Racing against time: leveraging preclinical models to understand pulmonary susceptibility to perinatal acetaminophen exposures

Abstract

Over the past decade, clinicians have increasingly prescribed acetaminophen (APAP) for patients in the neonatal intensive care unit (NICU). Acetaminophen has been shown to reduce postoperative opiate burden, and may provide similar efficacy for closure of the patent ductus arteriosus (PDA) as nonsteroidal anti-inflammatory drugs (NSAIDs). Despite these potential benefits, APAP exposures have spread to increasingly less mature infants, a highly vulnerable population for whom robust pharmacokinetic and pharmacodynamic data for APAP are lacking. Concerningly, preclinical studies suggest that perinatal APAP exposures may result in unanticipated adverse effects that are unique to the developing lung. In this review, we discuss the clinical observations linking APAP exposures to adverse respiratory outcomes and the preclinical data demonstrating a developmental susceptibility to APAP-induced lung injury. We show how clinical observations linking perinatal APAP exposures to pulmonary injury have been taken to the bench to produce important insights into the potential mechanisms underlying these findings. We argue that the available data support a more cautious approach to APAP use in the NICU until large randomized controlled trials provide appropriate safety and efficacy data.

CLINICAL CONSIDERATIONS: BEDSIDE CONCERNS LEAD TO THE BENCH

The potential relationship between acetaminophen (APAP) exposures and pulmonary morbidity has been interrogated in both children and adults. Although epidemiologic studies have linked childhood and adult APAP exposures with an increased risk of developing asthma and chronic obstructive pulmonary disease (COPD) (1–9), these relationships are far from conclusive (10–16). Similarly, several studies and meta-analyses of these data have linked prenatal APAP exposure following maternal intake to adverse pulmonary symptoms in childhood (6, 17–38). Despite the consistency in the observation that APAP exposure increases the risk of pulmonary morbidity, multiple coexposures, including family history, viral illness, and smoke/pollution exposure, may confound the observed associations (16, 23–25, 39, 40). Furthermore, many of these large epidemiologic studies use chart documentation of a physician diagnosis or medication prescription as a proxy for an asthma diagnosis, rather than through the use of objective diagnostic testing (19–37). Therefore, the relationship APAP exposure and lung disease remains debated.

Over the past decade, clinicians have exposed an entirely new cohort of patients to APAP. The initial indication for APAP use in the neonatal intensive care unit (NICU) was analgesia, with the goal of reducing opiate burden (41). Increasing use in the NICU was accompanied by greater exposures in premature neonates where safety, pharmacokinetic, and pharmacodynamic data are lacking (42–45). Critically, a 2011 case series identified a temporal relationship between APAP and patent ductus arteriosus (PDA) closure, establishing a potential new therapeutic target for APAP use (46). By 2018, APAP was the 10th most common medication used in US Pediatrix NICUs, the 17th most common medication used in extremely low birth weight (ELBW) infants, and the medication with the 8th greatest absolute increase in use between 2010 and 2018 (47). This finding is not limited to the United States. In French NICUs, only vitamin K and vitamin D are prescribed more frequently than APAP, with exposures among infants born at less than 27 wk of gestation approaching 70% (48).

PDA treatment is a primary driving factor increasing exposures among the most immature neonates. Recent surveys in the United Kingdom and Australia/New Zealand demonstrated that >80% of neonatologists had used APAP to treat the PDA (49, 50). Furthermore, it appears that the burden of APAP exposure falls on the most immature newborn infants. Data from France, Italy, and the United States demonstrate that in extremely low gestational age newborns (ELGANs), APAP as a primary agent to treat PDA ranges from ∼10%–50% (51–53). Given the increased use of APAP in immature neonates, it would be reasonable to robustly assess the relationship between exposure and pulmonary morbidities. Unfortunately, meaningful pulmonary outcomes have not been routinely reported. Although several published studies of APAP for PDA report rates of bronchopulmonary dysplasia (BPD), this is not because of any proposed injurious effect on the lung, but rather because BPD is a commonly assessed outcome in studies evaluation PDA treatments. With these limited outcomes reported, we have a very limited ability to interrogate the relationship between APAP and lung injury in the preterm infant.

