SEARCH

SEARCH BY CITATION

Keywords:

  • calpain–cathepsin hypothesis;
  • Hsp70;
  • hydroxynonenal;
  • lysosome;
  • neuronal death;
  • Niemann-Pick disease

Abstract

  1. Top of page
  2. Abstract
  3. Background: apoptosis versus necrosis
  4. Lysosomal stabilization
  5. Cleavage of carbonylated Hsp70.1 by calpain
  6. Lysosomal lipidosis underlying its rupture
  7. What remains unclarified in the ‘calpain–cathepsin hypothesis’?
  8. Summary
  9. Acknowledgements
  10. References

J. Neurochem. (2012) 120, 477–494.

Abstract

Necrosis has long been considered accidental and uncontrolled, but during the last decade, it became clear that necrosis is also a well-orchestrated form of cell demise, being as well programmed as apoptosis. To explain the mechanism of neuronal necrosis after ischemia/reperfusion, the ‘calpain–cathepsin hypothesis’ formulated in 1998 postulates that the post-ischemic μ-calpain activation compromises integrity of the lysosomal membrane, thereby leading to cathepsin spillage. Another cause of the lysosomal rupture occurring during reperfusion is reactive oxygen species (ROS) that generate 4-hydroxy-2-nonenal (HNE) by oxidation of membrane fatty acids such as linoleic and arachidonic acids. HNE is an endogenous neurotoxin, because HNE-induced carbonylation of the substrate protein shows loss of its function. However, the molecular mechanisms of lysosomal membrane breakdown are still poorly understood; especially, the biochemical cascade how μ-calpain and ROS work together to disrupt lysosomal membrane has remained unclarified. Three independent proteomic analyses of cerebral ischemia, glaucoma, or mild cognitive impairment in primates have altogether suggested that the common substrate of calpain and/or ROS is heat-shock protein 70.1 (Hsp70.1; simply Hsp70, also called Hsp72 or HSPA1), a major protein of the human Hsp70 family. Hsp70.1 serves cytoprotective roles as a guardian of the lysosomal membrane integrity by assisting sphingomyelin degradation or maintaining proper protein folding and recycling as a chaperone. However, calpain-mediated cleavage of Hsp70.1, especially after its carbonylation because of the oxidative stresses, can induce lysosomal rupture. Furthermore, Hsp70.1 dysfunction activates nuclear factor-kappaB (NF-κB) signaling that can also promote neurodegeneration. By focusing on Hsp70.1 and related lysosomal factors, this review describes rationale of lysosomal destabilization and rupture for executing programmed neuronal necrosis.


Abbreviations used:
AIF

apoptosis-inducing factor

ASM

acid sphingomyelinase

CA1

cornu Ammonis 1

BMP

bis(monoacylglycero) phosphate

ER

endoplasmic reticulum

Hsp70.1

heat-shock protein 70.1

HNE

4-hydroxy-2-nonenal

H2O2

hydrogen peroxide

IκBs

κB inhibitory proteins

KO

knockout

LMP

lysosomal membrane permeabilization

NF-κB

nuclear factor-kappaB

ROS

reactive oxygen species

Background: apoptosis versus necrosis

  1. Top of page
  2. Abstract
  3. Background: apoptosis versus necrosis
  4. Lysosomal stabilization
  5. Cleavage of carbonylated Hsp70.1 by calpain
  6. Lysosomal lipidosis underlying its rupture
  7. What remains unclarified in the ‘calpain–cathepsin hypothesis’?
  8. Summary
  9. Acknowledgements
  10. References

Lysosomes, which were first described by de Duve and colleagues in 1955, take care of cellular homeostasis and possibly differentiation by recycling cellular components (Turk and Turk 2009). Lysosomes degrade a wide variety of substrates such as proteins, glycosaminoglycans, nucleic acids, oligosaccharides, and complex lipids into their building blocks so as to be utilized for the re-synthesis of complex molecules or to be further degraded (Cuervo and Dice 1998; Vellodi 2005). Proteins destined for degradation enter lysosomes via endocytosis (for extracellular proteins), or via micro-, macro-, and chaperone-mediated autophagy (for intracellular proteins) (Turk and Turk 2009). During autophagy (indicating to eat oneself), long-lived proteins, damaged parts of the cytoplasm, or intracellular organelles are engulfed within autophagosomes and are ultimately delivered to lysosomes for the bulk degradation (Komatsu et al. 2005; Maiuri et al. 2007). As a pro-survival mechanism to maintain cellular energy homeostasis, autophagy provides catabolic substrates from the recycling of intracellular proteins and other components, and this can facilitate cell survival (Jin and White 2008).

The essential role of lysosome in the terminal degradation of cellular material engulfed within autophagosomes had been clarified well, before the molecular basis of its rupture was unraveled. Nowadays, it is becoming widely accepted that lysosomes are quite vulnerable to a number of insults including oxidative stresses; lysosomotropic detergents such as acridine orange, ceramide, and sphingosine; and aldhydes such as 3-amino propanal (Terman et al. 2006). Destabilized lysosomes may abolish the supply of catabolic substrates provided by autophagy, exacerbate metabolic stress, and prevent cells from recovering (Repnik and Turk 2010). Accordingly, it is in one sense true that cells may die actually not by autophagy, but with autophagy (Kroemer and Levine 2008).

The term ‘necrosis’ is derived from the Greek word ‘nekros’, meaning ‘death’ or ‘dead body’, and this term has been used for more than a century to describe all types of pathological or uncontrolled cell death. When the putative action of lysosomal enzymes in cell necrosis was first postulated by de Duve, lysosomal hydrolytic enzymes were thought to inevitably trigger necrotic cell death when released into the cytosol (de Duve et al. 1955; de Duve 1959), but its detailed mechanism had remained unexplored because of the two reasons. First, since Kerr et al. (1972) introduced the term ‘apoptosis’ (a Greek word indicating ‘falling leaves’) in 1972, over the last three decades it had been considered the sole form of programmed cell death during development, homeostasis, and disease. Second, the regulated cell death process of apoptosis was elegantly shown in Caenorhabiditis elegans by the Horvitz laboratory in 1980s, and apoptotic cell death has been well characterized at both the genetic and biochemical levels (Lettre and Hengartner 2006).

Although apoptosis has become increasingly well defined, our notions of necrosis are still vague. Nowadays, however, it is widely accepted that apoptosis and necrosis can often be initiated in response to the same types of insults with different doses (Zong and Thompson 2006). Many insults induce apoptosis at lower doses whereas necrosis at higher doses; a critical point to determine the cell death mode is present at the regulation of the lysosomal membrane integrity. Low-intensity stresses trigger lysosomal membrane permeabilization (LMP), a limited release of lysosomal hydrolytic enzymes into the cytoplasm. LMP can be triggered by a wide range of stimuli, including death receptor activation, endoplasmic reticulum (ER) stress, proteasome inhibition, oxidative stress, DNA damage, osmotic stress, and growth factor starvation (Guicciardi et al. 2004; Chwieralski et al. 2006; Stoka et al. 2007; Boya and Kroemer 2008; Kirkegaard and Jäättelä 2009), and it induces apoptosis and apoptosis-like cell death (Leist and Jäättelä 2001; Boya et al. 2003; Cirman et al. 2004; Guicciardi et al. 2004). In contrast, high-intensity stresses trigger lysosomal rupture, a generalized release of lysosomal hydrolytic enzymes into the cytoplasm, which is followed by necrosis (Kågedal et al. 2001). Necrosis participates in the pathogenesis of brain diseases, including ischemic and traumatic injuries, neurodegeneration, and viral or bacterial infection, thereby representing an attractive target for the avoidance of unwarranted cell death.

