Neuroprotection in neonatal Hypoxia-ischaemia: melatonin targets NCX1 to inhibit mitochondrial autophagy via the PINK1-Parkin pathway

Abstract

Objective

Hypoxic ischaemic (HI) damage is a major cause of white matter damage (WMD) in the brains of newborns, especially preterm infants; early neuroprotection is essential to improve cognitive outcomes. This study aimed to investigate the effect of melatonin on nerve injury by inhibiting mitochondrial autophagy.

Methods

We established a neonatal WMD model through HI induction in postnatal day 3 (P3) Sprague–Dawley (SD) rats. Following four days of intraperitoneal melatonin administration (10 mg/kg/d), temporal changes in expression of sodium-calcium exchanger 1 (NCX1), myelin integrity markers (myelin-associated glycoprotein [MAG]/proteolipid protein [PLP]), and mitophagy-related proteins (microtubule-associated protein 1 light chain 3β [LC3β], PTEN-induced kinase 1 [PINK1], and Parkin RBR E3 ubiquitin-protein ligase [Parkin]) were systematically quantified. Neuronal hyperexcitability was evaluated by whole-cell patch-clamp recordings, whereas myelin pathology was assessed by luxol fast blue (LFB) staining, and mitochondrial ultrastructures were evaluated by transmission electron microscopy. Cognitive recovery was determined using Morris water maze testing at postnatal day 28.

Results

Our results demonstrated that rats subjected to HI presented biphasic alterations in NCX1 expression, characterised by transient upregulation on day 7 followed by a progressive decline (P < 0.001). Concurrently, expression of mitochondrial autophagy markers (LC3β, PINK1, and Parkin) was significantly increased (P < 0.001). Histological analysis revealed distinct mitochondrial structural damage and autophagosome formation. Electrophysiological measurements revealed increased neuronal excitability (P < 0.05), which was correlated with spatial learning and memory deficits. Although melatonin treatment effectively attenuated these pathological alterations, subsequent pharmacological inhibition of NCX1 via SN6 administration in melatonin-treated rats resulted in the recurrence of mitochondrial ultrastructural abnormalities and the reactivation of autophagic pathways.

Conclusion

Melatonin attenuated activation of the PINK1-Parkin-dependent mitochondrial autophagy pathway in neonatal rats with HI-induced WMD through mediating the dynamic expression of NCX1. This intervention effectively reduced neuronal hyperexcitability, ameliorated demyelinating lesions, and improved long-term learning and cognitive functions.

Clinical trial registration

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10735-025-10601-5.

Introduction

Hypoxia–ischaemia (HI)-induced white matter damage (WMD) in neonates predisposes them to severe central nervous system complications, culminating in mortality or lifelong neurological sequelae (Hinojosa-Rodríguez et al. 2017; Locke and Kanekar 2022). While advancements in neonatal care have progressively reduced mortality rates over the past decade, epidemiological data have indicated that 40–60% of extremely preterm infants (26–29 gestational weeks) remain at high risk of fatal outcomes or severe neurodevelopmental disorders (Pascal et al. 2023). Longitudinal cohort studies have revealed that survivors frequently exhibit persistent neurological deficits, including cerebral palsy (GMFCS levels II-V), drug-resistant epilepsy, and intellectual disability (IQ < 70) (Yin et al. 2013; Lui et al. 2018; Pascal et al. 2023). Current clinical management methods lack evidence-based early intervention protocols targeting HI-induced WMD pathophysiology (Sizonenko et al. 2003).

Emerging neuroprotective strategies have translational potential: vagus nerve stimulation effectively suppresses lysolecithin-induced neuroinflammation while preserving myelin integrity in preclinical models (Bachmann et al. 2024). Noninvasive neuromodulation modalities—including transcutaneous direct current stimulation (tsDCS), repetitive transcranial magnetic stimulation (rTMS), and electromagnetic-pulsed gene (EPG) therapy—increase neuroplasticity and motor recovery postneural injury (Cywiak et al. 2020; Kahana et al. 2023). Nevertheless, technical limitations (e.g., skull impedance variations in preterm infants) and safety concerns regarding immature blood‒brain barrier permeability substantially restrict their neonatal application. This therapeutic impasse underscores the critical need for developing age-appropriate, mechanism-driven interventions to ameliorate HI-induced WMD-related disabilities.

Melatonin, an endogenous indoleamine synthesised primarily in the pineal gland, has unique pharmacokinetic properties, including the ability to penetrate all physiological barriers (e.g., the blood‒brain barrier and placental barrier) and ubiquitous distribution across tissues, cellular compartments, and subcellular structures (Welin et al. 2007; Robertson et al. 2012; Alonso-Alconada et al. 2013; Aly et al. 2014). Its multifaceted neuroprotective effects encompass potent antioxidant, anti-inflammatory, and antiapoptotic activities, positioning it as a promising therapeutic candidate for HI-induced WMD in neonates, particularly preterm infants (Tan et al. 2015; D'Angelo et al. 2020; Gou et al. 2020; Boutin et al. 2023).

Mechanistic studies have revealed that melatonin administration during HI events scavenges intracellular reactive oxygen species (ROS) via electron donation, preserving cellular energy homeostasis (NAD + /NADH ratio > 2.0) (Tan et al. 2015; Boutin et al. 2023); intensifies acute-phase oxidative damage in neural structures (dendrites, axons, and astrocytes) through the stabilisation of membrane integrity and prevention of organelle degeneration (Liu et al. 2024); modulates neuroinflammatory responses by suppressing P2X purinoceptor 7 (P2X7) receptor-mediated NOD-like receptor family pyrin domain containing 3 (NLRP3) inflammasome activation and promoting M2 microglial polarisation via Janus kinase 2-signal transducer and activator of transcription 3 (JAK2-STAT3) signalling (Jiang et al. 2021; Gelen et al. 2023); and enhances remyelination through cAMP response element-binding protein (CREB)-dependent synthesis of synaptic proteins (Alghamdi and Taleb 2020). Notably, the mitochondrial targeting capacity of melatonin enables the stabilisation of electron transport chain complexes and the inhibition of permeability transition pore opening, which are critical for maintaining ATP production during reperfusion (Riemersma-van der Lek et al. 2008; Wang et al. 2020). Preclinical evidence has further indicated its ability to rescue extracellular signal-regulated kinase 1/2 (ERK1/2)-mediated synaptic plasticity deficits, which are correlated with improved Morris water maze performance in animal models of ischaemic stroke (Chen et al. 2018). While these pleiotropic mechanisms underscore the therapeutic potential of melatonin for HI-induced WMD, key knowledge gaps persist regarding optimal dosing windows relative to HI insult phases, mitochondrial genome‒epigenome interactions, and long-term effects on oligodendrocyte lineage maturation.

