and recorded using an ozone analyzer (API Model 400; Teledyne Instruments).

Figure 1.

Timeline of ozone exposures and key life events. GD, gestational day; PND, postnatal day.

Necropsy and data exclusion criteria.

Within 4h after the 3^rd and final exposure, all offspring were euthanized with sodium pentobarbital (Fatal Plus, Virbac AH, Inc., Fort Worth, TX; >200 mg/kg, intraperitoneally) combined with exsanguination via the abdominal aorta to minimize retention of blood within the lungs. Body length was measured from the nuchal crest of the skull to the base of the tail. Due to maternal neglect by one ozone-exposed dam, there was early post-natal loss of the litter on PND3. Thus, there was one less female and male in the ozone dam groups (n = 11) compared to the air dam groups (n = 12). Group sizes indicated in figure legends and tables may be less due to data exclusion. Reasons for exclusion included technical issues with sample quality or processing (e.g., visual leakage of lung lobes during fixation, flipping of the lung section in processing, or poor RNA quality). Necropsy findings revealed that one F A:Ax3 rat had a large hydronephrotic kidney necessitating its exclusion. Lung data were excluded from one F O₃:Ax3 and one M O₃:Ax3 rat because they were inadvertently exposed to ozone on their 3^rd peri-adolescent exposure. Lastly, one F O₃:Ax3 rat had a visibly enlarged heart at necropsy. Based on outlier testing, it was necessary to exclude its data from F O₃:Ax3 group comparisons, but for select end points, her data are presented separately.

Left lung fixation, volume displacement, and histology.

Using a split lung approach as previously described,⁴⁴ the left lobe was collected and inflated with 10% buffered formalin at 25 cm H₂O pressure⁴⁵ for 10 minutes. The mainstem bronchus was ligated to ensure the lobe remained expanded and then the lobe was submerged in formalin overnight. The following day, lung lobe volumes were determined using the displacement method described by Scherle⁴⁶ (n = 9-12 per group). Formalin fixed lobes were then maintained at 4°C until further processing. A subset (n = 7-9 per group) was used to prepare three (3) transversely transected lung sections at the approximate branching for the 5^th (AW5) and 8^th (AW8) generation airways, and mid-way between AW8 and the lung tip. Sections were paraffin embedded, and 5 μm thick sections were stained with hematoxylin and eosin (H&E). A subset of lung slides (n = 6/group) was also stained with the collagen stain trichrome to aid in vascular assessments and with Alcian Blue/Periodic Acid Schiff (PAS) to assess for mucus-producing cells within large airways (Experimental Pathology Labs, Durham, NC). Stained slides were examined using a bright field microscope (Nikon Diaphot 300). Lung sections were further imaged with an Aperio AT2 high capacity scanner (Leica Biosystems, Buffalo Grove, IL).

Left lung airspace morphometric assessments.

The AW8 transverse lung sections were used for all airspace morphometry. The area of each H&E-stained AW8 lung section was determined by manually tracing its outer perimeter (Aperio ImageScope, 12.4.6, 64-bit OS, software). Within the AW8 section, lines were drawn across the width and length as guides for consistent placement of rectangular areas of interest within the (A) dorsal, (B) lateral, and (C) ventral lung (Fig. 2A). Based on a method for characterizing airspace size using intercept (chord) measurements,⁴⁷ five equidistant guard lines (each 850 μm in length) were drawn within each rectangle. In a blinded manner, structures crossing the guard lines were visually identified as alveoli (A) for smaller closed or nearly closed circular spaces or as conducting or “ductal” (D) spaces for larger, more irregularly shaped spaces (inclusive of alveolar sacs) (Fig. 2A). The number and chord distance across each type of airspace was quantified. Vessels and small airways crossing the guard lines were likewise assessed. Because it was not possible to identify very small (< 10 μm) airspaces, they were defined as “other” space. The combined A, B, and C rectangular regions represented 12,750 μm of total linear space within the AW8 lung section (3 regions x 5 guard lines x 850 μm per line). To account for edge effects wherein the guard line ended within an airspace, if the remaining distance was < 30 μm, it was attributed to the alveolar space. This approach allowed for inclusion of the total 12,750 μm of linear distance as well as identification of most structures present (totaling ≅ 200 - 250 alveoli-like spaces and 30 - 40 ductal spaces per lung). Each rectangle represented 730 μm x 900 μm = 660,000 μm², thus the total morphometric area assessed ≅ 2 x 10⁶ μm². This data, combined with the total area of the AW8 section, allowed for pro-rating of the number of structures defined with the combined A-to-C regions to that of the AW8 transverse section overall. This approach corrected, in part, for differences in lung shrinkage occurring during inflation, fixation, or staining steps.

Figure 2A-C.

For airspace morphometry, rectangular areas within dorsal (A), lateral (B), and ventral (C) regions were assessed. Within each rectangle, structures crossing guard lines were visually identified as alveolar (A) or ductal (D) space, and the number and chord length of the space is quantified (Fig. 2A, scale bar = 200 μm). Schematic of the small and medium-sized vessel medial wall thickness (MWT%) based on inner diameter (ID) and outer diameter (OD) measurements (Fig. 2B). Schematic of large vessel length and wall thickness (insert) for the central airway (AW) and pulmonary artery (PA) and pulmonary vein (PV) (Fig. 2C, scale bar = 400 μm).

Additional airspace assessments included a modified mean linear intercept (MLI_mod) obtained by: [12,750 μm – (total length of vessels + airways + “other”)] / [(# of alveoli identified) + (# of ducts identified)]. The mean alveolar width was estimated by: [total length of alveolar space/by the total number of alveoli identified] across the 12,750 μm of linear space. The mean ductal width was likewise quantified. Finally, by using the mean alveolar or ductal widths, and assuming each airspace represented a somewhat simplified circular or spherical structure, estimates of the corresponding airspace diameter, radius, area, and volume were calculated. Estimates were again prorated to the area of the AW8 section, providing assessment of the total number of alveolar or ductal airspaces within the lung section, and corresponding areas or volumes. Because the chord-based measurements do not represent the widest distance across the airspaces, the mean values presented are underestimates of their actual size and are meant only to define relative differences between exposure groups.

Lung small vessel morphometry.

Small arteries were identified based on their proximity to small airways using the trichrome-stained slides (n = 6 subjects/exposure group). The Medial Wall Thickness percentage % (MWT%) of smaller vessels was obtained based on the technique described by Bombicz et al.⁴⁸ In brief, vessels were binned into relatively small-sized (<125 μm), and medium-sized (125-350 μm) arteries based on their maximal width. The internal diameter (ID) and outer diameter (OD) distance across the narrowest part of the vessel were measured (Fig. 2B). The MWT% was calculated by the ratio of the [(OD - ID)/OD] x 100%. Results represent an average of 4-6 small and 4-6 medium-sized vessels for each subject. It was necessary to utilize all three lung sections for each rat to locate enough vessels for this assessment.

Lung large vessel morphometry.

Using both the AW5 and AW8 transverse lung sections, trichrome-stained slides were again utilized to identify the main pulmonary artery (stained in blue) and pulmonary vein (in red) located adjacent to the large central airway (n = 6 subjects/exposure group) (Fig. 2C). The approximate length of these structures was determined. For the pulmonary artery, the thickness of the intimal (inner) + medial (middle) layer (appearing lighter blue intermixed with purple on the trichrome-stained slides) and the adventitial (outer) layer (appearing darker blue) were measured (averaging 3 – 5 measurements across the large vessel), avoiding areas of oblique sectioning. For the pulmonary vein, the thickness of the deep red muscular layer was also measured (see inserts within Fig. 2C).

Lung protein content.

The right caudal lung lobe was snap frozen and stored at −80°C until analysis. In brief, the lung tissue was minced and homogenized in 1X radioimmunoprecipitation assay (RIPA) lysis buffer (Cell Signaling; Danvers, MA) containing protease and phosphatase inhibitors (Thermo Fisher; Waltham, MA). Homogenates were then centrifuged at 10,000 x g for 20 min. at 4°C to pellet cell debris. The protein content of the supernatants was measured using the DC Protein Assay (Bio-Rad; Hercules, CA) according to the manufacturer’s instructions.

Lung gene expression.

The right accessory lung lobe was snap frozen in liquid nitrogen and stored at −80°C until further processing. Total RNA was isolated from the lobe using the Direct-Zol RNA MiniPrep kit from Zymo Research (Irvine, CA). The Qubit RNA reagent (Life Technologies; Carlsbad, CA) was used to determine RNA quantity, and cDNA was generated from 1000 ng of RNA isolate using the QScript cDNA Supermix (Quanta Biosciences; Beverly, MA). Primers for selected genes were designed using published sequences or the National Center for Biotechnology Information database and Integrated DNA Technologies, Inc. (IDT, Inc; Coralville, IA) (Table 1). Quantitative real-time polymerase chain reactions (qRT-PCR) were performed in duplicate using Sybr Green Master Mix (Life Technologies; Carlsbad, CA) on a QuantStudio 7 Flex Real-Time PCR Instrument. Endogenous control genes were tested across all groups for sex or exposure influences, and the best control (Rps15a) was selected. For each gene assessed, the delta delta Ct method⁴⁹ was used to determine relative quantification (RQ) of change from the respective air control group.

Table 1.

qRT-PCR primer sequences designed for target genes.