During the past decade, ‘lysosomal rupture’ has emerged as a prominent area of research to elucidate the mechanisms of necrosis. LMP was first described in vitro in the oxidative stress-induced apoptosis of cultured non-neuronal cells by the Swedish group (Öllinger and Brunk 1995; Brunk et al. 1997, 2001; Roberg and Öllinger 1998; Brunk and Svensson 1999; Johansson et al. 2010). In contrast, using the monkeys undergoing transient global brain ischemia, the biological significance of ‘lysosomal rupture’ has been first demonstrated in vivo in necrosis of the hippocampal cornu Ammonis 1 (CA1) neurons (Fig. 1) by the author’s group (Yamashima et al. 1996, 1998, 2003; Yamashima 2000, 2004; Oikawa et al. 2009; Yamashima and Oikawa 2009; Sahara and Yamashima 2010). Although necrosis had been considered for a long time as a purely accidental and passive cell death subroutine, accumulating evidence from low-species animals (Syntichaki et al. 2002; Luke et al. 2007) to primates (Yamashima 2000) showed that necrosis is carried out by complex signal transduction pathways and execution mechanisms, in response to major cellular insults. Similar to apoptosis, necrosis is currently accepted to be tightly programmed by changes in the protein expression, protein–protein interactions, and proteolytic cleavage, in response to major cellular insults such as severe hypoxia and ATP depletion, massive free radical generation, or neurotransmitter-induced neuronal excitotoxicity (Zong and Thompson 2006; Golstein and Kroemer 2007). Furthermore, not only pathologic stressors but also physiological stressors, such as hormone withdrawal, can activate some forms of ‘physiological’ necrosis. Recently, for example, the signal transducer and activator of transcription 3-regulated, lysosome-mediated necrotic cell-death machinery was harnessed for the complex physiological necrosis of the mammary gland epithelial cells during the post-lactation mammary gland involution (Luke and Silverman 2007; Kreuzaler et al. 2011).

image

Figure 1.  A flow chart of the calpain–cathepsin cascade (upper), and spatial and molecular interactions of activated μ-calpain and Hsp70.1 (lower). Immediately after ischemia, Ca2+ mobilization occurring in the CA1 neuron provokes μ-calpain activation (1). During reperfusion, H2O2 molecules that easily diffuse from mitochondria into lysosomes will induce the formation of hydroxyl radicals (HO˙) by the Fenton-type reaction, because of abundant redox-active irons in lysosomes (2). Extremely reactive hydroxyl radicals generating within lysosomes oxidize membrane lipids (especially linoleic and arachidonic acids) to form hydroxynonenal (HNE) (2) that carbonylates Hsp70.1 at lysosomes (3). The immunofluorescence double-staining shows co-localization of Hsp70.1 and activated μ-calpain (lower left), both showing a lysosome-restricted punctate pattern, which was confirmed by co-staining with lysosome-associated membrane protein-2 (Sahara and Yamashima 2010; data not shown here). Then, carbonylated Hsp70.1 is efficiently cleaved by activated μ-calpain (4) (lower right) that leads to the lysosomal rupture (5). Consequently, release of cathepsins occurs (6) to provoke neuronal death (7). Cleavage of carbonylated Hsp70.1 by activated μ-calpain is obvious, because it was dose-dependently blocked by a specific calpain inhibitor N-acetyl-Leu-Leu-Nle-CHO (ALLN) (lower right). Cited from Sahara and Yamashima (2010).

Download figure to PowerPoint

Calpain and reactive oxygen species

Calpain is a Ca2+-dependent, papain-like, non-lysosomal protease that participates in various signaling pathways by modulating activity and/or function of the substrate proteins. Calpain is present in the cytosol as an inactive precursor, but in response to increased levels of cytosolic Ca2+ during extreme conditions or stimuli, it translocates to the intracellular membranes and is activated by autocatalytic hydrolysis. The most abundant and best-characterized calpains in the brain are two major isoforms: μ- and m-calpains (also called calpains I and II, respectively). In the hippocampus, μ-calpain is localized at the pyramidal and granular cells whereas m-calpain is localized at the interneurons (Rami 2003). Both isoforms consist of heterodimers made up of catalytic subunits containing four major domains and a shared, two-domain regulatory subunit. For the half-maximal activity, μ-calpain requires micromolar (3–50 μM) levels of Ca2+, whereas m-calpain requires nearly millimolar (400–800 μM) levels of Ca2+. As calpain is physiologically related to cell cycle regulation, differentiation, cell migration, adhesion, and signal transduction (Goll et al. 2003), its substrates exist across a wide range of functional categories including post-synaptic structural proteins, Ca2+ regulatory proteins, and signaling proteins.

In adult humans, the brain accounts for only a few percent of the body weight, but occupies approximately 20% of the basal O2 consumption. A main reason for such high O2 consumption is the vast amounts of ATP needed to maintain neuronal intracellular ion homoeostasis that is associated with propagation of action potentials and neurosecretion. Because the brain consumes a large percentage of inspired oxygen, and is rich in polyunsaturated fatty acids and inadequately equipped with antioxidant defense systems, this organ is especially vulnerable to the oxidative stress (Halliwell 2006). Neurons are especially sensitive to the oxidative stress, because they exist in a post-mitotic state, have a greater energy demand for excitability, and their metabolism is not by glycolysis in the cytosol but mostly by oxidative phosphorylation in mitochondrial matrix (Galluzzi et al. 2009). Then, excessive production of reactive oxygen species (ROS) such as superoxide (O2·−), hydrogen peroxide (H2O2), hydroxyl radical (HO˙), and nitric oxide, can damage virtually all biological molecules such as DNA, lipids, and proteins, because they include molecules with an unpaired electron, often termed free radicals (Zong and Thompson 2006).

ROS can damage DNA by causing cleavage of DNA strands, DNA–protein cross-linking, and oxidation of purines (Marnett 2000). As double bonds in polyunsaturated fatty acids are also excellent targets for ROS attacks, the lipid oxidation can lead to the loss of cell membrane integrity of lysosomes and ER, which are associated with an intracellular leak of proteases or a mobilization of Ca2+. Another target is amino acid residues with sulfur or proteins with sulfhydryl links. ROS can attack the disulfide bond, or break up the sulfhydryl links, thereby changing the function of the modified proteins. As proteins are major components of biological systems and play key roles in a variety of cellular functions, oxidative damage to proteins represents a primary event processing in neuronal injuries. As proteins have many amino acid residues that are susceptible to the oxidative stress, oxidative damages on the sensitive proteins may cause structural, functional, and stability modulations with a severe failure of biological functions. Protein oxidation in neurons is well known to occur in the age-related cognitive decline, depression, stroke, and neurodegenerative disorders such as Alzheimer’s and Parkinson’s diseases.

