Proteomic analysis of the effects of caffeine in a neonatal rat model of hypoxic‐ischemic white matter damage

Abstract Aim White matter damage (WMD) is the main cause of cerebral palsy and cognitive impairment in premature infants. Although caffeine has been shown to possess neuroprotective effects in neonatal rats with hypoxic‐ischemic WMD, the mechanisms underlying these protective effects are unclear. Herein, proteins modulated by caffeine in neonatal rats with hypoxic‐ischemic WMD were evaluated. Methods We identified differential proteins and performed functional enrichment analyses between the Sham, hypoxic‐ischemic WMD (HI), and HI+caffeine‐treated WMD (Caffeine) groups. Confirmed the changes and effect of proteins in animal models and determined cognitive impairment via water maze experiments. Results In paraventricular tissue, 47 differential proteins were identified between the Sham, HI, and Caffeine groups. Functional enrichment analyses showed that these proteins were related to myelination and axon formation. In particular, the myelin basic protein (MBP), proteolipid protein, myelin‐associated glycoprotein precursor, and sirtiun 2 (SIRT2) levels were reduced in the hypoxic‐ischemic WMD group, and this effect could be prevented by caffeine. Caffeine alleviated the hypoxic‐ischemic WMD‐induced cognitive impairment and improved MBP, synaptophysin, and postsynaptic density protein 95 protein levels after hypoxic‐ischemic WMD by preventing the HI‐induced downregulation of SIRT2; these effects were subsequently attenuated by the SIRT2 inhibitor AK‐7. Conclusion Caffeine may have clinical applications in the management of prophylactic hypoxic‐ischemic WMD; its effects may be mediated by proteins related to myelin development and synapse formation through SIRT2.


| INTRODUC TI ON
White matter damage (WMD) is a common type of brain injury occurring in premature infants and is among the main causes of cerebral palsy and cognitive impairment. [1][2][3] The main pathological features of brain WMD in preterm infants are brain atrophy, brain white matter area injury, and periventricular leukomalacia, which is characterized by oligodendrocyte maturation disorder and subsequent myelin sheath dysplasia. [4][5][6] In addition, WMD also damages axon growth, synaptogenesis, and neurogenesis. 7 Recent studies have suggested that the pathogenesis of WMD includes oxidative stress, inflammation, free radical damage, cytokine toxicity, and glutamate excitotoxic damage. 8 Among these, oxidative stress and inflammation are the two main causes of WMD in premature infants. 9 Although the etiology, pathogenesis, prevention, and treatment of WMD have been widely studied, effective neuroprotective strategies against WMD are still lacking.
Caffeine is a methylxanthine drug. Its early application in preterm infants can help prevent and treat apnea, can effectively reduce the incidence of bronchopulmonary dysplasia, and may exert long-term beneficial effects on the nervous system. 10 Caffeine can protect cells against oxidative stress and neuroinflammation, thereby reducing apoptosis and synaptic defects in developing neurons, reducing ventricular enlargement and white matter loss, and alleviating myelination disorders. 11,12 Additionally, caffeine exerts beneficial effects on neurodegenerative diseases, including Parkinson's and Alzheimer's disease. [13][14][15] Although caffeine has been shown to exert neuroprotective effects on WMD in premature infants, 11 the underlying neuroprotective mechanisms remain unclear.
Caffeine primarily acts as an antagonist of adenosine receptors at nontoxic doses/concentrations. 16 In the brain, the activity of caffeine involves A1 and A2a adenosine receptors 17 and indeed, "physiological" concentrations of caffeine only affect neuronal function through the action of these two adenosine receptors. 18 Furthermore, the brain neuroprotection afforded by caffeine is fully accounted for by the antagonism of adenosine A2A receptors, as recently extensively reviewed. 19 However, the molecular mechanisms engaged by the caffeine-mediated antagonism of A2A receptors to induce this neuroprotection have not yet been identified. Therefore, in this study, we examined the specific proteins and molecular mechanisms regulated by caffeine in a rat model of WMD caused by hypoxia-ischemia.

