Lifelong Hypoxia Produces Multigenerational Developmental Lung Disease and Pulmonary Hypertension in Mice

Abstract

Chronic hypoxia during development can impair lung and pulmonary vascular growth, but most mouse models of hypoxia-induced pulmonary hypertension expose adult animals for limited periods and do not capture lifelong or multigenerational hypoxic exposure. We developed a multigenerational model of lifelong moderate hypoxia by maintaining wild-type C57BL/6J mice at 13% inspired oxygen fraction (FiO₂), beginning before fertilization and continuing through gestation, postnatal development, and adulthood. This exposure permitted breeding over two generations. Pregnancy and delivery rates were largely preserved, but second-generation litters had fewer pups alive at birth and reduced overall survival to weaning, primarily due to fetal or perinatal loss. Among surviving adults, both first- and second-generation hypoxia-exposed mice developed pulmonary hypertension, right ventricular systolic dysfunction, right ventricular hypertrophy, and reduced exercise capacity compared with age-matched room-air controls. These functional abnormalities were accompanied by alveolar simplification, increased lung compliance, increased muscularization of small pulmonary arterioles, and reduced peripheral pulmonary vascular density. Several measures, including right ventricular systolic pressure and right ventricular remodeling, were modestly worse in second-generation mice. Hemoglobin concentration was increased in both hypoxia-exposed generations, whereas exploratory analysis of gastrocnemius muscle showed no detectable reduction in skeletal muscle capillary density in first-generation mice. Thus, lifelong exposure to 13% FiO₂ establishes a survivable but pathologic model of developmental hypoxia that permits multigenerational studies while producing persistent lung developmental abnormalities, pulmonary vascular disease, right ventricular dysfunction, erythrocytosis, and exercise limitation.

Graphical Abstract

Created in BioRender. Zazzeron, L. (2026) https://BioRender.com/fv5ip9y

New & Noteworthy

Lifelong exposure of C57BL/6J mice to 13% FiO₂ permits multigenerational breeding while producing persistent developmental lung disease, pulmonary vascular remodeling, pulmonary hypertension, right ventricular dysfunction, erythrocytosis, and reduced exercise capacity. Second-generation mice show reduced perinatal viability and modestly worse pulmonary vascular/right ventricular remodeling. This model enables investigation of chronic prenatal and postnatal hypoxia without the severe early mortality observed with more extreme hypoxic exposure.

Introduction

Approximately 140 million people worldwide live at high altitude, where chronic hypoxia represents a major environmental stressor that shapes cardiopulmonary physiology(1). Chronic hypoxia can interfere with normal lung and pulmonary vascular development and contribute to pulmonary hypertension (PH)(2). Experimental mouse models of chronic hypoxic exposure have been widely used to study pulmonary vascular remodeling and mechanisms of hypoxia-induced PH(3-5). These exposures are often equivalent to an inspired oxygen fraction (FiO₂) of approximately 10–12%, corresponding roughly to altitudes of 4,000–5,800 meters above sea level. Adult hypoxia models reproducibly induce muscularization of small pulmonary arteries, mild PH, and modest right ventricular (RV) hypertrophy, changes that are often reversible after return to normoxia(6).

The effects of developmental hypoxia on pulmonary vascular development, remodeling, and long-term cardiopulmonary outcomes have been studied less extensively than adult hypoxia models, but available evidence suggests that early-life hypoxia can durably impair lung development(7-10). Transient exposure of newborn C57BL/6 mice to 12% oxygen during early postnatal life produced persistent alveolar simplification and abnormal lung mechanics after return to room air, but did not result in sustained RV hypertrophy or pulmonary vascular abnormalities in adulthood(11).

Because alveolar and pulmonary vascular development extend beyond the early neonatal period, transient hypoxia models may not fully capture the consequences of chronic hypoxia throughout lung development(12, 13). This distinction is relevant to high-altitude exposure, where hypoxia begins before birth and continues through postnatal maturation. Although lifelong developmental hypoxia has been studied in high-altitude-adapted deer mice exposed to ~12 kPa O₂ from conception to adulthood, comparable exposure paradigms are poorly tolerated by lowland laboratory mice(14, 15). Consistent with these observations, we previously reported that lifelong exposure of C57BL/6 mice to 11% FiO₂ beginning at fertilization resulted in severe neonatal mortality, with no pups surviving beyond postnatal day 2. Initiation of hypoxic exposure later in gestation (embryonic day 15–18) reduced but did not eliminate mortality, with approximately 70% of pups dying before weaning. Postnatal exposure to 11% FiO₂ maintained until adulthood produced severe pulmonary hypertension, with RV systolic pressure (RVSP) reaching 70–80 mmHg, marked RV hypertrophy, reduced exercise capacity, RV failure, and increased long-term mortality. This phenotype was associated with impaired lung development, muscularization of small pulmonary arteries, and reduced distal pulmonary vascular density, indicating that prolonged developmental hypoxia causes both pulmonary vascular remodeling and impaired vascular growth(16, 17).

While this model provides a robust platform to study developmental hypoxia-induced cardiopulmonary disease, it also presents several limitations. Mortality was substantial, with 20–30% mortality before weaning and additional mortality in adulthood. Furthermore, an inspired oxygen fraction of 11% corresponds to extremely high altitude (~5,000 m), at which few humans permanently reside. In addition, adult mice exposed to this degree of hypoxia are unable to reproduce, limiting the ability to study multigenerational effects of hypoxic exposure. Finally, the exposure in this model begins postnatally, whereas humans living at high altitude are also exposed to hypoxia during prenatal development.