Primer	Forward Sequence	Reverse Sequence
Rps15α	5’-AGGTTGAACAAGTGTGGAGTTA-3’	5’-GAAACCAAACTGCCGTGATG-3’
Angpt1	5’-GACACCTTGAAGGAGGAGAAAG-3’	5’-GTTGTTGGTAGCTCTGCTAAGT-3’
Dusp1	5’-TGGTCTGCCCTCACAAATG-3’	5’-GCCTGCTCTGGGTCTATTTAC-3’
Et-1	5’-GAACATCTGTCCGGCTTCTAC-3’	5’-GGAACACCTCAACCTCTCTTG-3’
Hif-1α	5’-GAAGTTAGAGTCAAGCCCAGAG-3’	5’-CTCAGGTGAGCTTTGTCTAGTG-3’
Nos3	5’-TCCCAGCTGTGTCCAATATG-3’	5’-CCCTCATGCCAATCTCTGAA-3’
Pecam-1	5’-CCCAGTGACATTCACAGACA-3’	5’-ACCTTGACCCTCAGGATCTC-3’
Vegfa	5’-GCTCCTTCACTCCCTCAAATTA-3’	5’-GGTCTCTCTCTCTCTCTCTCTTC-3’
Vegfr1	5’-TACGTCACAGATGTGCCAAAC-3’	5’-GCAGTGCTCACCTCTAACGA-3’
Vegfr2	5’-GACGACCCATTGAGTCCAATTA-3’	5’-GTGAGGATGACCGTGTAGTTTC-3’

Statistics.

GraphPad Prism (6.07) was used to create graphs and to perform statistical comparisons. Data are expressed as the mean ± SEM. Unless otherwise noted, treatment effect differences were determined separately for female and male offspring. Group differences were assessed using a two-way analysis of variance (ANOVA). If significant dam exposure, pup exposure, or interaction effects were identified, differences between individual ozone-exposed groups were compared only to the corresponding A:Ax3 group (p ≤ 0.05). These comparisons were assessed by ANOVA testing with post hoc correction using the Dunnett’s multiple comparisons test. For serial assessments such as body weight, 2-way ANOVA repeated measures (RM) with the Dunnett’s multiple comparisons test adjustments were used (p ≤ 0.05). Then, 2-way ANOVA RM with the Holm-Sidak’s test adjustments were used to compare female and male peri-adolescent air-exposed groups only (p ≤ 0.05). For gene expression data, if data were not normally distributed based on Shapiro-Wilk testing, outliers identified via the Grubbs’ test (alpha = 0.05 %) were excluded. If data remained non-parametric, data log transformation was used. For a limited set of variables, Pearson one-tailed correlation coefficients were used to assess strength of data associations across select groups.

Results

Dam health and offspring growth.

As in previous studies with this FGR model, except for the early post-natal litter loss in one of the ozone-exposed dams herein, no significant ozone exposure effects were observed in dams for maternal weight gain, the number of pregnancies or births, the number of pups per litter, or litter weights shortly after birth (unpublished data). Onset of puberty in female offspring ranged from PND27–37 (based on vaginal opening) and in male offspring, from PND43-48 (based on preputial separation) (Fig. 1).

Based on body weights just prior to each peri-adolescent exposure, all offspring were experiencing rapid growth during this period (from 5 to 7 weeks-of-age). Accordingly, ANOVA RM assessments showed highly significant pup exposure effects (due to increasing age and maturity) across this period. Weights in females (Fig. 3A) increased by ≅ 50% and males nearly doubled their body weight (Fig. 3B). No significant effect of dam ozone exposure on body weight was observed. However, owing to a significant interaction effect (especially in males), we further examined the influence of dam exposure just within the peri-adolescent air-exposed groups. Body weight in males appeared to still be increasing linearly during this period, whereas the growth trajectory in females had already begun to plateau (Fig. 3C). Data showed that males weighed significantly more than females by the second and third but not the first exposure. By the third exposure, the O₃:Ax3 males weighed significantly more than A:Ax3 males, whereas no weight differences were apparent between the F A:Ax3 and F O₃:Ax3 groups (Fig. 3C).

Figure 3A-E.

Corresponding offspring group means (± SEM) values for body weight (gm) in females (Fig. 3A, n = 9-12/group) and males (Fig. 3B, n = 10-12/group) prior to the 1^st, 2^nd, and 3^rd peri-adolescent air or ozone exposure. Comparison of A:Ax3 to A:O₃x3 group-only body weights for both sexes (Fig. 3C). Comparison of final body length (cm) in males (Fig. 3D) and body mass index (kg/M²) in females (Fig. 3E). For Significant difference from the corresponding control group (*p<0.05, ***p<0.001, ****p<0.0001). In Fig. 3C, Sig. difference from females of the same age (****p ≤ 0.0001) and significant difference from M A:Ax3 males (λ p ≤ 0.05).

Measurements of offspring body length, a surrogate for height, were obtained because in human infants and toddlers, lung volumes increase linearly with body length⁵⁰ and in adults, lung volume and capacity are adjusted by age, sex, and height.⁵¹ A significant dam exposure effect was observed for body length in males (Fig. 3D) but not in females (unpublished data). Body weight and length measurements were further used to calculate body mass index (or BMI; kg/m²) (Fig. 3F). Data revealed a significant effect of peri-adolescent (i.e., pup) ozone exposure on BMI in females, indicative of reduced weight and/or adiposity relative to height. By the third exposure, male offspring of ozone-exposed dams appeared to have proportionate increases in height and to a lesser degree weight -- and thus were seemingly bigger overall but did not show differences in BMI (unpublished data). One would predict, therefore, that lung volume or ventilatory capacity in these males would be proportionately greater as well.

Left lung displacement volume.

To assess whether dam or pup ozone exposure affected lung capacity in offspring, left lung lobes were fixed under defined inflation pressure and lung displacement volume measurements were obtained. Clinically in humans, lung volume or capacity are adjusted by age, sex, and height.⁵¹ However, in rodent toxicology studies, if body size is variable, organ size is typically adjusted by body weight prior to assessing for treatment effects. Because the male offspring were much larger than females by 7 weeks-of-age, and because some O₃:Ax3 and O₃:O₃x3 male offspring were considerably larger than others, we compared two methods of normalizing lung volume to body size – one based on weight, and one based on height. Firstly, assuming total lung capacity (TLC) in rats ≅ 28 mL/kg and 40% of TLC is related to the left lung lobe,⁵² lung displacement volumes obtained for the left lobe alone (40%) were prorated to estimate TLC (100%). This prorated lung volume was then compared to that predicted by the weight-based, 28 mL/kg, estimate (Fig. 4A). Secondly, because lung capacity in humans is adjusted by sex and height, we used a similar approach⁵³ to calculate “height” adjustment factors for rats, by sex, based on the A:Ax3 group data [i.e., Factor = the air group mean lung displacement volume prorated to TLC (mL) ÷ air group mean body length (cm)]. Resultant factors for females (0.399) and males (0.386) were then used to predict lung volumes for offspring in the other exposure groups (Fig. 4A).

Figure 4A-C.

Left lung lobe displacement volumes prorated to TLC are depicted, including volumes predicted by body weight and by body height-adjustment factors (Fig. 4A). Corresponding group means (± SEM) values of the pro-rated total lung displacement volume adjusted by height for females (Fig. 4B, n= 9-12) and males (Fig. 4C, n = 10-12). Significant difference from the corresponding control group (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p<0.0001).

Results showed that in control (A:Ax3) females, both the weight-based and height-based estimates closely predicted the actual prorated-to-TLC lung volumes (white bars). However, FGR-prone female offspring had significantly smaller lung volumes (pink solid bars, pink hatched bars) than would be predicted based on their body height. In control (A:Ax3) males (white bars), lung volume predicted by weight significantly overestimated both the actual volume, as well as the value predicted by height. Likewise for all the other male exposure groups, the weight-based adjusted volume overestimated the actual volume, whereas height-adjusted volumes closely matched the prorated-to-TLC volumes. Therefore, lung prorated-to-TLC volumes were adjusted by height prior to performing group comparisons. Accordingly, ANOVA testing in females revealed a significant dam exposure effect on lung volume adjusted by height ― with the F O₃:O₃x3 group being marginally smaller than the air control group (p = 0.08) (n = 9-12) (Fig. 4B). No exposure effects were apparent for the height-adjusted lung volumes in males (n = 10-12) (Fig. 4C).

Area of the left lung transverse section at AW8.

It was not clear why the weight-based estimates exceeded the actual lung volumes to such a large degree in males, nor why lung volumes of A:Ax3 males were only 4% larger than that of A:Ax3 females considering that their body weight and height measurements were 24% and 11% greater, respectively. Therefore, the lung volume metrics above were compared to the AW8 lung section area measurements (Fig. 5A, 5B). In females, ANOVA testing again revealed a significant dam exposure effect on the AW8 lung section area (mm²); with the F O₃:Ax3 group having a significantly smaller area than the air control group (n = 7-8) (Fig. 5C). In males, the area measurements obtained for the A:Ax3 group were once again similar to the A:Ax3 females, and no significant differences were apparent across the male ozone-exposure groups (n = 7-9) (Fig. 5D). Taken together, data suggest that relative to females, 7-week-old-males were in a more active phase of body and lung growth, and thus their lung development was somewhat delayed (or less mature) compared to females of the same age. Importantly, these data also suggest that lung capacity of FGR-prone female offspring was diminished.

Figure 5A-D.

Representative transverse AW8 lung sections are depicted for females (Fig. 5A) and males (Fig. 5B). Corresponding group mean (± SEM) values for the AW8 sectional areas (mm²) are depicted for females (Fig. 5C, n = 7-8) and males (Fig. 5D, n = 7-9). Significant difference from the corresponding A:Ax3 group by sex (*p ≤ 0.05, **p ≤ 0.01).

Lung histology and modified mean linear intercept.

