Mechanism Of Necroptosis In Glial Cells And Protective Study After Spinal Cord Injury

One unique pathological change after primarily spinal cord injury(SCI) is the secondary injury, which is characterized by the formation of a spinal cavity. The cavity induced by primary insult was always surrounded by reactive a strogliosis and chronic inflammation. How reactive ast rocytes and microglia/macrophages, the major cell component of chronic inflammation, are eliminated during cavitation remain poorly investigated.Apoptosis of neurons was thought to be the main cause of the cavity formation. There are two peaks of spatial-temporal distribution of the neuronal apoptosis after SCI. One peak is at the 24 hours after SCI and the apoptotic cells locate at the epicenter, and the other is at 7 days after injury and the apoptotic cells distribute around the lesion center. It has been reported that AMPA receptor mediated apoptosis of oligodendrocytes also contributes to enlargement of spinal cavity. However, inhibiting apoptosis after SCI failed to prevent the expansion of spinal cavity. In our pilot experiments, there was no significant difference of the cavity size and functional recovery between the caspase-3-/- and wild type mice. These indicate that other types of cell death may be involved in the enlargement of the spinal cavity.There are three types of cell death, necrosis, apoptosis and autophagy, and the latter two has been studied in depth. Compared to the apoptosis, the autophagy of neurons and oligodendrocytes oc curs after SCI w ith f ewer nu mbers. H owever, the ne crosis w as considered as an acute and uncontrollable process, and it is characterized by the loss of integrity of c ell m embrane, organelle swelling, mitochondrial dysfunction, w ithout t he characteristic of pyknos is of nucleus in apoptosis or formation of a utophagosome in autophagy. Necrosis was deemed to oc cur i rreversibly at the acute phase of SCI and the mechanism of necrosis was poorly studied.Necroptosis, a type of programmed necrosis, with its signaling pathway mediated by receptor-interacting pr otein 1/3(RIP1/3) a nd m ixed 1 ineage ki nase dom ain-like pr otein (MLKL), has been recently unraveled. At present, the TNFa-induced necroptossi has been studied extensively. U pon activation of receptor of TNFa, t he dow nstream m olecules activate R IP1, t he 1 atter phos phorylates R IP3 a nd t hey f orm t he necrosome. T he necrosome then phosphorylates M LKL, and pM LKL f unctions in penetrating cell membrane, w hich mediates Ca2+ influx to i nduce necroptosis. T he e xpression and phosphorylation of RIP3 has been regarded as the key events during necroptosis execution. It h as be en r eported t hat ne croptosis can occur in d ifferent t ypes of c ells in t he pathological c onditions of he art, ki dney a nd 1 iver, a nd it plays a n i mportant r ole i n pancreatitis, tuberculosis and atherosclerosis. Inhibiting the activity of RIP1/3 effectively alleviates necroptosis, bringing hope for the above necroptosis-related diseases. Mounting researches ha ve focused on the molecular m echanism of ne croptosis a nd its relative protective measures in the field of cell death.Whether necroptosis occurs during the expansion of cavity and its mechanism after SCI ha s ne ver be en el ucidated. Previous s tudies h ave pa id attention to t he glutamate-induced a poptosis of ne urons a nd oligodendrocytes w ithin a nd a round t he lesion center. However, how reactive astrocytes and microglia/macrophages, which are the major component of the glial scar, play diverse roles in SCI and are particularly important in affecting neuronal survival, are eliminated remains poorly investigated. Understanding the m echanism o f glial death pos t-SCI ma y yield ne w i nsights into understanding t he mechanism of s econdary S CI and i mproving functional r ecovery. The present research was aimed to study the mechanism of necroptosis in glial cells and protective measures after SCI. This study is composed of the following six parts.Experiment 1:The death of reactive astrocytes after SCI in miceIn this part of experiment, we studied the expansion of the cavity and the death of cavity-surrounding r eactive a strocytes through a crush model of SCI in m ice. Immunohistostaining of GFAP was performed at 1 day,5 day,7 day,14 day and 28 day after SCI to observe the expansion of the cavity. We first te sted whether apoptosis or autophagy accounted f or t he 1 oss of r eactive a strocytes dur ing t he de velopment of secondary i njury. Double-immunostainings of GFAP/cleaved caspase-3, GFAP/TUNEL, GFAP/LC3, G FAP/Beclinl and G FAP/Lamp2a were pe rformed. Because immunohistochemistry of GFAP