| Experimental animals and study group establishment
All animal experiments were approved by the Animal Ethical Committee of China Medical University, Shenyang, China (approval no. 2017PS140K). Perinatal Sprague-Dawley rats were purchased from Liaoning Changsheng Biotechnology Co., Ltd. All rats were housed in a facility with a 12-h light/dark cycle and were provided with free access to food and water.
There were approximately 10-16 newborn rats in each litter, and the birth weight was approximately 7.3 g (7.26 ± 0.325 g). All the rats were cross-fostered between the dams. The animals were randomly divided into the following four groups: sham (Sham), hypoxic-ischemic WMD (HI), HI+caffeine-treated WMD (Caffeine), and caffeine with SIRT2 inhibitor (Caffeine+AK-7) groups. In total, 187 animals were used in this experiment. From days 2-6 after birth, an intraperitoneal injection of 20 mg/kg/day caffeine citrate (Casey Pharmaceuticals) or an equal volume of normal saline was injected daily for 5 consecutive days. Rats in the caffeine+AK-7 group were subjected to treatment with caffeine and the SIRT2 inhibitor AK-7 (20 mg/kg/day, 5 days, intraperitoneal injection) from days 2-6 after birth.

| Hypoxia-ischemia-induced brain WMD model in neonatal rats
Based on previously described methods, 7,20,21 a rat model of neonatal brain WMD caused by hypoxia-ischemia was established.
Three-day-old male and female Sprague-Dawley neonates were anesthetized via subjection to inhalation of isoflurane and were fixed on an operating table in the supine position. The left common carotid artery was exposed under a dissecting microscope and was permanently ligated in the HI group using 2.0 sterile needle sutures.
At both ends of the artery, the blood vessel was cut in the middle of the two ligature points, and the wound was sutured. The total operation time was 8-10 min. After completion of the operation, the rats regained consciousness and were housed with the mother to recover for 1 h; thereafter, they were placed in a hypoxic box in a constant temperature water bath (37°C). Mixed gas (8% O 2 + 92% N 2 ) was continuously provided as input for 2.5 h at a flow rate of 2 L/min, and the oxygen concentration was maintained at 8%. In the Sham group, the left common carotid artery was separated without ligation and hypoxia treatment.
The specimens were collected 14 and 21 days after model establishment, and 20 rats from each group were randomly selected at each time point for paraffin sectioning and western blot analysis; six randomly selected rats from each group were statistically analyzed.
Additionally, all six rats in each group were subjected to Morris water maze (MWM) test after model establishment.

| Sample preparation for proteomic analysis
Specimens were collected at 14 days after model establishment.
After stripping the cortex, the brain tissues were isolated from the ligated side midbrain paraventricular areas on ice, which was rapidly flash-frozen in liquid N 2 and stored at −80°C. Frozen brain specimens were weighed and then ground in liquid nitrogen. Each sample was transferred to a 5-ml centrifuge tube to which four volumes of lysis buffer (8 M urea, 1% Protease Inhibitor Cocktail) was added, and sonicated three times on ice using a high-intensity ultrasonic processor (Scientz). The remaining debris was removed via centrifugation at 12,000 g at 4°C for 10 min. Finally, the supernatants were collected, and the protein concentration was determined using a BCA Kit according to the manufacturer's instructions (Beyotime Biotechnology). For digestion, the protein solution was subjected to reduction with 5 mM dithiothreitol for 30 min at 56°C and was alkylated with 11 mM iodoacetamide for 15 min at room temperature (24-26°C) in the dark. The protein sample was diluted by adding 100 mM tetraethylammonium bromide (TEAB) to a urea concentration of <2 M. Finally, trypsin was added (1:50 trypsin-to-protein mass ratio) for the first digestion overnight and at a 1:100 trypsinto-protein mass ratio for the second 4-h digestion.

| Liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis
The tryptic peptides were dissolved in 0.1% formic acid in 2% acetonitrile (solvent A) and were directly loaded onto a reversed-phase analytical column (15 cm length, 75 μm internal diameter) packed with 1.9 μm/120 Å ReproSil-PurC18 resins (Dr. Maisch GmbH). The gradient was as follows: increase from 5% to 25% solvent B (0.1% formic acid in 90% acetonitrile) over 90 min, 25% to 35% in 22 min, increase to 80% in 4 min, and hold at 80% for the last 4 min. All steps were performed at a constant flow rate of 550 nl/min using an EASY-nLC 1000 UPLC system (Thermo Fisher Scientific).
The peptides were subjected to nanospray ionization source followed by tandem mass spectrometry (MS/MS) using the Q Exactive Plus (Thermo Fisher Scientific) coupled online to an ultraperformance liquid chromatography instrument. The electrospray voltage applied was 2.0 kV. The m/z scan range was 400-1500 for a full scan, and intact peptides were detected in the Orbitrap at a resolution of 60,000. Peptides were then selected for MS/MS by setting the normalized collision energy (NCE) to 20, and fragments were detected in the Orbitrap at a resolution of 15,000. A data-dependent procedure was then performed that was alternated between one MS scan followed by 20 MS/MS scans with 15.0-s dynamic exclusion.
Automatic gain control was set at 5E4, and the fixed first mass was set to 100 m/z. The MS/MS data obtained were processed using the MaxQuant search engine (v1.6.15.0) and were explored against the Rattus norvegicus database (uniport database, 29940 sequence. Download 20201214) concatenated with a reverse decoy database.
Extract the LFQ intensity in the database search result file, the row direction normalized by mean, and get the relative quantitative value of each sample.