On histologic examination, minimal lung inflammatory changes were present in the peri-adolescent ozone exposure groups (representative regional images depicted in Fig. 6A). However, in some F O₃:Ax3 offspring, multi-focal areas of increased heterogeneity in airspace size (i.e., emphysema-like changes) with increased septal thickening and increased mononuclear cellularity of the airspaces were observed (Fig. 6B). Such changes are not unlike that observed in humans with alveolar capillary dysplasia (ACD).⁵⁴ In some peri-adolescent ozone-exposed offspring, mild epithelial hyperplasia of the bronchoalveolar ductal zone was evident (Fig. 6C). In some O₃:O₃x3 offspring, notable heterogeneity in airspace size with knobby appearing septal tips (Fig. 6D) or thickened pulmonary vessels were observed (Fig. 6E). Based on Alcian Blue/PAS staining, no increase in mucus-producing cells within the epithelial layer of the central airway was apparent in any of the peri-adolescent ozone-exposed offspring (unpublished data).

Figure 6A-G.

Representative photomicrographs (Fig. 6A, Scale bar lines = 200 μm). 6B. Alveolar capillary dysplasia (ACD)-like changes are present with increased numerous of macrophages within airspaces (Scale bar = 200 μm), 6C. Mild bronchoalveolar ductal hyperplasia changes (Scale bar = 200 μm), 6D. Heterogeneous airspace size and thickened septal tips (black arrows) including an insert (Scale bar = 200 μm), and 6E. Cluster of thickened small vessels and thickened artery adjacent to a small airway (insert) (Scale bars = 200 μm). The mean (± SEM) of the modified mean linear intercept (AW8 lung section) for females (Fig. 6F, n = 7-8) and males (Fig. 6G, n = 7-9).

Next, a modified Mean Linear Intercept (MLI_mod) was used to assess relative differences in the dimensions of the distal airspaces. Larger MLI values may reflect a less mature state of lung development⁵⁵ and MLI increases are often used to indicate alveolar destruction secondary to chronic insults such as smoking.⁵⁶ Evaluation of the MLI_mod revealed that A:Ax3 males had a somewhat larger MLI_mod value (5.4%) compared to A:Ax3 females. Thus, data again suggest that lung maturation in control males was delayed compared to A:Ax3 females. Importantly, there was also a marginally significant pup exposure effect on MLI_mod in both females (p = 0.058) and males (p = 0.054) (Fig. 6F, 6G). Although differences in MLI cannot be directly interpreted as a change in alveolar size, results suggest that peri-adolescent ozone exposure disrupted lung maturation processes. To better assess the relative airspace occupied by mature alveoli, in the sections below we more closely examined the number and relative size of small, well-defined airspaces (i.e., alveoli) compared to the number and size of larger airspaces (i.e., ductal regions including alveolar sacs) across the AW8 lung sections.

Alveolar-related morphometrics.

During the later phases of lung development, larger airspaces are gradually subdivided into smaller airspaces through a complex and prolonged process of primary (or bulk) alveolarization.^{41, 42} We wanted to ascertain whether dam ozone exposure and resultant placental insufficiency may have disrupted these processes, thus limiting the number of mature alveoli present. Data in females showed a significant dam exposure effect on the number of alveoli pro-rated to the entire AW8 lung sectional area ― with the F O₃:Ax3 group (solid pink bars) having a significantly reduced number of alveoli; and F O₃:O₃x3 group (hatched pink bars) being marginally reduced (p = 0.07), relative to air controls (Fig. 7A). Next, assuming the mean chord width of the alveoli approximated the mean alveolar diameter, the mean alveolar area was calculated. This area was then multiplied by the number of alveoli in the AW8 lung section to estimate the total alveolar area of the AW8 section. ANOVA testing again revealed a significant dam exposure effect in females ― with the F O₃:Ax3 group having a marginally reduced (p = 0.08) alveolar area compared to air controls (Fig. 7C). By contrast, no exposure effect was detected in male offspring for these same alveolar metrics. Despite the greater body size and weight of the air control males, the mean number of alveoli (Fig. 7B) and alveolar area estimates (Fig. 7D) closely approximated that of A:Ax3 females. Results are consistent with gestational ozone exposure in some manner disrupting primary alveolarization processes in female, but not the male, FGR-prone offspring.

Figure 7A-D.

Airspace morphometrics related to alveolar number estimated for the AW8 lung section in females (Fig. 7A, n = 7-8) and males (Fig.7B, n = 7-9); with corresponding alveolar area estimations in females (Fig. 7C, n = 7-8) and males (Fig. 7D, n = 7-9). Group means (± SEM) are depicted. Significant difference from the corresponding A:Ax3 group (*p ≤ 0.05).

Ductal-related morphometrics and area ratios.

Ductal space metrics were evaluated next, with the prediction that any disruption of processes involved in subdividing larger airspaces into alveoli may have resulted in a greater number of ductal spaces or in a disproportionately larger ductal area within the AW8 lung section. No significant exposure effect was observed for the number of ductal spaces detected within the AW8 lung section in female (Fig. 8A) or male (Fig. 8B) offspring. It should be noted, however, that there was considerable variability in the ductal assessments, possibly owing to our ability to assess fewer ductal spaces within the AW8 section (i.e., only ≅ 1000 ducts total compared to ≅ 8000 alveoli). Nevertheless, data did reveal a significant pup exposure effect on the estimated total ductal area in both females (Fig. 8C) and males (Fig. 8D), with the M O₃:O₃x3 group (hatched blue bars) having a significantly increased ductal area compared to air control males (white bars).

Figure 8A-F.

Airspace morphometrics related to ductal number in the AW8 lung section in females (Fig. 8A, n = 7-8) and males (Fig. 8B, n = 7-9); corresponding ductal areas in females (Fig. 8C, n = 7-8) and males (Fig. 8D, males, n = 7-9); and ratio of the ductal area to alveolar area in females (Fig. 8E, n = 7-8) and males (Fig. 8F, n = 7-9). Group mean (± SEM) values are depicted. Significant difference from the corresponding A:Ax3 group (*p ≤ 0.05; **p ≤ 0.01).

The ratio of the ductal area to the alveolar area within each AW8 lung section was also assessed. The F A:Ax3 offspring had a ductal-to-alveolar area ratio of approximately 1.0 (Fig. 8E), much like that of the A:Ax3 males (Fig. 8F). Put another way, by area, nearly 50% of the airspace in air control offspring represented large airspaces that had yet to be subdivided into smaller alveoli. Data further revealed that in males, a significant pup exposure effect was again detected, and M O₃:O₃x3 males again had a significantly greater ductal-to-alveolar area ratio compared to air controls (Fig. 8F). Although no significant exposure effects were detected for the females (possibly due to the greater variability of this duct-based metric), relative to the air controls, all ozone-exposed female groups appeared to have some increase in this ratio. However, the only group exhibiting a significant increase in this ratio was the O₃:O₃x3 males.

Results thus far showed that the reduced lung volumes measured in FGR-prone females, in particular the F O₃:Ax3 group, were due to significant reduction in the number of mature alveoli present. In FGR-prone males, on the other hand, although lung volumes were not consistently reduced, peri-adolescent ozone-exposure appeared to increase the relative proportion of ductal space present. Hence, male offspring undergoing repeated ozone exposure during peri-adolescence appeared to incur greater disruption of secondary alveolarization processes, especially within the M O₃:O₃x3 group.

Smaller pulmonary vessel morphometry.

Experimental evidence implicates angiogenesis as being critical for primary alveolarization processes.⁵⁷ Therefore, to assess for corresponding vascular impairment in FGR-prone females, the medial wall thickness (MWT %) of small vessels (presumed to be arteries by their proximity to small airways) was quantified. Results of ANOVA testing in females again showed a highly significant dam exposure effect on the MWT %, with the F O₃:Ax3 group having a significantly increased MWT % for both small-sized (< 125 μm in length, Fig. 9A) and medium-sized (125-350 μm in length, Fig. 9C) vessels (n = 6/group). In males, results also showed a significant dam exposure effect on the MWT % of the small-sized vessels (Fig. 9B), but not the medium-sized vessels (Fig. 9D); and for small vessels, only the M O₃:O₃x3 group had a significantly increased MWT % (n = 6/group) (Fig. 9D). Representative H &E vessel photomicrographs are depicted for F A:Ax3 and F O₃:Ax3 groups (Fig. 9D) and M A:Ax3 and M O₃:O₃x3 groups (Fig. 9F). Notably in the F O₃:Ax3 group, the vessel medial layer appeared thicker, consistent with smooth muscle cell hyperplasia or hypertrophy and thus, increased muscularization of these small arteries.

Figure 9A-F.

Medial wall thickness (MWT%) of small-sized (<125 μm) vessels in females (Fig. 9A, n = 6/group) and males (Fig. 9B, n = 6/group); and medium-sized (>125 μm) vessels in females (Fig. 9C, n=6/group) and males (Fig. 9D, n = 6/group) from H&E-stained sections are depicted. Data are expressed as means ± SEM. Significant dam, pup, or interaction differences, and difference from corresponding A:Ax3 groups are indicated as *p ≤ 0.05; **p ≤ 0.01, ***p ≤ 0.001. Representative small- and medium-sized vessel H&E-stained images are provided for the F A:Ax3 and F O₃:Ax3 groups (Fig. 9E) and M A:Ax3 and M O₃:O₃x3 groups (Fig. 9F). Scale bar lines = 200 μm.

Large pulmonary vessel morphometry.