s tains p rimarily processes and seldom cell bodi es, we adopted GFAP-CreER:ROSA-YFP mice to visualize the cell bodies of astrocytes. Very rare could YFP/TUNEL-double positive cells be found at 5 day post injury, precluding the apoptosis of astrocytes. Double-immunostaining of YFP with autophagic markers showed that in bilateral areas from 400μm rostral and caudal to border of lesion center defined by YFP/GFAP-immunoreactivity, approximately 9.8% of Y FP-positive c ells were Lamp2a-positive,6.3% were LC3-positive and 8.1% were Beclinl-positive. These data suggested that i n m ice, during t he proc ess o f cavit ation, r eactive as trocytes m ay be eliminated by w ays ot her t han a poptosis a nd a utophagy. To test whether re active astrocytes could undergo necrosis, the third major type of cell death, we performed in vivo propidium iodide (PI)-labeling. From 3 to 14 dpi, about 80% of a 11 PI-labeled cells in these regions were GFAP-positive. These data indicate that reactive astrocytes around the spinal cavity undergo necrosis during the process of secondary injury.Experiment 2:Necroptosis of reactive astrocytes after SCITo test w hether n ecroptosis oc curs in the r eactive ast rocytes after S CI, we first examined the expression of RIP3 in mice by Western-blotting. The expression of RIP3 increased s ignificantly from 3 dpi t o 7 dpi. D ouble-immunostainings of GFAP/RIP3, GFAP/MLKL a nd GFAP/HMGB1 we re pe rformed. Immunohistochemistry de tected strong R IP3-immunoreactivity a round 1 esion c enter, w ith most expressed by GFAP-positive c ells. Q uantification s howed t hat a t all time poi nts examined, approximately 80% of the RIP3-positive cells were GFAP-positive, w hich was consist with the re sults of P I-labeling OX42-positive, N euN-positive and C C-1-positive c ells constituted for t he r emainder of R IP3-positive c ells. M LKL, a nother ke y m olecule i n execution of necroptosis, was also induced by SCI and expressed in reactive astrocytes. In addition, HMGB1, a member of high mobility group box p rotein that normally binds to chromatin a nd i s released by n ecrotic c ells, was de tected in the cytop lasm o f GFAP-positive cells, supporting the occurrence of necrosis in astrocytes. To confirm the necroptosis of astrocytes after SCI, we performed immune-electron microscopic study in spinal tissue within 5 mm around lesion center at 5 dpi in mice. Astrocytes with focal lysis of cytoplasm and RIP3 immunoreactivity on the cytoplasmic fibrils were frequently found. MLKL-immunoreactivity was found both at the cell membrane and within the bundles of cytoplasmic f ibrils, c onsistent with its role in penetrating the cell m embrane during necroptosis. N o apoptosis-like astrocytes w ere obs erved. T aken together, these da ta suggested that reactive astrocytes may undergo RIP3/MLKL-mediated necroptosis after SCI in mice.We then examined w hether ne croptosis c ould b e modeled i n r eactive a strocytes in vitro. The cells were challenged with tumor necrosis factor alpha (TNFa) and lipopolysaccharide (L PS) to mimic the inflammatory microenvironment in vivo, a nd z-VAD, a p an-caspase i nhibitor w as a dded t o i nhibit a poptosis a s r outinely us ed by researchers when i nducing ne croptosis. F orty-eight hours treatment of TNFa, LPS and z-VAD (TLZ) significantly i ncreased the expression of RIP3, MLKL, and cytoplasmic HMGB1, while de creased nuc leus 1 evels of HMGB1.U pon T LZ t reatment, t he intracellular 1 evel of r eactive oxy gen species (ROS) an d t he pe rcent of P I-labeled astrocytes were si gnificantly increased, and i ntracellular 1 evel of A TP si gnificantly decreased. These data indicate that necroptosis of astrocytes can be induced by TLZ in vitro. Further, Necrostatin-1, aw ell-used ne croptosis inhibitor tha tin hibits phosphorylation of R IP1, s ignificantly c ompromised the i ncrease o f R IP3, M LKL, cytoplasmic HMGB1, intracellular ROS, extracellular ATP a nd PI-labeling i nduced by TLZ treatment, indicating a rescuing effect of Nec-1 on a strocyte death. Because Nec-1 also i nhibits i ndoleamine-2,3-dioxygenase (IDO), wet hen us ed R IP3-/- astrocytes t o confirm the occurrence of astrocytic necroptosis. The effects of TLZ on ROS production, ATP1 evel a nd P I-permeability in wild-type ast rocytes w ere si gnificantly a bolished i n RIP3-/- astrocytes. These data demonstrated the necroptosis of mouse astrocytes in vitro.We ne xt ex amined the effects of Nec-1 and RIP3 knockout on necroptosis o f astrocytes after SCI. Both five successive days of treatment by Nec-1 and RIP3 knockout significantly de creased the pe rcent of P I-positive ast rocytes, the e xpression of RIP3, MLKL and HMGB1 in areas