Histone Methyltransferase Dot1L Contributes to RIPK1 Kinase‐Dependent Apoptosis in Cerebral Ischemia/Reperfusion

Abstract

Background

Neuron apoptosis is a pivotal process for brain damage in cerebral ischemia. Dot1L (disruptor of telomeric silencing 1‐like) is only known histone H3K79 methyltransferase. It is not clear whether the role and mechanism of Dot1L on cerebral ischemia is related to regulate neuron apoptosis.

Methods and Results

We use a combination of mice middle cerebral artery occlusion stroke and neurons exposed to oxygen‐glucose deprivation followed by reoxygenation to investigate the role and mechanism of Dot1L on cerebral ischemia. We find knockdown or inhibition of Dot1L reversed ischemia‐induced neuronal apoptosis and attenuated the neurons injury treated by oxygen‐glucose deprivation followed by reoxygenation. Further, blockade of Dot1L prevents RIPK1 (receptor‐interacting protein kinase 1)‐dependent apoptosis through increased RIPK1 K63‐ubiquitylation and decreased formation of RIPK1/Caspase 8 complexes. In line with this, H3K79me3 enrichment in the promoter region of deubiquitin‐modifying enzyme A20 and deubiquitinase cylindromatosis gene promotes the increasing expression in oxygen‐glucose deprivation followed by reoxygenation ‐induced neuronal cells, on the contrary, oxygen‐glucose deprivation followed by reoxygenation decreases H3K79me3 level in the promoter region of ubiquitin‐modifying enzyme cIAP1 (cellular inhibitors of apoptosis proteins), and both these factors ultimately cause K63‐deubiquitination of RIPK1. Importantly, knockdown or inhibition of Dot1L in vivo attenuates apoptosis in middle cerebral artery occlusion mice and reduces the extent of middle cerebral artery occlusion ‐induced brain injury.

Conclusions

These data support for the first time, to our knowledge, that Dot1L regulating RIPK1 to the apoptotic death trigger contributes to cerebral ischemia injury. Therefore, targeting Dot1L serves as a new therapeutic strategy for ischemia stroke.

Nonstandard Abbreviations and Acronyms

Dot1L

Disruptor of telomeric silencing 1‐like

Dot1L^+/‐

Dot1L heterozygous knockout

I/R

ischemia‐reperfusion injury

MCAO

middle cerebral artery occlusion

OGD/R

oxygen‐glucose deprivation followed by reoxygenation

WT

wild‐type

Clinical Perspective

What Is New?

• Using oxygen glucose deprivation‐induced primary cortical neurons and disruptor of telomeric silencing 1‐like (Dot1L) knockdown mice subjected to middle cerebral artery occlusion, we determine histone H3K79 methyltransferase Dot1L participates in cerebral ischemic diseases.

• Blockade of Dot1L prevents neurons from initiating apoptosis and promotes restoration of blood flow and provides neuroprotection against cerebral ischemia.

What Are the Clinical Implications?

• Our study reveals a previously unexplored role for Dot1L in cerebral ischemia disease, suggesting that Dot1L may represent a novel opportunity for the therapeutic manipulation of ischemic stroke.

Stroke is a major cause of disability and mortality around the world. ¹ At present, the most important treatment for ischemic stroke is to restore the blood supply quickly, which are believed to replenish nutrients and oxygen, as well as remove toxic metabolites. However, restoration of blood flow might also aggravate brain injury and functional damage, resulting in cerebral ischemia‐reperfusion injury (I/R). Many studies have proven several mechanisms involved in the pathological process of I/R injury, including excitatory toxicity, oxidative stress, inflammation, and apoptosis. ² , ³ Hence, prevention of apoptosis and restoration of blood flow are among the most important strategies, which could be a target for treatment of ischemic stroke.

The extrinsic pathway of apoptosis is initiated by the binding of death ligands of the TNF receptor superfamily. ⁴ Ligation of TNF (tumor necrosis factor) to TNFR1 (tumor necrosis factor receptor‐1) results in formation of a membrane protein complex (called Complex I). RIPK1 (receptor‐interacting protein kinase 1), a protein of the receptor‐interacting protein kinase family, is ubiquitinated in Complex I to induce the formation of a cell survival complex that promotes activation of nuclear factor‐κB. ⁵ , ⁶ The cIAP (cellular inhibitors of apoptosis proteins) 1/2 maintain RIPK1 in this complex to decrease the release of RIPK1 from this complex. But the degradation of cIAP1/2 would promotes the deubiquitination of RIPK1 by deubiquitinase cylindromatosis (CYLD) releases RIPK1 from this complex and results in the interaction of RIPK1 with FADD (Fas‐associated death domain protein), RIPK3, and Caspase 8 to form a cytosolic protein complex (Complex II), leading to the activation of Caspase 8. ⁷ , ⁸ Activated Caspase 8 can cleave and activate the executor caspases such as Caspase 3 and Caspase 7. This in turn triggers downstream effector mechanisms of apoptosis.