Meta-Analysis and Meta-Regression of Published APAP Trials for PDA Closure

Four published meta-analyses of APAP for PDA report no increase in BPD following APAP exposure (54–57). However, the trials included in these meta-analyses enrolled few infants born <28 wk’ gestation and were not well suited to examine a possible relationship between APAP exposure and pulmonary morbidity in a population of infants with low baseline risks of BPD (58). To overcome this limitation, we assessed whether the relationship between APAP exposure and BPD was modified by gestational age. To do this, we performed a systematic literature search in PubMed to identify all randomized controlled trials that evaluated acetaminophen for prevention or treatment of a PDA. Of 196 identified manuscripts, 12 trials comprising 1,001 subjects were included in the final analyses (Fig. 1) (59–70). Mean gestational ages of enrolled subjects ranged from 25.3 to 33.6 wk in the analyzed studies. The outcome of BPD, although variably defined, was reported for 930 subjects. Averaged across all trials, APAP exposure did not significantly increase the risk of BPD [Fig. 2A; RR 1.07 (95% CI 0.83, 1.37)]. However, meta-regression demonstrated a possible inverse relationship between GA and BPD risk. For each week decrease in mean study GA, the risk difference for developing BPD with APAP exposure increased by 1.2% (range 3% to −0.5%; P = 0.17; Fig. 2B). When limited to the eight studies that compared oral APAP to oral ibuprofen, this weekly risk increased to 1.7% (range 4.0% to −0.6; P = 0.14; Fig. 2C). These results bordered on statistical significance, but the power to detect true differences at lower gestational ages was limited by the fact that not a single trial limited enrollment to high risk neonates <28 wk gestation. The end result is that a very low number of high-risk subjects were randomized to receive APAP. It is impossible to determine the exact number, as not all trials reported the GA breakdown of participants. However, it can be estimated that from these 12 trials the outcomes of less than 150 neonates <28 wk randomized to receive APAP have been reported. Although these data from published randomized controlled trials (RCTs) show no significant effect of APAP on BPD risk in preterm infants, a possible and concerning inverse association between the risk of BPD with APAP exposure by GA suggests possible adverse effects among the most immature infants. Furthermore, our ability to assess pulmonary safety is limited by the number of pulmonary outcomes reported following exposure. Of the 12 RCT noted above, only three reported any pulmonary outcome besides BPD (days of mechanical ventilation, days of continuous positive airway pressure (CPAP), and days of supplemental oxygen) (61, 63, 69).

Figure 1.

PubMed search strategy and results.

Figure 2.

PubMed search strategy and results. A: BPD as reported in trials evaluating acetaminophen for prevention or treatment of a PDA. B: meta-regression on mean gestational age and risk difference for BPD in acetaminophen trial group using all 12 trials. Risk difference for BPD on y-axis, mean gestational age (wk) on x-axis. Size of the circle reflects the relative weight (using the inverse variance method) the individual trial data contribute to meta-regression analysis. C: meta-regression on mean gestational age and risk difference for BPD in acetaminophen trial group using 8 trials comparing oral APAP to oral ibuprofen. Risk difference for BPD on y-axis, mean gestational age (wk) on x-axis. Size of the circle reflects the relative weight (using the inverse variance method) the individual trial data contribute to meta-regression analysis. APAP, acetaminophen; BPD, bronchopulmonary dysplasia; PDA, patent ductus arteriosus.

The ability to make definitive conclusions with data from RCT is limited by the low number of extremely preterm infants enrolled in these trials. Of the cohort studies interrogating the use of APAP for PDA, we were able to identify five that met the criteria of having studying infants with a mean gestational age <28 wk and reported the outcome of BPD (71–75). Together, the outcomes of ∼350 subjects have been reported, far exceeding the numbers reported from RCT. Concerningly, the rates of BPD were higher among APAP-treated infants in four of the five studies including two that demonstrated a statistically significant association (Table 1, marked with asterisk). There are important limitations to these data including variation in treatment regimen and study design and the lack of a uniform BPD definition. However, these reports do not indicate that APAP exposures are definitively safe in extremely preterm infants. Again, robust reporting of pulmonary outcomes is largely missing, with only two trials reporting duration of mechanical ventilation in study participants (72, 73). Although APAP’s favorable safety profile related to hepatotoxicity has been lauded (41, 42, 76–78), its potential effects on the developing lung has been largely ignored. In fact, pulmonary outcomes are infrequently reported or ignored altogether in APAP studies (54, 79–84).

