Cerebral ischemia/stroke and small ubiquitin-like modifier (SUMO) conjugation – a new target for therapeutic intervention?

Authors

  • Wei Yang,

    1. Multidisciplinary Neuroprotection Laboratories, Department of Anesthesiology, Duke University Medical Center, Durham, North Carolina, USA
    Search for more papers by this author
  • Huaxin Sheng,

    1. Multidisciplinary Neuroprotection Laboratories, Department of Anesthesiology, Duke University Medical Center, Durham, North Carolina, USA
    Search for more papers by this author
  • H. Mayumi Homi,

    1. Multidisciplinary Neuroprotection Laboratories, Department of Anesthesiology, Duke University Medical Center, Durham, North Carolina, USA
    Search for more papers by this author
  • David S. Warner,

    1. Multidisciplinary Neuroprotection Laboratories, Department of Anesthesiology, Duke University Medical Center, Durham, North Carolina, USA
    2. Department of Surgery, Duke University Medical Center, Durham, North Carolina, USA
    3. Department of Neurobiology, Duke University Medical Center, Durham, North Carolina, USA
    Search for more papers by this author
  • Wulf Paschen

    1. Multidisciplinary Neuroprotection Laboratories, Department of Anesthesiology, Duke University Medical Center, Durham, North Carolina, USA
    2. Department of Neurobiology, Duke University Medical Center, Durham, North Carolina, USA
    Search for more papers by this author

Address correspondence and reprint requests to Professor Wulf Paschen, Multidisciplinary Neuroprotection Laboratories, Department of Anesthesiology, Duke University Medical Center, 130 Sands Building, Research Drive, Durham, NC 27710, USA. E-mail: wulf.paschen@duke.edu

Abstract

Transient cerebral ischemia/stroke activates various post-translational protein modifications such as phosphorylation and ubiquitin conjugation that are believed to play a major role in the pathological process triggered by an interruption of blood supply and culminating in cell death. A new system of post-translational protein modification has been identified, termed as small ubiquitin-like modifier (SUMO) conjugation. Like ubiquitin, SUMO is conjugated to the lysine residue of target proteins in a complex process. This review summarizes observations from recent experiments focusing on the effect of cerebral ischemia on SUMO conjugation. Transient global and focal cerebral ischemia both induced a rapid, dramatic and long-lasting rise in levels of SUMO2/3 conjugation. After transient focal cerebral ischemia, SUMO conjugation was particularly prominent in neurons located at the border of the ischemic territory where SUMO-conjugated proteins translocated to the nucleus. Many SUMO conjugation target proteins are transcription factors and sumoylation has been shown to have a major impact on the activity, stability, and cellular localization of target proteins. The rise in levels of SUMO-conjugated proteins is therefore likely to have a major effect on the fate of post-ischemic neurons. The sumoylation process could provide an exciting new target for therapeutic intervention.