In recent years, mounting evidence indicates that aberrant histone modification patterns can play a causative role in disease. Since epigenetic alterations are reversible in nature, histone (de) modifiers are attractive therapeutic targets. ⁹ Dot1L (disruptor of telomeric silencing 1‐like) is a histone H3K79 methyltransferase and produces mono‐, di‐, and tri‐methylated histone H3 lysine 79 (H3K79me1/2/3). ¹⁰ Recent genome‐wide analysis of the mammalian epigenome has shown that H3K79 methylation is mainly associated with actively transcribed genes. ¹¹ , ¹² Dot1L involved in the control of gene expression further supports the relationship between H3K79 methylation and gene expression. ¹³ , ¹⁴ , ¹⁵ Earlier studies have reported that Dot1L plays an important role in heart embryonic growth and cell cycle, ¹⁶ telomeric silencing, and the DNA damage response. ¹⁷ Histone modifications mediate gene expression, therefore, significantly influence functional outcome and neuronal survival after cerebral ischemia. ¹⁸ Nevertheless, it remains unclear whether Dot1L participates in cerebral ischemic diseases. Understanding the regulatory effect of Dot1L could not only uncover additional therapeutic opportunities, but also provide information about the fundamental function of Dot1L and H3K79 methylation in mammalian gene regulation.

In the present study, we discover Dot1L‐mediated cell death and inflammatory responses in cerebral I/R injury. Importantly, knockdown or inhibition of Dot1L mice is protected from I/R injury and also promotes better recovery after stroke. Importantly, we find that deubiquitinase A20 and CYLD and ubiquitinase cIAP1, as crucial factors for the activation of RIPK1/Caspase 8 signaling in apoptosis, are the target genes of Dot1L. The results of this study expand our knowledge in the molecular mechanisms of Dot1L contributing to RIPK1 kinase‐dependent apoptosis and demonstrate that blockade of Dot1L provides neuroprotection against cerebral ischemia.

Methods

The data that support the finding of this study are available within the article and in Data S1.

Experimental Animals

All animals and the experimental protocol conformed to the Animal Welfare Act Guide for Use and Care of Laboratory Animals and were approved by Institutional Animal Care and Use Committee, Fudan University, China. To investigate the role of Dot1L on neuroprotection in vivo, Dot1L heterozygous knockout (Dot1L^+/‐) mice were generated on a C57BL/6J background by using CRISPR/Cas9 system, for the sake of homozygous deletion of Dot1L led to embryonic lethality.

Cell Culture

Primary cortical neurons were obtained from SD (Sprague Dawley) rats after birth day 1 as previously described ¹⁹ and cultured at 37 °C in DMEM containing 10% fetal bovine serum (Gibco, USA). After 8 hours, the medium was changed into fetal bovine serum‐free DMEM‐F12 (Hyclone, USA) medium containing 2% B27. Cells were cultured for 7 days and were identified using specific neuron marker Map‐2 (Microtubule‐associated Protein 2) by immunofluorescent staining.

Oxygen‐Glucose Deprivation Model in Vitro

Oxygen‐glucose deprivation followed by reoxygenation (OGD/R) was established as follows: The prepared neurons were cultured in DMEM glucose‐free medium and were incubated for 4 hours in <0.1% oxygen in the hypoxia box. Then cultures were switched to completely normal conditions for 24 hours.

RNA Extraction and Quantitative Real‐Time Polymerase Chain Reaction

Total RNA was isolated from the cells or tissues using TRIzol reagent (Takara Bio Inc., China). Reverse transcription was performed according to the instruction of the Prime RT Master Mix kit (Takara Bio Inc., China). Quantitative real‐time polymerase chain reaction was performed in triplicate using SYBR Premix EX Taq II (Yeasen Biotech Co., Ltd. China). The sequences of primers were shown in Table S1.

Statistical Analysis

Data were expressed as mean±SD. Differences of means were analyzed by using 1‐way ANOVA with the Tukey‒Kramer post hoc test for multiple groups, and when comparing between 2 groups using unpaired Student t‐tests. P<0.05 was accepted as statistically significant.

Results

Dot1L Expression at Different Time Points After OGD in Cultured Cortical Neurons and After Middle Cerebral Artery Occlusion

Time course experiments clearly showed that the level of Dot1L and H3K79me3 was elevated at 1 hour after OGD in the primary cortical cultures (Figure S1A). OGD also induced mRNA expression of Dot1L (Figure S1B). To examine whether Dot1L contributes to damage after stroke, mice were subjected to 90‐minutes middle cerebral artery occlusion (MCAO) followed by 24 hours reperfusion. In line with the findings of in vitro OGD, we observed that the level of Dot1L was increased after MCAO; brain tissue from MCAO mice exhibited an increase in H3K79me3 abundance (Figure S1C). Therefore, we next determined the contribution of Dot1L to the cerebral I/R injury.