Table 1.

Cohort studies published since 2018 reporting APAP exposures and the outcome of BPD

Cohort Studies Ref. (Comparison Exposure)	Comparison Group		APAP Exposed		BPD Rate Difference (APAP—Comparison)
	Gestational age as reported mean	BPD % (n)	Gestational age	BPD % (n)
Karabulut (72) (ibuprofen, orally)	26.37 ± 1.34	23.5% (12/51)	26.78 ± 1.03	25% (9/36)	+1.5%
Vaidya (75) (indomethacin, iv)	25.8 ± 2.2	66% (12/18)	26.5 ± 2.5	60% (15/25)	−6%
Sehgal (74) (conservative management)	26.8 ± 1.7	16% (3/16)	25.4 ± 1.7	87% (27/31)	+71%*P < 0.001
Mashally (73) (ligation)	25.1 (24.7–26.1)	58% (25/43)	25.1 (24.2–26.3)	73% (36/49)	+15%*P < 0.04
Mashally (71) (ligation)	25.1 (24.7–26.0)	65% (22/38)	25.1 (24.2–26.3)	81% (34/44)	+16%
Total subjects		44.6% (74/166)		65.4% (121/185)	+20.8%

APAP, acetaminophen; BPD, bronchopulmonary dysplasia. *Statistically significant association.

For now, APAP use in the NICU continues whereas the clinical observation that the lung may be susceptible to APAP exposure remains speculative. Therefore, preclinical studies must be employed to identify potential mechanisms underlying APAP-induced lung injury, and to determine whether there are specific developmental windows or coexposures that increase this risk. These data may help guide the design of future clinical trials to enable thorough evaluation of the safety and efficacy of APAP therapy in the NICU. Empowered with these data, clinicians would be able to prescribe APAP more safely to high-risk patients both during pregnancy and in preterm and term infants after birth. With these objectives in mind, we review the mechanisms underlying APAP-induced cellular injury, and present data demonstrating that the lungs—and specifically the mesenchymal myofibroblast cells required for lung development—are adversely affected by APAP exposures.

PRECLINICAL LESSONS: SHEDDING LIGHT ON CLINICAL OBSERVATIONS

The conclusion that APAP exposures are “safe” during the neonatal period is largely supported by the observation that hepatotoxicity is rare, even following unintentional overdose (76, 78, 85–93). Hepatotoxicity is monitored as a potential adverse effect based on APAP metabolism data obtained in adults. In the mature liver, glucuronyl transferase and sulfotransferase activity coverts the majority (80%–90%) of ingested APAP to nontoxic metabolites (94). Activity of the xenobiotic enzyme CYP2E1 converts APAP into the mitochondrial toxin N-acetyl-p-benzoquinone imine (NAPQI). Glutathione will detoxify NAPQI; however, in the setting of glutathione depletion, NAPQI forms protein adducts resulting in cellular metabolic dysfunction and ultimately cell death. Thus, APAP-induced cellular toxicity is dependent on the exposure, cellular CYP2E1 protein expression, and the availability of detoxifying glutathione stores. In adults, hepatocytes—and specifically the pericentral hepatocytes—express the highest levels of CYP2E1 protein, thus explaining the zonated pattern of hepatic injury following toxic APAP exposure (95).

