Caffeine Treatment Promotes Differentiation and Maturation of Hypoxic Oligodendrocytes via Counterbalancing Adenosine 1 Adenosine Receptor-Induced Calcium Overload.

BACKGROUND We aimed to explore the involvement of adenosine 1 adenosine receptor (A1AR) in hypoxia-induced poor differentiation of oligodendrocytes (OLs), and the underlying mechanism of caffeine treatment in hypoxic injuries. MATERIAL AND METHODS Real-time polymerase chain reaction (RT-PCR) was used to assess the alterations of AR expression in cultured hypoxic OLs with or without caffeine treatment. Then, intracellular alterations of Ca²⁺ concentrations ([Ca²⁺) were detected by confocal Fluo-3 imaging. The subsequent changes of myelin related protein expression were determined by western blot and immunofluorescence. RESULTS Three hours after hypoxia, significantly upregulated expression of A1AR was observed, accompanied with significantly decreased expression of oligodendrocyte transcription factor (Olig2). In addition, either hypoxia stimulation or 100 μM adenosine induced apparent elevation of resting [Ca²⁺] in cultured OLs. However, pretreatment with DPCPX (A1AR selective antagonist) or caffeine abolished the [Ca²⁺] increase, and the subsequent adenosine of high dose induced Ca²⁺ activity in developing OLs. Furthermore, caffeine or DPCPX improved the expression MBP and CNPase proteins after hypoxia stimulation, which resulted in the morphological maturation of OLs. CONCLUSIONS Caffeine treatment exerted protective effects on neonatal hypoxia injuries. It prevented Ca²⁺ overload injury, kept Ca²⁺ homeostasis in hypoxic developing OLs, and facilitated optimal expression of myelin related proteins by inhibiting A1AR in vitro. This study also provided experimental evidence for clinical application of caffeine in early treatment of neonatal hypoxia, and highlighted the potential significance of A1AR in anti-hypoxic drug discovery.

Backlground

Neonatal hypoxia always causes severe and irreversible damage to the central nervous system (CNS) [1]. Studies have reported that poor differentiation of oligodendrocytes (OLs) and myelination disturbances were the main pathogenesis of hypoxic injury [2]. Thus, the development of OLs is a major concern in hypoxia. However, during distinct developmental stages of OLs lineage, early stage OLs (preOLs or developing OLs) are vulnerable to many factors, including hypoxia [3].

Caffeine, as a common therapeutic drug available in the neonatal intensive care unit, has been utilized for neonatal hypoxia treatment over 30 years [4,5]. Clinical studies have demonstrated that caffeine exerts therapeutic effects in neonatal hypoxic encephalopathy via counterbalancing the hypoxia-induced white matter injury and improving ventilation [6]. Additionally, caffeine has been shown to directly decrease hypoxia-induced ventriculomegaly and myelination disturbances in an animal model of chronic hypoxia [2]. However, until now, reports assessing the protective molecular mechanism of caffeine treatment are scarce.

Caffeine inhibits the function of extracellular adenosine, which is drastically elevated more than 100-fold under hypoxia condition [7,8]. It is also a non-selective agonist of adenosine receptors (ARs), which consist of A1, A2a, A2b, and A3 subtypes. All ARs are G-protein coupled receptors and associated with disease development in CNS, especially adenosine 1 adenosine receptor (A1AR) [9]. Cheng et al. reported the chronic ischemia-induced downregulation of A1AR contributed to white matter damage [10]. Activation of A1AR lead to white matter loss and ventriculomegaly [11], indicating that A1AR function was related to white matter damage. However, the role of A1AR in myelin forming cells, OLs, under hypoxic condition was still unclear.

Additionally, A1AR activation also mobilizes intracellular calcium signals (Ca2+) via inositol triphosphate receptor (IP3R) [12]. In physiological conditions, sleep deprivation induces the release of adenosine in cholinergic basal forebrain to stimulate A1AR, resulting in IP3R activation and alterations of transcription factors [13]. In pathological conditions, like hypoxia, abnormal elevation of adenosine might induce over-activation of A1AR, and lead to dysregulation of IP3R and Ca2+ signals. But whether this was the underlying mechanism of A1AR induced hypoxic injuries, requires experimental explorations.