| Bioinformatic analysis
The R package clusterProfiler was used for Gene Ontology (GO) analysis, and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed to predict the functions of differentially expressed peptides and their precursor proteins. Protein-protein interaction networks were mapped using the STRING database (https://strin g-db.org/) and UniProt database (https://www.unipr ot.org/). The isoelectric points and masses of peptides were determined using the tools available at https://www.expasy.org/ and https://smart.embl-heide lberg.de/. Calculate the average value of the quantitative values of each sample in multiple replicates, and then calculate the ratio of the average values between the two samples. The relative quantitative values of each sample were taken as log2 transform (so that the data conforms to the normal distribution), and p value was calculated by the two-sample two-tailed T-test method.

| Immunohistochemistry
The coronal tissue sections used for the detection of paraventricular white matter ranged from 1.5 mm before bregma to 0.5 mm after bregma. The detection areas were mainly the corpus callosum

| Luxol fast blue (LFB) staining
Brain slices were deparaffinized in xylene and were dehydrated in alcohol, immersed in 95% ethanol for 5 min at room temperature and then immersed in LFB solution (cat. no. G3245; Solarbio) overnight.
The sections were immersed in 95% and 70% hydrochloric acid ethanol until gray and white matter could be distinguished. 22

| Western blotting
At 14 and 21 days after HI injury, rats were anesthetized with 3% isoflurane inhalation, the brains were harvested, and the brain tissues were isolated from the ligation side paraventricular areas on ice and were stored at −80°C. Samples were processed for western blotting, as per methods described previously. 23 Brain tissue were lysed on ice in RIPA buffer containing phosphatase inhibitors (Sigma-Aldrich). Protein concentrations were determined using a BCA kit no. 10068-1-AP; Proteintech) was used as the loading control. After incubation with horseradish peroxidase-conjugated goat anti-rabbit antibodies (1:5000; cat. no. SA00001-2; Proteintech) or anti-mouse antibodies (1:5000; cat. no. SA00001-1; Proteintech) and development using enhanced chemiluminescence reagents (Thermo Fisher Scientific), the intensities of all bands were analyzed using the ImageJ software and were normalized against β-tubulin.

| Morris water maze test
The Morris water maze (MWM) experiment was conducted from days 28 to 33 after model establishment. 24 The test comprised the following: a circular pool (diameter: 160 cm; height: 60 cm), a black inner wall, a movable platform (diameter: 12 cm; height: 28.5 cm), a computer, a camera, and image-and data-acquisition and processing systems. Different graphics were used to mark the pool wall at the midpoints of the four quadrants ( Figure 2B). Before the commencement of the experiment, water was poured into the pool to a depth of 30 cm and the height of the platform was 1.5 cm below the water surface. The test included two phases, i.e., acquisition training and the probe trial. During the training period, the rats were subjected to training for 5 consecutive days, four times per day, to search for a platform for 120 s. Once a rat was found and remained on the platform for 5 s, the training was deemed complete. The time from entering the water to finding the platform was defined as the escape latency, and the system software automatically analyzed the swimming distance of the rat during this period. If the rat could not find the platform within 120 s, it was guided to rest on the platform for 20 s and the escape latency was recorded as 120 s. On day 6, the platform was removed and the probe trial was conducted, during which the rat was placed at the opposite quadrant and was allowed to swim for 120 s. The data were recorded using a video tracking system (Shanghai Mobile Datum Ltd.). All tests were performed by researchers who were blinded to the experimental groups. To evaluate the role of caffeine in cognitive impairment, the MWM test parameters, including the escape latency, time spent in the target quadrant, frequency of platform crossing (times), and moving velocity (moving distance/120) (cm/s), were evaluated.