However, hepatic expression of APAP-metabolizing enzymes is developmentally regulated. Specifically, glucuronidation is lower and sulfation is higher in term neonates compared with adults. Perhaps more importantly in regards to the lack of APAP-induced hepatotoxicity during the neonatal period, CYP2E1 expression is absent in fetal life, and only reaches adult levels after the neonatal period (96–103). These findings help explain the preclinical observation that APAP exposures that are hepatotoxic in adult mice fail to cause injury in similarly exposed neonatal (3-day-old) mice (104). These preclinical data support clinical observations, demonstrating a hepatic resistance to APAP toxicity in the neonatal period. However, it is important to note that cell type-specific CYP2E1 protein expression drives toxicity, and that this expression is developmentally regulated. Therefore, we must interrogate preclinical models to determine whether other cells and tissues express Cyp2e1 making them susceptible to APAP exposures, and if so, whether this expression is dynamic over development.

Importantly, CYP2E1 protein expression is not limited to the perivenous hepatocyte. In the adult murine and human lung, multiple cell types, including the pulmonary epithelium, club cells, and macrophages, express CYP2E1 (105–112). It is therefore not surprising that preclinical studies demonstrate that following exposure, APAP-protein adducts can be found in the lungs of adult animals (112–116) and that exposures that produce hepatotoxicity also injure the lung (112, 113, 117–120). With these data in mind, it is important to use preclinical models to assess the impact of APAP exposures on the developing lung. Furthermore, if the developing lungs are susceptible to APAP toxicity, more work is needed to understand the mechanisms underlying this susceptibility and the long-term implications of these early life exposures.

Recently, we demonstrated that early alveolar stage murine lungs are susceptible to the toxic effects of APAP (121). We assessed how an exposure sufficient to cause hepatic and pulmonary injury in adult mice (280 mg/kg ip × 1) impacted the developing lung. Interestingly, this exposure did not result in histological evidence of hepatotoxicity and instead caused respiratory/terminal bronchiole injury with peripheral lung emphysema and increased mean linear intercept. Importantly, pulmonary Cyp2e1 expression increased. Although this exposure exceeded the doses used in the NICU, this proof-of-concept study was the first to demonstrate that APAP could have a direct and acute effect on the early alveolar stage murine lung. With these data in mind, it is important to consider why the developing lung may be uniquely susceptible to the toxic effects of APAP and discuss how preclinical studies can answer three critical questions.

Are Secondary Crest Myofibroblasts That Drive Alveologenesis Susceptible to APAP?

Cellular injury following APAP exposure is dependent on its metabolic conversion by CYP2E1. Thus, determining the timing and cell type-specific expression of CYP2E1 is necessary to identify potential windows of APAP susceptibility and to understand its cellular mechanisms of injury. Murine data generated by the NHLBI-funded (U01HL122642) Molecular Atlas of Lung Development Program (LungMAP) and downloaded from their website (www.lungmap.net; March 31, 2021) demonstrate that Cyp2e1 expression peaks during the saccular stage of lung development (Fig. 3A). Importantly, Cyp2e1 expression appears limited to mesodermally derived secondary crest myofibroblasts (Fig. 3B) (128–130), a cell population necessary for formation of the gas exchange surfaces of the lungs during alveologenesis (123–125). Interrogation of a recent single-cell atlas of mouse lung development (126) demonstrates myofibroblast-specific expression of Cyp2e1 (Fig. 3C), and these findings were confirmed in a single-cell analysis of mouse lung fibroblast gene expression (127). In the lungs of human neonates, CYP2E1 is expressed in lung mesenchymal cells (Fig. 3D) and at 30 wk of gestation (128–130), CYP2E1 lung mesenchymal cell expression is highest in vascular smooth muscle and myofibroblasts (Fig. 3E; https://cmdga.org/consortium/ucsd-lungmap/) (131). Secondary crest myofibroblasts are characterized by expression of Pdgfrα during the saccular phase of lung development and Acta2 during the alveolar phase and direct septation of the distal lung air spaces by contraction (123, 132, 133). Loss of myofibroblast function either by inhibiting their differentiation, impairing their function, or by directly injuring this population results in failure of alveolar development and limits formation of surface area necessary for gas exchange (123, 133, 134). Whether the developmentally regulated and cell-type specific expression of Cyp2e1 has implications for APAP exposures during the perinatal period is unknown. Preclinical models must be employed to determine whether APAP exposures injure secondary crest myofibroblasts, and if so, the developmental and functional implications of this injury.

