Selenium supplementation provides potent neuroprotection following cerebral ischemia in mice

Abstract

Despite progress in reperfusion therapy, functional recovery remains suboptimal in many stroke patients, with oxidative stress, inflammation, dysbiosis, and secondary neurodegeneration constituting the major hurdles to recovery. The essential trace element selenium is emerging as a promising therapeutic agent for stroke. However, although several rodent studies have shown that selenium can protect against cell loss following cerebral ischemia, no study has yet examined whether selenium can enhance long-term functional recovery. Moreover, published studies have typically reported a single mechanism of action underlying selenium-mediated stroke recovery. However, we propose that selenium is more likely to have multifaceted actions. Here, we show that selenomethionine confers a potent neuroprotective effect in a canonical filament-induced transient middle cerebral artery occlusion (tMCAO) mouse model. Post-tMCAO selenium treatment significantly reduces the cerebral infarct volume, oxidative stress, and ferroptosis and enhances post-tMCAO motor performance in the acute phase after stroke. Moreover, analysis of the gut microbiota reveals that acute selenium treatment reverses stroke-induced gut dysbiosis. Longer-term selenium supplementation activates intrinsic neuroprotective mechanisms, prevents secondary neurodegeneration, alleviates systemic inflammation, and diminishes gut microbe-derived circulating trimethylamine N-oxide. These findings demonstrate that selenium treatment even after cerebral ischemia has long-term and multifaceted neuroprotective effects, highlighting its clinical potential.

Introduction

Stroke is the second leading cause of death and disability worldwide, ¹ with acute ischemic stroke accounting for more than 80% of cases. However, current stroke therapies, including aspirin or heparin to reduce blood clots or hemolytic products to dissolve clots and restore blood flow, have proven inadequate. This is largely because, despite successful reperfusion, oxidative stress, gut dysbiosis, inflammation, and secondary neurodegeneration still negatively impact functional recovery.^2–4 Attempts to develop neuroprotective agents for stroke have primarily failed at the clinical trial stage, with few therapeutic agents approved for acute ischemic stroke treatment. ⁵ However, the essential dietary trace element selenium is emerging as a promising therapeutic agent for stroke. ⁶

Deficiencies in selenium have long been implicated in stroke susceptibility, with epidemiological studies revealing an inverse association between circulating selenium levels and the occurrence of stroke.^7–9 Despite this, relatively few rodent studies have directly examined whether selenium supplementation can successfully protect against stroke-induced cell death and/or the associated functional deficits. In one such study, selenium treatment was shown to reduce cell death and improve functional recovery of spatial and sensory neglect in a mouse model of hemorrhagic stroke. ¹⁰ Several studies using the rodent middle cerebral artery occlusion (MCAO) model have also shown that selenium treatment can reduce cell loss^10–14 and improve motor co-ordination ¹³ following transient ischemic stroke. However, the mechanisms underlying this effect and whether selenium treatment can ameliorate ischemia-induced cognitive dysfunction and improve long-term functional recovery has not been investigated.

Although most studies typically investigate a single mechanism of action, we propose that selenium and selenoproteins have multifaceted neuroprotective actions following ischemia/reperfusion injury. Studies have shown that selenium attenuates ischemic damage by reducing DNA oxidation, ¹² restoring mitochondrial function, ¹² or inhibiting autophagy. ¹⁵ It was also recently reported that seleno-cystamine and methyl-selenocysteine attenuate MCAO-induced neurological deficits and reduce infarct volume via selenium-mediated protection against ferroptotic cell death. ¹¹ Alterations in gut microbiota have also been associated with ischemic stroke pathogenesis, with recent clinical studies finding an association between circulating trimethylamine N-oxide (TMAO), the most well-studied gut microbiota-derived metabolite, and adverse stroke outcomes.^16–18 Interestingly, selenium has recently been shown to have a positive effect on gut dysbiosis in mice.^19,20 However, whether it can protect against stroke-induced gut dysbiosis remains unknown. We have also recently shown that selenium increases adult hippocampal neurogenesis in a mouse model of stroke-like hippocampal injury. ²¹ However, whether selenium supplementation increases neurogenesis and protects against the associated spatial learning and memory deficit in the canonical mouse model of transient MCAO (tMCAO) remain unknown.

In the present study we examined whether delayed selenium treatment (2 h after successful reperfusion) affects cerebral infarct volume, oxidative stress, ferroptosis, gut dysbiosis and motor function in the acute phase after stroke. We also evaluated whether sustained selenium supplementation confers long-term benefits by examining the effects on secondary neurodegeneration, systemic inflammation, gut microbe-derived circulating TMAO, and spatial learning and memory 6 weeks after tMCAO.

Materials and methods

A more detailed description of all sections of the Methods used in this study can be found in the Supplementary Methods

Animals

C57BL/6J mice were purchased from the Vital River Laboratory Animal Technology Company. Animals were maintained under standard conditions on a 12 h light/dark cycle (lights on at 7:00 A.M.) with food and water provided ad libitum. All animal experiments were approved by the Animal Care Committee of the Southern University of Science and Technology and carried out in accordance with the regulations set by the Guidelines for the Care and Use of Laboratory Animals of Guangdong Province. The ARRIVE guidelines were followed when designing, performing and reporting animal experimentation. ²²

Middle cerebral artery occlusion surgery

Male C57BL/6 mice (3 months old, body weight 22.5 ± 0.5 g) were fasted overnight but had free access to water before surgery. The procedures used to generate tMCAO were essentially the same as previously described.^23–25

Neurological deficit scoring

Neurological deficits were scored as previously described.^23,26 A scale of 0, 1, 2, 3, 4 and 5 were used for quantitation with 0 as normal and 5 as comatose or moribund.