In the present study, we therefore tested whether moderate hypoxia, corresponding to 13% FiO₂ or an altitude of approximately 3,600 meters above sea level, initiated before fertilization and maintained throughout gestation and postnatal life, would support survival to adulthood and permit first-generation mice to reproduce under hypoxic conditions. We also sought to determine whether this survivable lifelong hypoxia exposure would produce persistent cardiopulmonary disease, including pulmonary hypertension, RV dysfunction, abnormal alveolar and pulmonary vascular development, and reduced exercise capacity; whether the phenotype differed between first- and second-generation hypoxia-exposed mice; and how it compared with our previously described, more severe postnatal 11% FiO₂ model.

Here, we report the successful development of a multigenerational colony of wild-type C57BL/6 mice maintained in chronic moderate hypoxia at 13% FiO₂. Pregnancy rates were relatively preserved under hypoxic conditions, although not all pregnant females delivered litters. Overall survival to weaning was reduced, particularly in the second-generation mice, primarily because of increased fetal or perinatal loss. Among mice that survived into adulthood, both first- and second-generation hypoxia-exposed animals exhibited alveolar simplification, pulmonary vascular abnormalities, pulmonary hypertension, RV dysfunction, and reduced exercise capacity, with modestly greater pulmonary vascular disease and RV remodeling in second-generation mice.

Methods

Animals and breeding

All animal procedures were approved by the Subcommittee on Research Animal Care and the Institutional Animal Care and Use Committee (IACUC) at Massachusetts General Hospital. Male and female C57BL/6J mice were studied. Animals were obtained from the Jackson Laboratory and bred in-house in the AAALAC-accredited Massachusetts General Hospital animal facility under specific-pathogen-free conditions. Mice were not genetically modified and had not undergone prior experimental procedures.

The parental generation (G0) included 43 female C57BL/6J mice paired for breeding at 8–12 weeks of age. Breeding pairs were placed in hypoxic chambers maintained at 13% FiO₂ at the time of pairing (Fig. 1A). Pregnant females were separated after confirmation of pregnancy. Pregnancy success rate, litter size, and pup survival were recorded for first-generation (G1 mice) offspring born and raised in hypoxia. G1 male and female mice from different litters were paired for breeding under the same hypoxic conditions, and pregnancy success rate, litter size, and survival of second-generation (G2 mice) offspring were recorded. Historical 21% room-air comparator data were obtained from C57BL/6J litters in our prior postnatal hypoxia study, before pups were assigned to hypoxic exposure, and were used as a reference for expected litter size and number of pups alive at birth under normoxic breeding conditions. Birth weight was measured at birth in 13%-G1 pups and compared with historical 21% room-air control pups. Birth weight was not measured in 13%-G2 pups because fewer pups were born alive, and newborn handling was minimized to avoid increasing early postnatal stress or mortality. For assessment of lung development, cardiovascular phenotype, and exercise capacity, age- and sex-matched mice maintained in normoxia (21% FiO₂) were studied as controls.

Figure 1. Multigenerational breeding outcomes and survival of mice exposed to moderate hypoxia (13% FiO₂).

(A) Experimental schematic illustrating breeding of wild-type mice under continuous hypoxia (13% FiO₂) to generate first-generation (G1) and second-generation (G2) offspring. (B) Pregnancy success rates were similar between G0 and G1 females (p = 0.67, Fisher’s exact test). (C) Delivery success among pregnant females was also comparable between groups (p > 0.99, Fisher’s exact test). (D) Birth weight in control and 13%-G1 mice was similar. (E) Litter size at delivery in 13%-G1 mice, and 13%-G2 mice was reduced compared to historical 21% room-air controls. (F) Number of pups alive at birth in historical 21% room-air controls, 13%-G1 litters, and 13%-G2 litters, showing reduced live-born pups in hypoxia-exposed litters and further reduction in G2. (G) Kaplan–Meier survival analysis demonstrated reduced overall survival in G2 mice compared to G1 (log-rank p < 0.0001), primarily driven by increased mortality at or around birth. (H-I) Adult body weight at 10–12 weeks of age. Because males were heavier than females overall, G1 and G2 mice were compared separately within each sex using unpaired t-tests. Body weight did not differ significantly between G1 and G2 mice in either males or females (n = 11–40 per group).

Because the number of available mice differed between generations, not all endpoints were measured in the same individuals. In G1 mice, larger colony size allowed different subsets of animals to be used for echocardiography, invasive hemodynamics, lung mechanics, histology, exercise testing, hemoglobin measurement, and skeletal muscle analysis. In G2 mice, because fewer animals survived to adult phenotyping, multiple measurements were performed in the same mice whenever technically feasible. Assignment of mice to endpoint analyses was based on animal availability, age, sex balance when possible, and tissue requirements.

Chronic hypoxia exposure

Mice were housed in standard cages placed inside semi-sealed normobaric hypoxia chambers connected to an oxygen-control system (Biospherix, Parish, NY, USA). Hypoxia was maintained by controlled nitrogen infusion, with chamber oxygen concentration continuously monitored and adjusted to maintain FiO₂ at 13% ± 0.2%. Internal fans promoted uniform gas mixing, and small ventilation ports allowed continuous gas exchange to prevent CO₂ accumulation. Each chamber was equipped with a CO₂ monitor, and CO₂ concentrations were maintained below 0.4% throughout the exposure period. Chamber temperature and humidity were monitored and maintained within standard housing ranges. Mice were kept on a 12-hour light/dark cycle with ad libitum access to food and water.

Transthoracic echocardiography

Transthoracic echocardiography was performed in lightly anesthetized mice breathing spontaneously under 1–2% isoflurane using a 30-MHz transducer and Vivid 7 ultrasound system (GE Healthcare, Milwaukee, WI, USA). Pulmonary artery acceleration time (PAAT) and the ratio of PAAT to pulmonary ejection time (PAAT/ET) were measured by pulsed-wave Doppler at the basal level in the parasternal short-axis view and used as noninvasive indices of pulmonary arterial pressure. Diastolic RV free-wall thickness (dRVfw) was measured from the parasternal long-axis view. RV systolic function was assessed by measuring tricuspid annular plane systolic excursion (TAPSE) from a focused right-sided four-chamber view obtained by positioning the transducer midway between the conventional apical four-chamber and subcostal views.