Melatonin exerts multifaceted protective effects on mitochondrial homeostasis through (1) potentiating endogenous antioxidant defences via free radical scavenging; (2) optimising energy metabolism by preserving complex I/III activity; and (3) inhibiting Bcl-2-associated X protein (BAX)-mediated apoptosis pathways (Lei et al. 2024). Mechanistically, melatonin accumulates in mitochondria via oligopeptide transporters peptide transporter 1/2 (PEPT1/2)-mediated uptake, where it directly neutralises ROS through electron donation while increasing glutathione peroxidase (GPX) activity, thereby maintaining the mitochondrial membrane potential (ΔΨm > 140 mV) under oxidative assault (Lei et al. 2024).

Notably, nanoparticle-encapsulated melatonin has enhanced bioavailability, stimulating mitophagy flux through PINK1‒Parkin axis activation to delay α-synuclein aggregation in Parkinsonian models (Biswal et al. 2024). Furthermore, sirtuin 3 (SIRT3)-dependent transcription factor A (TFAM) deacetylation facilitates mitochondrial genome stabilisation, which ameliorates sepsis-induced renal tubular damage in murine models of acute kidney injury (AKI) (Deng et al. 2024). Related data have revealed the calcium regulatory capacity of melatonin in pancreatic acini, where sarcoplasmic/endoplasmic reticulum calcium ATPase 2b (SERCA2b) expression upregulation and NCX1 membrane translocation collaboratively reduce [Ca2 +]i overload, mitigating cerulein-induced pancreatitis (Khananshvili 2022). While these findings underscore the pleiotropic organoprotective role of melatonin, its mitochondriotropic mechanisms in neonates with HI-induced WMD recovery remain elusive.

The sodium-calcium exchanger (NCX) is a bidirectional transporter localised to the mitochondrial membrane that plays a critical role in maintaining cellular ion homeostasis. As a membrane protein, NCX is widely expressed in the brain, heart, skeletal muscles, and pancreatic tissues (Castaldo et al. 2009; Valentim et al. 2022). In the central nervous system, three NCX subtypes have been identified: NCX1, NCX2, and NCX3. Notably, studies have demonstrated that the suppression of NCX1 expression reduces the ischaemic cerebral infarction volume, suggesting its predominant involvement in focal ischaemic brain injury and its pathophysiological relevance to central nervous system ischaemic disorders (Morimoto et al. 2012).

NCX operates in two distinct modes: forwards and reverse. Under physiological conditions, the forwards mode mediates the exchange of 3 Na⁺ ions into mitochondria for 1 Ca²⁺ extracted, utilising the Na⁺ electrochemical gradient to regulate Ca²⁺ efflux—a mechanism essential for maintaining cellular excitability. However, during pathological states such as ischaemia‒reperfusion injury, the reverse mode becomes activated, promoting Ca²⁺ influx into the cytosol concurrent with Na⁺ extrusion. This aberrant activity may induce intracellular calcium overload (Horváth et al. 2023), which disrupts mitochondrial membrane integrity, reduces the mitochondrial membrane potential, and triggers excessive mitophagy (Duchen 2000). While basal mitophagy contributes to cellular homeostasis, its dysregulation can exacerbate programmed cell death (Paz et al. 2015). In HI-related brain injury, both insufficient mitophagy and excessive mitophagy have been implicated in neuronal death (Dernie 2020). Exposure to HI activates the reverse mode of NCX1, leading to mitochondrial damage. First, decreased Na⁺/K⁺-ATPase activity results in energy depletion and an accumulation of intracellular Na⁺ ([Na⁺]i), which weakens the Na⁺ gradient. Second, the opening of ATP-sensitive potassium channels decreases K⁺ efflux and leads to membrane potential depolarisation, further promoting the reverse mode of NCX1. This induces efflux of 3 Na⁺ for influx of 1 Ca²⁺, resulting in intracellular Ca²⁺ overload, activation of calcium-dependent proteases such as calpain, opening of the mitochondrial permeability transition pore, and exacerbation of cell death (Iwamoto et al. 1996). Further research is warranted to investigate how melatonin modulates mitophagy dynamics, thereby optimising therapeutic strategies to mitigate severe neurological sequelae such as cerebral palsy and epilepsy in paediatric populations.

This study employed a neonatal rat model of HI-induced WMD (Cheng et al. 2015; Yang et al. 2022) to investigate the neuroprotective effects of melatonin against HI-induced neuronal dysfunction mediated through NCX1 expression regulation.

Methods

Animals

All animal procedures were approved by the Animal Ethics Committee of Qingdao University (Approval No. 20230907SD42020240328037). Perinatal Sprague–Dawley (SD) rats were obtained from SiPeiFu (Beijing) Biotechnology Co., Ltd., and housed under specific pathogen-free (SPF) conditions with a 12-hlight/dark cycle and free access to food and water.

HI-induced WMD model

Three-day-old SD rats (both sexes were randomly assigned to groups to minimise sex-related bias) underwent surgical procedures to establish a neonatal rat model of HI-induced WMD. Briefly, neonatal rats were anaesthetised with 2% isoflurane, and a 0.5–1.0 cm longitudinal incision was made along the left ventral cervical midline under a stereomicroscope. The left common carotid artery was isolated, doubly ligated with 7–0 absorbable sutures, and transected between the ligatures. After wound closure, the pups were returned to their dams for a 1-h recovery. The animals were subsequently exposed to hypoxia (8% O₂/92% N₂) for 2.5 h in a temperature-controlled chamber (gas flow rate: 2 L/min). Sham-operated controls underwent identical procedures, excluding artery ligation and hypoxia. Postoperatively, the pups were reunited with their dams (Fig. 1a).