Despite multiple attempts at the substrate sequence analysis, the in vivo calpain substrate has been completely unknown in all pathologic conditions (Bevers and Neumar 2008). Given the large number of calpain substrates, until recently it was unclarified in vivo how calpain-mediated cleavage of the substrate protein leads to execution of neuronal death. Heat-shock proteins (Hsp) are a molecular chaperone being essential for cells’ ability to cope with the environmental stresses, by preventing or reversing abnormal protein folding or aggregation. In cancer cells, Hsp70.1 (also called Hsp72/HSPA1, or commonly referred to as simply Hsp70; Daugaard et al. 2007) promotes cell survival by inhibiting LMP, a hallmark of stress-induced death (Kirkegaard et al. 2010). Both calpain activation and Hsp70.1 dysfunction are causally linked to neurodegeneration, but it has not been elucidated at all whether these two proteins work together for executing neuronal necrosis. It would be fascinating to focus on the outcome of Hsp70.1 after the modification by oxidative stresses with particular relevance to its susceptibility of calpain cleavage (Yamashima and Oikawa 2009).

Why focusing on Hsp70.1?

The pathophysiological significance of ROS-induced protein modifications has been reported not only in cerebral ischemia, but also in Alzheimer’s disease, atherosclerosis, cataract, glaucoma, aging, etc. Modifications such as carbonylation, nitration, and protein–protein cross-linking are generally associated with loss of the protein function, and this may lead to either degradation or aggregation of the damaged proteins and eventually to cell death (Stadtman and Berlett 1997; Halliwell 2006; Sultana et al. 2010). In response to ischemia/reperfusion, physical or chemical trauma, viral or bacterial infection, and neurodegenerative processes, a large number of dying neurons exhibits lysosomal rupture. The molecular mechanism linking ROS and the lysosomal rupture, however, has been missing until recently. Elucidation of the molecular mechanism of lysosomal rupture should aid in both understanding the mechanism of neuronal necrosis and developing novel therapies of intervention.

Oxidative stresses, together with redox-active irons, generate hydroxyl radicals within lysosomes through Fenton reaction. Hydroxyl radicals thereby cause oxidation of lysosomal membrane lipids and generate 4-hydroxy-2-nonenal (HNE), which can carbonylate key proteins in lysosomes. Redox proteomic analysis revealed elevated levels of Hsp70 carbonylation in the inferior parietal lobule of mild cognitive impairment as an early phase of Alzheimer’s disease (Sultana et al. 2010). Furthermore, two independent proteomic analyses using the primate samples have implied that the major stress-induced member of the human Hsp family, Hsp70.1, is a common substrate of calpain in the retina of glaucoma (Nakajima et al. 2006) and the hippocampus after ischemia/reperfusion (Oikawa et al. 2009; Yamashima and Oikawa 2009). Accordingly, for elucidating the molecular mechanism of the lysosomal rupture, it would be beneficial to know the mechanism of calpain-mediated cleavage of carbonylated Hsp70.1 and its effect upon the lysosomal destabilization.

This review is to provide a detailed description of the molecular mechanisms of lysosomal rupture, because elucidation of the cascade of the calpain- and ROS-mediated lysosomal rupture sheds new lights on the research of programmed neuronal necrosis. The author has focused on three aspects of Hsp70.1. First, the mechanism of stabilization and destabilization of the lysosomal membranes was discussed in its association with acid sphingomyelinase (ASM, EC 3.1.4.12) and bis(monoacylglycero) phosphate (BMP), by corresponding to Hsp70.1-mediated lysosomal stabilization in cancer cells (Kirkegaard et al. 2010) and Niemann-Pick type A pathology. Second, as a cause of lysosomal rupture, the proteolysis of carbonylated Hsp70.1 by activated μ-calpain was discussed by focusing on the post-ischemic monkey CA1 neurons. Third, the cascade of ischemic neuronal necrosis was discussed by focusing also on the interaction of Hsp70.1 upon nuclear factor-kappaB (NF-κB), because the latter is a well-known transcription factor of genes that can affect both neuronal survival and death (Dutta et al. 2006; Hayden and Ghosh 2008; Pizzi et al. 2009; Ridder and Schwaninger 2009).

Lysosomal stabilization

  1. Top of page
  2. Abstract
  3. Background: apoptosis versus necrosis
  4. Lysosomal stabilization
  5. Cleavage of carbonylated Hsp70.1 by calpain
  6. Lysosomal lipidosis underlying its rupture
  7. What remains unclarified in the ‘calpain–cathepsin hypothesis’?
  8. Summary
  9. Acknowledgements
  10. References

Role of acid sphingomyelinase

Lysosomes act as the terminal degradative compartment of the endocytotic, phagocytic, and autophagic pathways (Pryor and Luzio 2009). The degradation of macromolecules and smaller substances occurs in the acidic subcellular compartment. Inside this compartment, hydrolytic enzymes cleave the substrates under the very low pH optima. The lysosomal limiting (outer) membrane is covered with thick glycocalix that protects it from attack by the membrane-degrading enzymes. Sphingolipids and glycosphingolipids, the major constituents of plasma membrane, reach the lumen of endosomes as intra-endosomal vesicles or other lipid aggregates, and are digested on the surface of intra-lysosomal membrane (Kolter and Sandhoff, 2005). During maturation from endosomes to lysosomes, the luminal pH value decreases and the composition of intra-lysosomal membrane changes. When the intra-lysosomal membrane is prepared for degradation, its protein and lipid composition differs considerably from that of plasma membrane or the lysosomal limiting membrane. For example, the negatively charged BMP, a characteristic anionic phospholipid of the acidic compartment of the lysosome, is not present in the limiting membrane, but is required on the surface of intra-lysosomal membrane (Fig. 2). BMP is a well-described docking lipid for enzymes and one of cofactors that are involved in the lysosomal catabolism of sphingolipids (Schulze et al. 2009; Kolter and Sandhoff 2010). As a result of its isoelectric points, ASM is positively charged and adheres to the surface of intra-lysosomal membrane that bears a negative charge because of the presence of the acidic lipid BMP. In this form, ASM is protected from attack by acid hydrolases, but its premature release from the intra-lysosomal membrane, for example, by cationic amphiphilic drugs, leads to the premature degradation (Kolter and Sandhoff, 2010).

image

Figure 2.  The membrane-stabilizing effect of Hsp70.1 and its related molecules. Hsp70 can be efficiently endocytosed and localized to the lysosomes, and binding of Hsp70 to a lysosome-specific lipid stabilizes lysosomes (Kirkegaard et al. 2010). When recombinant Hsp70 is added (a), for example, to cancer cells (left: cited from Horváth and Vígh 2010), it is transported via late endosome (b) into the lysosome (c), then interacts with an anionic phospholipid, BMP (d). The Hsp70–BMP interaction enhances binding of BMP with ASM (EC3.1.4.12), which activates this enzyme. ASM catalyzes hydrolysis of sphingomyelin (ceramide phosphorylcholine) into ceramide (e) (left) and phosphorylcholine (right upper: cited from Smith and Schuchman 2008). The deficiency of ASM is responsible for the neurological disorder known as Niemann-Pick type A disease. In internal vesicles and membranes of lysosomes, BMP remains negatively charged even at pH 4.2. As ASM has an isoelectric point of around 6.8, it possesses positively charged regions in the acidic lysosomal environment. These contribute to the interaction of ASM with the membrane-bound BMP (Kölzer et al. 2004; Schulze et al. 2009). Changes in the lipid composition of the intra-lysosomal membrane by the inhibition of Hsp70 (within red rectangle) have a direct and potent influence on the stability of the entire lysosome (right lower: cited from Petersen et al. 2010). It is likely that ASM-dependent increase in the lysosomal ceramide content itself could have a positive impact on the stability of the outer membrane (Petersen and Kirkegaard 2010). Briefly, inhibition of Hsp70, being associated with low ASM activity, causes lysosomal destabilization (within red rectangle). Cited from Horváth and Vígh (2010) and Petersen et al. (2010).