Inhibition of Dot1L Reduces Infarct Volume and Neuronal Apoptosis After MCAO

EPZ5676, a specific inhibitor of Dot1L, was used to investigate the effects on MCAO mice. The design of the experiment was shown in Figure 1A. Firstly, we determined the content of EPZ5676 in brain tissue and plasma by HPLC (High Performance Liquid Chromatography) method, and found concentrations reached to certain level in serum and brain tissue, respectively (Figure 1B). Compared with the MCAO group, using EPZ5676 reduced infarct volume and Bederson score after MCAO (Figure 1C and 1D). After MCAO, the level of cell apoptosis was higher in vehicle treated‐mice brain sections relative to EPZ5676‐treated mice brain sections (Figure 1E). Further, our results revealed that cleaved‐Caspase 3, cleaved‐Caspase 9, cleaved‐PARP1, p53, and Bax (BCL2‐associated X protein) protein were induced by I/R injury, anti‐apoptotic factor Bcl‐2 was downregulated in these same samples, however, the altered regulation of these genes was markedly attenuated in brain tissues from EPZ5676‐treated mice (Figure 1F). There is now strong evidence that inflammatory processes may contribute to secondary brain damage after ischemic stroke. ²⁰ Consistently, EPZ5676 also decreased the expression of proinflammatory cytokines including VCAM‐1 (Vascular Cell Adhesion Molecule‐1), COX‐2 (Cyclooxygenase‐2), iNOS (Inductible Nitric Oxide Synthase) in the MCAO mice (Figure 1F). Breakdown of the blood‐brain barrier has been reported in stroke pathology. The decrease of tight junction proteins ZO‐1 (Zonula Occluden‐1) and Claudin‐1 in the penumbra of MCAO was reversed by EPZ5676 (Figure 1F). Double immunofluorescence staining with neuron marker Map‐2 further showed EPZ5676 inhibited pro‐apoptosis and inflammation protein level in neurons after MCAO (Figure 1G and Figure S2). Also, the images of double immunofluorescence staining with ZO‐1 and CD31 (an endothelial cell marker) further showed the level of ZO‐1 significantly was increased after treatment with EPZ5676 (Figure 1G). These results suggest that EPZ5676 inhibits the activation of apoptotic initiators and inflammation response, and then reduces infarct volume.

Figure 1. Inhibition of Dot1L (disruptor of telomeric silencing 1‐like) reduces infarct volume and neuronal apoptosis after middle cerebral artery occlusion (MCAO).

A, Experimental scheme. Mice required EPZ5676 treatment for 1 weeks before MCAO for 90 minutes, then brains are harvested 24 hours after MCAO injury. B, The contents of EPZ5676 in brain tissue and blood after MCAO were detected by high performance liquid phase. Data are presented as mean±SD, n=6, ^&&& P<0.001 compared with Sham group. C, TTC (2,3,5‐Triphenyltetrazolium chloride) staining was performed to show the ischemic regions. Data are presented as mean±SD, n=6, ^*** P<0.001 compared with Sham group, ^&&& P<0.001 compared with MCAO group. D, The neurological deficit scores of mice after MCAO. Data are presented as mean±SD, n=6, ^&&& P<0.001 compared with MCAO group. E, The apoptosis of neurons was evaluated by TUNEL. Scale bars: 20 μm. DAPI, 4,6‐diamino‐2‐phenyl indole; Map, Microtubule‐associated Protein; VCAM, Vascular Cell Adhesion Molecule; iNOS, Inductible Nitric Oxide Synthase; PARP, Poly (ADP‐ribose) polymerase. F, Immunoblot analysis of apoptosis‐related proteins (Bax, BCL2‐associated X protein; Bcl‐2, p53, cleaved‐PARP1, cleaved‐Caspase 3, cleaved‐Caspase 9), inflammation markers (iNOS, VCAM‐1, COX‐2), and tight junction‐associated proteins (Claudin‐1, ZO‐1) in brain of mice after MCAO. Data shown are presented as mean±SD from at least 4 independent replicates for 6 to 10/group, ^** P<0.01, ^*** P<0.001 compared with Sham group, ^&&& P<0.001 compared with MCAO group. The data are analyzed and counted through T‐test. COX, Cyclooxygenase; PARP, Poly (ADP‐ribose) polymerase 1; Bcl‐2, B‐cell lymphoma‐2; GAPDH, Glyceraldehyde‐3phosphate dehydrogenase; ZO, Zonula Occluden. G, Immunofluorescence staining of Map‐2 and Caspase 3 or CD31 and ZO‐1 in brain of mice after MCAO. Scale bars: 20 μm. MCAO indicates middle cerebral artery occlusion; and TUNEL, terminal deoxynucleotidal transferase–mediated biotin–deoxyuridine triphosphate nick‐end labeling.

Dot1L Knockdown Mice Are More Resistant to Focal Cerebral Ischemia

To investigate the role of Dot1L on neuroprotection in vivo, wild‐type (WT) and Dot1L^+/‐ mice were subjected to 90‐minutes MCAO followed by 24 hours reperfusion. Initially, Western blot determined that Dot1L levels remained low in the brains of Dot1L^+/‐ mice compared with the WT mice (Figure S3A). 2,3,5‐Triphenyltetrazolium chloride‐staining image and quantitative analysis of brain infarction indicated that knockdown of Dot1L led to smaller infarct size compared with the WT mice (Figure 2A). Likewise, compared with the WT‐MCAO mice, Dot1L^+/‐‐MCAO mice improved functional outcome, as assessed by the Bederson score (Figure 2B). Blood‐brain barrier permeability was evaluated by the Evans blue staining in the brains and found leaked Evans blue in Dot1L^+/‐‐MCAO mice was significantly reduced compared with WT‐MCAO mice (Figure 2C). Meanwhile, the knockdown Dot1L resulted in a significant increase in cerebral blood flow compared with WT at 24 hours reperfusion (Figure 2D). The TUNEL (terminal deoxynucleotidal transferase–mediated biotin–deoxyuridine triphosphate nick‐end labeling) staining showed that lacking Dot1L decreased the number of apoptotic cells in mice brain tissue after MCAO (Figure 2E). No obvious change was observed between WT‐Sham and Dot1L^+/‐‐Sham in the above experiments. These results show knockdown of Dot1L decreased MCAO‐induced brain infarct volume and brain tissue damage and improved behavioral outcomes.