Drug delivery

Selenomethionine (2 mg/kg body weight, i.p injection, dissolved in 0.9% w/v NaCl) was injected 2 h, 14 h, 36 h and 62 h after reperfusion. In the long-term selenium supplementation experiment, selenomethionine was added to the drinking water (10 μg/ml) from days 4–43. The equivalent amount of sodium chloride was added to the water of the control group. In another cohort of mice, N-acetylcysteine (40 mg/kg body weight) was combined with selenomethionine (2 mg/kg body weight) for injection following the same paradigm.

Infarct volume and edema assessment

Brains were removed and sliced in a mouse brain matrix into four 2 mm thick coronal slices and the slices were immediately stained with 5 ml of 2% 2,3,5-triphenyltetrazolium chloride (TTC) exactly as previously described.^23,25 Percentage of infarct = (volume of the contralateral hemisphere – (volume of intact ipsilateral hemisphere – the volume of ipsilateral infarct hemisphere)/volume of intact ipsilateral hemisphere) × 100; Percentage of edema = (volume of the ipsilateral hemisphere – the volume of the contralateral hemisphere)/volume of contralateral hemisphere × 100.

Motor function tests

Motor function was measured 24 and 72 h after the tMCAO surgery using an accelerating rotarod apparatus (47600, Ugo Basile). Mice were pre-trained to habituate with the accelerating rotarod task for 3 days (3 trials per day) prior to the surgery (day 1: 4 rpm for 5 min; day 2: 4–20 rpm for 5 min; day 3: 4–40 rpm for 5 min). One and 3 days after tMCAO, the mice were again tested for three trials of 5 min each, at a minimum of 4 rpm to a maximum of 40 rpm. The latency to the first fall was averaged over the three trials. Forelimb grip strength being assessed exactly as previously described.^25,27

Behavioral tests

Methods for spontaneous alternation Y maze and active place avoidance test are as previously described.^21,27 Mice were familiarized with the researchers and habituated to the training room prior to the testing. To minimize any difference due to experimenter handling, mice were always handled by the same experimenter.

Immunofluorescence staining

Mice were overdosed with pentobarbitone and transcardially perfused with NaCl (0.9% w/v). Whole brains were post-fixed in 4% PFA at 4°C for 48 h. Brains were subsequently cryoprotected in 30% sucrose solution at 4°C for 2–3 days. Coronal sections were cut. One series of every sixth section encompassing the entire hippocampus was subjected to fluorescence staining. ²¹

ELISA

Plasma and post-perfused brain tissue were collected at designated times after surgery. Tissues were snap-frozen, weighed, and homogenized. Optical densities were measured using a microplate reader (Model 680, Bio-Rad). ELISA was performed in accordance with the manufacturer’s protocol.

Inductively coupled plasma-mass spectrometry

Plasma and brain tissue were collected, lyophilized, digested in 5 ml of 65% (v/v) nitric acid (CNW Technologies), and diluted to a final volume of 10 ml using MilliQ-H₂O. Elemental analysis was performed using an Agilent Technologies 7850 Series inductively coupled plasma-mass spectrometry (ICP-MS). Plasma or tissue metal levels were expressed as a concentration, and metal levels were normalized to dry tissue weight or plasma volume.

Liquid chromatography-mass spectrometry

Plasma trimethylamine, trimethylamine N-oxide, choline, betaine, creatinine, and L-carnitine were quantified using liquid chromatography-mass spectrometry (LC-MS) methods as previously described. ²⁵

16S rRNA gene sequencing

Total genomic DNA was extracted from fecal samples and DNA was amplified using primers (27 F_(16S-F): 5′-AGRGTTTGATYNTGGCTCAG-3′ and 1492 R_(16S-R): 5′-TASGGHTACCTTGTTASGACTT-3′ targeting the full-length (variable 1–9) of 16S rRNA. All PCR reactions were carried out using Veriti96well9902 (Applied Biosystems). Sequencing libraries were generated using the TruSeq® DNA PCR-Free Sample Preparation Kit (Illumina, USA). Sequence analyses were performed and sequences with ≥97% similarity were assigned to the same operational taxonomic units (OTUs). Alpha diversity (Chao, Ace, Shannon, PD whole tree index) was analyzed by QIIME 2.

Statistical analysis

Data were analyzed using GraphPad Prism 9.3, and all statistical tests were two-tailed. All data are represented as mean ± SD. Kolmogorov–Smirnov tests were used to determine whether the data sets were normally distributed or not. The Mann–Whitney U test (non-normally distributed data) and the unpaired Student’s t-test (normally distributed data) were used to comparing continuous variables. Statistical significance was determined using a Student’s t test when the experiment contained two groups or ANOVA when comparing more than two groups. Post-hoc analysis was performed on ANOVAs using Bonferroni, Tukey’s, or Sidak post hoc tests as described in the figure legends. A confidence interval of 95% was used, and a P-value <0.05 was considered statistically significant.

Results

Delayed selenium treatment reduces cerebral infarct volume and motor impairment following tMCAO

Adult male C57BL/6 mice received a 30 min occlusion of the right MCA. A greater than 80% reduction of the cerebral blood flow in the MCA territory was confirmed using a laser Doppler flowmeter (Supplementary Figure S1). The tMCAO mice were then randomly assigned to two groups, with one group receiving selenium injections and the other group receiving saline. Selenium (in the form of selenomethionine) was dissolved in saline and delivered by intraperitoneal (i.p.) injection at a dose of 2 mg/kg (body weight) 2 h, 12 h, 36 h and 62 h after reperfusion (the tMCAO + Se group). An equivalent volume of saline was injected into tMCAO mice as a no-treatment control (the tMCAO group). Sham surgery was performed on a separate cohort of mice as a negative control (the sham group). Figure 1(a) outlines the experimental timeline.

Figure 1.