To establish whether the arterial medial hypertrophy-like changes in small vessels occurred in concert with large vessel changes, the thickness of the inner (intimal) + middle (medial) layer and the outer (adventitial) layer of the pulmonary artery were quantified in both the AW5 and AW8 lung sections. For comparison, the thickness of the muscular layer of the pulmonary vein and the approximate diameter of the central airway and lengths of the adjacent pulmonary artery and vein were also quantified (see Fig. 2C). As expected, the airway and pulmonary vessels were consistently smaller within the AW8 section compared to the AW5 section. In air controls, the airway diameter was reduced in females (by 23%) and males (by 22%); the pulmonary artery length was shorter in females (by 36%) and males (by 22%); and the pulmonary vein length was shorter in females (50%) and males (by 42%) (Table 2). Data support consistent sectioning of the lung at AW5 and AW8 for both sexes. No exposure effects were observed for airway diameter or pulmonary artery length in females or males, within either the AW5 or AW8 lung sections.

Table 2.

Morphometrics (in μm) of the large central airway (AW) and associated pulmonary artery (PA) and pulmonary vein (PV). Data are expressed as means ± SEM. Significant effects are bolded.

Groups	A:Ax3	O3:Ax3	A:O3x3	O3:O3x3	ANOVA
Females	n = 6	n = 6	n = 6	n = 6	Exposure Effect
AW5
Central AW diameter	1535 ± 110	1840 ± 114	1700 ± 102	1670 ± 81	n.s.
PA length	1280 ± 100	1200 ± 60	1230 ± 81	1060 ± 62	n.s.
Intimal + Medial layer	24.1 ± 1.4	33.6 ± 1.1 ***	26.2 ± 1.4	28.3 ± 1.4	*** Dam effect * Interaction
Adventitial layer	24.8 ± 1.9	25.9 ± 2.2	24.7 ± 1.3	21.2 ± 1.3	n.s.
PV length	1270 ± 120	1300 ± 114	1300 ± 115	1160 ± 146	n.s.
Muscular layer	50.4 ± 3.9	45.9 ± 4.9	42.5 ± 4.0	37.5 ± 2.4	* Pup Effect
AW8
Central AW diameter	1180 ± 69	1310 ± 98	1160 ± 85	1220 ± 77	n.s.
PA length	812 ± 45	821 ± 110	950 ± 83	801 ± 58	n.s.
Intimal + Medial layer	23.5 ± 1.2	30.5 ± 0.21 *	28.9 ± 2.6	27.1 ± 1.4	* Interaction
Adventitial layer	20.7 ± 2.1	25.5 ± 1.9	25.9 ± 0.95	22.2 ± 1.2	* Interaction
PV length	642 ± 120	773 ± 120	734 ± 96	650 ± 57	n.s.
Muscular layer	24.4 ± 5.6	26.0 ± 3.5	23.9 ± 4.1	15.0 ± 2.2	n.s.
Males	n = 6	n = 6	n = 6	n = 6	Exposure Effect
AW5
Central AW diameter	1520 ± 100	1620 ± 47	1760 ± 110	1450 ± 84	n.s.
PA length	1220 ± 100	1280 ± 75	1270 ± 140	1280 ± 51	n.s.
Intimal + Medial layer	34.1 ± 1.9	30.8 ± 1.2	31.7 ± 1.5	31.8 ±	n.s.
Adventitial layer	28.6 ± 3.2	27.4 ± 1.8	28.6 ± 2.3	27.8 ± 0.98	n.s.
PV length	1260 ± 120	1380 ± 75	1560 ± 110	1240 ± 82	* Interaction
Muscular layer	57.5 ± 4.1	58.2 ± 4.0	60.5 ± 3.0	49.4 ± 2.8	n.s.
AW8
Central AW diameter	1180 ± 75	1180 ± 120	1270 ± 56	1240 ± 84	n.s.
PA length	946 ± 51	920 ± 60	963 ± 130	1040 ± 110	n.s.
Intimal + Medial layer	32.8 ± 2.1	29.6 ± 0.74	30.5 ± 1.5	32.4 ± 0.89	n.s.
Adventitial layer	26.8 ± 1.1	25.7 ± 1.9	23.6 ± 1.3	27.7 ± 0.58	n.s.
PV length	726 ± 71	758 ± 70	957 ± 80	809 ± 25	* Pup Effect
Muscular layer	28.0 ± 4.1	26.2 ± 4.5	26.4 ± 3.3	20.7 ± 4.5	* Interaction

^*Indicates different than A:Ax3 group;

^*p ≤ 0.05;

^***p ≤ 0.001;

n.s., no significant effect.

In females there was a highly significant dam exposure effect and a lesser but significant interaction effect observed for the thickness of the (intimal + medial) layer of the pulmonary artery at AW5, and a significant interaction effect at AW8 (Table 2). Across both sections, the F O₃:Ax3 group had a significantly thicker medial layer, consistent with medial hypertrophy of the larger arteries as well as the smaller arteries in this group. By contrast, there was a significant pup exposure effect on the pulmonary vein thickness in females, due to a thinner muscular layer within the AW5 lung section.

In male offspring, no significant exposure effects were observed for any of the pulmonary artery parameters. The only significant exposure effect noted in males related to the pulmonary vein in the AW8 section, wherein there was a significant pup exposure effect on vein length (being increased) and an interaction effect on the thickness of muscular layer (being decreased). The effect on vein length, along with the minor increase in the MWT% (< 125 μm arteries only) noted in M O₃:O₃x3 males could occur if their lungs were more prone to shrinkage or collapse during processing, possibly owing to the significantly greater ductal-to-alveolar area ratio observed in this male exposure group.

Correlations between airspace and vascular morphometric assessments.

To assess whether the observed changes in alveolar number, duct size, or arterial wall thickness were interrelated, the strength of statistical correlations between these parameters was examined ― focusing on the F A:Ax3 (white circle) and F O₃:Ax3 (pink circle) groups; and on the M A:Ax3 (white square) and M O₃:O₃x3 (blue hatched square) groups (Fig. 10). Correlations inclusive of data from the F O₃:Ax3 pup with a visibly enlarged heart (large pink checkered circle) are shown in brackets. Data revealed that for these two groups of female offspring, there was a significant negative correlation between alveoli number and ductal space width (Fig. 10A); a significant negative correlation between alveoli number and the MWT % (Fig. 10C); and a significant positive correlation between ductal space width and pulmonary artery medial thickness (Fig. 10E). For these two male groups, although the control data were visually separated from M O₃:O₃x3 group, no significant correlations were observed for any of the data sets (Fig. 10B, 10D, 10F).

Figure 10A-H.

Correlations for F A:Ax3 (white circle) vs. F O₃:Ax3 (pink circle) groups, with and without the F A:O₃x3 subject with visibly enlarged heart (larger pink checkered symbol) (Fig. 10A, 10C, 10E; n = 6/group). The Pearson r correlation (r) and correlation significance are provided within graphs. Significance of correlations inclusive of pup with an enlarged heart are provided within brackets. Correlations for M A:Ax3 (white square) vs. M O₃:O₃x3 (blue hatched square) groups are depicted (Fig. 10B, 10D, 10F; n = 6/group). For the significant correlations, lines are provided only for ease of visualizing the direction of change. Histologic images of the F O₃:Ax3 offspring with an enlarged heart revealed greatly enlarged large airspaces (Fig. 10G) and extensive medial hypertrophy of the pulmonary artery (Fig. 10H). Arrow shows poor tethering of a small airway (AW). Scale bar = 200 μm.

Remarkably, the female with the enlarged heart had the fewest alveoli, the widest ducts (Fig. 10A, 10G), and the greatest pulmonary artery medial hypertrophy-like changes (Fig. 10C, 10H). The extreme hypoalveolarization in this pup appeared to reduce effective tethering of small airways by lung parenchyma, thus allowing the wall of the smaller airway to deform and its lumen to narrow (Fig. 10G, arrow). Such structural change can result in expiratory airflow obstruction, a key component of both asthma and COPD.⁵⁸ Results suggest that like others in her exposure group, this pup had impaired angiogenesis and disrupted primary alveolarization. However, for unexplained reasons, the pathologic changes were more severe such that by the time of necropsy, she had developed significant right ventricular hypertrophy causing her heart to be visibly enlarged. Despite these underlying changes, there were no clinically apparent differences in the body size or behavior of this pup.

Correlations between % alveolar airspace by volume and lung tissue protein content.

Next, two methods for evaluating lung maturation in animal models of FGR were assessed. The first was based on a study in rabbits showing that from age 3-days to 12-weeks, the % of airspace volume occupied by alveoli increased from ≅ 30 to 50%, the ductal space decreased from ≅ 50 to 30%, and the tissue volume fraction was unchanged at ≅ 20%.⁵⁹ Herein, assuming the individual alveolar and ductal spaces represented simple spheres, the relative % alveoli space by volume within the AW8 lung section was calculated (Fig. 11A). In females, no significant exposure effects on the % alveoli volume were evident, and yet on average, all ozone-exposed F groups appeared to exhibit comparable decreases (by 24 – 28 %) relative to air controls. In the F O₃:Ax3 and F O₃:O₃x3 groups, the % alveoli volume reduction was largely due to decreased alveolar volumes, whereas in F A:O₃x3 group it was due mainly to increased ductal volume. By comparison, the % alveoli volume for the F O₃:Ax3 pup with the enlarged heart was only 16.5%. In male offspring, a significant pup exposure effect was observed for the % alveolar space by volume (Fig. 11A). Across the male exposure groups, the % alveolar volume decreased in a stepwise fashion. Reductions were due to a combination of decreasing alveolar volume and increasing ductal volume, such that for M A:O₃x3 and M O₃:O₃x3 groups, the % alveoli volume was significantly reduced. On average, the M A:O₃x3 and M O₃:O₃x3 groups decreased by 31% and 42%, respectively. Although the M O₃:Ax3 group was not significantly different, the group mean value was reduced by nearly 20%.