Abbreviations used
APP

amyloid precursor protein

eIF2

eukaryotic initiation factor

ER

endoplasmic reticulum

HIF1α

hypoxia-inducible factor-1α

HSF1

heat-shock factor 1

IKK

IκB kinase

MAPK

mitogen-activated protein kinase

MCA

middle cerebral artery

NEMO

NFκB essential modulator

NFκB

nuclear factor κB

SENP

SUMO-specific proteases

SUMO

small ubiquitin-like modifier

When cells are exposed to a severe form of metabolic, thermal, physical, or toxic stress, various stress responses are activated that help cells to withstand the stressful conditions and to restore cell functions. Transient cerebral ischemia is a severe form of metabolic stress that interferes with many cellular pathways and activates various stress responses (Abe et al. 1995; Lipton 1996; Tymianski and Tator 1996; Kristian and Siesjo 1998; Paschen and Doutheil 1999; Fiskum 2000; Lewen et al. 2000; Paschen 2003a; Giffard et al. 2004; Koh et al. 2005; Ogawa et al. 2007; Paschen et al. 2007). These stress responses differ considerably with respect to underlying mechanisms and onset of action. For example, when brains are exposed to a short, non-lethal period of vascular occlusion that activates a state of tolerance, protein synthesis is required to establish the state of tolerance induced by that pre-conditioning. It takes up to 24 h for exposure to the pre-conditioning stimulus to be fully effective in protecting cells from the damage that would be induced by an otherwise lethal duration of ischemia (Dirnagl et al. 2003). On the other hand, post-translational protein modifications such as phosphorylation and ubiquitination are activated during early reperfusion after ischemia (Althausen et al. 2001; DeGracia and Montie 2004; DeGracia and Hu 2007; Paschen et al. 2007) and both processes are believed to be potentially protective. Phosphorylation of the alpha subunit of the eukaryotic initiation factor 2 (eIF2α), a key protein in the regulation of the initiation step of proteins synthesis (Schneider 2000), is activated during the first minutes of reperfusion (Burda et al. 1994; DeGracia et al. 1996; Althausen et al. 2001). Phosphorylation of eIF2α is brought about by double-stranded RNA-activated protein kinase-like endoplasmic reticulum kinase (Kumar et al. 2001, 2003), a protein kinase specifically activated under conditions associated with endoplasmic reticulum (ER) stress that impairs the ER-resident folding process (Harding et al. 1999). This supports the notion that transient ischemia impairs ER function (Paschen and Doutheil 1999; Paschen and Frandsen 2001). Besides these ischemia-induced changes in the pattern of phosphorylation of proteins playing a key function in the regulation of protein synthesis, various other phosphorylation/dephosphorylation processes have been associated with neuronal cell death induced by transient cerebral ischemia or protection of cells from ischemia-induced cell death. These includes kinases implicated in both ischemic brain pathology and in neuroprotection pathways, such as protein kinases A, B, and C, the Janus kinase/signal transducer and activator of transcription pathway, the p38 mitogen-activated protein kinase (MAPK), the extracellular signal regulated protein kinases, the c-Jun N-terminal kinases, and the phosphoinositide-3-kinase/Akt pathway (Bronstein et al. 1993; Domanska-Janik 1996; Barone et al. 2001; Nozaki et al. 2001; Irving and Bamford 2002; Chu et al. 2004; Bonny et al. 2005; Bright and Mochly-Rosen 2005; Kaminska 2005; Perez-Pinzon et al. 2005; Planas et al. 2006; Zhao et al. 2006).

Ubiquitin conjugation is a post-translational protein modification that targets proteins with a short half-life or mis- or unfolded proteins for degradation at the proteasome. Protein ubiquitination is dramatically activated after both transient global and focal cerebral ischemia (Hu et al. 2000, 2001). Like eIF2α phosphorylation, protein ubiquitin conjugation is a post-translational protein modification that is activated during early reperfusion. It is believed that impairment of proteasome function, preventing vulnerable neurons from degrading ubiquitinated proteins accumulated after transient ischemia, results in formation of protein aggregates (Ge et al. 2007). These sequester initiation factors and other components of the protein synthesis machinery, thus preventing recovery of protein synthesis (DeGracia and Hu 2007). Thus, a potentially protective stress response could eventually turn into a toxic process when the downstream reactions required to execute the entire response are blocked.

A prominent post-translational protein modification that has only recently been investigated in experimental models of cerebral ischemia is modification by the small ubiquitin-like modifier (SUMO). The SUMO family consists of four proteins, denoted as SUMO1–4. While ubiquitination targets proteins for degradation at the proteasome, SUMO conjugation modifies the interaction of target proteins with protein partners, and their subcellular localization, activity, and stability (Pichler and Melchior 2002; Gill 2003; Seeler and Dejean 2003; Hay 2005, 2007; Mukhopadhyay and Dasso 2007; Zhao 2007). It has been shown that even a moderate transient SUMO conjugation of the target protein can have a long-lasting effect (Hay 2005). As discussed below, the dramatic rise in levels of SUMO-conjugated proteins induced by transient cerebral ischemia is therefore likely to have a major effect on the fate of cells exposed to a transient interruption of blood supply. In this review, we summarize recent observations on the effects of permanent and transient cerebral ischemia on SUMO conjugation and discuss the possible consequences for the affected cells.