Delayed selenium treatment reduces brain infarcts and enhances functional recovery following tMCAO. (a) Experimental timeline. (b) Infarct analysis method. (c) Representative figures of TTC staining of coronal brain sections from mice in three groups at 1 and 3 days after tMCAO. The black dashed line encloses the infarct area. (d) Percentage of total corrected infarct volume, (e) edema. (f) neurological deficit scores of mice 1 and 3 days after 30 min tMCAO (n = 10–16 mice in each group). (g) Maximum time reached in the accelerating rotarod test and (h) maximum forelimb grip strength 1 or 3 days after ischemic stroke surgery (n = 10–12 mice in each group). Data represent mean ± SD. (d–f: unpaired two-tailed t test; g and h: one-way ANOVA with Bonferroni post hoc tests at each day).

Consistent with published data from both our group and others, we found that the cerebral infarct volume peaked at 24 h post-lesion and then decreased by 72 h.^25,28–30 We found that selenium treatment significantly reduced the volumes of cerebral infarction (Figure 1(b) to (d); 22.41 ± 2.22% vs. 41.55 ± 2.64%; t(26) = 5.572, P < 0.0001) and edema (Figure 1(e); 2.04 ± 0.45% vs. 12.74 ± 1.36%; t(25) = 8.136, P < 0.0001) at the 24 h timepoint compared with the tMCAO group. At 72 h, although the tMCAO mice had reduced lesion size compared to the 24 h timepoint, selenium treatment facilitated the recovery process by further reducing the infarct (Figure 1(c) and (d); 9.17 ± 2.11% vs. 22.06 ± 2.69%; t(21) = 3.725, P = 0.0013) and edema (Figure 1(e); 0.81 ± 0.27% vs. 7.39 ± 0.57%; t(20) = 9.840, P < 0.0001) volumes.

At 24 h, selenium treatment significantly reduced the neurological deficit score of the tMCAO + Se mice (Figure 1(f); 1.00 ± 0.13 vs. 1.58 ± 0.16; t(26) = 2.962, P = 0.0065). At 72 h, approximately 50% of mice in the tMCAO + Se group displayed no motor function impairments (neurological deficit score = 0), and the average score was significantly lower than that of the tMCAO group (Figure 1(f); 0.53 ± 0.17 vs. 1.18 ± 0.13; t(29) = 3.074, P = 0.0046), suggesting that recovery from the tMCAO insult was expedited by selenium treatment. All original TTC images used for the analysis of brain infarctions were shown in Supplementary Figure S6.

tMCAO mice showed a severe impairment in the motor function (Figure 1(g); F_(2, 34) = 65.59, P < 0.0001) and grip strength (Figure 1(h); F_(2, 31) = 32.51, P < 0.0001) tests, which lasted for at least 3 days after tMCAO. Selenium treatment significantly enhanced the post-tMCAO rotarod test scores after 1 day (Figure 1(g); 140.5 ± 11.71 s vs. 68.74 ± 9.96 s; P < 0.0001) and 3 days (Figure 1(g); 239 ± 28.55 s vs. 121 ± 17.16 s; P = 0.0009). Forelimb grip strength performance was also enhanced after 1 day (Figure 1(h); 57.43 ± 5.03 gf vs. 31.82 ± 4.71 gf; P = 0.0010) and 3 days (Figure 1(h); 76.02 ± 5.84 gf vs. 59.29 ± 5.25 gf; P = 0.1034) of selenium treatment. Together, these results demonstrate that delayed selenium treatment (2 h post reperfusion) confers a neuroprotective effect and leads to enhanced functional recovery after cerebral ischemia/reperfusion.

Delayed selenium treatment reduces ischemia/reperfusion-induced cell death, oxidative stress, and ferroptosis in the acute phase

To confirm that a single i.p. injection of selenium could increase selenium levels, plasma and brain samples were collected 24 h after the first selenium injection for ICP-MS. This revealed a dramatic 10-fold increase in circulating selenium (Figure 2(a)) and 14-fold increase in brain selenium (Figure 2(b)). Importantly, the increase in plasma selenium level correlated with reduced brain infarction (Figure 2(c); R² = 0.6250, F_(1, 14) = 23.33, P = 0.0003) and enhanced motor function (rotarod test in Figure 2(d); R² = 0.4982, F_(1, 14) = 13.90, P = 0.0022) 1 day after tMCAO as determined using simple linear regression analysis. These data demonstrate a beneficial effect of plasma selenium with stroke outcomes.

Figure 2.

Selenium treatment increases circulating and cerebral selenium levels while reducing oxidative stress, lipid peroxidation, and iron overload in the acute phase (24 h) after tMCAO. (a) Plasma selenium concentration and (b) selenium levels in the infarct hemisphere were measured by ICP-MS (n = 4 mice per group). (c, d) Simple linear regression analysis between plasma selenium and the corrected infarction volume (c) and maximum time reached in rotarod test (d) of tMCAO (red colored area) and tMCAO+Se (blue colored area) groups of mice 1 day after tMCAO. (e) Plasma potassium concentration was measured by ICP-MS (n = 9 mice per group). (f) Glutathione and oxidized glutathione ratio in the infarct hemisphere expressed as a percentage of sham (n = 8 mice per group). (g) Levels of MDA, and (h) 4-HNE in the infarct hemisphere (n = 8 mice per group). (i) Total iron levels in the plasma and (j) infarct hemisphere was measured by ICP-MS. (k) Plasma iron levels represented by the concentration of the iron-carrier protein ferritin (n = 8 mice per group). Data represent mean ± SD. (a, b, e–k: one-way ANOVA with Bonferroni post hoc comparisons).

In addition, the plasma ICP-MS analysis revealed a significant upsurge in circulating potassium 24 h after tMCAO (Figure 2(e)). Potassium is a pivotal ion for maintaining the electrical gradient of neurons, with a sudden increase in circulating potassium reflecting cell death-induced potassium release. In the sham group, the level of circulating potassium was around 3000 µg/l (Figure 2(e)), whereas that in the tMCAO group was 2 times higher. In comparison, the tMCAO + Se group had similar plasma potassium levels to the sham group.