| Statistical analysis
All data are presented as the mean ± standard errors of means (SEMs). For normality assessment, the Kolmogorov-Smirnov test with Dallal-Wilkinson-Lillie correction for p values was used. Data related to latency escape or time were analyzed using two-way repeated analysis of variance (ANOVA) followed by Bonferroni's post hoc test. All other results among three or more experimental groups were analyzed using one-way ANOVA followed by Bonferroni's post-hoc test if the data were normally distributed. If the data were not normally distributed, the Kruskal-Wallis test was performed as a nonparametric test. GraphPad Prism software (v.8.01; GraphPad Software) was used for the analyses, and results with p < 0.05 were considered statistically significant.

| Caffeine administration resulted in differential proteins involved in hypoxic-ischemic WMD in neonatal rats
At 14 days after hypoxic-ischemic model establishment, proteins were extracted from the brain tissues of three rats each from the Sham, HI, and Caffeine groups, and analyzed directly using LC-MS/ MS ( Figure 1A,B). In total, 47 differential proteins (Sham vs. HI, p < 0.05 and fold change >1.2; HI vs. Caffeine, fold change >1.2) were identified ( Figure 1C). Compared with those in the Sham group, the levels of 27 proteins were found to be upregulated in the HI group and downregulated in the Caffeine group, whereas the levels of 20 proteins were downregulated in the HI group and upregulated in the Caffeine group. We ranked these proteins according to their P values, and the proteins showing the greatest upregulation and downregulated are listed in Table 1.

| Bioinformatics analysis showed that caffeine protected against hypoxic-ischemic WMD in neonatal rats
To preliminarily explore the functions of the differential proteins in neonatal rats with hypoxic-ischemic WMD, we used the R package clusterProfiler to analyze GO and KEGG pathway enrichment of the 47 overlapping expressed genes of differential proteins. In total, 125 GO biological processes, 16 GO cellular components, 27 GO molecular functions, and 11 KEGG pathways were enriched. In particular, we focused on detecting enrichment for myelin-related pathways, including myelination, regulation of myelination, myelin sheath formation, and structural constituent of myelin sheath, and synaptic-related pathways, including axon development, main axon formation, long-term synaptic potentiation, and response to axon injury. We ranked the results according to the P values, and the top 10 enriched GO terms and KEGG pathways are shown ( Figure 1D).
To identify important proteins mediating the protective effects of caffeine against brain WMD in neonatal rats, we used the STRING and UniProt databases to further evaluate the functions of differential proteins. We generated a protein-protein interaction network ( Figure 1E) focusing on the interaction effects between MBP, PLP1, MAG, and SIRT2, which are related to axon formation and myelination.

| Caffeine attenuated cognitive impairment related to hypoxic-ischemic WMD in neonatal rats
To verify the effects of caffeine on cognitive impairment associated with hypoxic-ischemic WMD in neonatal rats, we performed a water maze experiment at 28 days after hypoxic-ischemic model establishment. In the MWM experiment, the escape latency in the HI group on days 3-5 was found to be significantly longer than that in the Sham group (p all <0.001); this increase was attenuated by caffeine ( Figure 2A). In spatial probe trials, rats in the HI group showed a lower proportion of exercise time spent in the target quadrant  of times they crossed the platform (p < 0.001) compared with those in the Sham group. However, caffeine could attenuate the alterations in time spent in the target quadrant and platform crossing caused by HI (p < 0.01, p < 0.001) ( Figure 2C,D,F). There were no significant differences in moving velocity within 120s among the three groups ( Figure 2E). This indicated that the above-mentioned differences could not be explained by differences in athletic ability among the groups. These results indicated that caffeine reduced the cognitive dysfunction of neonatal rats with brain WMD caused by hypoxia-ischemia.

| Caffeine prevented myelin development disorders related to hypoxic-ischemic WMD in neonatal rats
Next, we explored the effects of caffeine on myelin sheath development in neonatal rats with hypoxic-ischemic WMD by further evaluating differential proteins related to myelination, such as MBP, PLP,  Figure 3E). The AODs for MBP and LFB staining in the Caffeine group were higher than those in the HI group (p < 0.001, p < 0.01) ( Figure 3F,G). The nerve fibers were arranged in an orderly and dense manner, and the myelin sheath was partially remodeled ( Figure 3E). These data indicated that caffeine alleviated myelination disorders after hypoxia-ischemiainduced WMD.