Invasive hemodynamics

Mice were anesthetized with intraperitoneal ketamine (120 mg/kg) and fentanyl (0.09 mg/kg). After tracheostomy, rocuronium (1 mg/kg, intraperitoneal) was administered to achieve neuromuscular relaxation. Mice were mechanically ventilated in volume-controlled mode at a respiratory rate of 100 breaths/min, tidal volume of 10 mL/kg, positive end-expiratory pressure of 2 cmH₂O, and FiO₂ of 21%. A PE-10 catheter was placed in the carotid artery to measure mean systemic arterial pressure. After median thoracotomy, a small puncture was made in the right ventricular free wall near the apex using a 25-gauge needle, and a 1.4-F pressure-volume catheter (SPR-839, Millar Instruments) was introduced directly into the right ventricle (RV) to measure RVSP.

Fulton’s ratio

After euthanasia, hearts were excised and atrial tissue was removed. The RV was carefully dissected from the left ventricle plus septum (LV+S). RV hypertrophy was quantified by calculating the ratio of RV weight to LV+S weight (Fulton’s ratio).

Lung mechanical properties evaluation

Lung pressure-volume curves were generated at thoracotomy by inflating the lungs at a constant rate of 100 μL/s while continuously measuring airway pressure. Pressure-volume relationships were constructed from airway pressure values measured at 100-μL volume increments. Quasi-static lung compliance was calculated as the ratio of lung volume to airway pressure at an inflation volume of 1000 μL.

Hemoglobin measurement

Hemoglobin concentration was measured from tail-vein blood using an ABL90 FLEX blood gas analyzer (Radiometer Medical ApS, Brønshøj, Denmark). Blood was collected from a small tail-vein cut into a heparinized microcapillary tube and analyzed immediately.

Treadmill endurance test

Exercise capacity was assessed in 12-week-old mice using a six-lane motorized Exer Treadmill system (Columbus Instruments, Columbus, OH, USA; lane dimensions, 43.8 × 6 × 12.7 cm). Mice ran on a treadmill set at a 10° incline. After a 5-minute acclimation and warm-up period, during which running speed was gradually increased from 5 to 15 m/min, mice entered a graded exercise phase in which speed increased by 2 m/min each minute until exhaustion. Mice that drifted toward the rear of the treadmill were encouraged to continue running with gentle prodding. Exhaustion was defined as the inability to maintain running pace despite repeated encouragement. Running distance, exercise duration, and total work were recorded. The treadmill was cleaned with 70% ethanol between tests.

Lung tissue preparation and image acquisition.

After euthanasia, lungs were inflated through the airways with 4% paraformaldehyde in PBS, harvested, fixed in 4% paraformaldehyde for 24 hours, dehydrated, paraffin embedded, sectioned, and deparaffinized before staining. Hematoxylin–eosin staining was used to assess lung parenchymal architecture. For immunofluorescence, antigen retrieval was performed using Retrievagen A, pH 6.0 (550524, BD Biosciences). Tissue sections were blocked with 5% goat serum in SuperBlock Blocking Buffer in PBS (37517, Thermo Fisher Scientific) for 1 hour at room temperature and incubated overnight at 4°C with a rabbit polyclonal antibody against von Willebrand factor (vWF; GA527, Agilent/Dako), followed by a goat anti-rabbit immunoglobulin G secondary antibody conjugated to Alexa Fluor 488 (A-11034, Thermo Fisher Scientific; 1:100) for 1 hour at room temperature. Sections were then stained with a mouse monoclonal antibody against α-smooth muscle actin (α-SMA; clone 1A4, 14-9760-82, Thermo Fisher Scientific; 1:200), followed by a goat anti-mouse immunoglobulin G secondary antibody conjugated to Alexa Fluor 555 (A-32727, Thermo Fisher Scientific; 1:200). Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; D21490, Thermo Fisher Scientific; 1:3000). Slides were scanned using a NanoZoomer 2.0-RS digital slide scanner (Hamamatsu, Japan). For each microscopy experiment, images from control and hypoxia-exposed groups were acquired during the same imaging session using identical microscope settings, including objective, exposure time, illumination intensity, gain, and acquisition parameters.

Lung morphometric and vascular remodeling quantification

An investigator masked to treatment group performed all image analyses using NDP.view2 software. Mean linear intercept was quantified on hematoxylin–eosin-stained lung sections using a 1-mm test line placed in three randomly selected peripheral parenchymal fields per slide, avoiding large airways, large vessels, pleura, tissue artifacts, and non-representative enlarged airspaces. The mean value from the three fields was calculated for each mouse and used for statistical analysis. For pulmonary arteriole muscularization, vessels with an external diameter of 25–50 μm were identified across the whole lung section on the vWF channel, excluding vessels adjacent to conducting airways. At least 50 round-profile distal arterioles were analyzed per animal. Candidate vessels were first marked on the vWF channel, the image was then switched to the α-SMA channel, and each arteriole was classified as non-muscularized (NM), partially muscularized (PM), or fully muscularized (FM) based on circumferential α-SMA coverage of <25%, 25–75%, or >75%, respectively. For peripheral vascular density, three regions of interest from different lung lobes were selected using the DAPI channel. Regions of interest ranged from 0.7 to 1.9 mm² and were chosen in peripheral lung parenchyma while avoiding large airways, large vessels and sectioning artifacts. The image was then switched to the vWF channel, and vWF-positive vascular profiles <25 μm in diameter with a clearly detectable lumen were manually marked and counted. Vascular density was calculated as the number of vWF-positive vessels per mm² of analyzed tissue area.