Reperfusion-induced oxidative stress is a major factor that jeopardizes the intracellular redox balance, occurrence of subsequent lipid peroxidation and cell death. To evaluate cellular oxidative stress in the core infarct area, we measured the reduced glutathione (GSH) to oxidized glutathione (GSSG) ratio in the infarct brain sample 24 h post-tMCAO (Figure 2(d) to (f)). In the presence of oxidative stress, intracellular GSSG accumulates, and the GSH/GSSG ratio decreases. We found that the tMCAO mice had a significantly lower cellular GSH/GSSG ratio than the sham group (Figure 2(f); 41.95 ± 4.36% vs. 100 ± 5.09%; P < 0.0001) and that selenium treatment partially restored this ratio (Figure 2(f); 86.27 ± 6.76% vs. 41.95 ± 4.36%; P < 0.0001). The lipid peroxidation level was measured using two compounds, malondialdehyde (MDA) and 4-hydroxynonenal (4-HNE), both of which are by-products of the peroxidation of polyunsaturated fatty acids. We found that, 24 h after tMCAO, the level of MDA was increased approximately two-fold (Figure 2(g); 1.954 ± 0.118 µmol/g vs. 0.604 ± 0.059 µmol/g; P < 0.0001), and the level of 4-HNE was elevated approximately three-fold (Figure 2(h); 2.896 ± 0.17 ng/g vs. 1.034 ± 0.098 ng/g; P < 0.0001) in the core infarct area compared to the sham. With selenium treatment, the MDA and 4-HNE levels were reduced by 30% and 20%, respectively (Figure 2(g) and (h); MDA: 1.313 ± 0.136 µmol/g vs. 1.954 ± 0.118 µmol/g; P = 0.0014; 4-HNE: 2.147 ± 0.137 ng/g vs. 2.896 ± 0.17 ng/g; P = 0.0029).

Previous research has suggested that selenium treatment confers protective effects by inhibiting ischemia/reperfusion-induced ferroptosis. ¹⁰ Our finding that MDA and 4-HNE, two critical but not exclusive markers of ferroptosis, were significantly reduced after selenium treatment (Figure 2(g) and (h)) support this conclusion. To further corroborate the anti-ferroptotic effect of selenium, we tested another key marker of ferroptosis, the local accumulation of iron, by ICP-MS. When compared with the levels in the sham group, circulating iron and iron in the infarct brain tissue of the tMCAO group increased by 2- and 4-fold, respectively, while selenium treatment effectively reinstated iron homeostasis 24 h after tMCAO (Figure 2(i); Sham vs. tMCAO: P = 0.0056; tMCAO vs. tMCAO + Se: P = 0.0306; Figure 2(j); Sham vs. tMCAO: P = 0.0388; tMCAO vs. tMCAO + Se: P = 0.0352). This result was further corroborated by measuring plasma ferritin, an iron-carrier protein, 24 h post-tMCAO. This analysis revealed that the plasma ferritin levels were significantly higher in the tMCAO group and that selenium treatment counteracted this change (Figure 2(k); Sham vs. tMCAO: P = 0.0119; tMCAO vs. tMCAO + Se: P = 0.0106). Interestingly, in addition to being markers of ferroptosis, increased circulating ferritin and iron also correlate with stroke severity and the size of the infarct in clinical stroke patients. ³¹ Together, these results indicate that anti-oxidative and anti-ferroptosis pathways are both critical mechanisms regulating the protective effects of selenium against cerebral ischemia/reperfusion injury.

Delayed selenium treatment rescues post-tMCAO long-term cognitive dysfunction

It is now recognized that a cortical stroke can cause progressive loss of brain tissue, a phenomenon known as secondary neurodegeneration. ³ Unlike humans, mice can quickly recover from transient ischemia/reperfusion-induced motor impairment, but their cognitive performance, spatial learning, and memory have not yet been fully investigated. We therefore designed a battery of behavioral tests to evaluate cognitive function after tMCAO and to determine whether this is impacted by selenium supplementation.

A cohort of mice were allowed to recover for 7 days after tMCAO. Selenium was delivered by i.p. injection for the first 3 days and then switched to dietary supplementation in the drinking water (10 µg/ml, average daily intake is 30 µg per mouse) to reduce the daily intake of selenium and to prevent repetitive injection-induced stress (Figure 3(a)). This dosage was carefully selected according to our previous study in which we observed no obvious adverse reactions after 4 weeks of dietary supplementation in young or old mice. ¹³ Plasma selenium levels were also monitored for the first two weeks after tMCAO, revealing that these levels dropped by 50% when dietary selenium in the drinking water replaced the i.p. injections starting from day 3, then maintained a constant level until day 14 (Supplementary Figure S2A). The plasma iron content did not change after two weeks of selenium supplementation (Supplementary Figure S2B). In fact, most ion levels remained unchanged (Supplementary Figure S3) except for significantly increased levels of selenium and vanadium (Supplementary Figure S3A, E), and reduced levels of manganese, zinc, and cadmium (Supplementary Figure S3B-D).

Figure 3.

Selenium supplementation rescues post-tMCAO spatial learning and memory deficits. (a) Experimental timeline. (b) Y maze test to show spontaneous alternation (left), total path travelled (middle) and total arm entries (right) (n = 10 mice per group). (c) Schematic illustration the APA test. The red line encloses a 60^° area where an electric shock is delivered. (d) Representative trace graphic on day 5 of the test for three different groups of mice. (e) The improvement over the 5 days of the APA test, expressed as percentage improvement and (f) the mean number of entries to the shock zone (n = 10–12 mice per group). Data represent mean ± SD. (b, e: one-way ANOVA and Bonferroni post hoc comparisons; f: repeated-measures two-way ANOVA (effect of selenium treatment F_{(2, 31)} = 20.28, P < 0.0001) and Bonferroni post hoc comparisons: **P < 0.01, ***P < 0.001, ****P < 0.0001; $P < 0.05, $$P < 0.01, $$$$P < 0.0001; #P < 0.05).