This study was performed on cultured OLs in vitro. Oxygen and glucose deprivation (OGD) was utilized to simulate hypoxic conditions. We aimed to provided evidence for the involvement of A1AR in hypoxia-induced poor differentiation of OLs, and the underlying mechanism of caffeine treatment in hypoxic injuries.

Material and Methods

Animals and drugs

All procedures were conducted in accordance with the University Committee on Animal Care and Use. Sprague-Dawley (SD) neonatal rats (1 to 3 day postnatal, P1–P3, n=10) were purchased from the Animal Facility Center of the Army Medical University, P.R. China, and housed in separate cages in a controlled environment.

For adenosine receptor ligands, adenosine and caffeine (Sigma-Aldrich, St. Louis, MO, USA) were utilized; the selective antagonist DPCPX (8-cyclopentyl-1,3-dipropylxanthine) (Abcam, Cambridge, UK) was tested as well.

Culture of purified OPCs (OL progenitor cells)

The rat OL progenitor cells (OPCs) were propagated as previously described [14]. In brief, cell suspensions were collected from the cerebral cortex of postnatal 1 to 3 days old SD rats and then cultured in OPCs proliferation medium, containing: DMEM/F12 (Hyclone); 15% B104 conditioned medium; 1% N2 supplements (Invitrogen, USA). After 5 to 7 days, OPCs were isolated via OPCs-isolation medium (DMEM/F12; 0.01% EDTA; 5 mg/mL insulin, 15 to 30 minutes at room temperature and centrifuged at 1000 rpm for 5 minutes. Then, isolated OPCs were cultured with OPCs-proliferation medium in dishes or cover slips coated with poly-D-lysine. For OPCs differentiation experiments, OPCs were incubated with OPCs differentiation medium containing: DMEM/F12; 1% N2 supplement; 5 mg/mL N-acetyl-L-cysteine (Amresco, Solon, OH, USA); 1% fetal bovine serum (FBS); 5 mg/mL insulin. At the distinct time point of differentiation stage, cells were collected and fixed for experiments.

Hypoxic model of cultured oligodendrocytes

For oxygen and glucose deprivation (OGD) in OLs medium, cultures were maintained in DMEM with 10% FBS in incubator at 37°C in a humidified atmosphere with 5% CO2. Then, external O2 and glucose were replaced with N2 and sucrose (10 mM), respectively. Finally, 20 μM iodoacetate was used to block glycolysis. After 30 minutes in OGD conditions, OLs were returned to fresh media. The hypoxic damage was detected by immunocytochemistry and western-blot of CNPase (a maker of immature OLs), MBP (a maker of mature OLs), and GFAP (a marker expressed in immature oligodendrocytes and astrocytes) as well as Ca2+ alteration at different post stimulus times (1 day to 6 days).

For adenosine treatment, 100 μM adenosine [15] was used to mimic hypoxic injury. OLs were exposed to adenosine solution 20 minutes before returning to fresh medium with or without 0.1 mM caffeine, 100 nM DPCPX (selective A1AR antagonist). After adenosine stimulation, cells were cultured for 3 to 6 days for various experiments. For observation of Ca2+ dynamic alteration, we tested Ca2+ response of OLs to adenosine solution in concentration ranged from 100 μM and 300 μM. Adenosine 300 μM was utilized to induce temporal and drastic Ca2+ response in cultured OLs.