Skeletal muscle tissue preparation and image acquisition.

To assess peripheral skeletal muscle vascularization and fiber size, muscles from the right posterior distal hindlimb compartment were harvested at euthanasia, fixed in 4% paraformaldehyde for 24 hours, transferred to 70% ethanol, processed for paraffin embedding, and sectioned transversely. After antigen retrieval, sections were stained with wheat germ agglutinin (WGA), Alexa Fluor 594 conjugate (W11262, Thermo Fisher Scientific; 1:200), diluted in HBSS, for 10 minutes at room temperature to delineate muscle fiber membranes. Sections were then incubated overnight at 4°C with an antibody against cluster of differentiation 31/platelet endothelial cell adhesion molecule-1 (CD31/PECAM-1; AF3628, R&D Systems; 1:50), diluted in PBS, followed by a donkey anti-goat immunoglobulin G secondary antibody conjugated to Alexa Fluor 488 (A-11055, Thermo Fisher Scientific; 1:100) for 1 hour at room temperature. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; D21490, Thermo Fisher Scientific; 1:3000), diluted in PBS. Slides were scanned using a NanoZoomer 2.0-RS digital slide scanner (Hamamatsu, Japan).

Skeletal muscle capillary density and fiber size quantification.

Whole-slide images were reviewed using NDP.view2 software by an investigator blinded to experimental group allocation. Representative cross-sectional regions within the gastrocnemius muscle were selected from the whole-slide scan while avoiding large vessels, tissue folds, and areas with sectioning artifact. In each analyzed region, CD31-positive capillaries adjacent to muscle fibers and WGA-defined muscle fibers were manually counted. Capillary density was expressed as capillaries per muscle fiber and capillaries per mm² of analyzed tissue area. Muscle fiber cross-sectional area was quantified using CellProfiler software version 4.2.6 (Broad Institute) on WGA-stained transverse sections. Three 20× images were acquired from one section per mouse. Images were inverted, and WGA-positive sarcolemmal borders were used to segment individual myofibers using a global thresholding approach. Fiber cross-sectional area was measured using the MeasureObjectSizeShape module and converted from pixels² to μm² using the calibrated pixel size of the 20× NanoZoomer images. Objects with areas <500 μm² or >6000 μm² were excluded to remove debris, interstitial segmentation artifacts, and merged objects. Segmentation quality was assessed by visual review of object overlays. For each image, the median fiber cross-sectional area was calculated, and image-level medians were summarized for each mouse. Mouse-level values were used for statistical comparisons between groups.

Statistical Analysis

Statistical analyses were performed using GraphPad Prism version 10.3.1. Normality was assessed using the Shapiro–Wilk test. Pregnancy and delivery success rates were compared using Fisher’s exact test. Survival was analyzed as the time from birth to death. Kaplan–Meier curves were generated for each exposure group, and differences between groups were assessed using the log-rank test. Animals surviving to the study endpoint were censored at their final observation. As approved by the IACUC, deaths were counted as events in the survival analysis, whereas mice that remained alive at the end of the study were censored at their last recorded time point.

Echocardiographic parameters, including PAAT, PAAT/ET, TAPSE, and dRVfw, as well as RVSP, Fulton’s ratio, pulmonary vessel muscularization, pulmonary vascular density, MLI, lung compliance and hemoglobin concentration were analyzed using one-way analysis of variance (ANOVA) with Bonferroni post hoc correction for multiple comparisons. Pressure-volume curves were analyzed using two-way repeated-measures ANOVA followed by Bonferroni’s multiple-comparisons test. Body weight was compared using unpaired Student’s t-tests. Skeletal muscle capillary density, expressed as capillaries per mm² and capillaries per fiber, was analyzed using unpaired Student’s t-tests. Litter size was analyzed using the Mann–Whitney U test. Muscle fiber cross-sectional area was summarized as the median value for each animal and analyzed using the Mann–Whitney U test.

A P value <0.05 was considered statistically significant. Data are presented as mean ± standard deviation (SD) unless specified otherwise. Animals that died during hypoxia exposure were included only in the survival analysis, because early death precluded subsequent physiological or histological assessment. No other a priori inclusion or exclusion criteria were defined, and no animals were excluded from the remaining analyses.

Sample size estimates were based on our prior experiments and on published studies using comparable models of chronic low-oxygen exposure. Four primary outcome domains were predefined: 1) pregnancy success, delivery success, and offspring survival; 2) cardiovascular structure and function, assessed by echocardiography, Fulton’s ratio, and invasive hemodynamic measurements; 3) lung development and pulmonary vascular remodeling, assessed by histopathology and immunofluorescence; and 4) exercise capacity, hemoglobin concentration, skeletal muscle capillary density, and muscle fiber cross-sectional area.

Results

Pregnancy outcomes and survival of first- and second-generation mice in hypoxia

Pregnancy success rates were similar between G0 and G1 females (38/43 vs. 13/16; 88% vs. 81%; P = 0.67). Delivery success among pregnant females was also comparable between groups (28/38 vs. 10/13; 74% vs. 77%; P > 0.99) (Fig. 1B-C). The average time from initiation of breeding under hypoxic conditions to delivery was 46 ± 20 days for G0 females and 33 ± 12 days for G1 females (P = 0.09).