After recovering for 7 days following tMCAO surgery, mice were randomly assigned to two groups that underwent behavioral testing using either the Y maze or active place avoidance (APA) tasks. We found that the tMCAO induced a deficit in spatial working memory tested using the Y maze task, and that this deficit could be entirely rescued by selenium treatment (Figure 3(b) left panel; 56.89 ± 1.76% vs. 47.67 ± 2.257%; P = 0.0101; F_(2, 27) = 8.515, P = 0.0020). All groups of mice exhibited similar willingness and ability to explore new environments (Figure 3(b) the middle and right panel, respectively).

The APA test was performed on a separate group of mice to assess spatial navigation and memory. ²¹ Ischemia/reperfusion induced a severe deficit in the tMCAO group of mice which failed to learn to avoid the shock zone (Figure 3(c) to (e)). In contrast, the tMCAO + Se group performed significantly better on the last two days of the test (Figure 3(f); effect of selenium treatment F_(2, 31) = 20.28, P < 0.0001), suggesting an improvement in spatial learning ability. When the data for the improvement over the five days of the trial were expressed as a percentage, the tMCAO + Se group performed as well as the sham group, whereas the tMCAO group performed significantly worse (Figure 3(e); tMCAO vs. Sham: 13.03 ± 9.417% vs. 55.44 ± 6.707%; P = 0.0010; tMCAO + Se vs. tMCAO: 47.84 ± 5.599% vs. 13.03 ± 9.417%; P = 0.0053). To exclude the possibility of mobility differences amongst these groups of mice, the average velocity and total distance travelled during the 5 days of the APA test were recorded, which revealed no difference amongst the three groups of mice (Supplementary Figure S4).

To evaluate the long-term spatial learning and memory ability in these mice, a retrieval APA test was conducted 4 weeks after the last day of the first APA test (Figure 3(a)). In this session, the sham group had the lowest number of entries into the shock zone. Both the tMCAO and tMCAO + Se groups performed significantly worse than the sham group (Figure 3(f)). However, the number of shock zone entries for the tMCAO + Se group was significantly lower than that of the tMCAO group (Figure 3(f); 11.17 ± 1.353 vs. 16.83 ± 1.637; P = 0.0173), indicating a long-lasting beneficial effect of selenium supplementation.

Long-term selenium supplementation activates intrinsic neuroprotective mechanisms and prevents secondary ischemia/reperfusion neural degeneration

Hippocampal CA1 pyramidal neurons and fast-spiking parvalbumin (PV)-positive interneurons are particularly susceptible to an ischemia/reperfusion insult. ³² Therefore, we first determined whether long-term selenium supplementation (10 µg/ml in their drinking water for 6 weeks) prevented the loss of these vulnerable neurons. As shown in Figure 4, immunostaining of the mouse brain 6 weeks after tMCAO revealed a loss of cerebral cortex tissue and the appearance of scar formation in the ipsilateral hemisphere of both the tMCAO and tMCAO + Se groups (Figure 4(a) and (b)). In addition, a noticeable loss of CA1 NeuN-positive neurons was observed in the tMCAO group. In contrast, the corresponding CA1 neurons in the tMCAO + Se group were largely intact (Figure 4(a), (b) and (g); 394.4 ± 38.15 vs. 267.3 ± 44.93; P = 0.0448).

Figure 4.

Long-term selenium supplementation (6 weeks) following tMCAO prevents secondary neural degeneration and promotes neurogenesis in the infarct hemisphere. (a) NeuN (green) and PV (red) immunostaining of tMCAO mice or (b) tMCAO + Se mice 6 weeks after tMCAO. The infarct was in the right hemisphere. The red arrow indicates secondary hippocampal cell death, the yellow arrow indicates PV-positive neurons in the thalamic reticular nucleus (TRN), and the white arrow indicates a loss of intact cortical area due to the stroke insult. (c) c-Fos (red) and NeuN (green) immunostaining of the infarct DG for tMCAO or (d) tMCAO + Se mice 60 min after the APA retrieval test. (e) DCX (red) immunostaining of the infarct DG for tMCAO or (f) tMCAO + Se mice 6 weeks after tMCAO. (g) Relative NeuN fluorescence intensity in the CA1 region of the right hippocampus. (h) The number of c-Fos-positive cells per right DG following APA retrieval. (i) Relative PV fluorescence intensity in the TRN area of the right hemisphere. (j) The number of DCX-positive immature neurons per right DG. (g–j: unpaired two-tailed t-test, n = 5–10 mice per group). Data represent mean ± SD.

It has previously been reported that, during a memory retrieval task, the number of c-Fos-positive cells in the hippocampal dentate gyrus (DG) represents the activation of memory engram cells in the DG and mirrors their performance in the memory task. ³³ To further examine the histological changes in the hippocampus, we therefore also determined the number of c-Fos-positive cells. Our analysis revealed a higher number of c-Fos-positive cells in the tMCAO + Se group verses the tMCAO group in the ipsilateral DG (Figure 4(c), (d) and (h); 1496 ± 89.63 vs. 1026 ± 144.4 c-Fos⁺ cells, P = 0.045).

A significant loss of PV-positive interneurons also occurred in the thalamic reticular nucleus (TRN), which controls the information flow between the thalamus and the cortex and regulates several higher cognitive functions such as emotion, pain regulation and sleep rhythm,^34,35 in both the tMCAO and tMCAO + Se groups (Figure 4(a) and (b)). The quantification of PV fluorescence intensity demonstrated a significantly higher PV intensity in the tMCAO + Se group (Figure 4(i); 3639 ± 324.8 vs. 1901 ± 348.4; P = 0.0018). In contrast, the number of PV-positive cells in the contralateral hippocampus was similar between the tMCAO and tMCAO + Se groups (Supplementary Figure S5A).