Figure 11A-F.

Summary of the % space occupied by alveoli, as estimated by volume within the AW8 lung section (Fig. 11A, n = 7-8 females and n = 7-9 males). Lung protein content (normalized to lung mass) (Fig. 11B n = 9-12 females and n = 10-12 males). Data are expressed as means ± SEM. Simple plots of group mean values for % alveolar volume to lung protein content were assessed for females (Fig. 11C) and males (Fig. 11D) (*p ≤ 0.05; **p ≤ 0.01). Correlations of % alveolar volume to lung protein content for F A:Ax3 vs. F O₃:Ax3 groups (Fig. 11E, n = 7-8) and M A:Ax3 vs. M O₃:O₃x3 groups (Fig. 11F, n = 7-9). The Pearson r correlation (r) and correlation significance are provided within graphs. Lines are provided for ease of visualizing the direction of significant correlation.

The second approach was based on previous studies in FGR lamb and rat models using decreased lung protein content to assess severity of lung hypoplasia.^{60, 61} No exposure effects were observed for lung protein content (normalized to lung tissue mass) in either female or male offspring (Fig. 11C). However, lung protein content of all male groups was lower (by 18-26%) compared to the F A:Ax3 group, and furthermore, the range of lung protein content was quite variable in all groups except the F A:Ax3 group. To assess whether the estimated decreases in the % alveolar volume correlated with lung tissue protein decreases, simple plots of group mean values were assessed for females (Fig. 11C) and males (Fig. 11D). Then, correlations of individual pups were assessed, focusing again on the female A:Ax3 (white circle) and O₃:Ax3 (pink circle) groups (Fig. 11E); and on the male A:Ax3 (white square) and O₃:O₃x3 (blue hatched square) groups (Fig. 11F). Together data suggest by 7 weeks-of-age, the lower and generally more variable lung tissue protein content in males was consistent with their rather delayed, yet ongoing lung development relative to females of the same age and exposure group. Furthermore, assuming that reduced lung tissue protein is a valid index for lung hypoplasia, there was a reasonable correlation between the degree of hypoplasia with decreases in the % of alveolar volume in O₃:O₃x3 males, but not O₃:Ax3 females. Data continue to suggest that different mechanisms underlie the developmental impacts of early life ozone in female and male offspring. The significant pup exposure effect on the % alveolar volume in males (especially the O₃:O₃x3 group), but not in females, suggests that the relative immaturity of the males in some manner increased their sensitivity to ozone exposure across peri-adolescence.

Lung gene expression.

To further evaluate the observed sex differences in lung vascular development, mRNA expression of select genes and transcription factors involved in angiogenesis and vascular health were quantified in lung tissue. Somewhat surprisingly, many of the common so-called housekeeping genes (e.g., βactin, Gapdh, Hprt, Ppia, Rpl13a and 18S) did not prove sufficiently stable for both sexes across all exposure groups. After considerable screening, Rps15a was selected as the endogenous control for both sexes. Rps15a is a gene encoding a ribosomal protein that is a component of the 40S subunit. For genes of interest, results in females revealed a significant dam exposure effect on expression of hypoxia inducible factor-1 alpha (Hif-1α) and its downstream vascular endothelial growth factor (Vegfa), along with a marginal effect on Vegf trans-membrane receptor Vegfr2 (p = 0.06), but not Vegfr1. The F O₃:Ax3 group (solid pink bars) had significantly increased Vegfa expression and F O₃:O₃x3 offspring had decreased Hif-1α expression (Fig. 12A). Somewhat similarly, results in male FGR-prone offspring also showed a significant dam exposure effect on Vegfa and Vegfr2 expression, a marginal interaction effect on Vegfr2 expression (p = 0.06), whereas Hif-1α and Vegfr1 expression were unchanged. The M O₃:O₃x3 group (hatched blue bars) showed a marginal increase in Vegfr2 expression (p = 0.06) (Fig. 12B). Signaling via Hif-1α/Vegfa/Vegfr1 and Vegfr2 play important roles in angiogenesis and endothelial cell survival.^{62, 63}

Figure. 12A-B.

Using qRT-PCR, lung mRNA expression was assessed for female offspring (n = 7–11/ exposure group; Fig. 12A) from dams exposed to air (white bars) or ozone (pink bars) during gestation, and for male offspring (n = 8–11/group; Fig. 12B) from dams exposed to air (white bars) or ozone (blue bars) during gestation. Additional peri-adolescent ozone exposures are indicated by hatching of the bars. Genes included hypoxia inducible factor-1 alpha (Hif-1α), vascular endothelial growth factor (Vegfa), Vegf trans-membrane receptor 2 (Vegfr2), Angiopoietin1 (Angpt1), Nitric oxide synthase 3 (Nos3), Platelet endothelial cell adhesion molecule-1 (Pecam-1), Dual-specificity phosphatase 1 (Dusp1) and Endothelin-1 (Et-1). Data are expressed as means ± SEM. Significant dam, pup or interaction effects are noted as *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001. Significant difference from the corresponding A:Ax3 groups are noted as *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001.

Gene expression patterns for females and males were also quite similar for Angiopoietin1 (Angpt1), Nitric oxide synthase 3 (Nos3 or endothelial NOS) and Platelet endothelial cell adhesion molecule-1 (Pecam-1 or CD31). In females, a significant dam exposure effect for these 3 genes was observed, with expression being generally decreased. Specifically, in F O₃:Ax3 and F O₃:O₃x3 offspring, Angpt1 expression was significantly reduced. In males, Angpt1 expression in the O₃:Ax3 group was also significantly reduced (solid blue bar). Angpt1 is critical for maturation of newly formed vessels and for the integrity of adult vasculature. Nos3 expression was significantly reduced in M A:O₃x3 offspring. Pecam1 expression was reduced in both the F and M O₃:Ax3 and O₃:O₃x3 offspring. Notably, Pecam-1 was the only gene for which the peri-adolescent ozone exposure-only groups showed consistent change, with both F and M A:O₃x3 groups showing significant reductions. Pecam-1 is important for maintenance and restoration of vascular endothelial barrier function after injury,⁶⁴ and endothelium-derived NO is an important downstream mediator for a variety vascular growth factors.⁶⁵

For the remaining two genes, Dual-specificity phosphatase 1 (Dusp1) and Endothelin-1 (Et-1), female and male offspring showed dissimilar expression patterns. In females, significant dam and pup exposure effects were observed for Dusp1, with both F O₃:Ax3 and F O₃:O₃x3 groups being significantly reduced ― whereas in males, no change in Dusp1 expression was observed. Dusp1 is involved in deactivating mitogen-activated protein kinase (MAPK) signaling pathways.⁶⁶ Conversely, for Et-1, no exposure effect was observed in females, but in males there was a significant pup exposure effect with Et-1 expression being reduced (Fig. 12A &B). Et-1 is a potent vasoconstrictor expressed mainly in endothelial and vascular smooth muscle cells and is implicated in pathological processes of COPD and PAH.⁶⁷

Discussion

We report herein for the first time, sex-specific changes in rat lung development due to maternal ozone exposure during implantation only; changes due to peri-adolescent ozone exposure across a period of rapid body and lung growth; and combined effects of cumulative exposure across both periods. Results demonstrate the utility of this FGR model to provide sex-specific, biologically plausible mechanisms by which ozone exposure during critical developmental windows may differentially increase risk of pulmonary and cardiovascular disease later in life.

In this rodent FGR model, both peri-implantation and peri-adolescence appeared to be windows of increased sensitivity for the effects of early life ozone exposure on lung development. As hypothesized, data demonstrated that significant sex-based differences in lung growth and maturation occurred in the FGR-prone offspring. Notably, female offspring appeared to be more affected by peri-implantation ozone exposure alone, whereas males, especially the FGR-prone male offspring, were more affected by additional ozone exposure during peri-adolescence. The current study underscores the importance of examining both sexes and in evaluating results separately by sex.^{68, 69}

Ozone exposure and FGR in humans

Although FGR is clearly associated with impaired lung function in children, to date, epidemiological associations of FGR with exposure to air pollutants such as ozone during pregnancy are mixed. In areas more heavily polluted by ozone such as southern California, women appeared to be at increased risk for having babies with FGR if exposure occurred within the 2^nd or 3^rd trimesters.⁷⁰ Similarly in North Carolina, exposure to higher ozone concentrations in the third trimester was associated with low birthweight term births.⁷¹ On the other hand, in regions with few days of ozone exceedances, relative to fine PM exposure, associations of FGR with ozone exposure were null⁷² or negative.⁷³

Consistent with our study findings, Michikawa et al⁷⁴ reported that first trimester ozone exposure was independently associated with increased risk of poor fetal growth. Likewise, Hansen et al⁷⁵ used ultra-sound examination during pregnancy to show that reduced fetal abdominal circumference (an index for asymmetric FGR) was associated with ozone exposure in early pregnancy. Lee et al¹⁷ reported that first trimester exposure to both ozone and fine PM increased risk of fetal effects (e.g., small for gestation age, preterm birth) as well as other pregnancy complications (e.g., preeclampsia and gestational hypertension). Moreover, a recent study from southern California showed that maternal ozone exposure early in pregnancy increased risk of spontaneous premature membrane rupture, a serious pregnancy complication.⁷⁶ Conversely, Olsson et al⁷⁷ reported that increased ozone during the first trimester increased risk of pre-eclampsia and preterm birth but did not affect fetal growth per se. Klepac et al²¹ reported that ozone exposure over the entire pregnancy was significantly associated with higher risk for preterm birth, but not with FGR.