The SUMO conjugation pathway

Small ubiquitin-like modifier is a new group of proteins identified about a decade ago (Matunis et al. 1996; Mahajan et al. 1997). Like ubiquitin, SUMO is bound to the epsilon amino group of lysine residues of target proteins in a complex process involving activating, conjugating, and ligating enzymes (Desterro et al. 1997, 1999; Johnson and Blobel 1997; Johnson et al. 1997; Gong et al. 1999; Melchior et al. 2003; Hay 2005), as illustrated in Fig. 1. In the first step, SUMO is activated in an ATP-dependent reaction by thioester bond formation with the activating enzyme E1. This could become a critical step during ischemia when the energy state of cells is impaired. SUMO is then transferred to the conjugating enzyme (E2). The final step is formation of an isopropyl bond between the C-terminal glycine of SUMO and the ε-amino group of lysine in the acceptor protein, catalyzed by a SUMO ligase (E3). While several SUMO ligases have been identified that are specific to the respective target proteins, Ubc9 is believed to be the only SUMO conjugating enzyme. Ubc9 is therefore an ideal target to activate or block SUMO conjugation. Cells or animals over-expressing Ubc9 or a dominant negative form of Ubc9 have been used to investigate the possible role of SUMO conjugation under certain experimental conditions (Giorgino et al. 2000; Long and Griffith 2000; Mo et al. 2005; Nowak and Hammerschmidt 2006; Lee et al. 2007).

Figure 1.

 Scheme of the SUMO conjugation pathway. SUMO is synthesized as a precursor that is processed by SUMO-specific proteases (SENPs) to expose the two C-terminal glycine residues required for conjugation. Mature SUMO is activated in an ATP dependent reaction by thioester bond formation with the activating enzyme SAE (E1). SUMO is then transferred to the conjugating enzyme Ubc9 (E2). The final step is formation of an isopropyl bond between the C-terminal glycine of SUMO and the ε-amino group of lysine in the target protein. This modifies the subcellular localization, activity and stability of the target protein. SENPs cleave the isopropyl bonds between SUMO and its target protein resulting in free SUMO and un-modified protein.

Like ubiquitin, SUMO is synthesized as a larger precursor that is processed by SUMO-specific proteases (SENPs) to expose the two C-terminal glycine residues that are required for conjugation (Fig. 1). SUMO precursor processing and deconjugation of SUMO from sumoylated proteins are carried out by the same SENP. SENPs have distinct subnuclear and subcellular localization patterns and specific preferences for particular SUMO paralogues (Bailey and O’Hare 2004; Gong and Yeh 2006; Mukhopadhyay and Dasso 2007). In contrast to SUMO1, SUMO2 and SUMO3 both contain a consensus SUMO conjugation site and can therefore form polymeric chains (Vertegaal 2007), which attach to target proteins. The ability of SUMO2 and SUMO3 to form polymeric chains could be of key significance in pathological states of the brain. SUMO3 conjugation has been shown to modulate processing of the amyloid precursor protein (APP) to form Aβ. Generation of Aβ from APP is markedly reduced in SUMO3 over-expressing cells and is considerably activated in cells transfected with a K11R mutated SUMO3 construct that cannot produce SUMO3 polymeric chains (Li et al. 2003). These results have, however, been questioned in a recent study (Dorval et al. 2007) and it has been demonstrated that the modulation of Aβ generation by SUMO does not require conjugation to target proteins.

For many years, SUMO conjugation of proteins was believed to be a predominantly nuclear process, modifying transcription factors, nuclear pore proteins, and other nuclear proteins critical for genome integrity (Johnson 2004; Muller et al. 2004; Moschos and Mo 2006; Heun 2007). SUMO conjugation-induced modification of transcription factors could have a long-lasting effect on gene expression, because a transient sumoylation of transcription factors has been shown to alter their long-term fate (Hay 2005). It has been hypothesized that after SUMO conjugation the transcription factor is incorporated into a repression complex and will still be contained in the repression complex after SUMO deconjugation. While most of the investigated transcription factors are repressed by SUMO conjugation, hypoxia-inducible factor-1α (HIF1α), and heat-shock factor 1 (HSF1) are activated by SUMO conjugation (Hietakangas et al. 2003; Bae et al. 2004; Carbia-Nagashima et al. 2007). The exact role of SUMO conjugation in HIF1α activation is however still a matter of discussion (Cheng et al. 2007). Sumoylation can result in both activation and blocking of activation of nuclear factor κB (NFκB), depending on the conjugation site. Lysine K21 of the NFκB inhibitor protein IκB serves as a conjugation site for both SUMO and ubiquitin. The sumoylated pool of IκB is protected from ubiquitination and degradation. Sumoylation of IκB therefore blocks activation of NFκB. Stress-induced phosphorylation of IκB is catalyzed by the IκB kinase (IKK). IKK is composed of two catalytic subunits and the regulatory subunit NFκB essential modulator (NEMO). NEMO contains two SUMO acceptor sites at K277 and K309 (Huang et al. 2003). Stress associated with NFκB activation induces NEMO sumoylation. SUMO-conjugated NEMO is released from IKK and translocates to the nucleus. This causes activation of IKK as well as phosphorylation, ubiquitination, and degradation of the NFκB inhibitory subunit IκB. These observations are of particular interest for ischemia/stroke research, because transient ischemia induces a dramatic accumulation of SUMO-conjugated proteins (see below), and HIF1α, HSF1, and NFκB are markedly activated after ischemia (Higashi et al. 1995; Chavez and LaManna 2002; Kassed et al. 2004).