Birth weight did not differ between 13%-G1 pups and 21% room-air controls (Fig. 1D). Compared with historical 21% room-air litters from our prior postnatal hypoxia study, assessed before hypoxic exposure, both 13%-G1 and 13%-G2 mice had smaller litters at delivery (Historical 21%: 6 [5–8]; 13%-G1: 5 [4–7]; 13%-G2: 4 [2–6], median [IQR]; Fig. 1E). Litter size was smaller in 13%-G2 than in 13%-G1 mice, although this difference was not statistically significant. Moreover, the number of pups alive at birth was reduced in both hypoxia-exposed generations and was further decreased in 13%-G2 compared with 13%-G1 litters (Historical 21%: 6 [5–8]; 13%-G1: 5 [3–7]; 13%-G2: 3 [1–4], median [IQR]; Fig. 1F), indicating increased fetal or perinatal loss in the second generation.

Survival analysis demonstrated significantly reduced overall survival in G2 compared to G1 mice (p<0.0001). This difference was primarily driven by increased mortality at or before birth, as reflected by a marked drop in survival at day 0 in G2 pups (~75%) compared to G1 pups (>95%). Pre-weaning survival among pups born alive was similar between G1 and G2 mice. However, because of increased perinatal mortality in G2 litters, fewer pups ultimately survived to weaning in the G2 group compared to G1, with approximately 50% of G2 pups surviving to weaning compared to ~75% in G1 pups. Beyond the weaning period, survival remained relatively stable in both groups until age 120 days. Adult body weight at 10–12 weeks of age was similar between surviving G1 and G2 mice within each sex (Fig. 1G-H).

Overall, these findings demonstrate that chronic exposure to moderate hypoxia (13% FiO₂) supports multigenerational breeding with preserved pregnancy rates. However, second-generation offspring exhibit reduced perinatal viability, suggesting impaired reproductive fitness in G1 females, while postnatal survival of surviving pups remains largely preserved.

Lifelong 13% hypoxia induces pulmonary hypertension and RV remodeling, with greater severity in second-generation mice

To assess the impact of lifelong hypoxia on pulmonary vascular pressure and RV remodeling, we performed transthoracic echocardiography and invasive hemodynamic measurements at 8 weeks of age. RV hypertrophy was assessed noninvasively by measuring the diastolic RV free-wall thickness and confirmed by Fulton’s ratio. All measurements were compared with age-matched control mice maintained in room air.

After 8 weeks of exposure to 13% FiO₂, PAAT and PAAT/ET were reduced in both 13%-G1 and 13%-G2 mice compared with room-air controls, consistent with increased pulmonary vascular resistance (Fig. 2A-B). Invasive RV pressure measurements confirmed the development of pulmonary hypertension in both hypoxia-exposed groups compared with 21% controls, with RVSP further increased in 13%-G2 mice compared with 13%-G1 mice (RVSP, 21% control vs. 13%-G1 vs. 13%-G2: 28±4 vs. 45±5 vs. 56±4 mmHg; P < 0.0001 for each hypoxic group vs. control; P = 0.0053 for 13%-G2 vs. 13%-G1; Fig. 2C).

Figure 2. Lifelong exposure to 13% hypoxia induces pulmonary hypertension, RV systolic dysfunction, and RV hypertrophy.

At 8 weeks of age, mice exposed to lifelong 13% FiO₂ hypoxia in the first generation (13%-G1) or second generation (13%-G2) were compared with age-matched control mice maintained in room air (21% FiO₂). Transthoracic echocardiography demonstrated reduced pulmonary artery acceleration time (PAAT) (A) and reduced PAAT/ejection time ratio (PAAT/ET) (B) in both hypoxia-exposed groups compared with controls (n = 5–11 per group). Invasive hemodynamic measurements confirmed increased RVSP in both 13%-G1 and 13%-G2 mice, with RVSP further increased in 13%-G2 mice compared with 13%-G1 mice (C) (n = 5–16). Tricuspid annular plane systolic excursion (TAPSE) was reduced in both hypoxic groups, indicating impaired RV systolic function (D) (n = 5–11 per group). Lifelong hypoxia also induced RV hypertrophy, assessed by diastolic RV free-wall thickness (dRVfw) (E, n = 5–10 per group) and Fulton’s ratio (F, n = 9–19 per group). dRVfw and Fulton’s ratio were increased in hypoxia-exposed mice, with greater hypertrophy in 13%-G2 mice than in 13%-G1 mice. Closed circles indicate male mice; open circles indicate female mice. Statistical comparisons were performed by one-way ANOVA with Bonferroni correction for multiple comparisons.

RV systolic function was also impaired after lifelong hypoxia exposure, as shown by reduced TAPSE in both 13%-G1 and 13%-G2 mice compared with controls (Fig. 2D). In parallel, lifelong hypoxia induced RV hypertrophy. Diastolic RV free-wall thickness was increased in hypoxia-exposed mice, with a greater increase in 13%-G2 mice than in 13%-G1 mice (Fig. 2E). Fulton’s ratio confirmed marked RV hypertrophy in both hypoxic groups and was significantly higher in 13%-G2 mice than in 13%-G1 mice (Fig. 2F). Because the G1–G2 difference appeared larger by echocardiographic RV free-wall thickness than by Fulton’s ratio, we also examined absolute LV+S weight, the denominator of Fulton’s ratio. LV+S weight was slightly increased in 13%-G2 mice, suggesting that Fulton’s ratio may underestimate the magnitude of RV hypertrophy in this group and partially explaining the more modest G1–G2 separation compared with RV free-wall thickness (Fig. 2G).

These findings indicate that lifelong exposure to 13% hypoxia causes pulmonary hypertension, RV systolic dysfunction, and RV hypertrophy, with G2 mice showing modestly greater pulmonary vascular pressure and RV remodeling than G1 mice.

Lifelong 13% hypoxia impairs alveolar and pulmonary vascular development

We next assessed whether the sustained cardiopulmonary phenotype induced by lifelong 13% hypoxia was accompanied by abnormalities in alveolarization, lung mechanics, and distal pulmonary vascular remodeling.