We have recently shown that dietary selenium treatment also increases adult-neurogenesis in the DG. ²¹ In agreement with our previous data, long-term selenium supplementation after tMCAO significantly increased the number of DCX-positive new-born neurons in the ipsilateral DG (Figure 4(e), (f) and (j); 6749 ± 289.2 vs. 5358 ± 290.0 DCX⁺ cells, P = 0.0032), but not the contralateral DG (Supplementary Figure S5B).

Collectively, these data demonstrate that long-term selenium supplementation activates intrinsic neuroprotective mechanisms and prevents secondary ischemia/reperfusion neural degeneration.

Long-term selenium supplementation following tMCAO inhibits systemic inflammation and enhances the FGF-21 level

Clinical observation suggests that selenium treatment can effectively reduce the level of circulating C-reactive protein (CRP), supporting its anti-inflammatory capacity. ³⁶ Therefore, we next investigated how selenium affects the inflammatory response post-tMCAO in a separate cohort of mice. Plasma was collected at 5 different time points after tMCAO, and the CRP level was measured by ELISA. As shown in Figure 5(a), the sham group of mice exhibited a stable and low CRP level across time, whereas the tMCAO group had a significantly higher level of CRP on day 1 post-tMCAO (65% higher than the sham), which peaked on day 7 (200% higher than the sham) and then started to decline. In contrast, the extent of CRP changes in the tMCAO + Se group was significantly less than in the tMCAO group (Figure 5(a); tMCAO vs tMCAO + Se: day 3: P = 0.0038; day 7: P = 0.0059; day 14: P = 0.0492). By day 42, the levels of CRP in all three groups had returned to baseline levels, indicating the absence of inflammation at this timepoint.

Figure 5.

Long-term selenium supplementation inhibits systemic inflammation and enhances liver secretion of FGF-21 following tMCAO. (a) Plasma levels of CRP at different time points after tMCAO (n = 9 mice per group). (b) IGF-1 and (c) FGF-21 levels in the plasma 6 weeks after the tMCAO (n = 9–11 mice per group). Data represent mean ± SD. (a) repeated-measures two-way ANOVA (effect of selenium treatment F_(2, 24) = 41.06, P < 0.0001) and Bonferroni post hoc comparisons: **P < 0.01, ***P < 0.001, ****P < 0.0001; #P < 0.05, ##P < 0.01. and (b, c) one-way ANOVA and Bonferroni post hoc comparisons).

Clinical studies have also suggested that dietary selenium supplements confer protection against age-related cognitive decline by stimulating the liver secretion of insulin-like growth factor 1 (IGF-1) and fibroblast growth factor 21 (FGF-21).^37–39 To determine whether these proteins are involved in selenium-mediated neuroprotection, we collected plasma samples on day 42 after tMCAO and tested them by ELISA. The tMCAO group exhibited significantly reduced plasma IGF-1 levels, and long-term selenium supplementation did not alter its secretion (Figure 5(b); P = 0.016 vs. tMCAO; P = 0.0010 vs. tMCAO + Se). In contrast, the secretion of FGF-21 was reduced in the tMCAO group, and selenium treatment significantly augmented this secretion by 50% and 140% compared to the sham and tMCAO groups, respectively (Figure 5(c); 558.1 ± 34.48 pmol/ml vs. 364 ± 28.96 pmol/ml; P < 0.0001 vs. sham; 558.1 ± 34.48 pmol/ml vs. 233.5 ± 41.1 pmol/ml; P = 0.0017 vs. tMCAO).

Selenium treatment prevents tMCAO-induced acute dysbiosis and subsequent trimethylamine-N-oxide (TMAO) secretion

There is growing evidence that post-stroke dysbiosis and bacteria-derived compounds can exacerbate the outcome of stroke and jeopardize functional recovery.^40,41 It has also been reported that dietary selenium supplementation can modulate the composition of the gut flora. ⁴² We therefore investigated whether selenium treatment (i.p. injection in an experimental time scheme shown in Figure 3(a)) affects the composition of gut flora after tMCAO. Fecal samples were collected 24 h after tMCAO, and their bacterial composition was analyzed by full-length 16S rRNA sequencing. We found that the complexity of the gut microbes was considerably increased after tMCAO (Figure 6(a); 56.3 ± 1.324 vs. 47.37 ± 1.33; P = 0.0345) and that selenium treatment reversed this effect (Figure 6(a); 50.13 ± 3.153 vs. 47.37 ± 1.33; P > 0.99). Specifically, the Enterobacteriaceae and Ruminococcaceae families were significantly enriched, whereas Lachnospiraceae, Lactobacillaceae, Deferribacteraceae, and Erysipelotrichaceae were depleted in the gut of tMCAO mice (Figure 6(b)). Following selenium treatment, the enrichment of Enterobacteriaceae was significantly inhibited (Figure 6(b); P = 0.015).

Figure 6.

Selenium treatment prevents tMCAO-induced acute dysbiosis and subsequent TMAO production. (a) Alpha diversity of gut microbiota 24 h after tMCAO. (b) Relative abundance of Enterobacteriaceae, (c) Ruminococcaceae, (d) Lachnospiraceae, (e) Lactobacillaceae, (f) Deferribacteraceae, and (g) Erysipelotrichaceae amongst the sham, tMCAO, and tMCAO+Se groups 24 h following surgery (n = 5 mice per group). (h) Circulating TMA and (i) TMAO levels in the plasma at different time points after tMCAO (n = 9 mice per group). Data represent mean ± SD. [A–G: one-way ANOVA and Bonferroni post hoc comparisons, h, i: repeated-measures two-way ANOVA (effect of selenium treatment k: F_(1, 16) = 34.93, P < 0.0001; L: F_(1, 16) = 37.33, P < 0.0001) and Bonferroni post hoc comparisons]. Blood plasma levels of choline (j), TMA (k) and TMAO (l) and (m) The alpha diversity of gut microbiota following selenium supplementation in mice supplied a high choline diet. (n = 9–10 mice per group) Data represent mean ± SD. (unpaired two-tailed t-test: P > 0.05). *P < 0.05, **P < 0.01.