Recently, many SUMO conjugation target proteins have been identified in neurons that are not nuclear but cytosolic or cell membrane proteins (for a recent review see Martin et al. 2007a). These include the glutamate receptor subunit 6 where SUMO conjugation results in a decrease in kainate receptor-mediated currents (Martin et al. 2007b), proteins involved in axonal mRNA trafficking (Van Niekerk et al. 2007), proteins playing a role in the regulation of G protein, kinase and phosphatase signaling (Kadare et al. 2003; Mitra et al. 2005; Rodriguez-Munoz et al. 2007), and caspases 7 and 8 (Besnault-Mascard et al. 2005; Hayashi et al. 2006). SUMO conjugation of both caspases 7 and 8 induces nuclear translocation, indicating the strong propensity of many sumoylated proteins for nuclear localization. The nuclear translocation activity of SUMO conjugated to a target protein can be convincingly illustrated in cell culture experiments (Ayaydin and Dasso 2004). When cells are transfected with a construct expression yellow fluorescent protein, yellow fluorescent protein fluorescence is found throughout the entire cell. In contrast, when cells are transfected with a construct expressing a green fluorescent protein-SUMO fusion protein, green fluorescent protein fluorescence is highly concentrated in the nucleus, with the highest levels found in nuclear bodies (Ayaydin and Dasso 2004).

Ischemia dramatically increases levels of SUMO-conjugated proteins

In a recent publication, the group of John Hallenbeck presented data suggesting that activation of SUMO conjugation could be a protective response, shielding neurons from damage caused by low blood flow during hibernation torpor (Lee et al. 2007). Hibernation torpor is a physiological state induced by a highly controlled process where the body temperature is sharply reduced and blood flow, energy consumption, and protein synthesis are lowered to otherwise lethal levels (Frerichs et al. 1994, 1998; Frerichs and Hallenbeck 1998; Carey et al. 2003; Storey 2003). Hibernating animals are able to adapt to these extreme conditions of almost undetectable levels of energy consumption, blood flow, and protein synthesis without risking induction of cell damage. As protein synthesis is almost completely suppressed during hibernation, the authors concluded that new synthesis of neuroprotective proteins cannot be contributing to this adaptive process. Instead, they speculated that post-translational protein modifications might be involved. They did indeed observe a dramatic increase in overall SUMO1 and SUMO2/3 conjugation during the state of hibernation torpor and concluded that activation of SUMO conjugation could have induced a state of tolerance to the otherwise lethal levels of blood flow reduction. This assumption was corroborated by the observation that over-expressing the SUMO conjugating enzyme Ubc9 increased the tolerance of neuroblastoma SHSY5Y cells to transient oxygen/glucose deprivation whereas blocking SUMO conjugation, by expressing a dominant negative form of the SUMO conjugating enzyme Ubc9, increased the extent of cell death (Lee et al. 2007). As activation of SUMO conjugation might help cells to withstand the stress conditions caused by a transient period of low blood flow, it was of major interest to us to investigate the effects of transient cerebral ischemia on SUMO conjugation.

The effects of cerebral ischemia on levels of SUMO-conjugated proteins have been investigated in three independent studies (Cimarosti et al. 2008; Yang et al. 2008a,b). In the first set of experiments, the effects of transient severe forebrain ischemia on SUMO conjugation were investigated in mice (Yang et al. 2008a). Mice were subjected to 10 min transient global cerebral ischemia by bilateral common carotid artery occlusion and lowering of the mean arterial blood pressure to 30 mmHg. Transient cerebral ischemia caused a dramatic increase in levels of SUMO2/3-conjugated proteins and depletion of free SUMO2/3 protein levels, both in the hippocampus and cerebral cortex (Yang et al. 2008a). In contrast to SUMO2/3 conjugation, no major changes in SUMO1 conjugation could be identified after transient global cerebral ischemia (Yang et al. 2008a). This differs from the situation during hibernation where SUMO1 conjugation was found to be markedly activated in the state of hibernation torpor (Lee et al. 2007).