Fig. 1.

Experimental workflow and mechanistic schematic (Created in https://BioRender.com [CJ28MGN4JV])

Hematoxylin and eosin (H&E) staining

On the 7th, 14th, and 21st days post-birth, rats were anaesthetised with isoflurane and perfused transcardially with 0.9% physiological saline followed by 4% paraformaldehyde (prepared in 0.1 M PBS) for fixation. Brain tissues were then extracted and immersed in 4% formaldehyde solution (prepared in 0.01 M PBS) for 48 h. Subsequently, the tissues were dehydrated, embedded in paraffin, and sectioned at a thickness of 3 μm. Sections were de-paraffinised, hydrated, stained with haematoxylin for 10 min, rinsed under running water for 30 min, and stained with eosin for 3 min. Finally, sections were dehydrated, cleared, and mounted. The tissue morphology was observed using an Olympus optical microscope at 40 × magnification.

Experimental groups and drugs

Experimental groups

Overall, 420 pups were used in this study, and the neonates were randomly divided into the following groups (n = 10 per group): 1. Sham group: no ligation/hypoxia + vehicle (normal saline); 2. HI group: ligation/hypoxia + vehicle; 3. Mel group: HI + melatonin, from P3 to P6. 4. SN6 group: HI + melatonin + SN6 from P3 to P6.; 5. S + M group: Sham + melatonin from P3 to P6; 6. S + M + N group: Sham + melatonin + SN6 from P3 to P6; all treatments were initiated 2.5 h post-HI. Samples were collected at P7, P14, and P21 (n = 6 per group at each time point).

Experimental drugs

Melatonin (MCE; HY-B0075) is an endogenous hormone secreted by the pineal gland and is known for its significant anti-inflammatory and antioxidant properties. In this study, rats in the melatonin intervention group (Mel group) received intraperitoneal injections daily from postnatal days 3 to 6 (P3-P6) at a dose of 10 mg/kg/day using an injection solution concentration of 1 mg/ml.

SN6 (MCE: HY-107658) is a highly selective sodium-calcium exchanger (NCX) inhibitor used to explore the neuroprotective role of NCX. In the experimental design, rats in the SN6 group and the S + M + N group received melatonin pretreatment from P3 to P6, followed by immediate intraperitoneal injection of SN6. The SN6 dosing regimen was also 10 mg/kg daily, administered using a solution with a concentration of 2 µg/µl.

Brain slice preparation

Coronal brain slices (300 μm thick) from 14-day-old rats were prepared using a vibrating microtome (VT1200S, Leica Biosystems, Germany). Slices were incubated in oxygenated artificial cerebrospinal fluid (ACSF: 95% O₂/5% CO₂, pH 7.4) at 24 °C for 1–2 h prior to experimentation. Individual slices were transferred to a submersion recording chamber (Warner RC-26G) continuously perfused with normal ACSF (flow rate: 2 mL/min) for electrophysiological recordings.

Electrophysiological recording

Whole-cell current‒clamp recordings were performed on cortical pyramidal neurons using a dual-channel amplifier (HEKA EPC 10 USB, HEKA Instruments, USA). The signals were sampled at 20 kHz with a 3 kHz lowpass filter, and the pipette capacitance was fully compensated. Glass micropipettes (resistance: 8–10 MΩ) were filled with an intracellular solution (osmolarity: 295–305 mOsm). Neuronal excitability was assessed by injecting depolarising current steps (0 to + 150 pA, 10 pA increments, 1 s duration) to quantify action potential frequency versus stimulus intensity.

Transmission electron microscopy

Fresh brain tissues were fixed in 2.5% glutaraldehyde (4 °C, 24 h), postfixed with 1% osmium tetroxide, and dehydrated through a graded acetone series (30% → 100%). The samples were embedded in Epon 812 epoxy resin, and ultrathin Sects. (70 nm) were collected on copper grids. The sections were sequentially stained with uranyl acetate (30 min) and lead citrate (5 min) and then imaged using a JEM-1200EX transmission electron microscope (Hitachi, Japan) at 25,000 × magnification to analyse oligodendrocyte ultrastructures and myelin integrity.

Immunohistochemical assay

Paraffin-embedded brain sections were deparaffinised and rehydrated, and endogenous peroxidase activity was quenched by incubation in 3% H₂O₂ within a humidified chamber (37 °C, 30 min). Antigen retrieval was performed via microwave irradiation in 10 mM citrate buffer (pH 6.0), followed by blocking with 10% normal goat serum (37 °C, 20 min). The sections were incubated overnight at 4 °C with a rabbit anti-NCX1 monoclonal antibody (Proteintech, #26,835–1-AP, 1:300). After being rinsed with PBS, the sections were treated with biotinylated goat anti-rabbit IgG (1:500, 37 °C, 30 min) and then exposed to a streptavidin-HRP conjugate (1:1000, 37 °C, 30 min). Signal amplification was achieved using 3,3′-diaminobenzidine (DAB) chromogen under microscopic monitoring. Counterstaining with Mayer’s haematoxylin, ethanol dehydration, xylene clearing, and neutral resin mounting were sequentially performed.

Myelin sheath staining (LFB)

Deparaffinised sections were dehydrated through graded ethanol (70% → 95% → 100%, 2 h each step). LFB staining was conducted at 60 °C for 2 h, followed by different rinses in 95% ethanol and a lithium carbonate solution. The sections were counterstained with haematoxylin (10 min), washed in running tap water, dehydrated, cleared in xylene, and mounted with DPX medium. Images were visualised and obtained using a scanner (Pannoramic 250FLASH, 3DHISTECH, Hungary).