The one thing that all these studies had in common was that none reported associated risk of air pollutant exposure on maternal or infant health based on the sex of the child. Instead, all models were adjusted for fetal or infant sex. Importantly, however, studies show that sex may influence pregnancy complications (e.g., preeclampsia)^{78, 79} as well as infant health outcomes (e.g., males are more likely to be born prematurely^{80, 81} while females are more likely to develop FGR).^{71, 82} To better disentangle air pollutant exposure risks from that of other exposures and stressors occuring in early life, our study results suggest that infant sex should not only be a primary variable of interest ― but rather that air pollutant health associations should be reported separately for females and males.

Sex differences in lung maturation of FGR-prone rat offspring

The rat is a particularly useful animal model with which to evaluate effects of early life air pollution exposure on fetal growth and lung maturation for several reasons. The first reason is that rats and humans show similar stages of lung developmental progression⁴² with the final stage of lung development in both humans and rodents occuring principally after birth.⁸³ Another reason is that optimal implantation in rats and humans both occur during a narrow (≅ 2-3 day) window in early pregnancy (i.e., GD5-6 in rats and 8-10 days after ovulation in humans).⁸⁴ In addition, rats and humans have similar though not identical placentation types as both are classified as discoid and hemochorial.⁸⁵ Having similar placentation is important because the placenta is the central coordinator of fetal growth and the interface with maternal circulation. Thus, the placenta plays a critical role in determining sex differences in offspring health.⁸⁶ Despite these similarities, an obvious developmental difference between rats and humans is that because of their extremely short gestation (≅ 3 weeks), comparatively, rat pups are born very immature. Consequently, premature birth per se does not occur in rats. Furthermore, the later stages of lung development in rats, including bulk alveolarization, occur exclusively in early post-natal life⁴² (Fig. 1). Therefore, we allowed offspring to be ≥ 5 weeks-of-age before initiating peri-adolescent ozone exposures to better mimic exposures occuring in children and teenagers. With this approach, we can more clearly distinguish and infer potential lung developmental impacts of ozone exposure in people due to prenatal (i.e., pregnancy) exposure from that of postnatal (i.e., peri-adolescence) exposure.

Based on lung volume and morphometry changes observed at 7 weeks-of-age, the FGR-prone female offspring had a significantly reduced number of mature alveoli present, and thus had diminished lung volume or capacity. FGR-prone female offspring also exhibited thickening of their small and large pulmonary arteries, consistent with medial hypertrophy-like changes. Because the final stage of lung development in humans and rodents involves partitioning of large primary saccules (described herein as ductal spaces) into smaller air spaces by the inward protrusion of septa derived from the walls of the saccules, and knowing that coordinated angiogenesis is critical for the bulk alveolarization processes,⁸³ we postulate that the changes observed in FGR-prone females were due largely to impaired angiogenesis with corresponding decreases in primary alveolarization. We further propose that due to the underlying perfusion impairment, these female offspring experienced increased resistance to blood flow within the deep lung and responded through remodeling of the main pulmonary artery. The observed vascular remodeling changes are consistent with increased risk for developing pulmonary hypertension. Pulmonary hypertension represents a heterogeneous group of disorders that culminate in elevated pulmonary artery pressure and right ventricular failure. Pulmonary hypertension is classified clinically into 5 groups based on underlying pathophysiologic processes, (i.e., left heart disease, veno-occlusive lung disease, lung disease and/or hypoxia, multifactorial disorders, and familial or idiopathic PAH).⁸⁷ Idiopathic PAH is a progressive disease of the pulmonary vasculature and is strongly predominant in females, although how sex influences PAH development and progression remains poorly understood.⁸⁷

Male offspring, on the other hand, appeared to be less affected by the early impacts of maternal ozone exposure on lung angiogenesis. Yet upon repeated exposure to ozone during peri-adolescence, the males (especially the male FGR-prone offspring) had significantly greater lung hypoplasia, likely due to inhibition or disruption of secondary alveolarization processes. This possibly is supported by the significant increase in total ductal area observed in the M O₃:O₃x3 group, and the significantly lower % alveolar volumes in the M A:O₃x3 and M O₃:O₃x3 groups.

Superimposed upon these dominant effects of ozone exposure by sex (i.e., dam exposure effects in female offspring and pup peri-adolescent exposure effects in male offspring), there were also significant effects of peri-adolescent exposure in females, and vice versa, significant dam exposure effects in males. Specifically in females, significant pup exposure effects were observed for BMI reduction and for increased total ductal area. In males, significant dam exposure effects were observed for increased body length and for increased % MWT of small (<125 μm) arteries. Together, the combined dam and pup exposure effects culminated in all ozone-exposed offspring having some degree of reduced % alveolar volumes (up to 28% in females and 42% in males). Because effective gas exchange requires an adequate number of mature alveoli as well as coordinated ventilation and perfusion of alveoli, this magnitude of a decrease in alveolar volume could lead to significant hypoxemia or hypoxia, especially during exercise.

We hypothesized therefore that all ozone-exposed offspring, in particular the F O₃:Ax3 and M O₃:O₃x3 offspring, may have experienced significant hypoxia. Consequently, hypoxia-sensing factors may have been increased to further promote angiogenesis and improve gas exchange.⁸⁸ One such factor is Hif-1α. The Hif-1α subunit along with Hif-1β subunit comprise a heterodimeric transcription factor, Hif-1. Hif-1 responds to hypoxia by trans-upregulating downstream genes related to angiogenesis (e.g., Vegfa, angiopoietins). Vegfr2 is considered the main signaling receptor for Vegfa bioactivity.^{62, 63, 88} Herein, to assess whether vessel morphometry changes (e.g., % MWT or PA medial thickness) correlated with vascular gene expression changes, simple plots of group mean values were generated (Supplemental Figure S1). In females, significant positive correlations were observed for Vegfa and for Vegfr2 gene expression with the % MWT (< 125 um vessels) and pulmonary artery medial wall thickness. Consistently, A:O₃x3 females had the least change, O₃:Ax3 females had the most change, and O₃:O₃x3 females fell in between these two groups (Fig. S1A, C, E, G). In the male offspring, despite significant dam exposure effects on Vegfa and Vegfr2 expression, no significant correlations with vessel morphometry changes were detected (Fig. S1B, D, F, H). Although Vegfr2 increased with increasing % MWT (< 125 um vessels) in the M O₃:Ax3 and M O₃:O₃x3 groups, the A:O₃x3 males consistently showed lower gene expression for both Vegfa and Vegfr2 (relative to air controls), hence no significant correlations across all male groups were observed. Increased Vegfa mRNA expression and corresponding increases in Vegfa protein have been demonstrated in lung tissue from human PAH subjects.⁸⁹

Together, results support limited but increasing evidence that sex and lifestage influence endothelial cell (EC) biology and sensitivity to hypoxia. Shin et al⁹⁰ recently summarized sex-specific differences in EC responses to hypoxia, pro-oxidant, and angiogenic factors, concluding that differences were consistent with a vulnerable phenotype in females. Sex differences in EC biology reflect both intrinsic differences driven by chromosomes as well as differences driven by sex steroids.⁹¹ Some sex differences are already present at birth and are seemingly maintained throughout life, whereas others are acquired (or lost) across lifestages. Genes involved in acquired EC sex differences are often targets of sex steroids.⁹¹ For example, a study comparing cultured pulmonary arterial ECs from female and male donors under hypoxia (1% oxygen) or normoxia (20% oxygen) conditions showed that ECs from females were significantly more likely to become apoptotic, were less migratory under Vegf-enriched conditions, and with increasing donor age, became progressively less proliferative ― whereas male ECs were more proliferative with donor age.⁹² Relatedly, when pulmonary microvascular ECs were subjected to physiological shear stress under hypoxia vs. normoxia conditions, the rate of proliferation in female ECs was lower than male ECs; and thrombospondin-1, an inhibitor of proliferation, was more highly expressed in female cells. These differences between female and male ECs persisted even in the absence of sex hormone influences.⁹³

Endothelial responses may also be influenced by sex steroids. Ichimori et al⁹⁴ showing differential effects of β-estradiol supplementation or estrogen receptor blockade on pulmonary arterial ECs under hypoxia vs. normoxia conditions. Furthermore, in mice genetically mutated to induce loss-of-function in bone morphogenetic protein receptor 2 (BMPR2), a common mutation in familial PAH, chronic hypoxia (10% oxygen x 1 month) induced certain dichotomous gene (e.g., Nos3 and Et-1) and vascular collagen deposition responses in modified male and female mice, yet no sex difference in lung vessel thickness was observed.⁹⁵ Interestingly, in Sprague-Dawley rats hyperresponsive to treatment with the Vefgr2 blocker, SU5416, a disproportionate number of males (72%) developed severe PAH-like lung changes compared to females (27%) due in part to differences in persistent EC apoptosis. This difference was abolished when females were ovariectomized and furthermore, estradiol supplementation beginning 2 days prior to SU5416 treatment inhibited lung EC apoptosis and abrogated the severe PAH phenotype in males and ovariectomized females. Unfortunately, once the pathologic lesions were established, neither estradiol nor progesterone supplementation was able to reverse lung EC apoptosis or PAH-like changes.⁹⁶

In the present study, despite significant dam exposure effects in females for Hif-1a, Angpt1, Nos3, Pecam1, and Dusp1 genes, expression of these important hypoxia-sensing and endothelial growth factors was significantly but paradoxically reduced in FGR-prone females. Accordingly, significant but negative group correlations of Angpt1, Pecam1 and Dusp1 with pulmonary artery medial wall thickness were observed across the four groups of female offspring (Supplemental Fig. S2 A & C). Notably, the female with the enlarged heart consistently showed the greatest negative effect and A:O₃x3 females showed the least effect. Conversely, in male offspring, significant positive correlations of Angpt1 and Pecam1 expression with pulmonary artery medial wall thickness were observed. It may be relevant that in previous investigations of growth-restricted fetal lambs, alveolarization decreased by 20% whereas pulmonary vessel density decreased by over 40%.⁹⁷ We theorize that lung vessel density in FGR-prone females was likewise more markedly reduced than alveolarization. Consequently, fewer lung ECs in F O₃:Ax3 and F O₃:O₃x3 offspring may have been present, thus limiting their ability to increase expression and signaling of these critical vascular growth factors, relative to the air control offspring.