In this first experimental study designed to investigate ischemia-induced changes in SUMO conjugation, the entire hippocampus was used for analysis without further dissection into the vulnerable hippocampal CA1 subfield and the ischemia-resistant part of the hippocampus. The post-ischemic pattern of SUMO2/3 conjugation did not therefore permit any conclusions to be drawn on the possible association between the extent of post-ischemic increase in SUMO2/3 conjugation and the vulnerability of neurons to transient global cerebral ischemia.

In a second study designed to investigate ischemia-induced changes in SUMO conjugation, animals were subjected to various periods of focal cerebral ischemia by occluding the right middle cerebral artery (MCA; Yang et al. 2008b), an experimental stroke model. The extent of ischemia-induced changes in SUMO2/3 conjugation was evaluated after 30 min and 6 h of reperfusion. Like transient global cerebral ischemia, transient MCA occlusion induced a dramatic increase in levels of SUMO2/3-conjugated proteins. The post-ischemic pattern of SUMO2/3 conjugation suggests that this could indeed be a protective response of cells to the stress conditions associated with transient cerebral ischemia, because the increase was smallest in the striatum where the most pronounced cell damage occurs in this animal model. Furthermore, when animals were subjected to a pre-conditioning paradigm of 15 min vascular occlusion (Chen et al. 1996; Toyoda et al. 1997; Mackay et al. 2002; Malhotra et al. 2006), levels of SUMO2/3-conjugated proteins were also dramatically increased. This implies that this stress response is activated in cells that can withstand ischemia-induced metabolic stress.

If the post-ischemic activation of SUMO2/3 conjugation is indeed a neuroprotective response, we would expect it to be activated in neurons and for the activation to be most pronounced in cells that are not irreversibly damaged. Confocal microscopy did indeed confirm that levels of SUMO2/3-conjugated proteins were dramatically increased in neurons where ischemia induced translocation of SUMO2/3-conjugated proteins from the cytoplasm to the nucleus. The post-ischemic increase in levels of SUMO2/3-conjugated proteins and the nuclear translocation of these proteins was particularly pronounced in neurons located at the border of the MCA perfused territory.

In a recent publication, a moderate, up to twofold increase in levels of SUMO1- and SUMO2/3-conjugated proteins was found in the brains of animals subjected to permanent or transient focal cerebral ischemia (Cimarosti et al. 2008). After transient focal cerebral ischemia, the authors observed a significant activation of SUMO1 and SUMO2/3 conjugation in the infarcted striatum, and also an increase in levels of SUMO1-conjugated proteins in the contralateral striatum. The authors did not investigate whether transient MCA occlusion activates SUMO conjugation in the cerebral cortex, a brain region where levels of SUMO2/3-conjugated proteins increased dramatically (Yang et al. 2008b). In permanent focal cerebral ischemia, SUMO1 and SUMO2/3 conjugation was significantly increased 24 h after onset of vascular occlusion. The observation that both 6 and 24 h after transient MCA occlusion SUMO1 and SUMO2/3 conjugation was significantly activated in the infarcted striatum is surprising, given that SUMO conjugation is an ATP-requiring process that one would expect to be blocked in the infarcted ATP-depleted tissue.

As yet, no SUMO transgenic animals are available with which to investigate the role of SUMO conjugation in the pathological process triggered by transient cerebral ischemia and culminating in neuronal cell death or to establish whether it is a protective or toxic stress response, or simply an epiphenomenon not directly related to the pathological process at all. Furthermore, no chemical agents are available that could be used to modify the SUMO conjugation pathway and thus reveal the significance of this process for the fate of neurons exposed to a transient interruption of blood supply. Neonatal hypoxia/ischemia is an excellent experimental model with which to elucidate the possible role of SUMO conjugation in pathological states of the brain associated with critically reduced blood flow, because the brain is exposed to both a toxic stress that induces severe cell death (ischemic hemisphere) and a non-toxic stress that can activate a state of tolerance (non-ischemic hemisphere). Exposure of neonatal rats to 8% oxygen has been shown to induce tolerance to neonatal hypoxia/ischemia (Gustavsson et al. 2005, 2007). When we subjected 7-day-old rats to unilateral common carotid artery occlusion and exposed animals to 90 min hypoxia (8% oxygen) and 30 min to 3 h of recovery, levels of SUMO2/3-conjugated proteins were markedly increased in both hemispheres (Homi et al., in preparation), indicating that the pre-conditioning paradigm of hypoxia treatment is sufficient to activate this process and supporting the assumption that the post-ischemic activation of SUMO conjugation could be a stress response that helps cells to withstand the severe form of metabolic stress induced by a transient interruption of blood supply. An increase in levels of SUMO-conjugated proteins was also observed after exposing animals to 3 days of moderate hypoxia (10% oxygen), suggesting that hypoxia induces a long-lasting activation of this process (Shao et al. 2004). SUMO conjugation has furthermore been shown to reduce the sensitivity of cells to hypoxia or desferroxamine-induced injury (Nguyen et al. 2006).

