Blood and urine biomarkers associated with long-term respiratory dysfunction following neonatal hyperoxia exposure: implications for prematurity and risk of SIDS

Abstract

Former preterm infants, many of whom required supplemental O₂ support, exhibit sleep disordered breathing and attenuated ventilatory responses to acute hypoxia (HVR) beyond their NICU stay. There is an increasing awareness that early detection of biomarkers in biological fluids may be useful predictors/identifiers of short- and long-term morbidities. In the present study, we identified serotonin (5-HT), dopamine (DA) and hyaluronan (HA) as three potential biomarkers that may be increased by neonatal hyperoxia and tested whether they would be associated with an impaired HVR in a rat model of supplemental O₂ exposure. Neonatal rats (postnatal age (P) 6 days, P6) exposed to hyperoxia (40% Fio₂, 24hrs/day between P1-P5 days of age) exhibited an attenuated early (1 minute), but not the late (4-5 minutes) phase of the HVR compared to normoxia control rats; the attenuated early phase HVR was associated with increased levels of DA (urine and serum), 5-HT (platelet poor plasma only, PPP), and HA (serum only). At P21, both the early and late phases of the HVR were attenuated, but serum and urine levels of all 3 biomarkers were similar to age-matched control rats. These data indicate that changes in several serum and/or urine biomarkers (5-HT, DA, and HA) following short-term (days) neonatal hyperoxia can signify long-term (weeks) respiratory control dysfunction. Further studies are needed to determine whether early detection of similar biomarkers could be convenient predictors of increased risk of abnormalities in respiratory control including sleep disordered breathing in former preterm infants who had received prior supplemental O₂ and who might also be at increased risk of SIDS.

1.0. Introduction:

Intermittent hypoxia (IH) in preterm infants, a surrogate measure of apnoea of prematurity (AOP), has been associated with multiple morbidities including growth restriction, neurodevelopmental impairment and mortality (Di Fiore et al., 2019). Supplemental O₂ therapy is a common mode of respiratory support for preterm infants to help maintain adequate oxygenation and defend against prolonged and intermittent hypoxemia (Abu-Shaweesh and Martin, 2017). Recent studies have demonstrated a prevalence of sleep disordered breathing among adolescents who were born preterm (Raynes-Greenow et al., 2012; Tapia et al., 2016). Likewise, long-term persistent effects of prematurity on the respiratory neural control system have also been demonstrated in former preterm infants who exhibit attenuated ventilatory responses to acute hypoxia (HVR) as adults (Bates et al., 2013). An immature respiratory system and other co-morbidities commonly associated with preterm birth such as infection and IH are likely significant contributing factors underlying long-term consequences of prematurity. However, numerous animal studies have demonstrated that hyperoxia can have unintended effects on respiratory system development (Bavis et al., 2014; Bavis and MacFarlane, 2017). Specifically, chronic hyperoxia causes a long-term attenuation of the early (1-2 minutes) and late (>3-4 minutes) phase of the HVR, which have largely been attributed to changes in both peripheral (e.g. carotid bodies) and central mechanisms of respiratory neural control. Collectively, these data support the notion that prior hyperoxia exposure may be a contributing factor to the lingering disordered breathing and numerous other morbidities later in life of former preterm infants.