Download figure to PowerPoint

Sphingomyelin is a main lipid constituent of the plasma membrane and membranes of endosomes/lysosomes as well. ASM plays an important role in the turnover of membrane lipids by hydrolyzing sphingomyelin to ceramide and phosphorylcholine. Ceramide is further processed into sphingosine by ceramidase (Schutze et al. 1999). Owing to its N-terminal saposin-homology domain, ASM can hydrolyze membrane-bound sphingomyelin slowly even in the absence of an additional activator protein (Linke et al. 2001a,b). Ceramide serves as a backbone of all complex sphingolipids as an important structural lipid in the membrane macrodomains (Hannun and Obeid 2008). For example, ultraviolet irradiation of mammalian skin reduces barrier function; daily irradiation (0.5 MED) to the hairless rat skin for 15 days provoked dry and scaly skin by a reduced barrier function. Electron microscopic observation of such skin after the RuO4 fixation showed that compared with the non-irradiated control, the irradiated skin contained little sphingomyelinase activity, causing deficiency of ceramide (Meguro et al. 1999). For the cell survival, how does ceramide stabilize lysosomal membrane? ASM-mediated increase in the lysosomal ceramide facilitates fusion of lysosomes with other intracellular vesicles and the cell membrane (Heinrich et al. 2000) through modification of the conformation of lysosomal membrane. It is also likely that an ASM-dependent increase in the lysosomal ceramide content itself has a positive impact on the stability of the limiting membrane (Kirkegaard et al. 2010; Petersen et al. 2010). Both mechanisms may be related to the lysosomal stability (Horváth and Vígh 2010).

A considerable part of our knowledge about the mechanism of lysosomal lipid degradation is derived from a class of human disease ‘sphingolipidosis’, which is caused by inherited defects within the sphingolipid and glycosphingolipid catabolism. Niemann-Pick type A disease caused by a deficiency of ASM has an early onset in infancy with an acute course culminating in death at a few years of age (Gieselmann 1995). As the brain has the highest levels of ASM activity among all organs, large amounts of sphingomyelin and cholesterol accumulate in both the ASM-deficient human brain and ASM-knockout (KO) mouse brain (Horinouchi et al. 1995; He et al. 2002). As these lipids are major components of the plasma membrane and myelin sheaths, inability to breakdown them is deleterious especially to big neurons with wide membranous area, extensive dendritic arbors, and high metabolic demands. Thus, it is probable that Purkinje cells lacking ASM, being unable to maintain extraordinary membrane recycling, are the predominant cell type to degenerate in the Niemann-Pick type A brain (Macauley et al. 2008).

Role of Hsp70–BMP interaction

Lysosomes have been regarded as a kind of intracellular stomach that provides the cell with nutrients, and their major function is to take care of cellular homeostasis and possibly differentiation by recycling cellular components (Turk and Turk 2009). Degradation of dysfunctional or long-lived intracellular proteins in the lysosomes can occur through chaperone-mediated autophagy (Chiang et al. 1989). Heat-inducible chromosome puffs were discovered in the salivary glands of Drosophila larvae in 1962 by Ritossa (1962). This initiated a rapidly expanding research field on Hsp. Studying relation to chromosome puffs, Tissières et al. (1974) discovered the Hsp family as proteins that are induced in the salivary glands of the fruitfly Drosophila melanogaster during the heat-induced stress. Thirteen years later, the function of the Hsp family including Hsp90, Hsp70, and Hap60 as molecular chaperones was elucidated (Ellis 1987). Among the Hsp family, Hsp70 is, by far, structurally and functionally conserved protein in evolution, and is found in all organisms from archaebacteria and plants to humans (Daugaard et al. 2007). Hsp70 proteins consist of (i) a conserved ATPase (ATP-binding) domain, (ii) a protease-sensitive sites, (iii) a peptide-binding domain, and (iv) a G/P-rich C-terminal region containing an EEVD-motif (Freeman et al. 1995). The conserved ATPase domain consolidates the chaperone function and enables Hsp70 proteins to bind and release extended stretches of hydrophobic amino acids, exposed by incorrectly folded globular proteins in an ATP-dependent manner (Daugaard et al. 2007). Hsp70 is increased in response to various types of stress, and transports proteins between cellular compartments. It is involved not only in folding and transport of newly synthesized proteins but also in refolding of misfolded proteins and degradation of irreversibly damaged proteins. Molecular chaperones of Hsp70 protect the central nervous system from various insults by interacting with diverse client proteins to assist in their folding, stability, and function.

By generating lysosomal ceramide, ASM activation plays an important role for the stabilization of lysosomal membrane. Not only ASM but also Hsp70.1 have a positive effect on the lysosomal integrity by binding specifically to BMP (Petersen et al. 2010). In cancer cells, for example, Kirkegaard et al. (2010) demonstrated that, in acidic environments, Hsp70 binds with high affinity and specificity to BMP, facilitates activity of ASM, and thereby stabilizes lysosomes (Fig. 2). The Hsp70-mediated enhancement in the ASM activity has a protective function (Kirkegaard et al. 2010; Petersen et al. 2010), whereas the mere depletion of Hsp70 has a profound negative effect on cancer cell viability and resistance to cytotoxic drugs via ASM dysfunction (Rohde et al. 2005; Daugaard et al. 2007). Albouz et al. (1981) reported that the tricyclic antidepressant desipramine causes a decrease of ASM activity in cultured murine neuroblastoma cells and human fibroblasts. Experimental inhibitions of the Hsp70–BMP interaction similarly revert the Hsp70-mediated stabilization of lysosomes. The same may occur by anti-BMP antibodies, a point mutation in Hsp70 (Trp90Phe), or small-interfering RNA-mediated depletion of Hsp70 (Kirkegaard et al. 2010). These data altogether indicate that proteolysis of Hsp70.1, if occured by acquired brain injuries, can cause lysosomal destabilization through the inhibition of the Hsp70–BMP interaction and the resultant ASM dysfunction. The mechanisms of the Hsp70–BMP interaction and its impairment seem to be crucial to elucidate the mechanism of lysosomal rupture in neuronal death.

Kirkegaard et al. (2010) first demonstrated that Hsp70.1 stabilizes lysosomes by binding to BMP, because the Hsp70.1–BMP interaction enhances the activity of ASM by promoting binding of BMP with ASM (Fig. 2). They found that the ATPase domain of Hsp70.1 is a major determinant of the BMP–protein interaction, and that a point mutation of the ATPase domain decreases its binding capacity to BMP and abolishes lysosome stabilization. The ATPase domain of Hsp70 is positively charged at low pH (4.5) whereas BMP is negatively charged, then these opposite charges contribute to the strong interaction of Hsp70 with BMP under acidic conditions (Kirkegaard et al. 2010; Petersen et al. 2010). Autophagy is of special relevance for the membrane trafficking and degradation, and another role of autophagy is that it is a source of BMP. BMP is biosynthetically derived from phosphatidylglycerol generated in the ER or cardiolipin of mitochondria (Brotherus et al. 1974; Amidon et al. 1996; Heravi and Waite 1999; Sorice et al. 2004; Hullin-Matsuda et al. 2007), and reaches the lysosome as a component of mitochondria by macroautophagy. Unusual sn1 and sn10-configuration of BMP, compared with other phospholipids, leads to its higher resistance to the action of phospholipases (Matsuzawa and Hostetler 1979). Negatively charged lysosomal lipids such as BMP are stable on the surface of intra-lysosomal vesicles, and stimulate the interfacial hydrolysis of membrane-bound sphingomyelin by ASM, and of ceramide by acid ceramidase (Linke et al. 2001a,b).