Reverse transcription polymerase chain reaction (RT-PCR)

Total RNA was extracted and purified from cultured cells, according to the experiment protocol of TRIzol Reagent (Life Technologies, USA). Briefly, the reverse transcription of total RNA into single strand cDNA was performed using random primer (Takara) and Super Script III (Life Technologies, USA). The sequences of specific primers were listed as follows:

A1AR, 5′-TGCTCATTGCCTTGGTCTCT-3′ and

5′-GATCTTGACTCGGAGGTATCGA-3′;

A2aAR, 5′-CTGCCTCTTCTTCGCCTGTT-3′ and

5′-CGTGGAGTTCCCGTCTTTCT-3′;

A2bAR, 5′-GTCACTGGAACACGAGCAAGA-3′ and

5′-ATGAGCAGTGGAGGAAGGACA-3′;

A3AR, 5′-ACCTTCTATTTCATCGTCTCCCTA-3′ and

5′-CTTCTTTGAGTGGTAACCGTTCTA-3′;

Olig2, 5′-TCAAGTCATCTTCCTCCAGCAC-3′ and

5′-GGCTCAGTCATCTGCTTCTTATCTT-3′;

β-actin, 5′-CGTTGACATCCGTAAAGACC-3′ and

5′-CATCGTACTCCTGCTTGCT-3′.

Image Pro Plus image (Media Cybernetics, MD, USA) was used for statistical analysis of PCR products (normalized to β-actin).

Immunofluorescence staining

For immunofluorescence staining, cells were first treated in a mixture of 3% H2O2, methanol and phosphate-buffered saline (PBS) at the ratio of 1: 1: 1 to suppress endogenous peroxidase activity followed by 2 PBS washes. Then, sections were blocked with 5% bovine serum albumin (BSA) and incubated overnight at room temperature using primary antibodies: anti-A1R (1: 200, Boster, China); CNPase (1: 200, Santa Cruz, CA, USA); Olig2 (1: 200, Abcam, USA); MBP (1: 200, Santa Cruz, CA, USA). Afterwards, sections were washed with PBS and incubated with FITC-conjugated goat anti-rabbit secondary antibodies (1: 200, Boster, China) and TRITC-conjugated goat anti-mouse secondary antibodies at room temperature for 3 hours. For negative control sections, we used PBS to replace the primary antibody with the remaining steps being the same. The immunofluorescence images were conducted using confocal laser-scanning microscope (Olympus IX81, Japan). The differential interference contrast (DIC) image was also recorded to reveal the morphological features of OLs.

Calcium imaging

Firstly, OPCs were prepared and differentiated in glass-bottom dishes for 3 days. Then, cells were cultured in imaging wash buffer, containing: 2 mM glucose; 8 mM HEPES; 135 mM NaCl; 2 mM MgCl2; 3 mM KCl; 2.2 mM CaCl2. Fluo-3 AM (5 μM, Invitrogen) was added into buffer for 20 minutes at 37°C. After that, cells were transferred to room temperature for 20 minutes in fresh imaging wash buffer to allow complete dye de-esterification. At last, confocal laser scanning microscopy was utilized to determine the intracellular Ca2+ level in resting cells or the changes of intracellular Ca2+ activity. Fluo-3 images were recorded and measured using Fluoview Image Processing software (v2.1). Independent experiments with the same recording conditions were performed, with 10 fields recorded for each experiment.

Western blot analysis

After experimental intervene, total proteins were isolated from cultured OLs using the RIPA lysis buffer with 1% PMSF solution (Beyotime, China). Protein amounts were quantified by the Coomassie brilliant blue G250 method. Equal amounts of protein in different groups were denatured and loaded separately on SDS-PAGE gels. After electrophoresis, the target bands were transferred onto polyvinyldifluoride (PVDF) membranes. After BSA blocking for 1 hour, primary antibodies were used overnight at 4°C, including: anti-CNPase (1: 1000, Santa Cruz, CA, USA), anti-MBP (1: 1000, Santa Cruz, CA, USA), anti-GFAP (1: 1000, Santa Cruz, CA, USA), anti-IP3R2 (1: 500) and anti-β-actin (1: 1000, Boster, China). Then, HRP-conjugated secondary antibodies were used (1: 1000, Boster, China) and ECL plus detection kit (ECL plus, GE Healthcare) were conducted to reveal the immunoreactive bands. Image-Pro Plus (Media Cybernetics, MD, USA) was used for statistical analysis of western blot products (normalized to β-actin).