Figure 3.

Mouse lung expression of Cyp2e1. A: lung developmental timecourse (LDT) data obtained from the NIH-funded Molecular Atlas of Lung Development Program (LungMAP) and Lung Gene Expression Analysis (LGEA) demonstrate that Cyp2e1 expression in the mouse lung peaks at the late saccular stage on embryonic day 18.5 (E18.5). B: single-cell RNA-sequencing data collected from E18.5 mouse lungs demonstrate that Cyp2e1 expression is limited to myofibroblast/smooth muscle cells. These data are obtained from https://research.cchmc.org/pbge/lunggens/lungDTC/genequery_mouse_dtc.html?geneid=cyp2e1, https://research.cchmc.org/pbge/lunggens/genequery_E18_p3.html?geneid=cyp2e1. C: single-cell RNA-sequencing data collected from developing mouse lungs demonstrate that Cyp2e1 expression is limited to myofibroblast/smooth muscle cells. These data are obtained from https://lungcells.app.vumc.org. D: sorted cell-type analysis of CYP2E1 in newborn human lungs shows expression is highest in mesenchymal cells. These data are obtained from https://research.cchmc.org/pbge/lunggens/mainportal.html. E: single-cell sequencing of 30-wk gestation newborn human lung mesenchymal cells shows CYP2E1 expression in vascular smooth muscle and myofibroblast cells. These data are obtained from https://cmdga.org/consortium/ucsd-lungmap/. Relative expression of Cyp2e1 is expressed as reads per kilobase of transcript per million (RPKM) reads mapped in A and transcripts per million (TPM) in B. Adventitial FB, adventitial fibroblast cells; Prolif. Myo FB, proliferating myofibroblast cells; Prolif. Wnt2+ FB, proliferating Wnt2-expressing fibroblast cells; Wnt2+ FB: Wnt2-expressing fibroblast cells.

Do APAP Exposures Affect Mitochondrial and Metabolic Function in the Developing Lung?

The role of mitochondrial dysfunction in APAP-induced hepatocyte injury is well studied (135, 136). In the adult liver, changes in mitochondrial ultrastructure (137) and function (138–141) accompany toxic APAP exposures. Mitochondrial APAP adducts disrupt complex I and II activity (139–144). However, whether pulmonary mitochondria, from both developing and adult lungs, are susceptible to APAP-induced injury has not been thoroughly investigated. Previous work has shown that compared with adult lungs, neonatal lungs have different metabolic pathways activated at baseline and after hyperoxia-induced injury (145). Using fluorescence lifetime imaging microscopy (FLIM), we found that indeed normal early alveolar (P7) neonatal and adult lungs have different metabolic landscapes (Fig. 4A). Early alveolar neonatal tissues had significantly more glycolysis activation, compared with the adult lungs (Fig. 4B), despite having a similar mitochondrial load, as assessed by fluorescence lifetime redox ratio (FLIRR) as a measurement of oxidative phosphorylation (OXPHOS; Fig. 4C). Importantly, neonatal lungs have a significantly higher free flavin adenine dinucleotide (FAD) fraction (Fig. 4D). These data demonstrate that distinctive metabolites in addition to different ETC complexes are activated in neonatal compared with adult lungs.

Figure 4.

Pulmonary metabolic landscapes in normal neonatal and adult lungs. A: representative intensity images (top), NADH (middle), and FAD (bottom) fluorescent lifetime imaging microscopy (FLIM) maps for control lung tissues are shown. Quantified glycolytic index (B), fluorescent lifetime imaging redox ratio (FLIRR; C), and free FAD fraction (D). Data expressed as means ± SE [10–15 fields of view for each condition were measured in neonatal (n = 3) and adult (n = 3) lungs, respectively]. FAD, flavin adenine dinucleotide.

Consistent with these findings, we recently demonstrated APAP exposures have differential effects on early alveolar neonatal and adult pulmonary mitochondrial function. In the adult murine lung, we found that compared with untreated controls, a single APAP exposure of 280 mg/kg significantly increased glycolysis and significantly decreased free FAD fraction. Conversely, a single, nontoxic APAP exposure of 140 mg/kg led to a significant decrease in glycolysis and significant increase in FLIRR (OXPHOS) and free FAD fraction, suggesting mitochondrial overload as a first response to APAP in the lung eventually leading to mitochondrial dysfunction (146).