Stress-induced activation of SUMO conjugation

If activation of SUMO conjugation is a neuroprotective response of post-ischemic neurons, helping cells to restore functions that have been impaired by ischemia, it would be of tremendous interest to elucidate the signal transduction pathways that play a role in this stress response and thus to uncover potential ways to activate SUMO conjugation and make cells more resistant to stress conditions. Most of our understanding of the SUMO conjugation process derives from cell culture experiments, using cells over-expressing SUMO or target proteins. Furthermore, most experimental studies to date have focused on SUMO1 and not on SUMO2/3 conjugation. Various cellular stress conditions have been found to modulate SUMO conjugation, including genotoxic stress, thermal stress (both hypo- and hyperthermia), and oxidative stress (Manza et al. 2004; Shao et al. 2004; Anckar et al. 2006; Lee et al. 2007). It has been demonstrated that an increase in levels of SUMO-conjugated proteins can be induced in different ways. For example, HSF1 is rapidly SUMO conjugated at Lys298 in cells exposed to raised temperatures and sumoylation modulates HSF1 DNA binding ability to activate transcription of target genes (Hong et al. 2001). It has been shown that heat stress-induced phosphorylation of HSF1 at Ser307 stimulates sumoylation at Lys298 (Hietakangas et al. 2003; Hilgarth et al. 2003), and it has been proposed that heat shock-induced phosphorylation at Ser307 could stimulate HSF1 SUMO conjugation by inducing a conformational change that relieves the inhibitory effect of the C-terminal leucine zipper (Hilgarth et al. 2003).

Oxidative stress has been shown to modulate SUMO conjugation in different ways. At low concentrations, reactive oxygen species induce a rapid disappearance of SUMO1- and SUMO2-conjugated proteins by a direct inhibition of SUMO conjugating enzymes through the formation of disulfide bonds involving the catalytic cysteines of the SUMO activating enzyme and Ubc9 (Bossis and Melchior 2006). At higher concentrations, reactive oxygen species have been found to inhibit, by reversible disulfide formation, the SUMO protease SENP1 that is required to desumoylate SUMO-conjugated proteins (Xu et al. 2008). This results in a rise in levels of SUMO-conjugated proteins. As, however, millimolar peroxide levels are required for both the reactive oxygen species-dependent blocking of SUMO conjugation and desumoylation of SUMO-conjugated proteins, these processes are unlikely to play a role in the ischemia-induced modification of SUMO conjugation described above. We have used arsenite exposure to investigate the effects of oxidative stress on SUMO conjugation. Arsenite is a respiratory poison that induces oxidative stress and activates the heat-shock response (Bernstam and Nriagu 2000; Roybal et al. 2005). Exposure of cells to only micromolar concentrations of arsenite induced a marked increase in levels of SUMO1- and SUMO2/3-conjugated proteins (Yang et al. 2008a).

What are the possible targets of SUMO conjugation in cerebral ischemia?

As levels of SUMO2/3-conjugated proteins have been found to rise dramatically after cerebral ischemia, it would be of great interest to identify target proteins where the SUMO conjugation pathway is activated after ischemia. Identification of the target proteins would reveal the possible significance of this process for the final fate of neurons exposed to a transient interruption in blood supply. Identification by proteomic analysis is complicated by the fact that for most proteins with the SUMO binding motif, only a portion of the respective target protein is SUMO conjugated, yet the effect on cell metabolism tends to be significant and long-lasting, as discussed above. In most of the proteomic analyses designed to identify SUMO-conjugated proteins, either SUMO or the target protein was therefore over-expressed to increase SUMO-conjugated proteins to detectable levels. At present, it is therefore difficult to speculate on the possible protein targets where SUMO conjugation is induced after ischemia. A few SUMO conjugation target proteins that are believed to play a role in neuronal cell death induced by cerebral ischemia or in the resistance of neurons to a transient interruption of blood supply have been discussed above. These include the transcription factors HIF1α, HSF1, NFκB, and the glutamate receptor subunit 6, proteins that have been shown to play a role in the pathological process induced by transient cerebral ischemia. Besides identification of SUMO target proteins, another way to establish the possible relevance of the post-ischemic activation of the SUMO conjugation pathway for the fate of post-ischemic neurons is to silence expression of individual SUMO paralogues and thus to elucidate the role of the paralogue under investigation in the pathological process triggered by ischemia. If the post-ischemic activation of the SUMO conjugation pathway is indeed a protective response, aiding the recovery of cells from that severe form of metabolic stress, we would expect silencing of SUMO expression to result in more pronounced ischemic cell damage.