| Caffeine prevented reductions in synapse proteins induced by hypoxic-ischemic WMD in neonatal rats
We

| SIRT2 promoted the beneficial effects of caffeine on the prevention of hypoxic-ischemic WMD in neonatal rats
SIRT2 was one of the differential proteins identified in our proteomic analysis. We performed western blotting and immunohistochemistry to verify the changes in SIRT2 protein levels following treatment with caffeine to prevent the alterations caused by HI.
Western blotting results showed that at 14 days after hypoxiaischemia, SIRT2 protein levels were lower in the HI than that in the Sham group (p < 0.001), and caffeine prevented the HI-induced reductions in SIRT2 levels (p < 0.001) ( Figure 5A These results indicated that SIRT2 levels decreased in brain tissues after hypoxia-ischemic WMD and that caffeine can help alleviate myelination disorders and improve synapse proteins by preventing HI-induced reductions in SIRT2. Early studies have shown that caffeine exerts a neuroprotective effect on ischemic WMD. 11,12 Indeed, myelination is enhanced and ventriculomegaly is reduced in hypoxia-exposed neonatal pups 20 mg/kg daily maintenance dose) on the development of brain white matter in immature sheep also showed that there was no effect of daily high-dose caffeine on brain weight, oligodendrocyte density, neuronal density. 39 A recent study on the long-term effects of caffeine treatment on the nervous system of premature rabbits model also shows that caffeine treatment had no significant effect on behavior between preterm infants and full-term prepubescent individuals. Although the neuronal density of preterm infants is low, it is not affected by caffeine treatment. At the same time, there was no difference in synaptic density, astrocytes, and myelin sheath between the groups treated with or without caffeine. 40 Our study found that continuous use of caffeine before and after brain injury could alleviate long-term behavioral disorders caused by HI and played a neuroprotective role, which is consistent with the results of earlier studies.