Figure 2. Mice lacking Dot1L (disruptor of telomeric silencing 1‐like) reduces ischemia‐reperfusion‒induced‐brain infarct volume and edema and improved behavioral outcomes.

Wild‐type and Dot1L heterozygous knockout mice were subjected to 90‐minutes middle cerebral artery occlusion followed by 24 hours reperfusion. A, TTC (2,3,5‐Triphenyltetrazolium chloride) staining was performed to show the ischemic regions. B, The neurological deficit scores of mice after middle cerebral artery occlusion. C, Representative images of Evans blue leakage in each group at 24 hours of reperfusion after middle cerebral artery occlusion. D, Relative change in cerebral blood flow was assessed 24 hours of reperfusion after middle cerebral artery occlusion. Data are presented as mean±SD, n=6 to 8/group, ^** P<0.01, ^*** P<0.001. The data are analyzed and counted through ANOVA. E, The apoptosis of neurons was evaluated by TUNEL (terminal deoxynucleotidal transferase–mediated biotin–deoxyuridine triphosphate nick‐end labeling). Scale bars: 20 μm. Dot1L+/‐ indicates Dot1L heterozygous knockout; MCAO, middle cerebral artery occlusion; TUNEL, terminal deoxynucleotidal transferase–mediated biotin–deoxyuridine triphosphate nick‐end labeling; and WT, wild‐type.

Knockdown Dot1L Reduces Apoptosis and Inflammation Response

Considering the neuropathogenesis after reperfusion is generally associated with neural cell apoptosis, we further assessed the role of Dot1L knockdown in regulating neural cell apoptosis. We observed that the level of Dot1L was increased after MCAO (Figure 3A). Furthermore, knockdown Dot1L reversed the upregulation of pro‐apoptotic gene in I/R mice brain tissues, whereas it enhanced the expression of anti‐apoptotic factor Bcl‐2 in these same samples (Figure 3A). Notably, immunofluorescence staining analyses revealed that positive Caspase 3 staining was localized to Map‐2‐positive neuron in the cerebral cortex after MCAO injury, Caspase 3 expression was markedly attenuated in Dot1L^+/‐‐MCAO mice brain tissues relative to those from WT‐MCAO (Figure 3B). Meanwhile, MCAO‐induced upregulated protein levels of iNOS, COX‐2, and VCAM‐1 were inhibited by knockdown Dot1L (Figure 3C). Dot1L^+/‐‐MCAO mice brain tissues section showed a small degree of immunostaining for iNOS compared with WT‐MCAO mice (Figure 3D). Immunofluorescence staining further determined knockdown Dot1L resulted in the reduction of iNOS and COX‐2 expressions in neurons in penumbra after MCAO (Figure S3B). We also investigated the effects of knockdown Dot1L on the expression of ZO‐1 and Claudin‐1 at 24 hours after reperfusion and found knockdown Dot1L could upregulate the expressions of ZO‐1 and Claudin‐1 after ischemic stroke in mice (Figure 3E and 3F). mRNA expressions of p53, Cox‐2, CLDN1, and Tjp1 are consistent with protein expression trends that the altered regulation of these genes was markedly attenuated in Dot1L^+/‐‐MCAO mice brain tissues relative to those from WT‐MCAO mice (Figure S3C). These results demonstrate knockdown of Dot1L reduced apoptosis and inflammation response and upregulated the expression of tight junction proteins following I/R.

Figure 3. Mice lacking Dot1L (disruptor of telomeric silencing 1‐like) reduces apoptosis and inflammation response.

Wild‐type and Dot1L heterozygous knockout mice were subjected to 90‐minutes middle cerebral artery occlusion (MCAO) followed by 24 hours reperfusion. A, Immunoblot analysis of Dot1L and apoptosis‐related proteins in brain of mice after MCAO. B, Immunofluorescence staining of Map‐2 and Caspase 3 in brain of mice after MCAO. Scale bars: 50 μm. C, Immunoblot analysis of inflammation markers in brain of mice after MCAO. D, Immunohistochemistry staining of iNOS in brain of mice after MCAO. Scale bars: 50 μm. E, Immunoblot analysis of TJs‐associated proteins in brain of mice after MCAO. F, Immunofluorescence staining of CD31 and Claudin‐1 in brain of mice after MCAO. Scale bars: 50 μm. All blots data shown are presented as mean±SD from at least 4 independent replicates for 6 to 10/group, ^* P<0.05, ^** P<0.01, ^*** P<0.001. The data are analyzed and counted through ANOVA. Bax, BCL2‐associated X protein; Bcl, B‐cell lymphoma; COX, Cyclooxygenase; DAPI, 4,6‐diamino‐2‐phenyl indole; GAPDH, Glyceraldehyde‐phosphate dehydrogenase; iNOS, Inductible Nitric Oxide Synthase; Map, Microtubule‐associated Protein; PARP, Poly (ADP‐ribose) polymerase; VCAM, Vascular Cell Adhesion Molecule; ZO, Zonula Occluden. Dot1L+/‐ indicates Dot1L heterozygous knockout; MCAO, middle cerebral artery occlusion; TUNEL, terminal deoxynucleotidal transferase–mediated biotin–deoxyuridine triphosphate nick‐end labeling; and WT, wild‐type.