Morris water maze test

The Morris water maze test was performed at 28–33 days after modelling. The experimental environment was a circular pool (160 cm in diameter and 60 cm in height) with a black inner wall and four quadrants marked by obvious visual cues. The device consisted of a movable platform, 12 cm in diameter, placed 1.5 cm below the water surface (30 cm depth), and the water temperature was maintained at 25 ± 1 °C. Spatial Acquisition Phase (Days 1–4): Rats underwent four daily trials (60 s/trial) from randomised entry quadrants. The animals that located the hidden platform within 60 s remained there for 30 s; those that failed were gently guided to the platform. Intertrial intervals included towel drying and warming under a 37 °C heat lamp to prevent hypothermia. Probe Trial (Day 5): The platform was removed. The rats were released from a novel quadrant, and their swimming trajectories were recorded for 60 s to assess spatial memory retention (Revuelta et al. 2016). The target quadrant occupancy time, platform crossing frequency, and total swimming distance were recorded using a video tracking system (Shanghai Mobile Data Co., Ltd.). The software was used for maze calibration and parameter settings (e.g., area and sensitivity) and included a user interface for real-time tracking. It records parameters (latency, path length) and generates visual outputs (heatmaps, trajectories). Users can validate data through video playback, apply filters to exclude outliers, and export results in multiple formats for analysis.

Simple Western analysis

Brain tissues were homogenised in RIPA lysis buffer (Solarbio, Beijing, China) for protein extraction, and the total protein concentrations were determined using an enhanced BCA assay kit (Solarbio). Protein lysates were normalised to equal concentrations, and 3 μL of normalised lysate was loaded per capillary. Automated protein separation and immunodetection were performed using a Wes system (ProteinSimple, San Jose, CA, USA). The primary antibodies used included:

• Rabbit anti-NCX1 polyclonal antibody (1:20, 0.2 μg/μL, #28447-1-AP; Proteintech, Wuhan, China)

• Rabbit anti-PLP monoclonal antibody (1:10, 2 μg/μL, EPR23504-106, #ab254363; Abcam, Cambridge, UK)

• Rabbit anti-LC3B monoclonal antibody (1:100, 0.2 μg/μL, EPR18709, #ab192890; Abcam)

• Mouse anti-MAG monoclonal antibody (1:20, 2 μg/μL, sc-166849; Santa Cruz Biotechnology, Dallas, TX, USA)

• Mouse anti-PINK1 monoclonal antibody (1:100, 2 μg/μL, sc-517353; Santa Cruz Biotechnology)

• Mouse anti-Parkin monoclonal antibody (1:20, 2 μg/μL, sc-32282; Santa Cruz Biotechnology)

Statistical analysis

The data are presented as the means ± SEMs. Normality was assessed using the Kolmogorov–Smirnov test with the Dall-Wilkinson-Lillie approximation for p-value adjustment. Escape latency data (time-dependent measurements) were analysed using two-way ANOVA with Tukey’s post-hoc test. Other endpoints were evaluated using one-way ANOVA, followed by Tukey’s test. All analyses were conducted using GraphPad Prism v10.0 (GraphPad Software, San Diego, CA, USA), with statistical significance defined as p < 0.05.

Results

Effect of melatonin on pathological damage to neural tissues and cells in newborn rats with HI-induced WMD

HE staining revealed pathological changes in the white matter of the mouse brains at 7, 14, and 21 days in each group. The Sham group exhibited normal brain tissue structure characterised by tightly arranged, well-rounded cells with clearly visible nuclei (Fig. 2a–c). In contrast, the HI group displayed severe neural tissue and cell damage, including loose structures, sponge-like vacuolar changes, and noticeable nuclear pyknosis in numerous dead nerve cells (Fig. 2d–f). Following melatonin intervention, the HI group exhibited reduced vacuolar changes, improved cell membrane integrity, and significantly decreased nuclear pyknosis (Fig. 2g–i). However, in the SN6 group, the inhibition of NCX1 expression led to extensive liquefactive necrosis in tissues and cells, with the formation of some cavities (Fig. 2j-l).

Fig. 2.

H&E staining revealed pathological alterations in neural cells and tissues across all experimental groups. a–c, Normal brain tissue and cell morphology in the Sham group, characterised by tightly arranged tissue and plump cells. d–f, Cell necrosis in the HI group was accompanied by spongy vacuolisation and numerous instances of nuclear condensation. g–i, Intact cell membranes in the Mel group, with visible nuclei and dense tissue structures and no vacuolisation. (j–l). Cell necrosis and cavity formation in the SN6 group, with significant destruction of tissue structure. Scale bar = 100 µm, Sham group (n = 6); HI group (n = 6); Mel group (n = 6); SN6 group (n = 6)

Effect of melatonin on the excitability of nerve cells in newborn rats with HI-induced WMD

Whole-cell patch-clamp recordings revealed that HI injury significantly enhanced neuronal excitability compared to Sham. Under the same current stimulation, the action potential (AP) firing frequency increased (50 pA and 130–150 pA, p < 0.01; Fig. 3a–b), the threshold potential (HI: − 34.19 ± 2.61 mV vs. Sham: − 40.15 ± 3.20 mV, p < 0.05; Fig. 3c), and resting membrane potential (RMP) increased (HI: − 51.27 ± 5.95 mV vs. Sham: − 63.10 ± 2.92 mV, p < 0.001; Fig. 3d), suggesting that ion imbalance reduces the excitation threshold. Melatonin intervention effectively inhibited excitability, as indicated by a decrease in AP frequency (60–150 pA, p < 0.05–0.01, Fig. 3b), a significant reduction in threshold potential (Mel: − 47.26 ± 5.96 mV vs. HI: -34.19 ± 2.61 mV, p < 0.0001, Fig. 3c), and a significant reduction in RMP (Mel: − 58.85 ± 6.57 mV vs. HI: − 51.27 ± 5.95 mV, p < 0.05,Fig. 3d), approaching Sham levels. However, SN6 inhibition of NCX1 completely reversed the effects of melatonin, as evidenced by an increase in AP frequency (60–110 pA, p < 0.05, Fig. 3b), an increase in threshold potential (SN6: -38.51 ± 2.22 mV vs. Mel: − 47.26 ± 5.96 mV, p < 0.05, Fig. 3c), and an increase in RMP (SN6: − 50.51 ± 4.51 mV vs. Mel: − 58.85 ± 6.57 mV, p < 0.05, Fig. 3d), confirming that melatonin inhibits HI-induced neuronal hyperexcitability by regulating NCX1 activity.

Fig. 3.