The persistence of sleep disordered breathing in former preterm infants together with increased risk of long-term morbidities and mortality highlights the hypothesis of developmental origins of adult “disease” (Barker, 2007). Thus, there may be considerable benefit from early detection/prediction of heightened risk of later morbidities as well as risk of mortality later in life. Associations have been made between BPD¹ and biomarkers of oxidative stress and inflammatory markers in blood (infant and umbilical cord), urine and/or tracheal aspirates (see (Rivera et al., 2016; Millan et al., 2018) for review). However, we have identified serotonin (5-HT), dopamine (DA) and hyaluronan (HA) as three possible candidate biomarkers for several reasons. A recent study found increased serum 5-HT levels in a subset of Sudden Infant Death Syndrome (SIDS) cases (Haynes et al., 2017) and increased carotid body DA levels have also been implicated in SIDS (Porzionato et al., 2013). Plasma 5-HT levels in preterm infants are elevated in the first 2-3 postnatal days (Schumacher et al., 1987) and preterm infants are at increased risk of SIDS. Further, lung HA is increased by hyperoxia (Johnsson et al., 1998), which is also associated with severity of respiratory distress (RDS) in preterm animal models (Juul et al., 1994). Collectively, these data highlight the possibility that there may be useful biomarkers in biological fluids which may represent early predictors of later morbidities (respiratory disorders) or even mortality (e.g. SIDS). In the present study, we aimed to utilize a rat model of supplemental O₂ exposure (40% O₂) to screen blood and urine for increased levels of 5-HT, DA, and HA to test the hypotheses that they are associated with the short- (days) and long-term (weeks) impairments in the acute HVR. We further examined whether hyperoxia exposure affected gene expression of associated 5-HT, DA, and HA pathways in dorsal brainstem regions containing the nucleus tractus solitarius, an integrative region for carotid body (vagal) afferent inputs involved in mediating the ventilatory response to hypoxia.

2.0. Experimental procedures:

Ethical Approval:

Experiments were performed on neonatal male Sprague-Dawley rats (Charles River) maintained under standard housing conditions on a 12:12 hr light:dark cycle. All procedures were carried out in accordance with the National Institute of Health (NIH) guidelines for care and use of laboratory animals and were approved by the Animal Care and Use Committee at Case Western Reserve University.

Sustained hyperoxia exposure:

Time-pregnant rats were allowed to give birth naturally in the institutional animal facility. The day after birth (postnatal age 1 day, P1), the litters were randomly assigned to receive either room air or sustained hyperoxia (40% O₂) for 5 consecutive days (24hrs/day). Hyperoxia was premixed and administered to both the nursing dam and her pups. O₂ levels were monitored (TED 60T, Teledyne Analytical Instruments; CA, USA) and adjusted as necessary. At 5pm of the 5th day of exposure, the rats were removed from the chamber and measurements of whole-body plethysmography were performed the following day to assess the acute HVR and a modified Dejours test. A second group of hyperoxia-exposed rats were allowed to recovery in room air through to P21 days of age before also undergoing the same measurements. Treatment groups rat pups from at least 3 different litters/treatment for each age group. Hyperoxic rats were compared to age-matched normoxic raised rats. At the end of the measurements for each respective age group, the rats were euthanized and blood and urine samples were collected for later analysis of various biomarkers using appropriate ELISAs. Additional groups of rats were used to collect additional blood and urine samples as well as brainstem samples for assessing changes in gene expression using RT-PCR. These animals did not undergo whole-body plethysmography. Preliminary analysis of these blood and urine samples showed there was no effect of brief exposure to acute hypoxia/hyperoxia (plethysmography) on any of the biomarkers compared to animals which did not undergo plethysmography (data not shown).

Whole-body Plethysmography:

On the day of the measurements (P6 or P21), male rats were removed from the litter and placed inside a custom-made perspex plethysmograph chamber as described previously (Mayer et al., 2013). Airflow through the chamber was held constant using a mass flow controller (Aalborg, NY, USA).Temperature inside the chamber was maintained (~28-30°C) at all ages by adjusting a water bath (Isotemp 3013S, Fisher Scientific; PA, USA) that circulated water to a heat pad positioned underneath the plethysmograph. The chamber was assessed for adequate seal by observing stability of the square pressure change following injection of a calibration volume (50μl) using a glass micro-syringe (Hamilton, Harvard Apparatus; MA, USA). The same injection volume was used later for calibration of tidal volume changes associated with breathing. Rectal temperature was measured at the end of the experiment with a fine temperature thermocouple (Physitemp; NJ, USA). Rats were allowed approximately 30 minutes to acclimate to the plethysmograph before receiving 10% O₂ (5min), followed by rapid purging of the chamber with hyperoxia (100% O₂, 5min) for a modified Dejours test² Measurements of minute ventilation ($\stackrel{•}{\mathrm{V}}\mathrm{E}$) were made during baseline, during each minute of hypoxia and 100% O₂. Transition from hypoxia directly into hyperoxia was performed to mimic the rapid re- or hyper-oxygenation that can occur when an infant transitions from an O₂ desaturation to re-oxygenation associated with delivery of supplemental O₂. Ventilation was measured when the plethysmograph was sealed for 20 seconds between each minute of exposure to the test gas. Chambers were sealed by turning stopcocks upstream and downstream of the plethysmograph. The corresponding pressure signal associated with breathing during the time the chamber was sealed and was calibrated to calculate tidal volume. The HVR was further delineated into the early (1st minute) and late (4th-5th minute) phase to characterize a biphasic HVR.

ELISA’s:

Blood and urine samples were collected from rats there were euthanized at the end of the plethysmography, although most of the samples were obtained from additional rats that did not undergo plethysmography studies. The blood, obtained from cardiac puncture in euthanized rats, was spun at 4500g for 10 minutes to obtain serum; the top 2/3 of the serum sample was then spun a second time at 4500g for another 10minutes to obtain platelet poor plasma (PPP). Serum, PPP, and urine samples were frozen for later analysis using commercially available enzyme-linked immunosorbent assay (ELISA) validated for use in rats. When sufficient samples were collected, they were defrosted and analyzed according to manufacturers’ instructions for 5-HT (Novus Biologicals, Centennial, CO), hyaluronan (Echelon Biosciences, Salt Lake City, UT) and dopamine (MyBiosource, San Diego, CA). From the blood samples, 5-HT was measured in PPP, whereas HA and DA were measured in serum.

rtPCR:

The brainstem was removed ∼2 mm rostral and caudal to the obex and transected horizontally at the level of the central canal to separate the ventral and dorsal regions. The dorsal region which contained the nTS and DMNV, was frozen (−80°C) and stored for later analysis of mRNA using rt-PCR. Since hyperoxia affected DA, HA and 5-HT expression in the blood (and DA in urine), we investigated whether there were also changes in brainstem mRNA expression of associated signaling pathways. cDNA was generated by reverse transcription using qScript cDNA synthesis kit (Quanta Biosciences). rtPCR was performed using TaqMan reagents (Thermo Fisher). Genes of interest were hyaluronan synthases (HAS1 (Rn01455687g1) and HAS3 (Rn01643950m1)), 5-HT1A (Rn00561409s1) SERT (Rn00564737m1), and dopamine D1 (Rn03062203s1 ) and D2 (Rn00561126m1) receptors. Each sample was run in duplicate and normalized to gapdh (Thermo Fisher).

Statistical analysis:

Statistical comparisons of plethysmography measurements were made between treatment groups using Two-Way repeated measures ANOVA and a Student Newman-Keuls post-hoc analysis. Differences in gene expression were assessed using one-way ANOVA. Differences were considered significant at P < 0.05. All values are expressed as mean ± 1 SEM, although box and whisker plots were used for expression of biomarker levels and significant differences in median values were assessed using Kruskal-Wallis ANOVA on ranks.

3.0. Results:

Hyperoxic effects on resting conditions.

Hyperoxia did not have an effect on baseline minute ventilation (VE) or body weight at any age (Table 1). However, body temperature of hyperoxic P6 animals (34.4 ± 0.2°C) was higher than control rats (33.8 ± 0.3°C), whereas by P21 body temperature was similar between treatment groups.

Table 1.