Suppression or Knockdown of Dot1L Reduces Neuronal Apoptosis and Inflammation Response After OGD

We first investigated the effect of EPZ5676 in neuron subjected to 4 hours of OGD and 24 hours reoxygenation. EPZ5676 with a concentration‐dependent manner reversed the upregulation of pro‐apoptotic gene and inflammation modulators induced by OGD/R in neuron, whereas enhanced the expression of anti‐apoptotic factor Bcl‐2 in these same samples (Figure 4A). Immunofluorescence analyses revealed that positive Caspase 3 staining was inhibited by EPZ5676 (Figure 4B). Notably, OGD/R‐induced neuron apoptosis was significantly recovered by the pretreatment with EPZ5676 (Figure 4C). Consistent with these findings, immunofluorescence staining revealed a lack of iNOS and VCAM‐1 in EPZ5676 treated neuron as compared with those in OGD/R‐treated neuron (Figure S4A). Likewise, mRNA expression in apoptosis markers (p53, Bax), and inflammation markers (Nos2, Cox‐2, and Vcam‐1) were consistent with protein expression (Figure S4B). These results provide further support that EPZ5676 protects against cortical neuron OGD/R injury.

Figure 4. Suppression or knockdown of Dot1L (disruptor of telomeric silencing 1‐like) reduces neuronal apoptosis and inflammation response after oxygen‐glucose deprivation (OGD).

Rat neocortical neurons were pretreated with EPZ5676 (10, 20, 40 μmol/L) for 4 hours and subsequently treated with OGD 4 hours and 24 hours of reoxygenation. Immunoblot analysis of apoptosis‐related proteins and inflammation markers in cortical neurons after OGD (A); Immunofluorescence staining of Map‐2 and Caspase 3 in cortical neurons after OGD. Scale bars: 20 μm (B); The apoptosis of neurons was evaluated by TUNEL (terminal deoxynucleotidal transferase–mediated biotin–deoxyuridine triphosphate nick‐end labeling). Scale bars: 50 μm (C). Rat neocortical neurons were transfected with either control siRNA (siNC) or Dot1L siRNA cultured for 48 hours and subsequently treated with OGD followed by reoxygenation, immunoblot analysis of Dot1L and H3K79me3 protein expression (D) and apoptosis and inflammation markers (E); The apoptosis of neurons was evaluated by TUNEL. Scale bars: 20 μm (F). All data are presented as mean±SD of 4 independent experiments, ^*** P<0.001 compared with control or control+siNC, ^&& P<0.01, ^&&& P<0.001 compared with OGD followed by reoxygenation or OGD followed by reoxygenation +siNC. The data are analyzed and counted through ANOVA. OGD/R indicates oxygen‐glucose deprivation followed by reoxygenation; and TUNEL, terminal deoxynucleotidal transferase–mediated biotin–deoxyuridine triphosphate nick‐end labeling. Bax, BCL2‐associated X protein; Bcl, B‐cell lymphoma; COX, Cyclooxygenase; DAPI, 4,6‐diamino‐2‐phenyl indole; GAPDH, Glyceraldehyde‐phosphate dehydrogenase; iNOS, Inductible Nitric Oxide Synthase; PARP, Poly (ADP‐ribose) polymerase; SiNC, negative control siRNA; VCAM, Vascular Cell Adhesion Molecule; ZO, Zonula Occluden.

To further determine the effect of Dot1L on OGD/R‐treated neurons, we used siRNA to knock down Dot1L. When Dot1L was knocked down, cortical neurons subjected to OGD/R exhibited a decrease in H3K79me3 abundance (Figure 4D). In addition, pro‐apoptotic gene and inflammation modulators induced by OGD/R could be inhibited by knockdown Dot1L (Figure 4E). The knockdown of Dot1L also attenuated neuronal death after OGD by TUNEL assay (Figure 4F). Similarly, immunofluorescence assay further confirmed that silencing of Dot1L decreased p53 and COX‐2 expression (Figure S4C). These observations demonstrate that in vitro as well as in vivo, Dot1L contributes to ischemia damage.

Blockade of Dot1L Promotes K63‐Linked Ubiquitination of RIPK1 to Inhibit RIPK1/ Caspase 8‐Dependent Apoptosis Pathway

Receptor‐interacting protein kinase 1 (RIPK1, also known as RIP1) and Caspase 8 are key players in inflammation and cell death, positively and negatively regulating these processes depending on the cellular context. ²¹ In present study, we found OGD/R led to a significant elevation of RIPK1, RIPK3, and cleaved‐Caspase 8 protein levels. Importantly, inhibition or RNAi‐mediated depletion of Dot1L prevented RIPK1, RIPK3 and cleaved‐Caspase 8 protein expression induced by OGD/R in cortical neuronal cells (Figure 5A).