Enzymes such as ASM (Hurwitz et al. 1994) and acid ceramidase (Elojeimy 2006) are released from the intra-lysosomal membrane after treatment with desipramine, and are rapidly degraded. Consequently, the activity of ASM would be remarkably impaired after degradation, then non-degradable sphingomyelin accumulates as multilamellar bodies. As sphingomyelin is a cholesterol-binding lipid, storage of sphingomyelin indicates also storage of cholesterol within lysosomes, both leading to the lysosomal dysfunction (Liao et al. 2007). High amounts of BMP and low amounts of membrane-stabilizing cholesterol are essentially required at the intra-lysosomal membrane for the proper degradation of sphingolipids (Kolter and Sandhoff 2010). Without BMP, the storage of cholesterol-binding, non-degradable sphingomyelin appears to accelerate the formation of sphingolipids with inward budding of the lysosomal limiting membrane that leads to lysosomal destabilization. The author speculates that the same may occur after the impaired Hsp70.1–BMP binding because of calpain cleavage of the oxidative stress-injured Hsp70.1, so this issue will be discussed next.

Cleavage of carbonylated Hsp70.1 by calpain

  1. Top of page
  2. Abstract
  3. Background: apoptosis versus necrosis
  4. Lysosomal stabilization
  5. Cleavage of carbonylated Hsp70.1 by calpain
  6. Lysosomal lipidosis underlying its rupture
  7. What remains unclarified in the ‘calpain–cathepsin hypothesis’?
  8. Summary
  9. Acknowledgements
  10. References

Role of Hsp70.1 cleavage in lysosomal rupture

Regardless of whether the generation of ROS is primary or secondary, it is an important event for neurodegeneration and affects the pathogenesis of various brain diseases (Smith et al. 2000; Butterfield et al. 2001; Beal 2002; Butterfield and Lauderback 2002). Membrane proteins are one of the most important molecular signposts of the oxidative damage, because they have many amino acid residues that are more susceptible to ROS compared with deoxyguanosine in DNA. All amino acid residues are potential targets for oxidation by ROS, with methionine and cysteine residues being particularly sensitive. Membrane phospholipids are also modified by the oxidative stress; the overall effects of lipid peroxidation are to (i) decrease membrane fluidity, (ii) make it easier for phospholipids to exchange between the two halves of the bilayer, (iii) increase leakiness of the membrane to K+ and Ca2+, and (iv) damage membrane proteins such as receptors, enzymes, and ion channels (Halliwell 2006).

As mentioned above, proteins are oxidized directly by modification with ROS, but, the secondary protein modification may occur by lipid peroxidation products such as HNE and malondialdehyde, or by reactive sugars in glycation or glycoxidation reactions (Dunlop et al. 2009). HNE is a major product formed by lipid peroxidation from omega 6-polyunsaturated fatty acids such as linoleic acid and arachidonic acid (Esterbauer et al. 1991; Siems et al. 1992, 1998). HNE is one of the most powerful endogenous ‘neurotoxins’, because it increases Ca2+ levels, inactivates glutamate transporters, and damages neurofilament proteins (Mark et al. 1997; Ong et al. 2000). Most importantly, HNE-induced carbonylation can seriously affect the function and/or metabolic stability of the affected proteins (Levine 2002). Accordingly, lipid peroxidation followed by the protein carbonylation is widely accepted to be the principal causative factor of various neurodegenerative diseases. If lipid peroxidation occurs at the lysosomal membrane, which protein shown in Fig. 2 would be carbonylated by HNE to induce lysosomal rupture? Identifying the molecular mechanism of HNE-induced lysosomal rupture should be critical. As carbonylated proteins cannot be repaired, they are removed by proteolytic degradation or accumulate as damaged or unfolded proteins (Stadtman and Berlett 1998). In spite of the considerable number of papers reporting increased levels of protein carbonyls in the neurodegenerative diseases, the causal relationship between protein carbonylation and neuronal death has long remained unclear. It has been generally unresolved even whether the presence of carbonylated proteins has a causal role or simply reflects secondary epiphenomena.

Protein carbonyls are mostly derived from arginine, lysine, praline, and threonine residues under oxidative conditions (Toda et al. 2010). Only direct identification and characterization of the carbonylated protein and the affected sites can decipher the potential roles of protein carbonylation. Hsp70.1 mainly functions as a chaperone enabling the cell to cope with harmful aggregations of denatured proteins during and following the insults such as heat, ischemia, and oxidative stress (Hartl 1996; Jäättelä 1999). Significantly elevated levels (surprisingly, more than ‘10-fold’ compared with the non-ischemic controls) of cabonylated Hsp70.1 were detected from the proteomic analysis of the day 3 monkey CA1 tissue after 20 min ischemia/reperfusion (Oikawa et al. 2009). Furthermore, using matrix-assisted laser desorption ionization-time of flight/time of flight analysis of peptides obtained from the carbonylated Hsp70.1 spots, the author’s group first found that the critical carbonylation in the Hsp70.1 protein occurs at its key amino acid residue, Arg469 (Oikawa et al. 2009; Yamashima and Oikawa 2009). This finding is intriguing, if considering the report of Chang et al. (2001) that mutating Arg469 in the rat 70-kDa heat-shock cognate protein, a constitutive member of the heat shock-induced Hsp70 protein family, provoked a reduced affinity for peptide substrates and an impaired refolding activity, and this mutant 70-kDa heat-shock cognate protein (R469C) was more accessible to the proteolytic cleavage.

Carbonylation makes the substrate protein more susceptible to proteolysis as a result of unfolding of its targeted domains. Unfolding exposes amino acid residues that were normally hidden in the protein structure but became more prone for degradation in response to the proteasome and the Lon protease (Grune et al. 2004; Nyström 2005). Calpain-mediated proteolysis of substrates within the synaptic compartment is involved not only in synaptic signaling but also in neurodegeneration (Bevers and Neumar 2008). Although there are many studies concerning implications of calpain in neuronal death, the in vivo substrate of activated μ-calpain had been unknown until recently. A specific role for calpains in the degradation of oxidized proteins was not elucidated, but using the monkey CA1 tissue, Sahara and Yamashima (2010) first demonstrated that carbonylated Hsp70.1 by artificial oxidative stressors, such as HNE or H2O2, is much more vulnerable to the calpain cleavage (Fig. 1). Neurodegenerative diseases are often associated with protein misfolding and aggregation of protein outside and within the cells. Misfolded proteins and protein aggregation are controlled by molecular chaperones such as Hsp70.1. Then, inactivation of Hsp70.1 in the cytoplasm makes this protein unable to facilitate degradation of misfolded proteins by the lysosome and/or proteasome, and this would lead to the accumulation of toxic substance such as amyloid β. Furthermore, cleaved Hsp70.1 at lysosomes cannot bind with BMP to activate ASM and to generate ceramide (Fig. 2), which lead to destabilization of the lysosomal membrane and the accumulation of sphingomyelin as seen in Niemann-Pick type A disease.