surrounding lesion center. Moreover, increased numbers of neuronal survival and improved locomotion function were found and we then investigated the possible mechanism. Because astrocytes are well known for their supportive roles in neuronal s urvival a nd ne urons a round t he 1 esion c enter d ie a fter S CI, w e t hen t ested whether necroptosis could affect neurotrophic function of reactive astrocytes. Conditioned medium (CM) of normal astrocytes, TLZ treated necroptotic astrocytes, and necroptosis inhibited a strocytes w hich w as t reated by T LZ pl us N ec-1 (TLZN) w ere us ed to t reat primary neurons. More TUNEL-positive neurons were observed in cells treated by CM of necroptotic astrocytes i n comparison with those in c ells tre ated by C M of nor mal astrocytes. A significantly less TUNEL-positive neurons were observed in cells treated by CM of TLZN treated astrocytes, as compared to that of TLZ treated astrocytes. Further, we examined t he e ffects of ne croptosis inhibition on t he expression of glial c ell line-derived neurotrophic factor (GDNF) by reactive astrocytes after SCI. Both Nec-1 and RIP3 de pletion s ignificantly enhanced the expression of GDN F b y reactive ast rocytes. Taken together, these data suggest that inhibiting necroptosis after SCI not only attenuates astrocyte death but also rescues the neurotrophic function of reactive astrocytes, thereby promoting adjacent neuronal survival and functional recovery.Experiment 3:Mechanism of necroptosis of reactive astrocytes after SCIAs i nflammatory factors can induce ast rocytic ne croptosis in vitro and M 1 microglia/macrophages has be en thought to be the major source of t oxic i nflammatory factors after SCI, w e hypot hesized t hat M 1 microglia/macrophages m ight i nduce astrocytic necroptosis after SCI. Western-blotting showed that the expression of inducible nitric oxide synthase (iNOS), a commonly used marker for M1 microglia/macrophages, increased quickly from 1 day and peaked at 3 day post-SCI. The expression of MLKL and HMGB1 increased gradually and peaked at 5 dpi, just following the increase of iNOS. To directly investigate the effects of M1 microglia/macrophages on astrocytes, we polarized primarily cultured m icroglia/macrophages t oward e ither M1 or M2 phe notype, a nd stimulated astrocytes with conditioned medium c ollected f rom M 0 (normal c ultured microglia/macrophages), M1, and M2 microglia/macrophages. All three CMs increased the expression of RIP3, MLKL and HMGB1 in cultured astrocytes. Of note, CM of M1 microglia/macrophages (M1 C M) s howed t he s trongest i nduction of t hese ne croptosis markers. I n a ddition, i ntracellular ATP w as s ignificantly decreased, a nd t he num ber of PI-labeled cells increased in astrocytes treated by M1 CM.To investigate the death-inducing effects of M1 microglia/macrophages on astrocytes in vivo, we depleted M1 microglia by administration of gadolinium chloride (GdCl3) in the lesion si te as de scribed, w hich i nduces a poptosis of i nflammatory m acrophages vi a competitive inhi bition of C a2+ mobilization and damage t o plasma m embranes. In comparison with P BS c ontrol, G dCl3 treatment significantly r educed t he expression of RIP3, MLKL a nd HMGB1, a nd de creased the pe rcent o f P I-labeled as trocytes w ithin bilateral regions 400μm rostral and caudal to the epicenter. Consistently, smaller spinal cavity at 14 dpi a nd be tter 1 ocomotion r ecovery from 6 d pi w ere obs erved i n GdC13 treated mice. To further c onfirm th e e ffects of M1 microglia/macrophages on a strocyte death a nd s pinal c avity, w e pr epared pr imary macrophages f rom bo ne marrow, a nd transplanted MO and Ml macrophages into the injured spinal cord. Transplantation of M1 macrophages resulted in a dramatic increase of RIP3, MLKL and HMGB1 expression and the num ber of PI-labeled astrocytes, while t ransplantation of MO macrophages o nly showed minor e ffects, a s c ompared t o D ulbecco’s M odified Eagle’s Medium(DMEM) control. Likewise, the spinal cavity was significantly larger at 14 dpi, and the locomotion recovery w orse from 5 dpi in M1 macrophage-transplanted mice. Taken together, t hese data indicate that M1 micoglia/macrophages may play an inductive role in the necroptosis of astrocytes after SCI.Experiment 4:TLR/MyD88 pathway partially mediates necroptosis of reactive astrocytes after SCIAs shown above, LPS was required for the in vitro induction of astrocytic necroptosis. It ha s be en r eported that toll-like receptors (T LRs) are i nvolved i n necroptosis of macrophages and contributes to the activation of astrocytes. We thus speculated that TLR and t heir dow nstream molecules m ight be i