Histological analysis of lungs at 8 weeks of age demonstrated alveolar simplification in both 13%-G1 and 13%-G2 mice compared with room-air controls, as reflected by an increased MLI (Fig. 3 A-B). Alveolar simplification was accompanied by a significant increase in lung compliance (Fig. 3C-D).

Figure 3. Lifelong 13% hypoxia impairs alveolar development and increases lung compliance.

(A) Representative hematoxylin and eosin–stained lung sections from control mice maintained in room air (21% FiO₂) and from first- and second-generation mice exposed to lifelong 13% FiO₂ hypoxia (13%-G1 and 13%-G2). Lungs from hypoxic mice showed enlarged distal airspaces and simplified alveolar architecture compared with room-air controls. (B) Quantification of mean linear intercept (MLI) demonstrated increased airspace size in both 13%-G1 and 13%-G2 mice compared with controls, consistent with alveolar simplification (n = 3–8 per group). (C) Pressure–volume curves demonstrated increased lung distensibility in mice exposed to lifelong 13% hypoxia (n = 5–8 per group). (D) Quasi-static lung compliance, assessed at a lung volume of 1000 μL, was increased in hypoxic mice compared with controls (n = 5–8 per group). Closed circles indicate male mice; open circles indicate female mice. Statistical comparisons in panels B and D were performed by one-way ANOVA with Bonferroni correction for multiple comparisons. Pressure–volume curves in panel C were analyzed using two-way repeated-measures ANOVA followed by Bonferroni’s multiple-comparisons test: *P < 0.05, **P < 0.01, ***P < 0.001 vs. 13%-G1, ## P < 0.01, ### P < 0.001 vs. 13%-G2.

To determine whether lifelong hypoxia altered pulmonary vascular development and remodeling, we quantified muscularization of peripheral small pulmonary arterioles with external diameters of 25–50 μm. In parallel, peripheral microvascular density was assessed by counting vWF-positive vessels with luminal diameters <25 μm in predefined peripheral lung regions and normalizing the counts to the analyzed tissue area. Compared with room-air controls, both 13%-G1 and 13%-G2 mice showed a reduced proportion of NM arterioles, whereas the proportion of FM arterioles was increased in both 13%-G1 and 13%-G2 mice. In contrast, the proportion of PM arterioles was not markedly different between hypoxia-exposed groups. Peripheral microvascular density was reduced in both 13%-G1 and 13%-G2 mice compared with controls (Fig. 4A-D). Compared with 13%-G1 mice, 13%-G2 mice showed a higher proportion of FM arterioles (P = 0.0547) and lower peripheral microvascular density (P = 0.2684), although these differences did not reach statistical significance.

Figure 4. Lifelong hypoxia impairs pulmonary vascular development and remodeling across generations.

Pulmonary vascular structure was assessed in lungs from control mice maintained in room air (21% FiO₂) and from 13%-G1 and 13%-G2 mice at 8 weeks of age (n = 5–8 per group). Compared with 21% controls, the proportion of non-muscularized distal pulmonary arterioles was reduced in hypoxia-exposed mice, with a slight further reduction in 13%-G2 compared with 13%-G1 mice that did not reach statistical significance (P = 0.0531; A). The proportion of partially muscularized arterioles was not significantly different across groups (B). Fully muscularized arterioles were increased in both hypoxia-exposed groups compared with controls, with a slight further increase in 13%-G2 compared with 13%-G1 mice that did not reach statistical significance (P = 0.0547; C). Peripheral vascular density was reduced in both 13%-G1 and 13%-G2 mice compared with controls (D). Closed circles indicate male mice; open circles indicate female mice. Statistical comparisons were performed by one-way ANOVA with Bonferroni correction for multiple comparisons.

Together, these findings indicate that lifelong exposure to 13% hypoxia impairs alveolar development and pulmonary vascular growth and promotes small-vessel muscularization. Some pulmonary vascular endpoints appeared slightly more abnormal in 13%-G2 mice than in 13%-G1 mice, although these differences did not reach statistical significance.

Reduced exercise capacity despite preserved skeletal muscle capillary density

Reduced exercise capacity is a characteristic feature of maladaptation to chronic hypoxia and high-altitude exposure. Therefore, we investigated whether lifelong exposure to 13% hypoxia impaired functional exercise performance across generations. At 12 weeks of age, both 13%-G1 and 13%-G2 mice demonstrated reduced exercise capacity during endurance treadmill testing performed in room air, as reflected by shorter running distance, reduced exercise duration, and lower total work compared with room-air controls (Fig. 5A-C).

Figure 5. Lifelong hypoxia reduces exercise capacity and increases hemoglobin concentration.

Exercise capacity was assessed at 12 weeks of age by endurance treadmill testing in room air. Compared with 21% control mice, both 13%-G1 and 13%-G2 showed reduced exercise duration (A), running distance (B), and total work (C) (n = 5–10 mice per group). Hemoglobin concentration was increased after lifelong hypoxia exposure in 13%-G1 and 13%-G2 mice compared to controls (D, n = 12–16 per group). Closed circles indicate male mice; open circles indicate female mice. Statistical comparisons were performed using one-way ANOVA with Bonferroni correction for multiple comparisons.

To explore potential contributors to impaired exercise performance, we assessed systemic and peripheral determinants of oxygen delivery. Hemoglobin concentration was increased in both 13%-G1 and 13%-G2 mice compared with room-air controls, consistent with an erythropoietic response to chronic hypoxemia (Fig. 5D). We also quantified capillary density in gastrocnemius muscle from room-air controls and 13%-G1 mice to determine whether reduced exercise capacity was associated with overt changes in peripheral skeletal muscle vascularization. Capillary density, expressed as capillaries per fiber or capillaries per mm² of analyzed tissue area, did not differ significantly between groups. Mouse-level myofiber cross-sectional area was also not significantly different (Supplemental Fig. 1A-D).