Baseline ventilatory ($\stackrel{•}{\mathrm{V}}\mathrm{E}$), hypoxic (5^th min $\stackrel{•}{\mathrm{V}}\mathrm{E}$), hyperoxic (1^st min $\stackrel{•}{\mathrm{V}}\mathrm{E}$), and the magnitude of ventilatory depression ($\Delta \stackrel{•}{\mathrm{V}}\mathrm{E}$: 100% – 10% $\stackrel{•}{\mathrm{V}}\mathrm{E}$ value) following immediate transition from 10% Fio₂ breathing to 100% Fio₂ breathing (modified Dejours test) in P6 and P21 day old rats following 5 days of postnatal (P1-5 days) hyperoxia exposure.

Age group (days)	P1-P5 Treatment	N’s	Body weight (g)	Body temperature (°C)	Baseline $\stackrel{•}{\mathrm{V}}\mathrm{E}$ (ml/g/min)	10% Fio2 $\stackrel{•}{\mathrm{V}}\mathrm{E}$ (5th min, ml/g/min)	100% Fio2 $\stackrel{•}{\mathrm{V}}\mathrm{E}$ (1st min, ml/g/min)	$\Delta \stackrel{•}{\mathrm{V}}\mathrm{E}$ (ml/g/min) (100% $\stackrel{•}{\mathrm{V}}\mathrm{E}$ – 10% $\stackrel{•}{\mathrm{V}}\mathrm{E}$)
P6	Normoxia	10	16.4 ± 0.7	33.8 ± 0.3	2.03 ± 0.1	2.43 ± 0.15	0.81 ± 0.07	−1.63 ± 0.14
	Hyperoxia	14	15.0 ± 0.2	34.4 ± 0.2*	2.18 ± 0.1	2.27 ± 0.11	0.92 ± 0.07	−1.35 ± 0.09
P21	Normoxia	14	50.6 ± 1.6	36.7 ± 0.1	0.99 ± 0.06	2.19 ± 0.09	0.83 0.08	−1.36 ± 0.12
	Hyperoxia	16	52.1± 1.4	36.5 ±0.1	0.95 ± 0.06	1.66 ± 0.10*	0.68 0.08	−0.97 ± 0.08*

Values are mean ± 1SEM.

^*significant difference from normoxia treated rats (p<0.05). Abbreviations are indicated in the text.

Hypoxic ventilatory response:

P6 day old control (normoxia) rats exhibited an initial increase in $\stackrel{•}{\mathrm{V}}\mathrm{E}$ within the 1st minute (early phase) of acute hypoxia, followed by a ventilatory decline by the 2nd, through to the 5th minute (late phase) demonstrating a biphasic HVR (Fig. 1A). Hyperoxia treated rats also exhibited a similar biphasic HVR except the early phase (1st minute) was significantly attenuated compared to control rats. The late phase of the HVR up to the 5th minute of acute hypoxia was similar between groups. The attenuated early phase of the HVR in hyperoxia-exposed rats resulted from a reduced VT compared to control rats; however, fR remained elevated during acute hypoxia in both groups while the ventilatory depression of the late phase of the HVR resulted primarily via a gradual decline in VT (Fig. 1A). At three weeks of age, the control and hyperoxic group exhibited a sustained elevation in the late phase of the HVR (Fig. 1B) indicating maturation of the biphasic HVR. However, the hyperoxia treated rats had an attenuated early and late phase HVR, which was mediated primarily via an attenuated fR response throughout acute hypoxia.

Fig 1:

Ventilatory ($\stackrel{•}{\mathrm{V}}\mathrm{E}$), frequency (fr) and tidal volume (Vt) responses to acute hypoxia followed by 100 % Fio₂ breathing (modified Dejours test) in P6 (A) and P21 (B) day old male rats. Note in P6 rats, prior hyperoxia exposure (between P1-5 days of age) attenuated the early phase (1^st minute) of the HVR, whereas at P21 both the early and late phase (5th minute) of the HVR was attenuated in the hyperoxic treated group (A). Further, the ventilatory depression following rapid transition from acute hypoxia to breathing 100% Fio₂ (modified Dejours test) at P6 was similar between both control and hyperoxia treated rats (A); the ventilatory depression (below baseline) still persisted in P21 day old hyperoxic, but not control rats (B). Values are expressed as means ± 1 S.E.M. Horizontal dashed lines designate baseline values for respective treatment group. *indicates significant difference from 1^st minute of hypoxia; ^#indicates significant difference between treatment groups for a given time-point; ^$indicates a given 100% Fio₂ value is significantly different from baseline value (P<0.05). Treatment groups: P6: RA, N = 10, 40% O₂, N = 14; P21: RA, N = 16, 40% O₂, N = 16.