Figure 5. Blockade of Dot1L (disruptor of telomeric silencing 1‐like) promotes K63‐ubiquitylation of RIPK1 (receptor‐interacting protein kinase 1) to inhibit RIPK1/Caspase8‐dependent apoptosis pathway.

A, Rat neocortical neurons were transfected with Dot1L siRNA or pretreated with EPZ5676, and subsequently treated with oxygen‐glucose deprivation followed by reoxygenation (OGD/R), immunoblot analysis of RIPK1, RIPK3, and cleaved‐Caspase 8 expression. HT22 cells were pretreated with EPZ5676 for 4 hours and subsequently treated with OGD/R, immunoblot analysis of RIPK1, RIPK3, and cleaved‐Caspase 8 expression (B) and quantitative real‐time polymerase chain reaction to quantify mRNA level (C); Data are presented as mean±SD of 3 independent experiments, ^*** P<0.001 compared with Control or Control+siNC, ^&&& P<0.001 compared with OGD/R or OGD/R+siNC. D, Immunoblot analysis of RIPK1, RIPK3 and cleaved‐Caspase 8 protein expression in brain of mice after middle cerebral artery occlusion. Data shown are presented as mean±SD from at least 3 independent replicates for 6–10/group, ^* P<0.05, ^*** P<0.001. The data are analyzed and counted through ANOVA. E, Blockade of Dot1L enhances the level of K63 chains on RIPK1 compared with OGD/R‐treated HT22 cells. F, Co‐immunoprecipitation experiments show blockade of Dot1L suppressed the interaction of RIPK1 and Caspase 8. Dot1L+/‐ indicates Dot1L heterozygous knockout; MCAO indicates middle cerebral artery occlusion; OGD/R indicates oxygen‐glucose deprivation followed by reoxygenation; RIPK1, receptor‐interacting protein kinase 1; RIPK3, receptor‐interacting protein kinase 3; WT, wild‐type.

HT22 cell, an immortalized mouse hippocampal progenitor cell, is an excellent in vitro model for unveiling the molecular mechanisms of ischemic insults. Hence, by using the HT22 hippocampal mouse cell line, we further unveiled the role of Dot1L. Consistent with the observation in cortical neuron, OGD/R resulted in the upregulation of pro‐apoptotic gene in HT22 cells, whereas decreased the expression of anti‐apoptotic factors Bcl‐2 in these same samples, which can be inhibited by EPZ5676 (Figure S5A). Similarly, immunofluorescence assay further confirmed EPZ5676 decreased p53 expression (Figure S5B). In addition, we demonstrated Caspase 3 activity was decreased sharply in EPZ5676‐treated HT22 cells (Figure S5C). Likewise, EPZ5676 prevented protein and mRNA expression of RIPK1, RIPK3 and Caspase 8 induced by OGD/R in HT22 cells (Figure 5B and 5C). Further, treatment with EPZ5676 or knockdown of Dot1L prevented MCAO‐induced increase of RIPK1, RIPK3 and cleaved‐Caspase 8 (Figure 5D). Those results show that blockade of Dot1L downregulates RIPK1/Caspase 8‐dependent apoptosis pathway.

K63‐linked ubiquitination of RIPK1 represents a further control point that restricts the lethal potential of RIPK1. In OGD/R‐treated HT22 cells, K63‐ubiquitination of RIPK1 was reduced and its cleavage increased in complex I, reflective of increased Caspase 8 activity and promote apoptosis, on the contrary, EPZ5676 recovered the decreased K63 chains attached to RIPK1 caused by OGD/R (Figure 5E). Additionally, we analyzed RIPK1/Caspase 8 interactions in HT22 cells and found EPZ5676 prevented formation of Caspase 8/RIPK1 complexes (Figure 5F). Those results show that blockade of Dot1L promotes K63‐ubiquitylation of RIPK1, inhibiting RIPK1/Caspase 8‐dependent apoptosis pathway.

Dot1L Transcriptionally Activates CYLD and A20, and Inhibits cIAP1 to Promote Complex IIb

cIAP acts as E3 for K63‐linked ubiquitination‐mediated ubiquitylation of RIPK1, and cIAP1/2‐mediated RIPK1 K63 ubiquitination in complex I is believed to prevent the transition of RIPK1 from complex I to complex II. ²² Recent studies have also indicated the formation of a RIPK1‐dependent apoptosis‐inducing complex (complex IIb), which is formed in the absence of cIAP. ²³ A notion supported by the fact that deubiquitinating enzymes (DUBs) such as CYLD and A20 were suggested to promote complex IIb assembly by dismantling the ubiquitin chains remaining on RIPK1 at complex I. ⁸ , ²⁴ Finally, to further investigate whether the protective role of Dot1L inhibition against RIPK1‐dependent apoptosis was depending on deubiquitinase and ubiquitinase. We found that OGD/R induced an increased deubiquitinase CYLD and A20 in cortical neurons, but that could be inhibited by either Dot1L siRNA or EPZ5676 (Figure 6A). By contrast, silencing or inhibiting of Dot1L could recover the decreased ubiquitinase cIAP1 upon OGD/R injury (Figure 6A). Similarly, EPZ5676 prevented the altered protein expression of CYLD, A20 and cIAP1 induced by OGD/R in HT22 cells (Figure 6A). The immunostaining of cIAP1 in the OGD/R‐treated HT22 cells also showed a decreased expression of cIAP1, which was thus recovered by EPZ5676 treatment (Figure 6B). Similar to the in vitro experiments, ischemia triggered increase of the level of CYLD and A20 and the decrease of cIAP1 in the ischemic cortex after MCAO, nevertheless, the altered regulation of these genes was markedly attenuated in EPZ5676‐treated or Dot1L^+/‐ mice brain tissues relative to those from WT or vehicle‐treated mice (Figure 6C). As Dot1L is the only methyltransferase targeting H3K79, we performed ChIP assays across the CYLD, A20 and cIAP locus using an anti‐H3K79me3 antibody. Remarkably, OGD/R treatment promoted recruitment of H3K79me3 to near the promoter of Cyld (CYLD) and Tnfaip3 (A20) as validated by ChIP‐PCR in HT22 cells, instead, OGD/R treatment decreased H3K79me3 level near the promoter of Birc2 (cIAP1) (Figure 6D). Quantitative real‐time polymerase chain reaction validated the increased Cyld and Tnfaip3 mRNA abundance, also decreased Birc2 transcription level upon OGD/R treatment, by contrary, the altered regulation of these genes was markedly attenuated by treatment with EPZ5676 (Figure 6D and 6E). In summary, these data suggest that Dot1L transcriptionally activates CYLD and A20 and inhibits cIAP1 via its H3K79 methylation dependent function.