These results are consistent with the clinical observation that Enterobacteriaceae enrichment after stroke is a key marker of gut dysbiosis and dictates functional recovery. ⁴⁰ To investigate how this enrichment produces poor functional outcomes after tMCAO, we focused on bacteria-derived trimethylamine (TMA) and its oxidized product TMAO, which is known to damage the cerebrovascular system, induce atherosclerosis, exacerbate ischemic stroke outcomes, and increase the risk of stroke recurrence.^41,43,44 A recent study suggested that Enterobacteriaceae harbor the high choline-TMA conversion enzyme, and that the enrichment of Enterobacteriaceae is associated with higher urine TMA and TMAO content. ⁴⁵ Intriguingly, selenium treatment significantly reduced circulating TMA and TMAO from day 1 to day 42 after tMCAO (Figure 6(h) and (i); effect of selenium treatment H: F_(1, 16) = 34.93, P < 0.0001; I: F_(1, 16) = 37.33, P < 0.0001). This result suggests that selenium treatment delivered after tMCAO can prevent ischemia/reperfusion-induced gut dysbiosis and inhibit the detrimental secretion of gut bacteria-derived TMAO. However, it is unclear whether selenium directly or indirectly affects the composition of the gut microbiome. To examine this, we supplied adult mice with a high choline diet (1% choline in chow) to induce TMAO production and supplemented their drinking water with selenium for 3 weeks. We found that, in healthy mice, selenium supplementation neither reduced the circulating TMAO concentration, nor altered the richness or diversity of the gut bacteria (Figure 6(j) to (m); P > 0.05, unpaired two-tailed t test). This suggests that selenium does not directly target the gut bacteria to reduce TMAO production, but instead protects the brain against cerebral ischemic insult and indirectly alleviates subsequent gut dysbiosis and TMAO production.

The combined antioxidative properties of selenium and N-acetylcysteine significantly enhances the neuroprotective effect

Finally, we questioned whether a combinatorial treatment could be used to increase the neuroprotective effects of selenium. N-acetylcysteine (NAC) is a clinically used mucolytic agent, which is also known to possess powerful anti-oxidative and anti-ferroptosis capacities. ⁴⁶ NAC has been shown to cross the blood-brain barrier and increases the intracellular content of glutathione, the substrate of the selenoprotein GPX4. ⁴⁷ This suggests that it could be used in conjunction with selenium to optimize its anti-oxidative efficacy. To determine the protective effect of NAC in combination with selenium, either selenium or selenium with NAC were injected to tMCAO mice as described in the timeline shown in Figure 7(a). We found that, compared to selenium alone, the combination of NAC and selenium significantly enhanced the neuroprotective effect in the acute phase of ischemic/reperfusion. The infarct volume of the NAC + selenium group (Se+NAC) was significantly reduced at 1–3 days post-tMCAO compared to the selenium alone group (Figure 7(b) and (c); day 1: 14.62 ± 1.884% vs. 23.44 ± 2.738%; P = 0.0192; day 3: 4.283 ± 1.44% vs. 9.411 ± 2.311%; P = 0.076). However, the edema volumes were not affected (Figure 7(d)). We then used the APA test to assess spatial navigation and memory 3 days after tMCAO. The tMCAO group of mice treated with the combination of NAC and selenium (tMCAO + Se +NAC) were significantly better at avoiding the shock zone than the selenium only group. When the data for the improvement over the five days of the trial were expressed as a percentage value, the tMCAO + Se +NAC group performed significantly better than the tMCAO + Se group (Figure 7(e) and (f), P = 0.0227). Together these findings highlight that the neuroprotective effect of selenium can be further enhanced in combination with another antioxidative agent.

Figure 7.

The combination of NAC with selenium enhances the neuroprotective effect of the latter. (a) Experimental timeline. (b) Representative TTC staining of tMCAO brain sections 1 and 3 days after tMCAO. (c) Percentage of total edema-corrected infarct volume, and (d) edema treated with selenium (Se) or selenium and NAC together (Se + NAC), 1 day and 3 days after 30 min tMCAO (n = 10–12 mice in each group). (e) The improvement over the 5 days of the APA test expressed as percentage improvement, and (f) the mean number of entries to the shock zone. Data represent mean ± SD. (c–e: unpaired two-tailed t test; f, two-way ANOVA with Tukey’s post hoc comparisons: ****P < 0.0001, ####P < 0.0001 compared with Day 1, respectively).

Discussion

The indispensable trace element selenium is emerging as a versatile, multi-target therapy that has both direct and indirect effects on ischemia/reperfusion recovery. Here, we demonstrate that delayed selenium treatment effectively reduces the cerebral infarct volume and rescues motor function impairment in the acute phase after tMCAO. We also show that longer-term selenium supplementation during the convalescence phase rescues the subsequent spatial learning and memory deficits, decreases secondary neurodegeneration, alleviates systemic inflammation, and reduces gut dysbiosis.

Despite several epidemiological studies reporting an inverse correlation between blood selenium level and the prevalence of stroke, few animal studies have investigated selenium supplementation as a viable intervention. Selenium is significantly decreased in the ischemic brain, and it has been shown that selenium pre-treatment protects neurons against ischemic damage by reducing DNA oxidation. ¹² Recently, it was reported that a single dose of selenium delivered directly into the brain protects neurons and improves sensory neglect behavior in a mouse hemorrhagic stroke model via inhibition of ferroptotic cell death. ¹⁰ The hallmark feature of ferroptosis is the iron-dependent accumulation of oxidatively damaged phospholipids. Given its antioxidant capacity, selenium treatment rapidly reinstated the cellular redox balance, prevented lipid peroxidation, and protected neurons against ischemia/reperfusion insult.