What are the mechanisms underlying ischemia-induced activation of the SUMO conjugation pathway?

The mechanisms underlying the rise in levels of SUMO-conjugated proteins observed during and after cerebral ischemia remain to be established. Various scenarios could be involved in the dramatic increase in levels of SUMO-conjugated proteins in cells exposed to a severe form of stress. Elucidating the mechanisms underlying ischemia-induced activation of the SUMO2/3 conjugation pathway could help us to develop strategies to block or activate this reaction and thus to establish the significance of the process for the fate of post-ischemic neurons. Data have been presented suggesting that other post-translational protein modifications, including phosphorylation, ubiquitination, and acetylation, may modify SUMO conjugation (for a recent review see Guo et al. 2007). An example is given above for HSF1 where heat shock-induced phosphorylation activates SUMO conjugation. Another example is SUMO conjugation of Smad4, which involves p38 MAPK activation (Ohshima and Shimotohno 2003). This is of particular interest for ischemia research, because the p38 MAPK pathway has been shown to be activated after ischemia and to contribute to the pathological process leading to cell death and also to be activated by ischemic pre-conditioning (Irving and Bamford 2002; Ferrer et al. 2003; Sun et al. 2006). The marked increase in levels of SUMO-conjugated proteins observed in experimental models of cerebral ischemia could also arise from a blocking of SUMO deconjugation by SENPs. Ischemia could cause a decrease in SENP activity or an increase in SENP degradation, both of which would raise levels of SUMO-conjugated proteins. Whatever processes are involved in the dramatic rise in levels SUMO-conjugated proteins, it is highly likely that the post-ischemic activation of this pathway defines the final fate of neurons exposed to a transient interruption of blood supply.

Small ubiquitin-like modifier conjugation and neurodegeneration

Various pathological processes are believed to play a role in neuronal cell death in both transient cerebral ischemia and neurodegenerative diseases, including formation of protein aggregates, impairment of ER function, disturbances in cellular calcium homeostasis, and oxidative stress (Chan 2001; Orth and Schapira 2001; Paschen and Frandsen 2001; Paschen 2003b; Chung et al. 2005; Halliwell 2006; Mattson 2006; DeGracia and Hu 2007; Ge et al. 2007; Giacomello et al. 2007; Hegde and Upadhya 2007; Trushina and McMurray 2007). We therefore need to understand the significance of activation of SUMO conjugation associated with degenerative disorders of the brain and the possible implications for brains exposed to transient ischemia. SUMO conjugation has been investigated in various neurodegenerative diseases, using postmortem brain sections from patients or samples from animal or cell culture models where the mutated gene, characteristic of the respective disorder, was expressed (for recent reviews see: Dorval and Fraser 2007; Martin et al. 2007b). These include Alzheimer’s, Parkinson’s, and Huntington’s disease, other polyglutamine diseases, amyotrophic lateral sclerosis, alpha-synucleinopathy diseases such as multiple system atrophy and dementia with Lewis bodies, Down’s syndrome, and neuronal intranuclear inclusion diseases (Chan et al. 2002; Terashima et al. 2002; Ueda et al. 2002; Li et al. 2003; Pountney et al. 2003, 2005; Steffan et al. 2004; McFadden et al. 2005; Riley et al. 2005; Dorval and Fraser 2006; Fei et al. 2006; Takahashi-Fujigasaki et al. 2006; Um and Chung 2006; Zhong et al. 2006; Dorval et al. 2007; Gibb et al. 2007). In most of these studies, conjugation of SUMO to a target protein playing a key role in the degenerative disease under investigation is believed to be involved in the pathogenesis of the respective disorder. Using Drosophila as a model system to investigate mechanisms underlying polyglutamine [poly(Q)] repeat toxicity, the role of SUMO conjugation in the pathological process resulting in neuronal cell death was further elucidated (Chan et al. 2002; Steffan et al. 2004). Expression of a pathogenic androgen receptor with an expanded poly(Q) repeat, associated with the neurodegenerative disease spinal and bulbar muscular atrophy (La Spada et al. 1991), resulted in cellular degeneration, preferentially in neuronal tissues (Chan et al. 2002). Degeneration was enhanced when a non-functional SUMO activating enzyme was co-expressed with the pathogenic androgen receptor protein, implying a modulation of poly(Q) induced pathogenesis by SUMO conjugation. A similar effect of SUMO conjugation has been observed in a Drosophila model of Huntington’s disease where sumoylation of a pathogenic fragment of huntingtin resulted in exacerbated neurodegeneration (Steffan et al. 2004). The exact role of SUMO conjugation in degenerative disorders and the implications for brains exposed to transient ischemia remains to be verified.