nvolved i n t he M 1 microglia/macrophages-induced astrocytic necroptosis. Among TLRs members, TLR2 and TLR4 respond to same type of stimulus and have been thought to be involved in astrocyte activation. D ouble-staining of G FAP w ith T LR2 or T LR4 s howed t hat only am inor portion of reactive astrocytes surrounding spinal cavity express TLR2, while most reactive astrocytes express TLR4. We then focused on the expression of TLR4 and the intracellular adaptor molecule MyD88. The immunoreactivity of TLR4 overlapped very well with that of RIP3 in the injured mouse spinal cord, suggesting that TLR4/MyD88 signaling was activated in necroptotic astrocytes after SCI. In vitro, TLZ treatment stimulated a dramatic increase of TLR4 and MyD88 in astrocytes. A TAT-tagged MyD88 inhibitory p eptide which could block the homodimerization of MyD88 was added into culture medium of astrocytes t o a ntagonize t he f unction of M yD88. T he MyD88 i nhibitory pe ptide significantly suppr essed t he i ncrease of RIP3 and PI-labeling i n a strocytes i nduced by both TLZ and M1-CM treatment, suggesting that TLR/MyD88 may b e required for the activation of necroptotic signaling in astrocytes.We then examined the effects of M 1 m icroglia/macrophages on the expression of TLR4 and MyD88 in astrocytes after SCI. In vivo, GdCl3 treatment significantly decreased the percent of astrocytes expressing TLR4 or MyD88 by approximately 39% and 47.4%, respectively. In contrast, tra nsplantation of M1 macrophages si gnificantly increased the number of a strocytes expressing T LR4 a nd M yD88, c ompared t o t he D MEM c ontrol. These data indicate t hat i n m ice, TLR/MyD88 s ignaling may be i nvolved i n t he M 1 microglia/macrophage-induced necroptosis of astrocytes.Experiment 5:Mechanism of necroptosis in astrocytes of injured human spinal cordThe above expe riments i n mice i ndicated that after S CI, reactive as trocytes died through necroptosis, which was induced by M1 microglia/macrophages, partially through TLR/MyD88 m ediated s ignaling. We ne xt t ested w hether s imilar pa thological c hanges could occur in human after SCI. Immunohistochemistry showed a very weak expression of RIP3 and MyD88, mainly by a strocytes in normal human spinal cord. No expression of MLKL, HMGB1 and TLR4 was detected in normal human spinal cord (data not shown). However, a fter S CI, s trong R IP3-, MLKL-, phosphorylated-MLKL-(pMLKL) a nd HMGB1-immunoreactivities w ere detected in G FAP-positive c ells a round t he 1 esion center in one patient at 5 dpi. Interestingly, TLR4-and MyD88-immunoreactivities were also detected in RIP3-positive or GFAP-positive cells. Similar expression patterns of RIP3, pMLKL, HMGB1, TLR4 and MyD88 were observed in another patient at 15 dp i. These results sugg ested that in hum an, r eactive a strocytes m ay a lso unde rgo TLR/MyD88 mediated necroptosis after SCI.Experiment 6:Microglia/macrophages undergo endoplasmic reticulum(ER) stress involved necroptosis after SCIIn this part of experiment, we investigated the necroptosis of microglia/macrophages after S CI, and the i possible m echanism. Double-immunostainings of Iba-1/MLKL, GFAP/MLKL, NeuN/MLKL and CC1/MLKL were performed to observe the expression of M LKL i n di fferent t ypes of c ells. A pproximately 57% of t he M LKL po sitive c ells around lesion area are Iba-1 positive. In vivo PI-labeling a nd Necrostatin-1 treatment confirmed t he ne croptosis of microglia/macrophages. Furthermore, triple-staining of MLKL wi th Iba-1 a nd CXCR4, a c hemokine r eceptor t hat i s hi ghly expressed on t he membrane of microglia/macrophages, revealed that ML KL-immunoreactivity was distributed both i n t he c ytoplasm a nd on t he membrane of I ba-1-positive c ells. In the electron-microscopic study, M LKL i mmunoreactivity was f ound on c ell membrane. Surprisingly, immunogold particles were also observed on endoplasmic reticulum(ER). Glucose-regulated protein 78 (GRP78), an ER stress sensor, was up-regulated in MLKL positive microglia/macrophages after SCI, suggesting a possible link between necroptosis and ER stress. In vitro, oxygen-glucose deprivation (OGD) stress induced ER stress and necroptosis in microglia. Inhibiting ER stress by 4-phenylbutyrate (4-PBA) significantly blocked t he O GD-induced ne croptosis of microglia. In the end, our data showed that, GRP78 and phos phorylated MLKL were co-expressed by the microglia/macrophages in the i njured hu man s pinal c ord. Taken together, t hese r esults sugge sted t hat microglia/macrophages unde rgo a n E R-stress i nvolved ne croptosis after S CI, implying that ER st ress and n ecroptosis c ould be m anipulated for m odulating i nflammation post-SCI.