Figure 6. Dot1L (disruptor of telomeric silencing 1‐like) transcriptionally activates deubiquitinase cylindromatosis (CYLD) and A20 and inhibits cIAP (cellular inhibitors of apoptosis proteins) to promote complex IIb.

Rat neocortical neurons or HT22 (Hippocampal Neuronal Cell) cells were transfected with Dot1L siRNA or pretreated with EPZ5676, and subsequently treated with oxygen‐glucose deprivation followed by reoxygenation (OGD/R), immunoblot analysis of CYLD, A20 and cIAP1 protein expression. Data are presented as mean±SD of 3 independent experiments; ^*** P<0.001 compared with control or control+siNC, ^&&& P<0.001 compared with OGD/R (Oxygen‐glucose Deprivation/Reperfusion) or OGD/R+siNC (A); Immunofluorescence staining of cIAP1 was assayed. Scale bars: 100 μm (B). C, Immunoblot analysis of CYLD (Cylindromatosis), A20 and cIAP1 (Cellular Inhibitor Of Apoptosis Protein 1) protein expression in brain of wild‐type, Dot1L^+/‐, EPZ5676‐treated mice after middle cerebral artery occlusion. Data shown are presented as mean±SD from at least 3 independent replicates for 6–10/group, ^* P<0.05, ^** P<0.01, ^***P<0.001 compared with Sham or middle cerebral artery occlusion group, ^&&& P<0.001 compared with middle cerebral artery occlusion group. D, ChIP‐PCR (Chromatin immunoprecipitation‐PCR) showing H3K79me3 lack at promoter of Tnfaip3 and Cyld and gain at promoter of Birc2 in EPZ5676‐treated HT22 cells. Data are presented as mean±SD of 3 independent experiments; ^*** P<0.01 compared with Control, ^& P<0.05, ^&& P<0.01, ^&&& P<0.001 compared with OGD/R. E, Quantitative real‐time polymerase chain reaction to quantify mRNA level of Tnfaip3, Cyld and Birc2 after EPZ5676 treatment in OGD/R‐treated HT22 cells. Data are presented as mean±SD of 3 independent experiments, ^*** P<0.001 compared with control, ^&&& P<0.001 compared with OGD/R. The data are analyzed and counted through ANOVA. F, Working model for Histone methyltransferase Dot1L contributes to RIPK1 (receptor‐interacting protein kinase 1)‐dependent apoptosis in cerebral ischemia/reperfusion. Dot1L participates in the regulation of neuronal programmed apoptosis and cerebral infarction through targeting RIPK1/Caspase 8 dependent apoptosis in condition of decreased RIPK1 K63‐ubiquitylation obtained by increased CYLD and A20 also cIAP1 depletion. DAPI, 4,6‐diamino‐2‐phenyl indole; GAPDH, Glyceraldehyde‐phosphate dehydrogenase; MCAO, Middle Cerebral Artery Occlusion. cIAP1 indicates cellular inhibitors of apoptosis proteins; IgG, immunoglobulin G; MCAO, middle cerebral artery occlusion; OGD/R indicates oxygen‐glucose deprivation followed by reoxygenation; RIPK1, receptor‐interacting protein kinase 1; TNF, tumor necrosis factor; and WT, wild‐type.