Quantitative image analysis

For each experiment, 10 randomly chosen fields at 20× magnification were recorded for cell counting. Image-Pro Plus software (n=10) was used to quantify the immunofluorescence intensity of immunostaining, Fluo-3 staining and cell counting. The relative optical density of PCR or western blot results were quantified by Image-Pro Plus image analysis system (normalized to β-actin).

Statistical analysis

Statistical analysis was performed through SPSS18.0 for Windows (SPSS Inc, Chicago, IL, USA) in one- or two-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test. Comparisons of 2 experimental groups were carried out using Student’s t-test. Data was presented as mean ± standard deviation (SD) and P<0.05 was considered statistically significant.

Results

To explore whether the expression of ARs were altered in hypoxic OLs and whether caffeine functions as the antagonist of ARs, mRNA expressions of ARs were detected by RT-PCR in normal and hypoxia conditions, with or without 0.1 mM caffeine treatment (Figure 1). After hypoxia, in cultured developing OLs, a rapid elevation of A1AR mRNA was found from 1 to 12 hours, peaking at 3 hours in the OGD group (P<0.05; Figure 1A, 1B). However, treatment with 0.1 mM caffeine early abolished the elevation of A1AR observed in the OGD group. In contrast, A2aAR, A2bAR, and A3AR did not display significant alterations in the early stage of hypoxia as well as in caffeine treatment (Figures 1D–1F). Besides, in normal condition, OLs expressed sufficient oligodendrocyte transcription factor 2 (Olig2) (Figure 1A). PCR results revealed that the expression of Olig2 mRNA was notably decreased in the OGD group from 1 hour to 12 hours (P<0.05; Figure 1C). After caffeine treatment, OLs displayed similar expression of Olig2 mRNA with control (CTL) group. It indicates hypoxia interfered the differentiation of OLs, which was the hallmark of injury and caffeine treatment reversed this hypoxic injury.

Figure 1

Caffeine affected the expression of A1AR and Olig2 mRNA in hypoxic OLs. (A) RT-PCR showing that A1AR, A2aAR, A2bAR, A3AR and Olig2 mRNA expression in OLs in vitro in CTL, OGD and 0.1 mM caffeine treatment. (B) OGD (30 minutes) caused significant increase of A1AR mRNA from 1 hour to 12 hours, peaking at 3 hours in OLs in vitro and 0.1 mM caffeine treatment abolished this elevation. * P<0.05 versus CTL. (C) OGD caused significant decrease of Olig2 mRNA in OLs in vitro and 0.1 mM caffeine treatment recovered Olig2 mRNA expression. ** P<0.01 versus CTL. Values represent mean ± standard deviation. (n=3 for each group). (D–F) Statistically analysis of PCR results displayed OGD or caffeine did not influence the expression of A2aAR, A2bAR and A3AR in OLs (G) Immunostaining showing expression and distribution of A1AR (red) and Olig2 (green) in cultured OLs of 1 day in CTL, OGD and 0.1 mM caffeine treatment. OGD enhanced fluorescence intensity of A1AR and fluorescence signals were notably increased on cell protuberance compared with CTL group. After caffeine treatment, expression and distribution of A1AR were similar to the CTL group. However, OGD significantly decreased the percentage of Olig2 positive cells while caffeine treatment recovered this effect. Scale bar=25 μm. (H, I) Statistically analysis of A1AR and Olig2 immunofluorescence results were presented in bar graph. * P<0.05 versus CTL or caffeine group. A1AR – adenosine 1 adenosine receptor; OLs – oligodendrocytes; OGD – oxygen and glucose deprivation; RT-PCR – real-time polymerase chain reaction; CTL – control.