Whole-cell patch-clamp experiments to evaluate the effect of melatonin on neuronal excitability in newborn rats with HI-induced WMD, a Voltage–time curves: Action potentials were normal in the Sham group. The frequency of delivery significantly increased in the HI group. The waveform of the Mel group was similar to that of the Sham group, with a lower frequency than that of the HI group. In contrast, delivery in the SN6 group was more intensive than that in the Mel group, b Number of peak potentials: The number of peak potentials was significantly greater in the HI group than in the Sham group when the same current was used to stimulate the same time points. The number decreased in the Mel group but increased again in the SN6 group. Data for each group are expressed as the mean ± SEM, n = 7–10. Statistical significance: HI vs. Sham: **p < 0.01, ***p < 0.001, and ****p < 0.0001; Mel vs. HI: #p < 0.05 and ##p < 0.05; SN6 vs. Mel: &p < 0.05, c–d Threshold potentials and resting membrane potentials: The results revealed that both the threshold potential and the resting membrane potential were greater in the HI group than in the Sham group. In the Mel group, the HI-induced increase was inhibited, whereas in the SN6 group, the effect of Mel was partially reversed. The data for each group are expressed as the mean ± SEM (n = 7–10,; *p < 0.05, ***p < 0.001, ****p < 0.0001)

Effect of melatonin on long-term learning and cognitive deficits in newborn rats with HI-induced WMD

Long-term neurofunctional damage caused by HI insult and its potential amelioration by melatonin treatment were evaluated using the Morris water maze test. During the spatial acquisition phase (Days 1–4), the escape latency and time spent in the target quadrant were normal in the Sham control group. Compared with the Sham control rats, the HI group rats exhibited significantly longer escape latencies (p < 0.001; Fig. 3c). Spatial memory deficits in HI rats were further evidenced by a reduced dwell time in the target quadrant (p < 0.05; Fig. 3d). Melatonin treatment markedly shortened escape latency (p < 0.01 vs. the HI group; Fig. 3c) and increased target quadrant preference (p < 0.01; Fig. 3d). However, SN6-mediated NCX inhibition reversed these improvements by prolonging escape latency (p < 0.05 vs. the Mel group; Fig. 3c) and decreasing target quadrant exploration (p < 0.01; Fig. 3d) (Fig. 4).

Fig. 4.

Water maze experiment to evaluate the effects of melatonin on learning and cognitive deficits in newborn rats with HI-induced WMD, a Schematic diagram of the water maze, b Path diagram of 28-day-old rats on Day 4 of the spatial navigation stage of the water maze, c Avoidance latency (sec) in the spatial navigation stage of the water maze in 28-day-old rats: The avoidance latency in the HI group was significantly shorter than that in the Sham group, and the time spent in the Mel group was significantly shorter than that in the HI group after drug administration for brain damage; the time spent in the SN6 group was significantly longer than that in the Mel group (n = 6), d Distance swam by the rats in each group during the spatial navigation phase (m). n = 6, data represent the mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ns > 0.05

Protective effect of melatonin on myelin development in newborn rats with HI-induced WMD

LFB staining revealed significant differences in myelin integrity at 7, 14, and 21 days post-intervention. In the Sham group, the myelin integrity was excellent at all time points (7, 14, and 21 days), with the myelin fibres exhibiting deep blue staining and displaying dense, well-organised structural arrangements (Fig. 5a). Compared with that in the Sham group, the development of myelin in the HI group was significantly impaired, the staining pattern was blurred (Fig. 5a), the structure was disrupted, and the average optical density (AOD) value of myelin was significantly lower on Days 7, 14, and 21 (P < 0.01, P < 0.001, and P < 0.001, respectively; Fig. 5b). In contrast, the development of myelin in the Mel group presented an orderly arrangement of nerve fibres, dense structural integrity, and strong staining (Fig. 5a). Quantitative analysis revealed that the AOD value of the Mel group was significantly greater than that of the HI group at each time point (p < 0.01, p < 0.001, and p < 0.001, respectively; Fig. 5b). However, compared with the Mel group, the SN6 treatment group presented structural abnormalities, such as myelin disorder and loosening (Fig. 5a), and the myelin staining intensity was lower. (p < 0.001, p < 0.05, p < 0.001; respectively; Fig. 5b). These findings further confirmed that the activity of NCX1 was essential for the protective effects of melatonin on myelin development.

Fig. 5.

Evaluation of the protective effect of melatonin on myelin development in HI-induced neonatal rats by LFB staining of myelin sheaths and the expression of myelin marker proteins. a Myelin LFB staining (scale bar = 50 µm) was performed on rats aged 7, 14, and 21 days. The Sham group exhibited darker blue staining, indicating a dense myelin structure. In the HI group, the blue staining was relatively weak, reflecting a loose structure with disorganised nerve fibres. Compared with the HI group, the Mel group presented darker staining, indicating a denser myelin structure, b The mean optical density values of myelin LFB staining for the various groups (n = 6), c–e The expression of the myelin proteins PLP and MAG decreased in the HI group, increased in the Mel group, and decreased again in the SN6 group (n = 4). The data represent the mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001

The Wes system approach was employed to assess the therapeutic efficacy of melatonin on HI-induced myelination impairment (Fig. 5c–e). In the Sham group, both myelin proteolipid protein (PLP) and myelin-associated glycoprotein (MAG) were highly expressed. Compared with that in the Sham group, the expression of the myelin-associated markers PLP and MAG in the HI group was markedly lower postinjury (PLP: p < 0.001, p < 0.05, and p < 0.001, respectively; Fig. 8c and h; MAG: p < 0.0001 and p < 0.001, respectively). The expression of PLP and MAG at each time point after melatonin was significantly upregulated compared with that in the HI group (PLP: p < 0.001, p < 0.05, and p < 0.01; Fig. 5c–d; MAG: p < 0.01 and p < 0.0001; Fig. 5c and e). In contrast, the SN6 intervention group demonstrated a significant reversal of the effect of melatonin, with the expression levels of both PLP and MAG decreasing markedly (PLP: p < 0.001, p < 0.05, and p < 0.01, respectively; Fig. 5c–d; MAG: p < 0.01, and p < 0.0001, respectively; Fig. 5c and e). These results confirmed that the promotion of myelin-associated protein expression by melatonin was related to NCX1 activity at the molecular level.

Fig. 8.