| DISCUSS ION
At the same time, we identified the specific proteins regulated by caffeine in hypoxic-ischemic WMD, including MBP, PLP, and MAG.
Regulation of the expression levels of myelination-associated proteins is important for preserving the structural integrity and functionality of oligodendrocytes. Caffeine regulated the expression levels of myelination-related proteins and attenuated the reduction in oligodendrocyte maturation caused by hypoxic-ischemic WMD. In this study, the levels of myelination-associated proteins (PLP, MAG, and MBP) were found to be decreased in hypoxic-ischemic WMD, and caffeine alleviated this decrease.
Simple myelination disorder development cannot fully explain the diversity of long-term sequelae in preterm infants with WMD. Studies have suggested that synaptic damage is an important pathological basis for this diversity. 41 Synapse formation and remodeling are vital for brain development in premature infants. 42 Researchers have reported synaptic damage in animal models of hypoxic-ischemic encephalopathy, 43 resulting in abnormal limb movement and cognitive memory, among other issues. 44 There are numerous studies showing that caffeine controls synaptic dysfunction associated with neuropsychiatric disorders. [45][46][47] A recent study identified the main targets of nontoxic doses of caffeine as a key controller of synaptic stability 48 and showed that the exposure to caffeine during brain development affects susceptibility to brain dysfunction later in life. 49 In addition, caffeine can also affect hippocampal synaptic plasticity through A1 and A2A receptors. 50 Also, it has been shown that caffeine has profound effects on synaptic markers in younger age in rodents 51 and prevents memory dysfunction later in life after challenges applied in early life. 52 Caffeine treatment also prevents the sleep deprivation-induced downregulation of synaptophysin and synapsin I proteins with no change in PSD-95 protein in hippocampus. 53 Prenatal caffeine intake increases the number of NeuN-stained nuclei in the fetal brains of rats, and the maturation of telencephalon may be accelerated. At the same time, prenatal caffeine intake temporarily affects synaptic proteins and changes fetal brain development. 54 We also found that although caffeine did not significantly enhance the protein levels of the presynaptic protein Syp or PSD-95 at 14 days after hypoxia-ischemia, the levels of both Syp and PSD-95 were significantly increased at In our proteomics analysis, in addition to identifying the myelination-related protein regulated by caffeine in hypoxicischemic WMD, it is more important to identify the protein interacting with myelination-related protein-sirtuin 2 (SIRT2). As a nicotinamide adenine dinucleotide-dependent deacetylase, SIRT2 is mainly expressed in myelinating glia of the central nervous system (CNS) and plays an important role in neurological diseases. [55][56][57][58] Pathological features, such as neuroinflammation, synaptic dysfunction, metabolic abnormalities, and oxidative stress, are affected by SIRT2 expression. 55 Early studies have shown a counterbalancing role of SIRT2 against a facilitatory effect of tubulin acetylation on oligodendroglial differentiation. Selective SIRT2 availability to oligodendroglia may have important implications for myelinogenesis, myelin-axon interaction, and brain aging. 59 Meanwhile, proteolipid protein is required for the transport of sirtuin 2 into CNS myelin. 60 In recent years, research has proved that SIRT2 expression enhances OL differentiation and arborization in vitro, 61 and plays a critical role in OL differentiation by promoting both arborization and downstream expression of myelin-specific genes. 62 In addition, SIRT2 also affects synaptic plasticity.
SIRT2 can act as a deacetylase for α-amino-3-hydroxy-5methyl-4-isoxazole propionic acid (AMPA), which interacts with the GluA1 subunit of AMPA receptor (AMPAR) and regulates AMPAR transport and protein homeostasis, thereby affecting learning and memory ability. 63 Although SIRT2 is widely thought to accelerate the development of neuropathology, 55 recent studies have shown that middle-aged mice lacking SIRT2 exhibit axon degeneration and motor dysfunction. 64 These contradictory findings indicate that SIRT2 level is necessary for normal axon formation; however, this protein functions differently after neurological injury. SIRT2 may have different effects on various brain injuries owing to the activation of diverse signaling pathways. Therefore, SIRT2 may function as a metabolic coordinator through deacetylation of substrate in disease. Our study found that SIRT2 was decreased in hypoxic-ischemic induced WMD, and that caffeine could prevent the HI-induced reductions of SIRT2. In the presence of the SIRT2 inhibitor AK-7, caffeine failed to enhance the expression of myelinating proteins and synaptic proteins after hypoxiaischemia. We speculate that the beneficial effects of caffeine on myelinating and synaptic proteins may be achieved by preventing the reduction in SIRT2 induced by HI. Although we did not include a control group in which AK-7 was applied alone in the HI group, many previous studies have revealed the safety of AK-7. In order to determine the dose-effect relationship of AK-7, Chopra et al.
first conducted tests using 10 and 20 mg/kg AK-7 tests in male C57BL/6 mice and found that these doses had no obvious side effects on Huntington's disease model mice. 65 In addition, a study on the inhibition of SIRT2 by AK-7 in rats with traumatic brain injury showed that AK-7 alone had no adverse effect in the control group, whereas AK-7 after brain injury could aggravate neuroinflammation and blood-brain barrier damage. 66 We speculate that SIRT2 may be necessary for damage repair, but that its function may differ before then. Therefore, the exact roles and mechanisms of SIRT2 in biological processes should be further studied in order to make rational use of SIRT2 as a potential therapeutic target.

| CON CLUS ION
Our results indicated that caffeine may interfere with oligodendrocyte maturation and with synaptic function, conferring protection to neonatal brain tissue against damage caused by hypoxia and ischemia. However, a major finding was the alteration of SIRT2, a major metabolic coordinator. These findings led to obvious conclusion that metabolic rebalancing may be another candidate mechanism to explain the neuroprotective effects of caffeine, in accordance with the well-established ergogenic effects of caffeine in sports physiology. However, we did not test the effects of AK-7 in control rats or of AK-7 alone in HI rats, and our study had some other limitations.
Based on our findings, we hypothesize observed effects were just the summation of the two independent effects caused by caffeine and AK-7. Therefore, further studies are needed to fully elucidate the molecular mechanisms involved. Based on the results obtained herein, we suggest that caffeine may be considered an effective drug for preventing WMD in premature infants; further studies on the molecular mechanisms and biological functions of caffeine may provide new insights into targets for promoting the use of caffeine as a neuroprotective agent.

ACK N OWLED G M ENTS
We would like to thank the families in the first neonatal ward of Shengjing Hospital Affiliated to China Medical University.

CO N FLI C T O F I NTE R E S T
The authors have no conflicts of interest to declare.

AUTH O R CO NTR I B UTI O N S
Jianhua Fu and Xindong Xue made substantial contributions to conception and design of the experiment. Liu Yang, Xuefei Yu, Yajun Zhang, Na Liu, and Danni Li contributed to the acquisition, analysis, and interpretation of data. All authors approved the final version of the manuscript for publication.

CO N S ENT TO PA RTI CI PATE
All animal data reported in this study were obtained following ARRIVE guidelines.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.