A summary of proposed signaling between vascular ECs and vascular smooth muscle cells (VSMC) is shown in Figure 13. Significant dam (↓ or ↑) and pup (↓* or ↑*) exposure effects on gene expression are shown using pink arrows for females and blue arrows for males. Based on the direction of gene expression change observed herein, predicted EC and VSMC health impacts for offspring by sex are likewise depicted. In this schematic, we propose that due to insufficient EC niche populations in female FGR-prone offspring, increased Vegfa expression and signaling failed to translate into increased expression of downstream genes necessary to promote angiogenesis (i.e., Nos3, Angpt1, Pecam1). Limited Angpt1 and Pecam1 expression/release would, in turn, compromise EC migration, proliferation, and survival. Likewise, limited Hif-1α and Nos3 expression would reduce NO production. In turn, decreased NO release would further limit EC survival and would fail to inhibit VSMC proliferation. Additionally, without sufficient Dusp1 expression to counteract MAPK signaling, VSMC proliferation would be further perpetuated. Together, these proposed impacts on VSMC are consistent with the significantly increased % MWT observed in small- and medium-sized arteries of female FGR-prone offspring.

Figure 13.

Schematic of gene expression changes for significant dam or pup exposure effects with predicted health impacts in endothelial cells (EC) and vascular smooth muscle cells (VSMC). Significant dam (↓ or ↑) and pup (↓* or ↑*) exposure effects on gene expression and predicted EC and VSMC health impacts are depicted are shown using pink arrows for females and blue arrows for males. Arrows indicate direction of change.

New roles for the DUSP family of proteins in the pathogenesis of lung disease are emerging due to their ability to regulate MAPKs.⁹⁸ Thus, the reduced Dusp-1 expression observed in FGR-prone female offspring is of interest not only because pathologic MAPK pathway signaling may contribute to lung vasoconstriction and vascular remodeling,⁶⁶ but because it may also serve to perpetuate chronic inflammatory processes underlying inflammatory airway diseases such as asthma and COPD.⁹⁸ Additionally, Dusp-1 upregulation appears to be a central mechanism by which glucocorticoid treatment resolves inflammation.⁹⁹ Moreover, inhaled glucocorticoid treatment has been shown to be generally more effective in men (i.e., 44% better for asthma and 28% better for COPD) than in women.¹⁰⁰

It may also be relevant that women appear to have increased incidence and unique disease mechanisms underlying lung diseases such as asthma, COPD, and PAH.¹⁰¹ The FGR-prone female offspring had lung changes that could increase risk of developing 2 of the 4 hallmark features of COPD, namely, emphysematous change, and pulmonary arterial hypertension. The other key features of COPD, airway remodeling and chronic bronchitis, may require additional exposures (e.g., cigarette smoking).¹⁰² Even after adjusting for pack-years of smoking, women appeared to have a 50% increased risk of chronic obstructive pulmonary disease (COPD) compared to men.¹⁰³ The pathologic basis or mechanism(s) of these sex-based differences in COPD remain unknown.

In male offspring, as in the female offspring, there was a significant dam exposure effect on Vegfa and Vefgr2 gene expression. However, without the limited Hif-1α or Nos3 expression (as likely occurred in females), NO production may have been maintained or increased, thus improving EC survival and minimizing VSMC proliferation in FGR-prone male offspring. In males, as in females, Pecam-1 expression was reduced in all ozone-exposed groups. A previous study in neonatal rat pups demonstrated that administration of anti-PECAM-1 antibodies disrupted normal alveolar septation by inhibiting EC migration without affecting proliferation.⁸³ Herein, diminished Pecam-1 expression may have impacted EC migration in both sexes, possibly contributing to the knobby-appearing septal tips depicted in Fig. 6D.

Unique to the male offspring, there was a significant pup exposure effect observed for Et-1 production. Reduced Et-1 expression may have resulted in mixed, possibly counteracting, effects. On one hand, if an insufficient EC population was the reason for reduced Et-1 expression, then reduced signaling in EC via the ET_B type receptor would reduce NO production, thereby limiting future EC survival and may have promoted VSMC proliferation (Fig. 13). This scenario would be consistent with the minor increase in % MWT noted in the M O₃:O₃x3 group. On the other hand, if the EC population was adequate, then decreased Et-1 expression may have reflected down-regulation of Et-1 production to reduce signaling to the VSMC. Reduced Et-1 signaling via VSMC ET_B receptors would limit VSMC constriction and reduced signaling via ET_A receptors would limit VSMC proliferation. This latter situation is consistent with the lack of change observed in the M A:O₃x3 group regarding % MWT of small and medium-sized arteries.

Key questions on sex differences in ozone-induced lung developmental change

Overall, these results raise two key questions. First, why were female offspring more affected by dam exposure during implantation? Secondly, why were males, especially the FGR-prone offspring, more affected by peri-adolescence ozone exposure?

To address the first question, it is important to review the pathogenesis of asymmetrical FGR. In people, asymmetrical FGR is believed to arise from reduced fetal nutrient and oxygen supply late in pregnancy.³⁹ Under normal conditions, fetal growth is regulated by placental size and rate of nutrient transfer to the fetus via changes in uterine artery blood flow and vascular resistance.¹⁰⁴ If maternal stressors during pregnancy significantly limit fetal nutrient availability, the placenta adapts to better ensure adequate nutrient and oxygen supply. However, with increasing severity of such stressors, placental adaptation may be insufficient to sustain normal fetal growth. Under these circumstances, signaling from the fetus appears to further prioritize nutrient allocation. So-called “brain sparing” occurs as the feto-placental unit seeks to maintain development of critical organs, in particular the brain (and thus head size) by limiting adiposity and liver development, thereby decreasing abdominal circumference.¹⁰⁵ In preterm neonates, cerebral blood flow was shown to be significantly affected by FGR; and notably, males with FGR had consistently greater cerebral blood flow than females with FGR.¹⁰⁶ Yet, in redistributing blood flow towards the brain, other organs (e.g., the liver) may receive excessively reduced blood, oxygen, or nutrient supply, hence the fetal dilemma of “spare the brain and spoil the liver” described by Nathanielsz and Hanson.¹⁰⁷

In this rat FGR model, dam peri-implantation ozone exposure led to a series of events including changes in uterine arterial blood flow resistance, utero-placental insufficiency, and ultimately, asymmetrical late-term growth-restricted fetuses.^{36, 37} Also in this model, notable fetal sex differences in placental efficiency and metabolism were observed.⁴⁰ Placentae of male FGR-affected fetuses showed increased mitochondrial metabolism such that male fetuses not only had reduced body weight and adiposity, but they also had significantly greater down-regulation of hepatic metabolic signaling along with comparatively minor hypothalamic gene changes. By contrast, placentae of female FGR-affected fetuses adapted by increasing autophagy-related processes, and although female fetuses showed fewer hepatic gene and pathway changes, they had significantly greater hypothalamic changes.⁴⁰

Taken together, our findings suggest that by late gestation, male feto-placental units prioritized shunting of blood (and thus oxygen and nutrients) to the brain and less to the liver. Such change likely occurred at the ductus venosus, one of the three main shunts occuring in fetal vasculature during development.¹⁰⁷ In so doing, we further speculate that as with the brain blood supply in males, the blood flow to the lungs in late gestation was at least partially maintained. By contrast, female feto-placental units did not appear to reallocate blood supply in a similar manner. Consequently, by late gestation, FGR-affected female fetuses likely incurred proportionate reductions in blood flow, thereby experiencing greater hypoxia in all body organs to some degree. In so doing, blood and oxygen supply to the lung may have fallen below a critical threshold required for normal angiogenesis and thus inhibiting primary alveolarization. We further speculate that female fetuses did not engage in “brain-sparing” responses because any further reduction of blood supply to the liver or to other abdominal organs (e.g., developing ovaries and oocytes) may have adversely affected the future fertility of the female offspring.

These proposed sex-based responses are supported by recent studies showing that reduced blood supply to the developing neonatal lung (induced in mice by direct banding of the pulmonary artery on PND1), resulted in severe impairment of alveolarization and vascularization, the main characteristics of pulmonary dysplasia.¹⁰⁸ Mechanistically, pulmonary hypoperfusion induced by PND1 banding appeared to disrupt cell-cell communication and axon guidance necessary to achieve an adequate number of alveoli by PND21.¹⁰⁹ Results are not unlike that of FGR-prone females whose limited lung capacity appeared to be due, in large part, to a reduction in the number of mature alveoli present.