Together, these findings indicate that lifelong hypoxia reduced exercise capacity during room-air testing despite increased hemoglobin concentration and without a detectable reduction in skeletal muscle capillary density in the sampled regions.

Discussion

In this study, we developed and characterized a multigenerational mouse model of lifelong moderate hypoxia. Continuous exposure to 13% FiO₂ from conception through adulthood permitted breeding over two generations, while producing a sustained cardiopulmonary phenotype in surviving offspring. Both G1 and G2 mice developed pulmonary hypertension, RV systolic dysfunction, RV hypertrophy, and reduced exercise capacity. These functional abnormalities were accompanied by abnormal lung structure, including alveolar simplification, increased lung compliance, increased muscularization of small pulmonary arterioles, and reduced peripheral pulmonary vascular density. In contrast, reduced exercise capacity was not accompanied by a detectable reduction in skeletal muscle capillary density. Although both generations were affected, G2 mice showed evidence of modestly greater pulmonary vascular disease and RV remodeling than G1 mice, without increased post-weaning or adult mortality during the period of observation.

In this study we did not detect consistent sex-dependent differences, however this finding should be interpreted cautiously. Prior mouse and rat studies have shown that sex can modify the long-term effects of perinatal hypoxia on pulmonary vascular function and remodeling(18, 19). Our sex-stratified analyses were exploratory and limited by sample size, particularly in G2 mice, and therefore were not powered to exclude biologically meaningful sex-specific effects.

A central observation of this study is that reproductive and early-life survival outcomes differed markedly between lifelong exposure to 11% and 13% FiO₂. In our prior lifelong 11% FiO₂ experiments(16), breeding under hypoxic conditions remained possible, with pregnancy rates exceeding 80%, but approximately half of pregnancies underwent complete resorption and no pups survived beyond postnatal day 2. In contrast, lifelong exposure to 13% FiO₂ allowed a substantial proportion of pups to survive to weaning and adulthood, and permitted G1 mice to reproduce under the same hypoxic conditions. These findings suggest that developmental hypoxia may have a steep threshold effect at the level of pregnancy maintenance, fetal oxygen delivery, placental function, and/or early neonatal cardiopulmonary transition. Approximate altitude equivalents illustrate that these exposures represent distinct levels of hypoxic stress: 11% FiO₂ corresponds to roughly 5,000–5,200 m, whereas 13% FiO₂ corresponds to roughly 3,600–3,900 m above sea level. The 13% FiO₂ exposure remains physiologically relevant, as an estimated 14.4 million people live at elevations ≥3,500 m and 6.4 million at elevations ≥4,000 m(23).

Several species-specific differences should be considered when interpreting this model. Mice and humans differ in pulmonary vascular structure and hypoxic pulmonary vascular responses, and murine and human pregnancy differ in placentation, fetal number, developmental timing, and compensatory physiology. Nevertheless, the reproductive and fetal outcomes observed in this model are conceptually consistent with human high-altitude pregnancy: limited oxygen delivery during pregnancy can reduce fetal growth and litter size (20), while high-altitude hypoxia in humans is associated with reduced birth weight and a modestly increased risk of spontaneous preterm birth(21, 22). Despite these limitations, mice provide a practical platform for lifelong and multigenerational hypoxia studies because relatively large breeding cohorts can be housed inside oxygen-controlled chambers under controlled experimental conditions.

An important finding of this study is that the survival phenotype was not uniform across developmental stages or generations. Pregnancy rates and delivery success were relatively preserved, but G2 litters had fewer pups alive at birth, indicating increased fetal or perinatal loss. The mechanisms responsible for this early loss were not determined. In addition to impaired fetal oxygen delivery during gestation, abnormalities of the immediate neonatal cardiopulmonary transition could contribute to death shortly after birth. For example, failure of normal ductus arteriosus closure under chronic hypoxic conditions could impair postnatal circulatory adaptation(24). However, such a mechanism would be expected to affect both G1 and G2 pups exposed to the same hypoxic environment, whereas mortality was greater in G2. This pattern suggests that maternal factors may also contribute to the generation-specific vulnerability. A key difference between G1 and G2 offspring was the developmental history of their mothers: G1 mice were born to G0 females that had been introduced into hypoxia as adults, whereas G2 mice were born to G1 females that had themselves developed under lifelong hypoxic conditions. Because G1 mice exhibited abnormal lung development after lifelong hypoxia, impaired maternal gas exchange or reduced cardiopulmonary reserve during pregnancy may have further limited oxygen delivery to G2 fetuses. Such a mechanism could contribute not only to increased fetal or perinatal loss in G2 litters, but also to the modestly worse pulmonary vascular disease and RV remodeling observed in surviving G2 mice. Future studies assessing maternal oxygenation, placental structure and perfusion, fetal growth, and fetal cardiopulmonary development will be needed to define the mechanisms underlying this generation-specific vulnerability. Among pups born alive, however, pre-weaning survival was similar between G1 and G2 mice, and survival after weaning remained largely stable in both generations. Although the cause of pre-weaning death was not determined, mortality clustered between postnatal days 14 and 21, a period when maternal milk production normally declines as pups approach weaning(25); chronic hypoxia may further compromise maternal lactation and thereby reduce nutritional support during this vulnerable transition(26).