Hypoxic-hyperoxic respiratory depression (modified Dejours test):

The rapid transition from breathing acute hypoxia (10% FIO₂) into 100% FIO₂ (shaded area, Fig. 1) caused an immediate ventilatory depression in both treatment groups at P6 days of age, decreasing $\stackrel{•}{\mathrm{V}}\mathrm{E}$ significantly below the baseline ($p<0.05; Fig. 1A) and the hypoxic value (Table 1). The ventilatory depression of the P6 group was mediated by a decrease in both fR and VT in both treatment groups. Table 1 quantifies the magnitude of the ventilatory depression specifically between acute hypoxia (5th minute) to the 1st minute of 100% FIO₂. The magnitude of the hypoxic-hyperoxic ventilatory depression was similar between treatment groups at P6 whereas at P21, prior neonatal hyperoxia exposure attenuated the magnitude of the ventilatory depression compared to control rats (Table 1).

Further, recovery of the ventilatory depression over the course of 5 minutes of breathing 100% FIO₂ was also similar between treatment groups, although $\stackrel{•}{\mathrm{V}}\mathrm{E}$ remained significantly below baseline even by the 5th minute of O₂. At P21 the hypoxic-hyperoxic transition decreased $\stackrel{•}{\mathrm{V}}\mathrm{E}$ in both control and hyperoxic treatment groups. Unlike the P6 control group response, however, at P21 $\stackrel{•}{\mathrm{V}}\mathrm{E}$ did not fall below baseline suggesting maturation of the ventilatory depressant response of the modified Dejours Test. In contrast, $\stackrel{•}{\mathrm{V}}\mathrm{E}$ fell below baseline in the hyperoxic treated group, but had recovered to baseline $\stackrel{•}{\mathrm{V}}\mathrm{E}$ levels by the 5th minute (Fig. 1B). The ventilatory depression of the P21 hyperoxic group was caused by reduced respiratory frequency compared to baseline levels.

Blood and urine biomarkers (5-HT, dopamine and hyaluronan):

Hyperoxia exposure significantly increased serum DA and HA as well as PPP 5-HT levels at P6 compared to age matched normoxic rats (Fig. 2). The increased DA, HA and 5-HT, however, had returned to normoxic levels by P21 (ie. after ~2 weeks recovery from hyperoxia). Urine DA levels were elevated at P6 following hyperoxia, which had also normalized to normoxic levels by P21 (Fig. 2), whereas neither HA nor 5-HT levels in the urine were affected by hyperoxia exposure at any age.

Fig. 2.

Box and whisker blots of blood (A) and urine (B) levels of the biomarkers dopamine (DA), hyaluronan (HA) and serotonin (5-HT) in P6 and P21 day old rats following normoxia or hyperoxia (40% O₂, P1-5 days) exposure. Lower and upper error bars indicate 10th and 90th percentiles respectively; boundary of the box closest to zero indicates 25th percentile; line within box indicates the median; open symbols designate values exceeding 1.5x the SD. *significantly different from normoxia control rats p<0.05. PPP, platelet poor plasma. Treatment groups: P6: N = 11-16 and P21: N = 8-16 pups/treatment depending on the biomarker.

Brainstem mRNA:

Hyperoxia did not affect dorsal brainstem mRNA expression for any gene of interest at P6 whereas at P21, hyperoxia increased HAS3, 5-HT1A receptor, as well as D1 and D2 receptor mRNA expression.