In contrast, APAP exposures result in mitochondrial overload in the early alveolar stage murine neonatal lung. Using FLIM we detected that compared with untreated controls, APAP exposure (280 mg/kg ip × 1) decreased glycolytic index and increased OXPHOS, similar to findings in the liver following nontoxic exposures in adult mice. Importantly, unlike in adult mice, free FAD fractions were unchanged following APAP exposure. Because the electron transport chain (ETC) (147) is the primary source of free FAD, it is likely that the unchanged free FAD fraction and increased FLIRR ratio observed in the developing lung following APAP indicates an impact on the tricarboxylic acid (TCA) cycle (121).

Taken together, these results demonstrate that compared with the developing early alveolar stage lungs, adult lungs have significantly increased glycolytic index (Fig. 5A), significantly decreased mitochondrial function (Fig. 5B), and significantly decreased free FAD fraction (Fig. 5C) after toxic APAP exposures (280 mg/kg ip × 1). These results are consistent with reduced complex II activity and decreased mitochondrial load, which aligns with previously published findings in the adult liver (141).

Figure 5.

Pulmonary metabolic changes in neonatal and adult lungs following toxic APAP exposure. Quantified glycolytic index (A), fluorescent lifetime imaging redox ratio (FLIRR; B), and free FAD fraction (C). Data expressed as means ± SE [50–80 fields of view for each condition were measured in neonatal (n = 3) and adult (n = 5) lungs, respectively]. APAP, acetaminophen; FAD, flavin adenine dinucleotide.

Based on these findings, we conclude that the developmental stage of the lung must be considered before making conclusions regarding the impact of APAP on pulmonary mitochondrial function. Furthermore, these findings in the early alveolar stage lung and the mature lung reflect the impact of exposures known to cause hepatic and pulmonary toxicity. We have recently reported how APAP exposures that do not cause hepatotoxicity impact the adult murine lung. Following a single-dose exposure (140 mg/kg ip × 1), we did not detect any histologic are serologic evidence of hepatotoxicity (146). In contrast, this level of exposure resulted in alveolar wall thinning and acute inflammatory cell infiltrate in the bronchoalveolar lavage fluid (BALF) and pulmonary parenchyma. More work needs to be done to understand if and how the pulmonary metabolic changes following a nonhepatotoxic APAP exposure (140 mg/kg ip × 1) results in alveolar wall thinning. Furthermore, these studies must be extended to include exposures during lung development and correlate these findings with the timing and cell-specific expression of CYP2E1 to determine whether metabolic and morphologic changes occur.

Do Coexposures Unique to The Neonatal Period Compound Pulmonary Susceptibility to Toxic APAP Metabolites?

The neonatal lung is vulnerable to oxidative stress, and therefore may be uniquely susceptible to APAP exposure (148). Birth includes the transition from a relatively hypoxic environment in utero to breathing room air, and as a result, even the healthy term baby undergoes an immediate oxidative challenge. Many infants experience additional oxidative stressors including oxygen therapy, mechanical ventilation, sepsis, hyperglycemia, and poor nutrition (148). These oxidative challenges occur more frequently in babies born preterm, where baseline antioxidant defenses such as glutathione are low (149).