In addition, immunofluorescence was used to reveal the expression of A1AR and Olig2 in cultured OLs of 1 day (Figure 1G). In the CTL group, A1AR was found distributed in the cytoplasm and the fluorescence intensity was weak on cell protuberance. In the OGD group, the fluorescence intensity of A1AR in the cytoplasm and cell protuberance were apparently increased (P<0.05, Figure 1G, 1H). After caffeine treatment, the fluorescence intensity of A1AR was decreased and similar to the CTL group. Additionally, the expression of Olig2 was significantly decreased in the OGD group while this effect was also recovered by caffeine treatment (P<0.05, Figure 1G, 1I).

To investigate whether caffeine reversed hypoxia-induced disturbance of OLs differentiation, we observed the morphological differentiation of OLs after OPCs were cultured for 3 days (Figure 2A, 2B). During the maturation process, immature OLs passed through a series of morphological changes. Compared with immature OLs, mature OLs formed more primary and secondary cell processes to present a net-like structure. In the CTL group, more than 60% of OLs displayed mature morphological features and arborization was observed. However, in the OGD group, only 18.4% displayed mature features (P<0.01). OLs were swelling and cell processes were ruptured. Then, we used 100 μM adenosine to simulate hypoxic injury and the mature percentage of OLs was decreased to 24.1%, which was similar to the OGD group (P<0.01). In contrast, when treated with 0.1 mM caffeine or 100 nM DPCPX (A1AR selective antagonist), in the OGD group, differentiation disturbance was morphologically recovered and the percentage of mature OLs both reached approximately 57%.

Figure 2

Caffeine protects differentiation of hypoxic OLs and inhibits hypoxia-induced intracellular resting [Ca2+] elevation. (A) DIC images of OLs cultured for 3 d showing morphological differentiation of OLs in CTL, OGD, 100 μm adenosine, 0.1 mM caffeine and 100 nM DPCPX (selective A1AR antagonist). (B) OGD or 100 μm adenosine caused significant decrease of the percentage of mature OLs (18.4%, 24.1%, respectively versus 61.2%). 0.1 mM caffeine or 100 nM DPCPX treatment recovered the percentage of mature OLs in hypoxia (57.1%, 57.2%, respectively) ** P<0.01 versus CTL. (C, D) Fluo-3 pseudo-color images showing intracellular resting [Ca2+] in CTL, OGD, 100 μm adenosine, 0.1 mM caffeine and 100 nM DPCPX. Fluorescence intensity of Fluo-3 was detected as reflection of intracellular resting [Ca2+]. OGD or 100 μm adenosine induced significant elevation of resting [Ca2+]. 0.1 mM caffeine or 100 nM DPCPX treatment inhibited resting [Ca2+] uptake in hypoxia. values represent mean ± standard deviation. (n=10 for each group) ** P<0.01 versus CTL. (E) Western analysis showing that IP3R2 protein expression in CTL, OGD and 0.1 mM caffeine treatment in OLs. (F) OGD significantly increased the expression of IP3R2 protein and 0.1 mM inhibited the elevation of IP3R2 in hypoxia. ** P<0.01 versus CTL or caffeine group. Scale bar, 25 μm. OLs – oligodendrocytes; DIC – differential interference contrast; CTL – control; OGD – oxygen and glucose deprivation; DPCPX – A1AR selective antagonist; A1AR – adenosine 1 adenosine receptor; RT-PCR – real-time polymerase chain reaction; IP3R2 – inositol triphosphate receptor.

To explore the mechanism of hypoxic induced poor differentiation of OLs and caffeine effects, confocal Ca2+ imaging was utilized to detect intracellular resting [Ca2+] via quantitating the fluorescence intensity of Fluo-3 labelling OLs (Figure 2C, 2D). Fluorescence signals were normalized to average values measured before drug application. In the CTL group, the fluorescence intensity of Fluo-3 pseudo color was stable and not intense. In the OGD group, the fluorescence intensity of Fluo-3 was significantly increased compared to the CTL group (P<0.05) and similar elevation was recorded after 100 μM adenosine treatment. In contrast, 0.1 mM caffeine or 100 nM DPCPX treatment prevented elevation of the fluorescence signal in the OGD group. This indicated that 0.1 mM caffeine or 100 nM DPCPX treatment prevented abnormal elevation of intracellular resting [Ca2+] in developing OLs in hypoxia.