4.0. Discussion:

Overall, the primary results of the current study demonstrated that neonatal hyperoxia affected peripheral (particularly blood) levels of HA, 5-HT, and DA biomarkers, whereas there was a “delayed” effect on their corresponding genes of interest in the dorsal brainstem. The significance of these findings are discussed below in the context of neonatal hyperoxic effects on the respiratory neural control system (HVR) and detection of various biomarkers as early predictors of long-term changes in brainstem neurochemistry and respiratory control dysfunction.

Short- and long-term effects of neonatal hyperoxia on respiratory neural control:

Neonatal hyperoxia exposure, which mimics the clinical application of supplemental O₂ administration to preterm infants, and its effects on the respiratory neural control system have been fairly well described in recent years. Our data are consistent with the findings from animal studies of a short- and long-term impairment in the early phase of the HVR (Ling et al., 1997; Bavis et al., 2013), which has been attributed largely to impaired carotid body (CB) chemoreceptor development and O₂ sensitivity (Ling et al., 1997; Erickson et al., 1998; Prieto-Lloret et al., 2004; Donnelly et al., 2005; Dmitrieff et al., 2012). As adults, the HVR is also attenuated in former preterm infants who had received prior supplemental O₂ while in the NICU (Bates et al., 2013). Whether changes in CB size, morphology or sensitivity is responsible for the attenuated HVR remains inconclusive (Bates et al., 2018), although infants with lung disease and prior supplemental O₂ use express long-term impairments in chemoreceptor drive in response to inspired FIO₂ (Calder et al., 1994; Katz-Salamon et al., 1996). The latter finding is in part consistent with the results of the present study in which the magnitude of ventilatory depression following transition from hypoxia (10% FIO₂) to 100% FIO₂ (modified Dejour’s Test; Fig. 1) was dampened in hyperoxic treated rats particularly at 3 weeks of age (i.e. 2 wks post-hyperoxia exposure).

In addition to the peripheral (carotid body) determinants of the HVR, several brainstem neurochemical factors modulate the late phase of the HVR including BDNF, platelet-derived growth factor beta (PDGF-β; (Gozal et al., 2000b; Vlasic et al., 2001), neuronal nitric oxide synthase (nNOS), and NMDA glutamate receptors (Gozal et al., 2000a; Ohtake et al., 2000; Vlasic et al., 2001; Simakajornboon and Kuptanon, 2005), some of which have been shown to be affected by neonatal hyperoxia exposure (Chavez-Valdez et al., 2012; Bavis et al., 2014). In the present study, we chose to assess changes in HA (a component of the ECM), and the neurotransmitters 5-HT and DA due to the changes observed in blood and urine samples at P6 (Fig. 2). HA and other components of the ECM play an important role in development of the CNS including modulation of axon guidance and synaptogenesis. HA is synthesized from different synthases (HAS1, 2, or 3) and the synthesis of low molecular weight HA from HAS3 has known pro-inflammatory properties, which may be of significance in the context of inflammatory and/or oxidative challenges. Increased HA levels were associated with RDS severity in a preterm primate model (Juul et al., 1994) and neonatal hyperoxia exposure also increased lung HA levels in premature rabbits (Johnsson et al., 1998) and rats (Juul et al., 1995). Similarly, preterm infants that died within 32 weeks of birth exhibited significantly higher pleural HA levels, which were even higher in infants exposed to intrauterine infection (Johnsson et al., 2003). The results of the current study demonstrated a delayed (ie. weeks after hyperoxia exposure ended) increase in brainstem HAS3 (but not HAS1) mRNA expression, which could have implications for disrupted neural development and sleep disordered breathing seen in former preterm infants. Similarly, brainstem 5-HT1A and DA receptor mRNA expression were also increased which could have implications for the brainstem 5-HT1A (Paterson et al., 2006) and by extension the CB DA abnormalities seen in SIDS cases. It is possible that neonatal hyperoxia exposure could be a driving risk factor for the increased risk of SIDS associated with prematurity. However, these data should be interpreted with caution given the multitude of co-morbidities associated with preterm birth, which were not replicated in the hyperoxic model of the present study.