Our data support the finding that selenium suppresses oxidative damage-induced ischemic neuronal death, which underlies the early and acute phase of ischemic/reperfusion damage to the brain. A recent clinical trial has demonstrated that the combination of selenium with NAC, one of the most widely used antioxidant drugs in preclinical and clinical studies, effectively reduced the perihematomal edema in patients with intracerebral hemorrhage. ⁴⁸ However, no studies have investigated the effects of this combinational therapy following ischemic stroke. We show that combining selenium with NAC, results in a further significant decrease in infarct volume 1 d after ischemia compared to selenium alone, which also translates to an improvement in functional recovery. Interestingly, no additional effect of NAC was observed on edema. This is likely due to the 30 min tMCAO model used in this study, which does not induce severe brain edema together with the fact that the selenium supplementation itself had already mitigated this resulting edema. Both selenium and NAC are widely available and relatively inexpensive compared to other ROS scavengers such as edaravone and glutathione. The mechanism of NAC suggests that it may enhance selenium-dependent antioxidant systems by replenishing the intracellular cystine pool. Although our study is the first to investigate the potential effect of combining NAC and selenium for the treatment of ischemic stroke, the resulting data strongly support the translational prospect of this combined therapy.

We have recently shown that selenium supplementation can enhance hippocampal neurogenesis levels and rescue the hippocampus-associated spatial learning and memory deficits observed in mouse models of physiological ageing and hippocampal injury. ²¹ In the present study we show that long-term selenium delivery (7 or 28 days) protects against the loss of mature (NeuN⁺) neurons. In particular, the PV-positive interneurons in the TRN. Preserving the integrity of the TRN and PV-positive inhibitory interneurons is critical for stroke recovery as the damage to the interneuron system is likely to result in post-stroke seizures and epilepsy, which are highly correlated with post-stroke mortality. ⁴⁹ We also observed a significant increase in adult hippocampal neurogenesis (DCX⁺ cells) in the tMCAO + Se group compared to the tMCAO group, concomitant with an improvement in spatial learning and memory function. This is not surprising given that it is these newborn neurons that are known to contribute to this form of memory.

The primary mode of cell death by which neurons die following ischemia remains unclear. Although apoptosis is clearly involved in this process, it is becoming increasingly recognized that additional cell death mechanisms, including ferroptosis, are also involved in mediating this response. Our findings that lipid peroxidation levels and local accumulation of iron, two key markers of ferroptosis, were significantly reduced in stroke brains after selenium treatment support this conclusion. FGF-21 is known to act as a potent neuroprotective protein,^50,51 and was recently identified as an intrinsic ferroptosis suppressor. ¹⁹ We found that the secretion of FGF-21 was reduced in the tMCAO group, and selenium treatment significantly augmented this secretion. It is therefore, intriguing to consider whether selenium might also function indirectly by stimulating the secretion of FGF-21.

Modulation of the gut microbiota can be mediated by the host’s selenium status ⁵² with dietary selenium deficiency resulting in increased selenium uptake by the gut microbiota and decreased selenoprotein expression by the host resulting in the increased release of pro-inflammatory cytokines and a reduction in immune cell action. ⁴² Studies have also linked changes in the gut microbiota with ischemic stroke pathogenesis via the gut-brain axis. Alterations in the diversity, abundance and functions of the gut microbiome, termed gut dysbiosis, result in dysregulated gut-brain signaling, thereby inducing intestinal barrier changes, endotoxemia, systemic inflammation and infection, all of which affect stroke outcomes. ⁵³ Stroke can lead to an increased abundance of opportunistic pathogens such as the Enterobacteriaceae bacteria, ⁴⁰ which produce a variety of harmful substances, such as lipopolysaccharide (LPS), ⁵⁴ which can trigger neuroinflammation via TLR-4 and further aggravate stroke injury. Despite several human clinical studies showing associations between gut dysbiosis and cardiovascular disease, whether selenium plays a direct effect in mediating this effect remains unknown. ⁴² In our study, we found that selenium treatment, significantly inhibited the stroke-induced enrichment of Enterobacteriaceae. However, whether this contributes to the reduction in stroke-induced systemic inflammation that we observed following selenium supplementation remains unclear.

Gut dysbiosis is mediated by microbially derived metabolites including TMAO and short-chain fatty acids, bacterial components such as lipopolysaccharide (LPS) and immune cells such as T helper cells. TMAO has been shown to lead to cardiovascular disease and stroke.^41,43,44 TMA is released from the metabolism of choline, lecithin, and L-carnitine, found in foods such as eggs, dairy, fish and meat, by the gut microbiota and is converted to TMAO in the liver. Our results demonstrate that, in healthy mice, additional selenium supplementation does not alter the gut microbe composition or reduce the level of circulating TMAO. This observation is in agreement with the results of a recent clinical study which showed that 3 months of selenium supplementation (via the consumption of Brazil nuts) did not reduce the levels of plasma TMA or TMAO in patients with coronary artery disease. ¹⁶ Hence, we propose that under an ischemic attack, selenium directly reduces the magnitude of the brain insult, thereby indirectly decreasing stroke-induced TMAO production. Whether inhibition of Enterobacteriaceae leads to a direct reduction in circulating TMAO remains unknown, and the causal relationship with stroke outcomes requires further investigation.

The current study provides new insight into the impact of delayed selenium supplementation on neuroprotection and functional recovery following ischemic stroke. We show that selenium can simultaneously inhibit multiple detrimental mechanisms induced by cerebral ischemia/reperfusion, including increased ferroptosis, decreased neurogenesis, systemic inflammation and gut dysbiosis. However, it remains unclear whether selenium is independently targeting these processes or whether these actions are influenced by each other. Nonetheless, these mechanistic insights broaden the understanding of how selenium supplementation enhances ischemia/reperfusion recovery and suggest beneficial clinical outcomes for stroke patients, including in combination with the clinically used mucolytic agent NAC.