To further our understanding of hypoxia-induced resting [Ca2+] elevation, we detected the protein expression of potential [Ca2+] release channels, IP3Rs. Among 3 subtypes of IP3Rs, only IP3R2 has been reported in OLs [16]. Our western blot results showed that IP3R2 was expressed in the CTL group. However, hypoxia induced significant increase in the expression of IP3R2 protein (P<0.01; Figure 2E, 2F). Treatment of 0.1 mM caffeine notably decreased its expression.

To understand the mechanism of caffeine treatment in regulating Ca2+ response, confocal Ca2+ imaging was used to detect dynamic intracellular [Ca2+] activity upon ATP or adenosine stimulation (Figure 3). In OPC culture of 3 days, as a positive control, 300 μM ATP triggered overt [Ca2+] transient wave in developing OLs, which peaked sharply and descended quickly (Figure 3A). In contrast, 300 μM adenosine induced apparent [Ca2+] response. The Ca2+ transient wave underwent sustained elevation and remained at plateau phase after the descent of ATP-induced [Ca2+] uptake (Figure 3B). When pharmacologically blocked, the activation of A1AR in developing OLs by 0.5 mM caffeine or 300 nM DPCPX pretreatment, Ca2+ transient wave did not undergo intensive elevation and [Ca2+] wave remained at relatively lower and stable levels after adenosine stimulation (Figure 3C, 3D). The area under the curve (AUC) and peak value of Ca2+ transient waves were calculated (Figure 3E, 3F). Significant differences were found with or without caffeine treatment. Caffeine treatment not only inhibited adenosine-induced Ca2+ release at the peak (P<0.01), but decreased intracellular [Ca2+] release after adenosine stimulation (P<0.01). This phenotype was similar with DPCPX treatment.

Figure 3

Caffeine decrease adenosine-induced intracellular Ca2+ response in developing OLs via inhibiting A1AR. (A) Ca2+ transient wave of 300 μm ATP in cultured OLs. ATP induced overt [Ca2+] response in developing OLs, which reached peak value sharply and then descended quickly as a positive control. (B) Ca2+ transient wave of 300 μm adenosine in cultured OLs. Adenosine induced sustained elevation of [Ca2+] and remained at plateau phase. (C, D) After 0.5 mM caffeine or 300 nM DPCPX treatment, Ca2+ transient wave did not undergo intensive elevation and [Ca2+] wave remained at relatively lower and stable level after adenosine stimulation. (E, F) AUC and peak value of Ca2+ wave. Significant differences were found between adenosine and caffeine, DPCPX treatment. Basal line of Ca2+ transient waves were normalized to the averaged measured in cultures in standard solution. Values represent mean ± standard deviation. (n=10 for each group) ** P<0.01 versus adenosine group. OLs – oligodendrocytes; A1AR – adenosine 1 adenosine receptor; DPCPX – A1AR selective antagonist; AUC – area under the curve.

To verify the effect of caffeine in regulating the differentiation and myelination of hypoxic OLs, firstly we detected quantified CNPase and GFAP protein expression levels by western blot (Figure 4A). Our data showed that hypoxia induced a significant reduction of CNPase proteins (P<0.05, Figure 4B) while inducing an elevation of GFAP proteins (P<0.05; Figure 4C). The alteration was also observed after treatment of OLs with 100 μM adenosine (P<0.05). In contrast, treatment with 0.1 mM caffeine or 100 nM DPCPX reversed the changed expression of CNPase and GFAP proteins in the OGD group.