Glutathione is one of the most abundant intracellular antioxidants, essential for maintaining redox homeostasis and regulating oxidative stress (150, 151). In the adult, glutathione concentration is higher in the epithelial lining fluid than the circulation, indicating an important role for glutathione defense in the lung (152). Glutathione detoxifies the mitochondrial toxin NAPQI after APAP exposure (94). Pulmonary glutathione levels decrease rapidly in the adult lung following toxic APAP exposure (118, 153, 154). Whether APAP exposures similarly deplete glutathione in the developing lung remains to be tested. However, there are multiple reasons why the neonatal lung may exhibit increased vulnerability to APAP exposure specifically related to glutathione. First, clinical and preclinical work demonstrates that neonatal pulmonary glutathione levels are lower than those of adult, in the whole lung as well as BALF (155–158). Furthermore, the developing lung has lower glutathione stores than the term lung, as glutathione concentrations in the bronchial alveolar lavage fluid increase with gestational age (155–157). Second, the activity of the enzyme for glutathione recycling is developmentally regulated (159). Glutathione exists primarily in a reduced state (GSH). After it is oxidized to glutathione disulfide (GSSG), it can be reduced back to GSH by the enzyme glutathione reductase (GSSG-R). Decreased GSSG-R may limit the ability of GSH stores to be reutilized during oxidative challenges. Clinical evidence demonstrates that the GSSG-R activity measured in tracheal aspirate cells is lower in preterm infants than those at term (157). Preclinical work also demonstrates that whole lung GSSG-R activity correlates with age (159). Finally, glutathione is a tripeptide requiring a supply of three amino acids including cysteine. Cysteine uptake is developmentally regulated in tracheal aspirate cells of female infants, thereby lower in babies born preterm (160). Interestingly, there may be sex differences in the glutathione metabolism and abundance that render male infants at greater risk to oxidative stress. Male infants exhibit lower glutathione levels and decreased cysteine uptake in tracheal aspirate cells compared with females infants (156, 160). These data predict an increased pulmonary susceptibility to toxic APAP metabolites in neonates, and that this risk increases with preterm delivery.

Oxidative challenges experienced by preterm infants could augment APAP induced injury by acting as coexposures. Several oxidative stressors are known to similarly deplete glutathione levels. In preclinical work, hypoxia, sepsis, and mechanical ventilation decrease pulmonary and epithelial lining fluid glutathione levels (161–164). Cumulative glutathione depletion in the lung may worsen pulmonary outcomes after oxygen exposure, as the abundance of pulmonary glutathione regulates the degree of hyperoxia-induced injury (165–167). Hyperoxia increases pulmonary glutathione levels in both neonatal and adult rodent models, and is considered to be a protective adaptation (165–167). Preclinical in vivo studies demonstrate that decreasing pulmonary glutathione before hyperoxia, by protein restriction or pharmacologic glutathione depletion, exacerbates hyperoxic induced lung injury (168, 169). Furthermore, glutathione depletion increases alveolar type 2 call susceptibility to hyperoxia-induced injury (170, 171). We speculate that cumulative glutathione depletion from APAP and coexposures could have particular clinical relevance for preterm infants who frequently require oxygen therapy.

Conclusions

The association between APAP exposures and increased risk of pulmonary morbidity is a consistent clinical observation. This includes antenatal, perinatal, early life, and adult exposures. Intriguingly, an increased risk of BPD in infants born to mothers whose breast milk contains APAP metabolites has recently been reported (172). Questions remain about whether there is a direct, causal link between APAP exposures and pulmonary injury. Here, we present the data supporting the hypothesis that the lung is susceptible to APAP exposures due to cell-specific CYP2E1 expression. Importantly, there are reasons why the developing lung may be particularly susceptible to these exposures. To this point, despite growing use in the NICU, pulmonary outcomes have not been adequately followed nor reported. We urge further preclinical and clinical study before adopting more widespread use during this vulnerable period.

GRANTS

This work was supported by National Institutes of Health (NIH) Grants R01HL132941 and R01HD107700 to C.J.W., NIH Grant R01HL146859 to D.J.M., NIH Grant K08HL143051 to J.M.S.S., and a K12 Child Health Research grant from the University of Colorado Department of Pediatrics to L.G.S.

DISCLAIMERS

The contents are the authors’ sole responsibility and do not necessarily represent official NIH views.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

D.J.M., E.A.J., J.M.S.S., S.M., E.D., and C.J.W. prepared figures; L.G.S. and C.J.W. drafted manuscript; D.J.M., E.A.J., J.M.S.S., S.M., L.G.S., E.D., and C.J.W. edited and revised manuscript; D.J.M., E.A.J., J.M.S.S., S.M., L.G.S., E.D., and C.J.W. approved final version of manuscript.