Significance of blood and urine biomarkers:

A significant finding of the present study was the increased PPP 5-HT levels following neonatal hyperoxia exposure. We chose 5-HT as one potential biomarker that may be affected by hyperoxia for several reasons. Anecdotal evidence suggests supplemental O₂ increases blood 5-HT levels in preterm infants (Schumacher et al., 1987). Similarly, serum 5-HT levels were elevated in SIDS infants and has been proposed as a possible pre-mortem identifier of infants at risk for SIDS (Haynes et al., 2017). The cause for the increased blood 5-HT is unclear although it is of interest that hyperoxia disrupts platelet 5-HT turnover and may be one possible source (Mais and Bosin, 1984). The detection of elevated 5-HT could also have an impact on respiratory neural control since the CB is sensitive to 5-HT (Peng et al., 2006). The significance of this finding is unclear given that the effect on 5-HT levels had resolved by P21, whereas the attenuated HVR had persisted to this age. Rather we propose that the changes in PPP 5-HT levels in the context of the current study could represent an early blood biomarker of neonatal hyperoxia exposure in a way that could predict long-term disturbances in respiratory neural control. From a translational standpoint, this may be of considerable significance for the early identification of preterm infants at increased risk of long-term sleep disordered breathing as well as those who may be at risk of SIDS. The same rationale could be applied to the increased serum HA and DA levels. Increased dopamine levels have been observed in the CB of SIDS (Porzionato et al., 2013). DA is a well-known modulator of CB excitability and, although there is evidence that it has excitatory properties, it is considered largely as an inhibitory neurotransmitter (Powell, 2007). The only long-term effect of neonatal hyperoxia was a delayed increase in urine DA, which was not seen at the time of hyperoxia exposure. Thus, collectively, these data could have implications for the detection of circulating 5-HT, HA, and DA as early biomarkers of preterm infants who may go on to develop sleep disordered breathing.

Conclusions:

We speculate that circulating HA, 5-HT and DA levels could be useful and conveniently detectible O₂ “sensitive” biomarkers which could be used to identify preterm infants (on supplemental O₂) who may be at increased risk of developing long-term neurological impairments associated with (among other possible morbidities) sleep disordered breathing. Further, the long-term changes in gene expression in the brainstem could underlie these neurological abnormalities and could contribute our understanding of their increased risk of SIDS. Future studies are needed to address the clinical role that these biomarkers could play as well as the significance of sleep disordered breathing in former preterm infants particularly those who had received prior supplemental O₂. This is of considerable concern given the emerging evidence demonstrating a link between intermittent hypoxic events in adults with sleep apnea and acceleration of cancer progression (Almendros and Gozal, 2018; Sillah et al., 2018). The impact of early life experiences for the premature infant and their long-term consequences, therefore, could extend far beyond the risk of poor neurodevelopmental outcome and it is imperative to take into account the potential impact of other morbidities associated with prematurity.

Fig. 3.

Effects of neonatal (P1-6 days) hyperoxia (40% Fio₂) on brainstem mRNA expression in P6 (A) and P21 (B) day old rats. Genes of interest (hyaluronan synthases, HAS1 and 3); serotonergic (SERT, 5-HT1A receptor); and dopaminergic (D1 and D2 receptors)) were chosen based on biomarkers identified in blood and urine at P6 days of age. Note, there was no effect of hyperoxia on genes of interest at P6, whereas at P21 hyperoxia increased HAS3, 5-HT1A and dopamine receptor mRNA. Values are means ± SEM and expressed as fold change from normoxia treated rats; symbols for individual values are also provided; *significantly different from normoxia control rats p<0.05. Treatment groups: P6: N = 11-14 and P21: N = 13-16 pups/treatment.