Figure 4

(A–F) Caffeine promotes differentiation of hypoxic developing OLs. (A) Western blot showing the expression of CNPase and GFAP proteins in cultured OLs of 3 days in CTL, OGD, 100 μm adenosine, 0.1 mM caffeine and 100 nM DPCPX. (B) OGD or 100 μm adenosine caused significant decrease of CNPase protein. 0.1 mM caffeine or 100 nM DPCPX recovered its expression in hypoxia. (C) OGD or 100 μm adenosine caused significant increase of GFAP protein. 0.1 mM caffeine or 100 nM DPCPX inhibited the elevation in hypoxia. ** P<0.01 versus CTL group. (D) Immunostaining of CNPase (red) in cultured OLs of 3 days in CTL, OGD and 0.1 mM caffeine treatment. OGD overtly decreased the fluorescence intensity of CNPase protein and OLs displayed immature features in OGD group. 0.1 mM caffeine treatment recover the fluorescence intensity of CNPase protein and differentiation of OLs. Immunostaining of MBP (green) in cultured OLs of 6 days in CTL, OGD and 0.1 mM caffeine treatment. OGD decreased the fluorescence intensity of MBP protein and cells were swollen with few processes left. 0.1 mM caffeine treatment recover the fluorescence intensity of MBP protein and arborization was abundant and dense similar with CTL group. Values represent mean ± standard deviation. ** P<0.01 versus CTL or caffeine group. Scale bar, 25 μm. OLs – oligodendrocytes; CNPase – a maker of immature OLs; GFAP – a marker expressed in immature oligodendrocytes and astrocytes; CTL – control; OGD – oxygen and glucose deprivation; DPCPX – A1AR selective antagonist; MPB – a maker of mature OLs.

Three days after hypoxia, CNPase proteins were expressed in OLs and cells presented nearly mature features. However, the fluorescence intensity of CNPase protein was overtly decreased and OLs displayed immature features in the OGD group. Caffeine treatment was observed to counterbalance the hypoxic effects on OLs in the expression of CNPase and also in morphological differentiation (Figure 4D, 4E). Besides, in 6 days, OLs in the CTL group and the caffeine group presented high fluorescence intensity of MBP protein and arborization was abundant and dense. The OGD group displayed immature features of OLs with swollen cytoplasm and fewer cell processes. In addition, the fluorescence intensity of MBP protein in the OGD group was weak than in the CTL group, while caffeine treatment recovered the expression of MBP protein (Figure 4D, 4F).

Discussion

A1AR is widely expressed in CNS. It is also found in OLs, astrocytes [17], and microglia cells. Turner et al. reported a correlation between A1AR activation and hypoxia-induced white matter loss [11]. Our results also revealed the importance of A1AR in hypoxia. In the early stage of hypoxia (<12 hours), A1AR underwent a rapid upregulation and peaked at 3 hours in developing OLs. This response was associated with stark changes of adenosine concentration, which increased by more than 100-fold within minutes after acute hypoxia. High concentrations of adenosine were also proven to induce hypoxia-like injury in cultured OLs in our experiments via interfering with morphological differentiation. In addition, we did not observe apparent relationships between the other 3 adenosine receptors and cellular hypoxic response. These findings provide evidence that A1AR is involved in hypoxic injury in developing OLs. However, A1AR has been reported to protect the brain in embryo [18,19], which was different from our results. This indicates that A1AR activation might lead to different effects before versus after birth [18].

For years, A1AR has been regarded as an anti-inflammatory in the CNS [20], indicating its protective role in hypoxia. However, this effect has not been considered to be critically important in hypoxic and ischemic models [9]. Our results also strongly indicated that A1AR activation was associated with the developmental disorder of OLs and white matter loss. As caffeine has been used to treat hypoxia-induced white matter loss for over 30 years, we pretreated OLs with 0.1 mM caffeine after hypoxia. PCR results showed a significant reduction in A1AR expression in the early stage of hypoxia. In addition, we also observed a significantly decreased Olig2 expression in early hypoxia and later decreased protein expression of CNPase protein. These findings indicated that in hypoxia, developing OLs are kept in an immature stage, and it is difficult to initiate myelination. However, caffeine treatment reversed this potential injury. In conclusion, caffeine exerts antagonistic effects on A1AR and prevents its overexpression in hypoxia. Caffeine also protects the expression of differentiation related factors of OLs in hypoxia.