Role of Hsp70.1 on NF-κB activation

A dimeric transcription factor, NF-κB, controls the expression of genes that regulate normal tissue development, homoeostasis, and physiological functions of various cells within the immune, hepatic, epidermal, and nervous tissues. In the central nervous system, NF-κB regulates a broad range of biological processes such as synaptic plasticity, neurogenesis, and differentiation. It can protect neurons against oxidative stresses or ischemia-induced degeneration. NF-κB signaling was identified in neurons after cerebral ischemia (Clemens et al. 1997; Nurmi et al. 2004) and also in the brains of Alzheimer’s or Parkinson’s diseases (Hunot et al. 1997; Kaltschmidt et al. 1997). NF-κB activation or inactivation has been proposed to serve a dual role as a regulator of neuronal death or survival (Dutta et al. 2006; Ridder and Schwaninger 2009). For example, NF-κB stimulation with the release of p65/RelA and p50 contributes to the neurodegenerative process and neuronal death (Pizzi et al. 2009; Ridder and Schwaninger 2009), whereas NF-κB protected neurons from ischemic injury in the middle cerebral artery occlusion model of mice (Duckworth et al., 2006; Li et al., 2008). Death or survival of neurons was previously thought to depend on the cell type and the timing of NF-κB stimulation, but recently it was proposed that within the same neuronal cell, activation of diverse NF-κB dimers drives opposite effects on the cell fate (Sarnico et al. 2009).

In mammalian cells, the NF-κB/Rel family is comprised of five members, p50, p52, p65/RelA, RelB, and c-Rel, which can diversely combine to form the active transcriptional dimer. In the naive state, NF-κB complexes normally remain in the cytoplasm in association with the κB inhibitory proteins (IκBs) such as IκBα, IκBβ, or IκBε. However, in response to stimuli, IκB proteins are phosphorylated, poly-ubiquitinated, and then degraded, and these result in p65/RelA translocation into the nucleus to stimulate gene transcription (Hayden and Ghosh 2008). NF-κB breakdown with the simultaneous release of p65/RelA and p50 contributes to the neurodegenerative process and neuronal death (Pizzi et al. 2009; Ridder and Schwaninger 2009).

NF-κB is a major regulator of programmed cell death via apoptosis or necrosis (Dutta et al. 2006). Although NF-κB generally protects cells by inducing the expression genes encoding anti-apoptotic and anti-oxidizing proteins, its role in apoptosis and necrosis can vary markedly in different cell contexts, and Hsp70.1 can affect cell death fate by interacting with the NF-κB signaling pathway. Anti-inflammatory role for Hsp70 has been suggested to be because of the inhibition of NF-κB destabilization in cerebral ischemia by interaction of Hsp70.1 with NF-κB proteins or other proteins in the NF-κB regulatory pathway (Giffard et al. 2008). Although the role of the lysosomal rupture in neuronal necrosis has been grossly elucidated by the calpain-mediated cleavage of Hsp70.1, implication of Hsp70.1 in regulating NF-κB signals following cerebral ischemia in primates is not fully elucidated.

Regional differences in the Ca2+ mobilization within the brain during ischemia/reperfusion (Yamashima et al. 1996; Bevers and Neumar 2008) may explain why the pathologic calpain activation occurred restrictively in the hippocampal CA1 sector but not in the other cerebral cortex. Zhu et al. (2012) recently studied the underlying mechanism of the differential neuronal vulnerability by comparing the ischemia-resistant motor cortex and the ischemia-vulnerable CA1 sector against ischemia/reperfusion (Fig. 3). They found that the motor cortex after the ischemic insult did not show a remarkable activation of μ-calpain after the ischemic insult, compared with the non-ischemic cortex. Accordingly, intact Hsp70.1 can interact and bind with NF-κB p65/RelA and IκBα in the cytoplasm of motor cortex after ischemia. This binding leads to inhibiting degradation of IκBα and nuclear translocation of p65/RelA, and effectively protects neuronal cells from ischemic damages. In contrast, the dysfunction of Hsp70 after calpain-mediated cleavage in CA1 fails to bind NF-κB p65/RelA with IκBα for stabilization of this complex in the cytoplasm, and this results in a sustained NF-κB p65/RelA activation with nuclear translocation and in neurodegeneration. It is likely that the Hsp70.1-mediated cascades affecting stabilization not only of lysosomes but also of NF-κB may influence neuronal survival/death after the ischemia/reperfusion.

image

Figure 3.  The mechanism how Hsp70.1 dysfunction activates NF-κB signaling leading to neurodegeneration. As carbonylation of Hsp70.1 (Oikawa et al. 2009) in the post-ischemic CA1 leads to its cleavage by activated μ-calpain (Sahara and Yamashima 2010), the resultant Hsp70.1 dysfunction induces NF-κB destabilization that was indicated by the absence of Hsp70.1 binding with p65/RelA or IκBα on immunoprecipitation (a, red arrow). The nuclear transfer of NF-κB p65/RelA contributes to neurodegeneration via the transcription of pro-inflammatory/apoptotic genes. In contrast, the motor cortex after the ischemic insult did not show a remarkable activation of μ-calpain after the ischemic insult, compared with the non-ischemic control (Zhu et al., 2012). Accordingly, up-regulation of Hsp70.1 (b) without calpain cleavage contributes to neuroprotection by stabilizing both lysosomes and NF-κB in the motor cortex after the transient ischemia followed by reperfusion (Zhu et al. 2012). NF-κB stabilization was indicated by the Hsp70.1 binding with both p65/RelA and IκBα on immunoprecipitation (c, red arrow). The inhibition of nuclear transfer of NF-κB p65/RelA fails to transcribe pro-inflammatory/apoptotic genes, which leads to neuroprotection. Cited from Zhu et al. (2012).

Download figure to PowerPoint

Lysosomal lipidosis underlying its rupture

  1. Top of page
  2. Abstract
  3. Background: apoptosis versus necrosis
  4. Lysosomal stabilization
  5. Cleavage of carbonylated Hsp70.1 by calpain
  6. Lysosomal lipidosis underlying its rupture
  7. What remains unclarified in the ‘calpain–cathepsin hypothesis’?
  8. Summary
  9. Acknowledgements
  10. References

Experimental ischemia models

Multilamellar bodies (also called multivesicular bodies or myeloid bodies) with multiple concentric membrane layers are generated as lysosomal structures that frequently exhibit an electron-dense core. They form via cellular autophagy, and their lysosomal nature has been suspected by the involvement of various lysosomal enzymes (Hariri et al. 2000). The formation of multilamellar bodies starts with inward budding of the limiting endosomal membrane and requires the sequential action of three endosomal sorting complexes such as ESCRT-I, -II, and -III for transport (Hopkins et al. 1990). Multilamellar bodies are generated under both physiological and pathological conditions; initially, single or multiple foci of lamella appear within an autophagic vacuole, then transform into multilamellar structures. They are ubiquitously found in numerous cell types functioning in the lipid storage and secretion (Schmitz and Müller 1991). Physiologically, multilamellar bodies are responsible for the surfactant secretion in type II alveolar cells to prevent alveolae from collapse during respiration, but accumulate in other cell types under pathological conditions such as cerebral ischemia (Fig. 4) and lysosomal storage diseases (Lajoie et al. 2005). Transfection of lung type II alveolar cells with ß1–6-N-acetylglucosaminyl transferase V, an enzyme regulating the expression of polylactosamine oligosaccharides, was associated with the formation of numerous multilamellar bodies. Putative b1–6 branching and polylactosamine glycosylation of their glycoproteins might enhance resistance to degradation by lysosomal proteases or modify interactions between the lamellar components, thereby favoring lamella formation (Hariri et al. 2000). Treatment of the cells with leupeptin, an inhibitor of many lysosomal proteases, results in the reduction of multilamellar bodies, whereas treatment with the phosphatidylinositol 3-kinase inhibitor 3-methyladenine, a specific inhibitor of autophagy (Seglen and Gordon1982; Blommaart et al. 1997), results in the disappearance of multilamellar bodies. These data indicate that ‘selective impairment of lysosomal degradation’ within the ‘autophagic vacuole’ is indispensable for the formation of multilamellar bodies (Hariri et al. 2000).