Together with our previous work(16, 17), the present findings support the concept that both the timing and severity of hypoxic exposure shape the developmental cardiopulmonary phenotype, including both lung developmental injury and survival. In our prior studies, postnatal exposure to 11% FiO₂ produced the most severe phenotype, with marked disruption of lung development, RVSP values of approximately 70–80 mmHg, Fulton’s ratio of approximately 55–60%, and reduced survival and failure to reproduce. Adult-onset 11% FiO₂ hypoxia produced a milder response, with RVSP values of approximately 30–40 mmHg and Fulton’s ratio of approximately 35–40%, consistent with pulmonary vascular remodeling in the absence of developmental lung injury. In comparison, lifelong 13% FiO₂ hypoxia beginning before birth produced an intermediate phenotype, with impaired alveolar and pulmonary vascular development, RVSP values of approximately 50–60 mmHg, and Fulton’s ratios of approximately 40–45%, compared with control values of approximately 25–30 mmHg for RVSP and 25–30% for Fulton’s ratio. Importantly, unlike the more severe 11% developmental exposure, lifelong exposure to 13% FiO₂ remained compatible with survival to adulthood and reproduction across generations. Thus, lifelong 13% hypoxia provides a useful intermediate developmental model: severe enough to impair lung maturation and produce persistent pulmonary vascular disease, but survivable enough to study prenatal exposure, multigenerational effects, and long-term cardiopulmonary outcomes.

Exercise capacity provides an integrated functional readout of cardiopulmonary and peripheral oxygen transport and utilization and is commonly used as an indirect measure of whole-body aerobic capacity(27). Maximal exercise performance is determined by multiple components of the oxygen cascade. Key determinants include arterial oxygen saturation and PaO₂, hemoglobin concentration and oxygen-carrying capacity, hemoglobin–oxygen affinity and oxygen unloading, cardiac output, tissue perfusion, capillary density, and mitochondrial oxygen utilization(28, 29). In the present study, lifelong 13% hypoxia reduced treadmill performance despite increased hemoglobin concentration; in an exploratory analysis of 13%-G1 mice, this was not accompanied by detectable skeletal muscle capillary rarefaction in the sampled gastrocnemius regions. The gastrocnemius is a major locomotor muscle recruited during treadmill running, but it does not capture the full heterogeneity of hindlimb skeletal muscle. In addition, our analysis was limited to capillary density and myofiber cross-sectional area in selected cross-sectional regions and did not assess mitochondrial function, fiber-type composition, muscle oxygen extraction, or exercise-induced blood flow. While these findings suggest that exercise limitation was not primarily explained by a failure to increase oxygen-carrying capacity or by reduced gastrocnemius capillarization, the possibility of other skeletal muscle or systemic contributors cannot be excluded. The combination of alveolar simplification, reduced peripheral pulmonary vascular density, pulmonary hypertension, RV hypertrophy, and reduced TAPSE points toward impaired pulmonary oxygen uptake and limited cardiopulmonary reserve as likely contributors. This interpretation is consistent with our prior 11% postnatal hypoxia model, in which impairment of lung development was associated with lower PaO₂ under hypoxic conditions(16). Future studies measuring arterial oxygen saturation during exercise, cardiac output, hemoglobin–oxygen dissociation properties, and tissue oxygenation will be important to define the relative contributions of impaired gas exchange, altered oxygen unloading, and circulatory limitation to reduced exercise capacity in developmental hypoxia.

This study has several limitations. First, the mechanisms underlying reduced G2 perinatal viability remain undefined and may involve maternal, placental, or fetal adaptations to chronic hypoxia. Second, because only mice that survived to adulthood were studied, we could not determine whether animals dying before or shortly after birth had more severe lung, pulmonary vascular, or cardiac abnormalities. Third, although treadmill endurance provides an integrated measure of functional capacity, oxygen consumption, arterial oxygen saturation, and cardiac output were not measured during exercise; therefore, the relative contributions of impaired gas exchange, pulmonary vascular disease, RV dysfunction, and peripheral oxygen utilization could not be directly defined. Fourth, tissue and mitochondrial oxygenation were not directly assessed in lifelong 13% hypoxic mice. An additional limitation is that the 21% room-air controls used for adult phenotyping were age- and sex-matched C57BL/6J mice obtained from The Jackson Laboratory and maintained in our facility before testing, rather than contemporaneously bred in-house normoxic controls. Finally, the colony was followed through the second generation only, and additional studies will be needed to determine whether the phenotype progresses, stabilizes, or adapts in subsequent generations.

In conclusion, lifelong exposure to 13% FiO₂ establishes a survivable but pathologic model of developmental hypoxia. This model permits multigenerational breeding while producing impaired alveolar and pulmonary vascular development, pulmonary hypertension, RV dysfunction, erythrocytosis, and reduced exercise capacity. These features make lifelong 13% hypoxia a useful platform for investigating mechanisms that link chronic developmental hypoxia to persistent cardiopulmonary disease, and for testing therapeutic strategies aimed at improving lung development, pulmonary vascular remodeling, RV function, and oxygen transport. This model may be particularly relevant to children exposed to chronic hypoxemia during development, including those living at high altitude or affected by developmental lung disease.

Supplementary Material

Supplemental Figure 1: https://doi.org/10.6084/m9.figshare.32860406.

Abbreviations

α-SMA

alpha-smooth muscle actin

AAALAC

Association for Assessment and Accreditation of Laboratory Animal Care

dRVfw

diastolic right ventricular free-wall thickness

ET

ejection time

FiO₂

inspired oxygen fraction

FM

fully muscularized

G0

parental generation

G1

first generation

G2

second generation

IACUC

Institutional Animal Care and Use Committee

LV+S

left ventricle plus septum

MLI

mean linear intercept

NM

non-muscularized

O₂

oxygen

PAAT

pulmonary artery acceleration time

PH

pulmonary hypertension

PM

partially muscularized

RV

right ventricle

RVSP

right ventricular systolic pressure

TAPSE

tricuspid annular plane systolic excursion

vWF

von Willebrand factor

WGA

wheat germ agglutinin

Data Availability

The datasets generated and analyzed during the current study are available from the corresponding author upon reasonable request.