Additionally, fetal cardio- and cerebro-vascular adaptations occurring in response to FGR and associated chronic intrauterine hypoxia, can lead to organ-specific oxidative stress. Excessive oxidative stress can in turn interfere with the dynamic and spatial signaling required for proper angiogenesis and microvascular maturation.¹¹⁰ Of relevance, we previously showed that when neonatal rats were exposed to ozone, the youngest (PND14) female pups were the most affected; and females appeared to incur greater oxidative stress compared to males of the same age.¹¹¹ In the present study, the more severe changes noted in the O₃:Ax3 female with the enlarged heart may have been due to more severe reduction of blood flow to the developing lung or alternatively, that more severe hypoxia and/or oxidative stress culminated in pulmonary venous stenosis¹¹² or premature closure of the ductus arteriosus. In a lamb model of PAH, in utero ligation of the ductus arteriosus was shown to produce pulmonary hypertension and vascular remodeling owing to increased formation of reactive oxygen species.¹¹³

One factor to explain the second key question, why males were more affected by peri-adolescent ozone exposure, relates to their ongoing and more active state of body and lung growth. Herein, data consistently suggested that at 7 weeks-of-age, males had delayed, yet ongoing lung development, relative to the females. Especially during the 2^nd and 3^rd 0.8 ppm ozone exposures, males appeared to be experiencing greater and more rapid growth than females. These findings are not unlike the sex differences observed in lung growth and capacity in adolescent humans. For example, Li et al¹¹⁴ demonstrated that age influences on forced lung vital capacity (FVC) were largely mediated by height growth. In addition, in adolescence, lung growth appeared to lag behind body height increases; and based on allometric constants between height and forced expired volume in one-second (FEV₁) measurements, girls peaked earlier (at 12 years-of-age) compared to boys (at 16 years-of-age).¹¹⁵ We therefore propose that the relative immaturity of lung development of the males in some manner increased their sensitivity to ozone exposure during peri-adolescence. Moreover, because some of the FGR-prone males were much bigger (especially in body length), these males may have been at still greater risk for exposure, thus causing the M O₃:O₃x3 group to have a significant increase in ductal area compared to the air control males (Fig. 8D).

Another factor for why males may have been more affected by peri-adolescent ozone exposure was that most female offspring had entered puberty by the 2^nd and 3^rd ozone exposures, whereas most (but not all) males had only entered puberty by the 3^rd but not the 2^nd exposure. Thus, protective influences of hormones such as estrogen or progesterone¹¹⁶ in the female offspring during their 1^st 0.4 ppm and 2^nd exposure 0.8 ppm ozone exposures may have further influenced these results. Massaro et al¹¹⁷ previously demonstrated that by 8 weeks-of-age, female rats tended to have smaller alveoli than age-matched males, and that estrogen, not testosterone, was largely responsible for modulating sex differences in alveolar size and gas-exchange surface area in young adult rodents.¹¹⁸ In addition to the aforementioned protection provided by female sex hormones on EC health,^{91, 94, 96} sex differences in sensitivity to oxidative stress also exist.¹¹⁹ Healthy adult females are alleged to be more proficient in maintaining redox homeostasis¹¹⁹ such that under physiologic conditions, females are less susceptible to oxidative stressors.¹²⁰ Antioxidant properties of estrogen are frequently implicated as a key reason contributing to the female sex advantage in maintaining redox balance, at least until menopause.¹²⁰ Accordingly, if one focuses on signaling responses due to oxidative insult in the healthy adolescent offspring only (i.e., offspring from air-exposed dams exposed to ozone only during secondary alveolarization processes), A:O₃x3 males had significantly decreased Vegfa (p = 0.023), marginally decreased Vegfr2 (p = 0.054), and significantly decreased Et-1 expression (p = 0.045) when compared to A:Ax3 males (Fig. 12A, based on 2-tailed t tests). Notably, these trends were reversed in the O₃:O₃x3 males. By contrast, compared to A:Ax3 females, A:O₃x3 females did not shown significant change for these same genes (Fig. 12B). These results complement existing evidence⁹⁰ that sex and lifestage differences in sensitivity to hypoxia, oxidative stress, and redox imbalance influence vascular signaling pathways and thus contribute, in a sex-specific manner, to adverse pulmonary vascular health outcomes. Results clearly emphasize the need to include sex or sex steroid responses as biological variables to better understand and translate animal model findings on EC and VSMC signaling to that of humans with pulmonary hypertension, PAH, or COPD.

Study limitations.

All observations and proposed mechanisms for the effects of early life ozone exposure in the rat offspring were based on changes observed at 7 weeks-of-age. We acknowledge the limitation of using a single age to define health impacts, and that additional time-course studies are necessary to assess earlier changes and potential differences in adaptive responses. Furthermore, morphometric changes at 7 weeks would have primarily reflected only the 1^st (0.4 ppm) and 2^nd (0.8 ppm) ozone exposures, as more time after the 3^rd exposure would be necessary for subsequent morphometric change. A 2-4 h post-exposure necropsy time was used to better evaluate lung gene expression changes acutely after the 3^rd ozone exposure.

It is also important to acknowledge that the ozone concentrations used herein were higher than the current U.S. National Ambient Air Quality Standard (NAAQS). As in our previous investigations, a 0.8 ppm ozone concentration was used to expose the dams during implantation, and the pups were similarly exposed to 0.8 ppm ozone for their 2^nd and 3^rd exposures. For the 1^st exposure, owing to their immaturity, a reduced 0.4 ppm concentration was used initially. These concentrations were selected based on data from Hatch et al comparing isotopic ¹⁸O-labeled ozone reaction product concentrations in rat and human lungs, revealing a four- to five-fold lower lung ozone dosimetry in the rat.¹²¹ This occurred in part because rats were exposed during the daytime when nocturnal rodents are much less active and in part because the human subjects were intermittently exercising during their exposures. By extension, the 0.4 and 0.8 ppm ozone concentrations used herein would correspond to ≅ 0.08-0.10 and ≅ 0.16-0.20 ppm ozone, respectively, for healthy adult humans actively exercising outdoors. Comparatively, the 2015 U.S. NAAQS for ozone, averaged over 8-h, is 0.07 ppm (U.S. Environmental Protection Agency. 2015. National ambient air quality standards for ozone EPA-HQ-OAR-2010-0885 https://www.federalregister.gov/documents/2015/10/26/2015-26594/national-ambient-air-quality-standards-for-ozone (accessed 5-2-2024).

Unfortunately, exceedances of the ozone NAAQS are not uncommon, particularly in warmer regions with higher levels of ozone precursors (e.g., tailpipe emissions from heavy traffic). Climate change-related extreme heat events and wildland fires are anticipated to further increase the number of “high ozone” days (i.e., days exceeding the NAAQS for ozone).^{122, 123} Other factors, including species differences in lung antioxidants, further influence the lung toxicity of inhaled ozone.¹³ Relatedly, in pregnancy, increased minute ventilation (up to 48%) occurs within the first trimester and is then maintained throughout pregnancy.¹²⁴ Such changes in ventilation may increase lung dosimetry of ozone and other air pollutants across pregnancy. For these reasons, a 2-consecutive day “high ozone” exposure could potentially occur, thus impacting a sensitive window such as peri-implantation. Future epidemiologic studies in pregnant women are warranted to begin to address this possibility in people.

Furthermore, children, and by analogy peri-adolescent rats, are more sensitive to ozone exposure compared to adults because they are actively growing, and their lungs are still undergoing development. Children and teenagers also spend more time outdoors participating in sports or active play during which they breathe faster, deeper, and engage in open mouth breathing ― all factors known to increase air pollutant lung dosimetry.¹²⁵ Relatedly, a recent systematic review of ozone effects on human lung function concluded that in children (but not adults), low-level (below the current NAAQS) ozone exposure, particularly over longer periods, may adversely affect lung function.¹²⁶ As highlighted in the review, one study by Hwang et al¹²⁷ evaluated longer-term (2 yr) ozone exposure in 10 to 12-year-olds and reported associations with decreased lung function and lung function growth. Moreover, the peri-adolescent boys were reportedly more susceptible than the girls.¹²⁷ Authors of the systematic review question the adequacy of existing ozone short-term (daily) ozone NAAQS to protect children’s respiratory health in the absence of a long-term standard.

Conclusions and future direction

This FGR model provides sex-based biologic plausibility for potential consequences of early life exposure to ozone, a ubiquitous urban air pollutant. For obvious reasons, air pollutant human clinical studies only use adult, non-pregnant subjects, and in many, only one sex (often males) is used or alternatively, responses in males and females are simply combined. Similarly, many animal toxicological investigations use only one sex of adult rodents. As such, the primary exposure groups in a typical toxicology study are equivalent to this study’s A:Ax3 vs. A:O₃x3 groups, and yet these F and M A:O₃x3 exposure groups incurred in the least effects on lung health overall despite using peri-adolescent rats instead of adult rats. Thus, our study reveals the importance of (1) evaluating results separately by sex, and (2) that animal models to better simulate exposures in at-risk populations such as pregnant women or growing children are essential to defining vulnerable windows of exposure.

Lastly, a recent epidemiology study showed that in multiple U.S. cities, ozone was associated with pediatric respiratory morbidity; and notably within each city, children living in low socio-economic status (SES) environments appeared to be especially at risk based on high underlying rates of respiratory morbidity.¹²⁸ With this FGR model, we can specifically incorporate defined social determinants of health inequities in health and well-being. It is increasingly recognized that growth restriction during infancy and childhood, especially when combined with additional early life exposures or stressors, can alter developmental programming for risk of disease later in life. As described at length herein for lung disease, important sex-differences also exist for risk of cardiovascular disease,¹²⁹ cardiometabolic disease,¹³⁰ and neurodevelopmental or learning disorders.¹³¹ Using this FGR model, investigations are in progress to assess differential early life ozone exposure effects, by sex, on lung injury, inflammation and antioxidant response, metabolic disease, and cardiovascular outcomes. Together, epidemiological and animal model investigations may help to define underlying factors and mechanisms that will allow more targeted and effective prevention and intervention strategies to be enacted.