Discussion

The pathogenesis of neuronal damage in the brain after ischemic stroke is highly complex, involving necrotic, apoptotic and autophagic processes. To suppress neuron apoptosis and necrosis, it is essential to elucidate the underlying molecular mechanisms and discover impactful therapeutic target. Epigenetic processes are involved in stroke pathogenesis and recovery, including the deployment of stress responses that modulate cell viability and promote tissue repair and functional reorganization. ²⁵ , ²⁶ , ²⁷ Dot1L is the only known H3K79 methyltransferase and located at the globular domain of nucleosome. ²⁸ Dot1L is broadly known to play roles in the pathogenesis of heart embryonic growth and cell cycle, DNA damage response. ¹¹ While the functional role of Dot1L in gene expression regulation has been studied extensively, however, not much is known about Dot1L recruitment to specific sites in regulating programmed apoptosis in the cerebral ischemia. Our present study demonstrated that Dot1L participated in the regulation of neuronal programmed apoptosis and cerebral infarction through targeting RIPK1/Caspase 8 dependent apoptosis in condition of decreased RIPK1 K63‐ubiquitylation obtained by increased CYLD (deubiquitinase cylindromatosis) and A20 also cIAP1 depletion (Figure 6F). Importantly, inhibition or knockdown of Dot1L in I/R mice significantly attenuated neurological deficit, decreased brain infarct volume and neuronal apoptosis. Our study has provided important evidence about the function of Dot1L in the programmed apoptosis pathway, a step toward better understanding the Dot1L‐based therapy for I/R injury and stroke.

It is well established that apoptosis contributes to neuronal cells death following I/R injury. This was supported by our results wherein transient cerebral ischemia could significantly induce neuron apoptosis, meanwhile, Dot1L is upregulated after cerebral hypoxic‐ischemic injury, where Dot1L‐mediated apoptosis is assumed to participate. Further studies with a specific inhibitor and knockdown of Dot1L to detect the apoptosis assay are warranted to address this issue. We found that knockdown or inhibition of Dot1L alleviated cell apoptosis in OGD neuronal cells, and significantly attenuated neurological deficit, decreased brain infarct volume and neuronal apoptosis in I/R mice, showing that Dot1L activity contributes to brain I/R injury. It has been demonstrated that an inflammatory response occurs centrally after ischemic injury. ²⁹ Our results also confirmed that ischemic injury induced the increase of inflammation mediators in the brain. Interestingly, Dot1L knockdown or inhibition could simultaneously inhibit inflammatory response in the brain following ischemic stroke and showing blockade of Dot1L may serve as a novel therapeutic strategy for ischemic brain injury.

Covalent conjugation of ubiquitin molecules to target proteins, termed ubiquitination, regulates protein stability and protein‐protein interactions in signaling complexes. ³⁰ In particular, K48‐linked poly‐ubiquitin chains generally target proteins for proteasomal degradation, whereas linear and K63‐linked chains primarily serve as a scaffolding network for the formation of signaling complexes. ³¹ , ³² The status of RIPK1 K63‐ubiquitination determines cell fate by creating either a prosurvival signal (ubiquitinated RIPK1) or a death signal (deubiquitinated RIPK1). ³³ When RIPK1 is deubiquitinated, RIPK1 forms a complex with TRADD, FADD, and Caspase 8 to activate Caspase 8 and subsequently trigger apoptosis. ⁵ This was supported by our results wherein activation of RIPK1 kinase was almost undetectable in sham‐operated animals. Transient cerebral ischemia promotes the K63 deubiquitination of RIPK1, and nonubiquitinated RIPK1 binds with Caspase 8 to form complex II and promotes the death receptor pathway of apoptosis. Interestingly, RIPK1/Caspase 8 complex formation is inhibited in Dot1L depletion OGD cells, suggesting that Dot1L might promote RIPK1/Caspase 8 complex formation and activates the apoptotic pathway. However, in dissecting the role of Dot1L in the cell death pathway, we were not able to exclude the role of Dot1L in necroptosis.

It is clear that RIPK1‐mediated apoptosis can occur under conditions of IAP (Inhibitor Of Apoptosis Protein) inhibition. ³⁴ , ³⁵ The overexpression of IAPs can inhibit cell death through ubiquitination of RIPK1. We found K63 ubiquitination level of RIPK1 was reduced upon OGD, in parallel, the expressions of deubiquitinases CYLD and A20 were increased, whilst cIAP1 expression was decreased, which leads to the cleavage and activation of Caspase 8 and the induction of apoptosis. More importantly, an apparent decrease of H3K79me3 signal was shown on promotors of Bric2 (cIAP1) gene in OGD‐induced neuronal cells and consequently inhibiting its transcription. Meanwhile, Dot1L initiated H3K79me3 enrichment in the promoter of Cyld and Tnfaip3 (A20) and consequently activated transcription, consistent with Dot1L’s methyltransferase activity on H3K79. Taken together, blockade of Dot1L inhibited K63 de‐ubiquitination of RIPK1 by downregulated expressions of CYLD and A20, and upregulated cIAP1, thus prevented RIPK1/Caspase 8 complex formation and inhibited apoptotic pathway.

In conclusion, our study was the first to demonstrate that Dot1L was increased in the brain after ischemia, which then activated the RIPK1/Caspase 8‐dependent apoptosis signaling pathway to cause neuronal death. EPZ5676 treatment was able to protect against ischemic neuronal death by inhibiting RIPK1‐mediated apoptosis in mice brains following ischemic stroke. Because of the specific activation of Dot1L in cerebral ischemia, we suggest further studies to explore its feasibility as a potential therapeutic target for neuroprotection.

Sources of Funding

This work was supported by grants from the National Natural Science Foundation of China (No. 81872861); project supported by Shanghai Municipal Science and Technology Major Project (Grant No. 2017SHZDZX01).

Disclosures

None.

Supporting information

Table S1

Figures S1–S5

References36, 37, 38, 39, 40, 41, 42, 43, 44