SUMO conjugation – a new target for therapeutic intervention?

Cerebral ischemia/stroke is a challenging disease because of its rapid onset, often without warning, and its devastating consequences. The high incidence of cerebral ischemia/stroke-induced death and disability presents a major burden for public and private health systems today and will become an even more pronounced problem in the future with an increasingly ageing population. Since most of the clinical stroke trials have failed, except thrombolysis that is suitable only for a small fraction of stroke victims, it would be of great interest to elucidate mechanisms of neuronal cell death induced by transient ischemia to uncover possible new targets for therapeutic intervention. At present SUMO research is not far enough advanced to predict the significance of stress-induced activation of SUMO conjugation for the fate of neurons exposed to a transient interruption of blood supply. Most SUMO studies are based on cell culture experiments and it still needs to be verified how the respective observations translate to the mechanisms underlying stress-induced changes in SUMO conjugation/deconjugation in intact organs. Furthermore, many SUMO studies focus on one target protein, using cells over-expressing the target protein or the SUMO paralogue under investigation. This approach may however lead to controversial conclusions, as discussed above under the role of SUMO3 in APP processing. Blocking/activating SUMO conjugation in cells located at the border of the ischemic tissue should reveal the significance of this stress response for the fate of neurons exposed to transient energy depletion. Furthermore, establishing safe ways to activate the sumoylation process could help to verify whether non-toxic stress-induced activation of SUMO conjugation can indeed make cells more tolerant to stress conditions.

Conclusions and future perspectives

Transient cerebral ischemia is the first pathological state in which it has been convincingly demonstrated that levels of SUMO-conjugated proteins are dramatically increased in an intact organ in vivo. Transient cerebral ischemia specifically results in accumulation of SUMO2/3-conjugated proteins with only minor changes in levels of SUMO1-conjugated proteins. Post-ischemic accumulation of SUMO2/3-conjugated proteins is a neuronal response to the severe form of metabolic stress induced by a transient interruption of blood supply to the brain parenchyma. Furthermore, transient ischemia induces a massive translocation of SUMO2/3-conjugated proteins from the cytoplasm to the nucleus. Whether the marked post-ischemic accumulation of SUMO2/3-conjugated proteins is caused by an activation of the SUMO conjugation process or by a suppression of desumoylation of SUMO-conjugated proteins or both, needs to be verified in future experiments.

The significance of the dramatic post-ischemic accumulation of SUMO2/3-conjugated proteins for the fate of neurons remains to be established. Results from studies focusing on SUMO conjugation in various degenerative diseases imply that this could be a stress response toxic to affected cells, as discussed above. On the other hand, the marked increase in SUMO conjugation observed during hibernation torpor suggests that it could be a protective response, enabling cells to withstand the otherwise lethal extent of blood flow reduction. This assumption is corroborated by results from focal cerebral ischemia studies using the experimental model of transient MCA occlusion where the most pronounced increase in SUMO2/3-conjugated proteins is found in neurons located at the border of the MCA territory and not in neurons located in the parietal cortex in the center of the tissue supplied by the MCA or in the lateral striatum where most neurons are irreversibly damaged (Yang et al. 2008b). If our hypothesis proves valid and ischemia-induced activation of SUMO conjugation is shown to be a protective response of post-ischemic neurons, and if activation of SUMO conjugation is shown to make cells more tolerant to a severe form of metabolic stress, the sumoylation process could become an exciting new target for therapeutic intervention by providing a means of increasing the resistance of neurons to the severe form of metabolic stress imposed by transient cerebral ischemia.

Ancillary