image

Figure 4.  Electron micrographs of the monkey CA1 neurons after the in vitro or in vivo ischemic insults. The left and middle columns show CA1 neurons in the hippocampal slice after 20 min hypoxia–hypoglycemia, whereas the right column shows CA1 neurons of the post-ischemic day 1 monkey. In both, accumulation of multilamellar bodies (curved white arrows) is seen within lysosomes and in the cytoplasm. As shown by immunoelectron microscopy (left and middle), activated μ-calpain (black deposits shown by black arrows) is localized at lysosomes, but not to mitochondria. Both multilamellar bodies and calpain activation are not seen in the control lysosomes (Cont; white arrow). Multilamellar bodies, abundant non-degradable lipid accumulation, increase as early as on day 1 after the ischemia/reperfusion in the monkey neurons (right). After the oxidative stress, CA1 neurons conceivably internalize parts of the damaged cell membrane, in part together with foreign materials, via endocytosis, and deliver them to early endosomes. However, impaired recycling of the early endosomal/lysosomal contents during hypoxia/hypoglycemia or cerebral ischemia may cause accumulation of multilammer bodies within and/or in the vicinity of lysosomes (curved white arrows). Although an intimate relationship between lysosomes and mitochondria has been suggested in the execution of cell death, most of the mitochondria in the CA1 neurons appear to be intact at the electron microscopic level. After the immunohistochemistry using anti-activated μ-calpain antibody and diaminobenzidine, the in vitro hypoxia–hypoglycemia slices (left and middle) were fixed with glutaraldehyde and OsO4, whereas the in vivo ischemia tissue (right) was fixed with glutaraldehyde containing 0.1% tannic acid and OsO4. Both specimens were observed without ultrastructural stainings, but the reaction products with diaminobenzidine look black, whereas lipids reacting with tannic acid appear as electron-dense multilamellar bodies (rectangle magnified). Cont; control; Hypo; hypoxia-hypoglycemia, Day 1; the day 1 monkey after 20 min whole brain ischemia followed by reperfusion, N; nucleus, bar = 1 μm. From the author’s unpublished data.

Download figure to PowerPoint

Cell survival depends upon its ability to adapt to environmental changes by regulating protein synthesis and degradation. In response to the serum starvation, for example, cells adapt by activating their lysosomal proteolytic pathway such as macroautophagy, which starts with the formation of autophagosomes by invaginations of the rough ER (Dunn 1990a; Ueno et al. 1991). Macroautophagy is a process by which organelles and cytosol are sequestered into autophagosomes, which then mature into autolysosomes to degrade the sequestered component by acquiring lysosomal hydrolases (Dunn 1990b; Lawrence and Brown 1992). Autophagy is essentially the process for recycling of organelles and parts of the cytoplasm, and removal of superfluous or damaged organelles provides nutrient supplies to the cell to help it to survive. However, an excess level of autophagy is detrimental to the cell as shown in the developmental cell death of Drosophila salivary glands (Berry and Baehrecke 2007). Stimulation of autophagy results in an increased size and concentric lamellar morphology of multilamellar bodies (Hariri et al. 2000; Lajoie et al. 2005). Intriguingly, autophagy with the concomitant multilamellar body formation is occurring in the CA1 neurons in both the sliced hippocampal tissue after transient hypoxia-hypoglycemia in vitro (Fig. 4, left and middle) and the living monkeys after ischemia/reperfusion in vivo (Fig. 4, right). In both in vitro (Fig. 4 left and middle, black arrows) and in vivo (Yamashima et al. 1996, 2003) experimental paradigms, the immunoelectron microscopic analysis reveals immunoreactivity of activated μ-calpain at the lysosomal limiting membrane. It is easily suggested from these data that an excess autophagy of cell membrane, if combined with calpain-induced cleavage of carbonylated Hsp70.1 at the lysosomal membrane, may lead to the lysosomal membrane destabilization and its rupture (Oikawa et al. 2009; Yamashima and Oikawa 2009; Sahara and Yamashima 2010).

As polar heads of sphingomyelin and phosphatidyl choline lack primary amines found in such phospholipids as phosphatidyl ethanolamine and phosphatidyl serine, they cannot directly react with glutaraldehyde or OsO4. However, tannic acid reacts with the choline ‘base’ of sphingomyelin and phosphatidyl choline to form a complex that can react with OsO4 to become insoluble during the dehydration process with ethanol or acetone (Kalina and Pease 1977a,b). Accordingly, the conventional ultrastructural fixation by glutaraldehyde and OsO4 cannot preserve multilamellar bodies well, but this becomes possible by introducing glutaraldehyde mixed with 0.1–1% tannic acid before osmication. Using tannic acid, highly ordered, multilamellar bodies containing abundant sphingomyelin can be appreciated even without uranyl and lead citrate stainings in the day 1 CA1 neuron after 20 min ischemia (Fig. 4, right). Storage of sphingomyelin within lysosomes appears as large subcellular vesicular structures at both light and conventional electron microscopic observation (Fig. 5d, e, and j) because of the ethanol or acetone dehydration. However, with the aid of tannic acid, lysosomal lipidosis can be preserved during the dehydration procedure, and observed as stacked membrane-like multilamellar bodies at the electron microscopic level (Fig. 4, right).

image

Figure 5.  The affected neurons in ASM-KO mice, Niemann-Pick calf, and post-ischemic monkeys, showing similar light and electron microscopic features. Both the degenerating Purkinje cells in the ASM-KO mice (upper left; a, b, c, d: Macauley et al. 2008) and the monkey CA1 neurons after ischemia/reperfusion (lower left; f, g, h, i: Oikawa et al. 2009) are microscopically characterized by eosinophilic coagulation necrosis with many cytoplasmic tiny vacuoles (red circle in d, arrows and red circle in i) and pyknotic nuclei (d, i). Electron micrographs of a Purkinje cell in the 7-month-old Niemann-Pick calf (e: Saunders and Wenger 2008) and the monkey CA1 neuron on day 7 after ischemia (j) show large vesicular structures (e, j, arrows) indicating presumably remnants of lysosomal lipidosis. As both Niemann-Pick calf neuron (e) and post-ischemic CA1 neuron (j) ultrastructurally show disrupted cell membranes and punctuated chromatin condensation, but no blebbing and hypercompacted nuclei, the cell death pattern of each should be ‘necrosis’ despite of the distinct etiology each other. Intriguingly, similar cytoplasmic tiny vacuoles measuring 1–2 μm can be appreciated in the Alzheimer’s CA1 neurons with light microscope, if one observes carefully at the high-power magnification (data not shown, here). It is possible that most of the accumulated lipids within lysosomes were dissolved with ethanol or acetone during the dehydration procedure. Bars in e, j = 1 μm. c: non-ischemic control; d3, d5, d7: days 3, 5 and 7 after ischemia-reperfusion. Cited from Macauley et al. (2008), Saunders and Wenger (2008), and Oikawa et al. (2009).