Simple Western blot analysis to evaluate the effect of melatonin on the expression levels of mitochondrial autophagy-related proteins. (a and e), Simple Western blot analysis of target protein expression, b–d and f–h Peak area under the curve; a–d The expression of the autophagy-related proteins LC3β, Parkin, and PINK1 was upregulated in the HI group significantly suppressed in the Mel group but partially reversed in the SN6 group, indicating a mitigation of the effect of melatonin (n = 4); e–h There were no significant differences in the expression of autophagy-related proteins among the S, S + M, and S + M + N groups (n = 3). Data represent the mean ± SEM; ns p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

The ultrastructures of the cells were observed in detail by transmission electron microscopy (Fig. 6). Histological examination revealed significant morphological differences among the experimental groups. In the Sham group, the myelin sheath structure was intact, with tightly and orderly arranged lamellae presenting a typical concentric circular lamellar structure, uniform myelin thickness, and no vacuoles or separation between lamellae (Fig. 6a). In contrast, the HI group exhibited severe damage to the myelin sheath structure, with loose and broken lamellar structures, vacuolisation in some areas, and myelin sheath delamination in some regions (Fig. 6b). The myelin sheath structure in the melatonin-treated group significantly improved, with closely rearranged lamellae and reduced vacuolisation (Fig. 6c). However, in the SN6 intervention group, after NCX1 expression was inhibited, the myelin sheath structure exhibited significant abnormalities. Compared with those in the Mel group, the lamellae in the SN6 group were disordered and presented wavy distortions, and complete separation of the myelin lamellae was observed in some areas (Fig. 6d). These results confirmed that alterations in NCX1 activity could affect myelin development at the ultrastructural level and that melatonin might maintain the integrity of the myelin sheath structure through an NCX1-dependent mechanism.

Fig. 6.

Transmission electron microscopy was used to evaluate the protective effects of melatonin on myelin development in newborn rats with HI-induced WMD. a–d Myelin ultrastructure (scale bar = 200 nm): a In the Sham group, the myelin plate layers were dense and well arranged, b In the HI group, the myelin plate layers are twisted, and interplate vacuoles are formed, c In the Mel group, the myelin plate layers are well arranged among the structures, d In the SN6 group, the myelin plate layers are lax and disorganised, e–h Mitochondrial ultrastructure (scale bar = 500 nm), e Mitochondria in the Sham group were intact with visible intracellular mitochondrial cristae, f Mitochondrial cells in the HI group were necrotic and liquefied, with vacuoles formed and mitochondrial cristae disappeared, g The mitochondrial chromatin in the Mel group was clear, with some visible mitochondrial cristae, h Mitochondria in the SN6 group were disrupted, and vacuole-like degeneration occurred, i–j Mitochondrial autophagosomes (scale bar = 500 nm): Autophagolysosomes with a double-layered membrane structure

Melatonin regulates NCX1 expression to alleviate mitophagy via the PINK1-Parkin pathway

Immunohistochemical and Wes system analyses revealed marked upregulation of NCX1 expression during the acute phase of HI-induced WMD (AOD: p < 0.01; Fig. 7a–b; protein: p < 0.01, Fig. 7c–d). However, HI markedly reduced NCX1 expression at 14 and 21 days postinjury (AOD: p < 0.001; Fig. 7a–b; protein: p < 0.0001 and p < 0.01; Fig. 7c–d). Notably, melatonin treatment effectively attenuated the early HI-induced increase in NCX1 expression (AOD: p < 0.01; Fig. 7a–b; protein: p < 0.05, Fig. 7c–d) and reversed the decrease in NCX1 expression on Days 14 and 21 (AOD: p < 0.05 and p < 0.01; Fig. 7a-b; protein: p < 0.001 and p < 0.01; Fig. 7c–d). The drugs used in the SN6 group inhibited NCX1 activity and reversed the effects of melatonin. At 7 days of age, NCX1 expression was upregulated compared with that in the Mel group (AOD: p < 0.01; Fig. 7a–b; protein: p < 0.05; Fig. 7c–d). However, on Days 14 and 21, NCX1 expression was significantly decreased (AOD: p < 0.01, p < 0.001; Fig. 7b; protein: p < 0.001 and p < 0.01; Fig. 7c–d). These results confirm that melatonin bidirectionally regulated NCX1 activity.

Fig. 7.

Immunohistochemical and simple Western blot analysis to evaluate the effect of melatonin on NCX1 expression activity. a Immunohistochemistry was used to assess the expression levels of NCX1 in neonatal rats at 7, 14, and 21 days. In the Sham group, normal NCX1 protein expression was observed in the brain. Compared with the Sham group, the HI group showed significantly greater NCX1-positive staining at 7 days of age after the brain injury. However, the positive results for NCX1 were diminished at 14 and 21 days of age. In contrast, strong positive expression was demonstrated in the Mel group at 14 and 21 days of age. Compared with the Mel group, the SN6 group exhibited high expression at 7 days of age and low expression at 14 and 21 days of age; scale bar = 20 µm, b Average optical density (AOD) of NCX1 staining in each group (n = 6), c–d NCX1 protein expression exhibited dynamic changes in the HI group, with an increase observed on postnatal Day 7 compared with that in the Sham group but a significant decrease on Days 14 and 21. This trend was reversed in the Mel group, whose expression was lower than that in the HI group on Day 7 but significantly greater on Days 14 and 21. In the SN6 group, expression was upregulated on Day 7 but decreased at subsequent time points (n = 4), e There were no significant differences in the expression of NCX1 or autophagy-related proteins among the S, S + M, and S + M + N groups (n = 3); the data represent the mean ± SEM; ns p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

However, when sham-treated rats were treated with melatonin alone (S + M group) or in combination with SN6 (S + M + N group), NCX1 expression did not significantly differ from that in the Sham group (p > 0.05; Fig. 7c and e). These findings suggested that melatonin did not influence NCX1 expression in newborn rats in the absence of HI-induced brain damage.