Our results also provided evidence that hypoxia and adenosine treatment induced significant elevation in resting intracellular [Ca2+]. When we treated cells with caffeine or the A1AR antagonist DPCPX, resting [Ca2+] decreased to normal levels in hypoxia. This indicated that caffeine was able to regulate intracellular [Ca2+] and keep Ca2+ homeostasis in developing OLs during hypoxia by antagonizing A1AR. Furthermore, adenosine stimulation of OLs resulted in long-lasting elevation of [Ca2+], indicating that adenosine induces sustained elevation of intracellular [Ca2+]. However, the adenosine-induced [Ca2+] uptake could be inhibited by treatment with caffeine or DPCPX. In conclusion, caffeine treatment inhibited adenosine induced Ca2+ activity and maintained Ca2+ homeostasis. Our findings unveiled the underlying mechanism of OLs injury in hypoxia by observing [Ca2+]: as adenosine increased rapidly and violently in hypoxia within minutes, A1AR was stimulated by high adenosine concentrations, which resulted in excessive Ca2+ release. Moreover, this excessive Ca2+ release can directly lead to Ca2+ overload in hypoxia [21,22], and is considered with initiate cell injury. However, previous studies have also reported that A1AR reduces [Ca2+] by inhibiting voltage-gated Ca2+ channels in excitable cells, including motor neurons [23], and retinal ganglion cells [24]. This effect is considered an important way to inhibit the release of excitatory neurotransmitters, such as glutamate [25], in hypoxia. In contrast, our results demonstrated that A1AR contributed to Ca2+ overload in developing OLs. During the whole developmental process of OL lineage, OLs exhibit obvious plasticity to Ca2+ signal at different stages of development and are sensitive to the change of [Ca2+] [26]. [Ca2+] overload in hypoxic developing OLs directly leads to apoptosis, mitochondrial damage, production of reactive oxygen species, and activation of Ca2+-dependent proteases [27]. In addition, A1AR activation, in a concentration-dependent manner, mobilizes intracellular [Ca2+] via the IP3Rs located on the endoplasmic reticulum membrane [13,28]. This mechanism can explain our results and the findings of enhanced [Ca2+] in isolated OLs via A1AR activation [29].

In cultured developing OLs, 100 μM adenosine was able to decrease the protein expression of CNPase and show a compensatory increase in the expression of GFAP, resulting in morphologically poor differentiation of OLs. When cells were exposed to hypoxia conditions, similar effects were observed. Interestingly, 0.1 mM caffeine or 100 nM DPCPX reversed these effects. It becomes increasingly clear that caffeine inhibits A1AR activation to upregulate the expression of myelin-related proteins and promote the morphological differentiation of OLs. Our findings were in agreement with previous works by Turner and collaborators [11,30], who generated a gene knockout model of A1AR and observed increased expression of MBP proteins and myelination in hypoxic brain. Together, A1AR plays a crucial role in the differentiation of developing OLs, caffeine treatment exerts significant protective effects in developing OLs by antagonizing A1AR to promote cell differentiation.

Conclusions

Our results demonstrated that caffeine treatment was able to prevent Ca2+ overload and keep Ca2+ homeostasis in developing OLs. In addition, in hypoxia conditions, this mechanism antagonizes A1AR and IP3R2 activation and results in the expression of myelin-related proteins to initiate maturation and myelination. In terms of potential clinical use, this mechanism might be able to reverse hypoxia-induced neonatal periventricular leukomalacia. In addition, A1AR displayed a potential significant role in anti-hypoxic drug discovery.