Prolonged HI injury led to ATP depletion in neurons, significantly altering mitochondrial structure, as confirmed by transmission electron microscopy (Fig. 6). In the Sham group, the mitochondrial structure was intact, the cells were plump, and the chromatin was evenly distributed with visible mitochondrial cristae (Fig. 6e). In the HI group, the cytoplasmic part of the mitochondria dissolved, leading to liquid necrosis and the formation of numerous cavities, with the disappearance of the mitochondrial cristae (Fig. 6f). This ultimately led to the formation of autophagosomes with visible double-membrane structures (Fig. 6i-f). Conversely, in the melatonin group, the mitochondrial structure of the rats remained intact, the cell morphology was good, and the nuclear chromatin was evenly distributed (Fig. 6g). In the SN6 group, the inhibition of NCX1 expression resulted in damaged mitochondrial structures and vacuolar degeneration (Fig. 6h). These experiments confirmed that melatonin had a protective effect on mitochondria after HI-induced brain injury.

We measured the expression of mitochondrial autophagy-related proteins in each group of rats on Days 7, 14, and 21 (Fig. 8). The results revealed that the expression of the mitochondrial autophagy-related proteins LC3β, Parkin, and PINK1 in the HI group was significantly upregulated at 7, 14, and 21 days (LC3β: All P < 0.001; Fig. 8a–b; Parkin: P < 0.01, P < 0.001, and P < 0.01, respectively; Fig. 8a and c; PINK1: All P < 0.0001; Fig. 8a and d). However, compared with that in the HI group, the expression of these autophagy-related proteins in the Mel group was significantly downregulated after melatonin treatment (LC3β: All P < 0.001; Fig. 8a-b; Parkin: P < 0.01, P < 0.001, and P < 0.01; Fig. 8a and c; PINK1: All P < 0.0001; Fig. 8a and d). In contrast, in the SN6 group, the expression of autophagy proteins increased significantly after treatment with both melatonin and NCX1, which reversed the results observed in the Mel group (LC3β: All P < 0.05; Fig. 8a-bb; Parkin: P < 0.01, P < 0.05, and P < 0.01; Fig. 8a and c; PINK1: P < 0.05, P < 0.05, and P < 0.0001; Fig. 8a and d).

In summary, melatonin downregulated the expression of mitochondrial autophagy proteins, indicating a protective effect against HI-induced mitochondrial autophagy. However, melatonin did not influence NCX1 expression when SN6 inhibited NCX1 expression. These findings demonstrate that the protective effect of melatonin on mitochondria was closely associated with NCX1 activity. However, the expression levels of Parkin, PINK1, and LC3B in the S + M group and S + M + N group did not significantly differ from those in the Sham group (p > 0.05; Fig. 8e–h). These findings indicate that melatonin and SN6 treatments did not influence the mitophagy phenotypes of neonatal rats in the absence of HI-induced brain damage.

Discussion

The results of the present study indicate that melatonin attenuated HI-induced mitochondrial autophagy by regulating the expression of NCX1, and the protective effects of melatonin on the nervous system following WMD were explored. To the best of our knowledge, this is the first study to elucidate the effects of HI injury on mitochondrial function, particularly the dynamic changes in NCX1 activity and mitochondrial autophagy. NCX1 is a key transmembrane protein in the mitochondria that regulates the intracellular calcium ion concentration primarily through reverse transport and is thus essential for maintaining intracellular calcium ion homeostasis (Bano and Nicotera 2007; Yao et al. 2023). We observed that the mitochondrial membrane potential was disrupted in the early stages of HI injury. To maintain intra- and extracellular homeostasis, the expression of NCX1 was increased in a compensatory manner. However, we observed a marked decrease in the expression of myelin-associated proteins and in LFB-positive myelin staining. An increase in NCX1 expression is an early compensatory response to HI; however, rather than mitigating damage, it exacerbates energy crisis and oxidative stress, ultimately leading to impaired myelin synthesis and cellular dysfunction. With prolonged HI, ATP depletion worsens, activating the reverse mode of NCX1 activity and leading to the translocation of calcium ions into the mitochondria and sodium ions out of the mitochondria, thereby exacerbating intracellular (especially intramitochondrial) calcium overload (Hirota et al. 2007; Khananshvili 2014). Calcium overload is the key initiator of cellular injury.

Mitochondrial autophagy is an important cellular protective mechanism that plays a crucial role in maintaining cellular homeostasis and survival by removing damaged mitochondria and preventing further deterioration of the intracellular environment (Senft and Ronai 2015; Yamashita and Kanki 2017). As the injury progresses, ATP depletion in the mitochondria intensifies, and NCX1 expression significantly decreases, leading to reduced intracellular calcium ion clearance and continuous calcium ion accumulation. This high-calcium environment enhances the activity of the mitochondrial respiratory chain, generating excessive amounts of reactive oxygen species (ROS) that further damage the mitochondrial membrane and DNA. This results in a sustained loss of membrane potential and ultimately triggers severe mitochondrial dysfunction (Sennoune et al. 2015).

Mitochondrial dysfunction is a crucial signalling pathway that initiates mitochondrial autophagy. The pathways associated with mitochondrial autophagy include mainly the PINK1-Parkin, BNIP3/NIX, and FUNDC1 pathways, with the PINK1-Parkin pathway recognised as a canonical pathway (Ashrafi and Schwarz 2013; Tur et al. 2020). Our study confirmed that the protein expression of the mitochondrial autophagy markers LC3β, PINK1, and Parkin significantly increased after HI injury, and transmission electron microscopy revealed an increased number of cytosolic lysosomes involved in the autophagy pathway. These findings clearly indicate that the PINK1-Parkin-dependent mitochondrial autophagy pathway is strongly activated after HI injury (Yang et al. 2013; Onishi et al. 2021). Autophagosomes recognise and engulf damaged mitochondria to form autolysosomes, which then fuse with lysosomes to digest damaged components and complete the mitochondrial clearance process (Lazarou et al. 2015; Choong et al. 2021).

Through comprehensive whole-cell brain slice membrane clamp experiments, we observed that HI injury led to an abnormal increase in neuronal excitability, as evidenced by a significant increase in the resting potential and threshold potential. The biological basis of hyperexcitability was closely associated with NCX1 dysfunction under HI conditions. The resting potential reflects the potential difference between the inside and outside of the cell membrane, whereas the threshold potential is the critical level required to trigger action potentials. Under HI conditions, when cells were subjected to the same intensity of electrical stimulation, the frequency of action potential occurrence increased significantly compared with that under normal conditions. Both the resting potential and the threshold potential were elevated, indicating that the cells required a smaller stimulus to reach the threshold potential from the resting potential, thus demonstrating increased excitability post-HI injury. Notably, the attenuation of neuronal hyperexcitability in the central nervous system (CNS) due to early HI injury is crucial for preventing subsequent epileptic development (Pissas et al. 2023).

The central finding of this study was that melatonin exerted neuroprotective effects by regulating mitochondrial NCX1 expression. Melatonin has multiple biological functions, including antioxidant, anti-inflammatory, and antiapoptotic effects, as well as protective effects against WMD (Revuelta et al. 2016; Cardinali 2019). The administration of melatonin following neonatal HI injury has been shown to reduce cell death and reactive astrocyte proliferation, protect the CNS (D'Hooge and De Deyn 2001), restore changes in auditory evoked potentials, and mitigate hypothalamic morphological damage in perinatally asphyxiated rats (Alonso-Alconada et al. 2012). Melatonin aids in maintaining normal NCX1 function and neuronal survival by stabilising mitochondrial function, promoting continuous mitochondrial energy metabolism in the brain, and alleviating imbalances in the mitochondrial membrane potential.

The protective effect of melatonin on mitochondria significantly improves the neural structure and function. It modulates neuronal excitability by regulating NCX1 activity, a mechanism distinct from previously reported mechanisms involving voltage-gated ion channels or MT1 receptors (Xu et al. 2017; Oliveira-Abreu et al. 2018). Additionally, melatonin effectively protected myelin from HI damage, as evidenced by the upregulated expression of myelin-related proteins (e.g., MAG and PLP) and improved myelin ultrastructure observed by transmission electron microscopy. These findings confirm the role of melatonin in promoting CNS myelin formation (Núñez et al. 2024), which is consistent with studies showing an increase in the number of PLP + , MBP + , and MAG + cells (Jiang et al. 2021; Tapia-Bustos et al. 2021). Moreover, the effect of melatonin on myelin sheath formation in early life differs from that in later life. Importantly, sustained melatonin intervention early in life significantly ameliorated long-term learning and cognitive dysfunction due to HI injury, a cognitive protective effect reported in other disease models (e.g., Alzheimer's disease and depression) (Taleb and Alghamdi 2019; Chen et al. 2021), which has important clinical implications.

To verify the mechanism underlying the neuroprotective effect of melatonin through NCX1, we conducted experiments using the NCX inhibitor SN6. The results revealed that in neonatal rats with hypoxia-reperfusion injury, SN6 treatment negated the protective effects of melatonin on mitochondrial structure and function and significantly increased the expression of LC3β and the PINK1/Parkin pathway, similar to those in the HI group. These findings indicate that inhibiting NCX1 activity exacerbated mitochondrial autophagy. SN6 impaired the ability of melatonin to stabilise mitochondrial calcium homeostasis and energy metabolism by blocking NCX1, leading to mitochondrial damage and autophagy activation. These findings confirm that melatonin inhibited excessive mitochondrial autophagy and exerted neuroprotective effects by regulating NCX1 activity.

Conclusion: We systematically investigated the effects of HI injury and melatonin intervention on the expression of NCX1, a key regulator of mitochondrial function and autophagy, and assessed neurological outcomes. HI injury induced abnormalities in mitochondrial NCX1 expression (elevated in the early stage and decreased in the late stage) by causing an overload of calcium, bursts in reactive oxygen species, and overactivation of mitochondrial autophagy. This triggered abnormal neuronal excitability (elevated resting/threshold potential), myelin damage (MAG/PLP downregulation and ultrastructural changes), and cognitive dysfunction. Melatonin effectively inhibited aberrant NCX1 expression and attenuated mitochondrial damage and autophagy overactivation through its antioxidant, anti-inflammatory, and calcium homeostasis-stabilising effects, thereby improving neuronal function, protecting myelin, and enhancing cognitive performance. These findings highlight the neuroprotective role of melatonin in neonatal HI-induced WMD.

As an endogenous hormone with low toxicity and high efficiency, melatonin has great potential for clinical application and may be a promising and relatively safe protective agent for treating neonatal HI-related brain injury, especially WMD. The results of this study provide a crucial scientific basis for the future development of melatonin-based neuroprotective strategies for treating neonatal HI-induced WMD.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

HI

Hypoxia–ischaemia

WMD

White matter damage

Mel

Melatonin

P3

Postnatal day 3

SD

Sprague‒Dawley

NCX1

Sodium-calcium exchanger 1

MAG

Myelin-associated glycoprotein

PLP

Proteolipid protein

LC3B

Light chain 3B

PINK1

PTEN-induced kinase 1

LFB

Luxol fast blue

tsDCS

Transcutaneous direct current stimulation in the spine

rTMS

Repetitive transcranial magnetic stimulation

EPG

Electromagnetic-pulsed gene

ROS

Reactive oxygen species

P2X7

P2X purinoceptor 7

NLRP3

NOD-like receptor family pyrin domain containing 3

JAK2-STAT3

Janus kinase 2-Signal transducer and activator of transcription 3

ERK1/2

Extracellular signal-Regulated Kinase 1/2

BAX

Bcl-2-associated X protein

PEPT1/2

Peptide Transporter 1/2

SIRT3

Sirtuin 3

TFAM

Transcription Factor A, Mitochondrial

AKI

Acute Kidney Injury

SERCA2b

Sarcoplasmic/Endoplasmic Reticulum Calcium ATPase 2b

Author contributions

Tongfei Cheng and Yingjun Xu conceptualized and designed the study, drafted the initial manuscript, and critically reviewed and revised the manuscript. Tongfei Cheng and Ziyan Lu designed the data collection instruments, collected data, carried out the initial analyses, and critically reviewed and revised the manuscript. Shanlong Du designed the data collection instruments, collected data, carried out the initial analyses, and critically reviewed and revised the manuscript. Yi Cao conceptualized and designed the study, coordinated and supervised data collection, and critically reviewed and revised the manuscript for important intellectual content. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

Funding

This study was supported by grants from the Medical and Health Science and Technology Project of Shandong Province (No. 202406031337).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Conflict of interest

The authors declare no competing interests.

Ethical approval

All animal procedures were approved by the Animal Ethics Committee of Qingdao University (Approval No. 20230907SD42020240328037).