Download figure to PowerPoint

Niemann-Pick type A disease

Niemann-Pick type A disease is a rare, autosomal recessive, lysosomal storage disorder that is characterized by neurovisceral accumulation of sphingomyelin as a result of the inherited deficiency of ASM activity. Loss of ASM activity results in the intracellular accumulation of sphingomyelin within the distended lysosomes throughout the body. In addition to sphingomyelin accumulation in the peripheral organs, storage pathology is severe in neurons and glia throughout the nervous system, which results in the progressive, global deterioration of neurological function and death by the age of 3 years (Schuchman and Desnick 2001). Descriptions of Niemann-Pick type A neuropathology are extremely limited in the human patients (Elleder 1989; Lyon et al. 1996). However, murine models with targeted disruption of the ASM gene (Horinouchi et al. 1995) are appropriate for the detailed analysis. In the ASM-KO mouse model mimicking Niemann-Pick type A disease, the affected swollen neurons in the cerebral cortex, hippocampus, thalamus, cerebellum, brain stem, spinal cord, etc. are characterized by the accumulation of distended lysosomes within the cytoplasm, and most of the affected neurons showed cell death (Fig. 5b, arrows) or eosinophilic degeneration with tiny vacuoles (Fig. 5d) (Macauley et al. 2008).

Despite of the widely distributed storage of sphingomyelin in neurons and glia, only the Purkinje cells show the remarkable degeneration in the Niemann-Pick type A disease, ASM-KO mice (Fig. 5b and d), and Niemann-Pick calf (Fig. 5e) (Otterbach and Stoffel 1995; Kolodny 2000; Macauley et al. 2008; Saunders and Wenger 2008). Then, such a question emerges that why Purkinje cells selectively manifest neuronal death in Niemann-Pick type A disease. As the size of Purkinje cells exceeds by far that of most other cells owing to their prominent role in motor plasticity (Konnerth et al. 1990), an excess amount of membrane-derived sphingomyelin accumulates within lysosomes of Purkinje cells so that becoming detrimental to the lysosomal membrane. In the case of post-ischemic CA1 neurons, because of the remarkable Ca2+ mobilization and the resultant calpain activation, the calpain-mediated cleavage of Hsp70.1 presumably exacerbated lysosomal destabilization (Yamashima and Oikawa 2009; Sahara and Yamashima 2010). Although speculative at present, the similar light (Fig. 5d vs. g, h, i) and electron microscopic (Fig. 5e vs. j) features of neurons that were observed in ASM-KO mice, Niemann-Pick calf, and post-ischemic monkeys may implicate lysosomal lipidosis in the pathogenesis of not only ischemic but also degenerative neuronal death (Fig. 5).

What remains unclarified in the ‘calpain–cathepsin hypothesis’?

  1. Top of page
  2. Abstract
  3. Background: apoptosis versus necrosis
  4. Lysosomal stabilization
  5. Cleavage of carbonylated Hsp70.1 by calpain
  6. Lysosomal lipidosis underlying its rupture
  7. What remains unclarified in the ‘calpain–cathepsin hypothesis’?
  8. Summary
  9. Acknowledgements
  10. References

Ischemic neuronal death

Christian de Duve introduced the term ‘suicide bags’ to describe a characteristic of lysosomes in 1959, but the mechanism of cell suicide (Turk and Turk 2009) long remained unknown until the formulation of the ‘calpain–cathepsin hypothesis’ in 1998. Since then, accumulating evidence starting from the C. elegans (Syntichaki et al. 2002) has paved the way to the concept of programmed cell necrosis in diverse experimental paradigms (Golstein and Kroemer 2007; Luke et al. 2007; Giusti et al. 2009; Vandenabeele et al. 2010). Calpain and cathepsin are main players during the propagation and execution phases of the programmed cell necrosis; these two cysteine proteases directly or indirectly (Yamashima 2000) provoke damages to the lysosomal membrane, which culminate in the disruption of cell integrity. Although almost the whole cascade of delayed CA1 neuronal death after ischemia/reperfusion was unraveled now, nobody knows exactly the in vivo substrates of calpain and cathepsin in the post-ischemic CA1 neurons, because it is extremely difficult to substantiate in vivo.

Wei et al. (1995) first reported a role for Hsp70 as an essential factor for cancer cell survival in the Hsp70 depletion study, which has since been substantiated in a series of experiments (Kirkegaard and Jäättelä 2009; Kirkegaard et al. 2010). As shown in the monkey experimental paradigm (Yamashima et al. 1996, 1998), the cancer cell death induced by the Hsp70 depletion was also characterized by the release of cathepsins into the cytosol, and inhibition of them provided a significant cytoprotection (Nylandsted et al. 2000, 2004). Furthermore, exogenous Hsp70 effectively inhibited lysosomal destabilization induced by various stresses (Nylandsted et al. 2004; Gyrd-Hansen et al. 2006; Bivik et al. 2007). It is suggested from these data that the cytoprotective effect of Hsp70 is partly because of the stabilization of lysosomal membranes. Interestingly, using homogenates of the monkey hippocampal tissue treated by HNE or H2O2, Sahara and Yamashima (2010) recently found that Hsp70.1, especially after the artificial oxidative stress, becomes vulnerable to the calpain cleavage (Fig. 1). Even if the calpain-mediated cleavage of carbonylated Hsp70.1 is closely related to the occurrence of lysosomal rupture, the detailed molecular mechanism still remains to be clarified as to how lysosomal membrane destabilization occurs by the Hsp70.1 cleavage. Moreover, the molecular cascade is not well understood how the accumulation of multilamellar bodies in the lysosomes occurs so quickly, within hours after the insult (Fig. 4, left and middle).

Monkeys undergoing transient global brain ischemia followed by reperfusion were shown to be protected against delayed CA1 necrosis even by the post-ischemic administration of cathepsin B or L inhibitor, CA-074 or E64c (Yamashima et al. 1998). Luke et al. (2007) also found in C. elegans that intracellular serpins capable of neutralizing lysosomal cysteine proteases have a crucial role in blocking lysosomal-mediated necrotic cell death, and that a serphin-related srp-6 gene believed to be a natural protease inhibitor may have a role in preventing necrosis. These findings strongly support the pathophysiological relevance of the ‘calpain–cathepsin hypothesis’, suggesting cathepsin inhibitors as potential neuroprotective agents from low-species animals to humans. Lysosomal cathepsins released into the cytoplasm may continue to damage the remaining still-intact lysosomal membrane from outside or activate phospholipases that degrade all types of cellular membrane. Furthermore, the cleavage of BID and degradation of anti-apoptotic BCL-2 proteins by lysosomal cysteine cathepsins were identified in vitro (Repnik and Turk 2010). They attack mitochondria to release cytochrome c and other pro-apoptotic factors from the outer membrane, and to induce more H2O2 production by interfering with the mitochondrial electron-transporting complexes (Terman et al. 2006). Despite decades of work on the cathepsins, however, very little is known about the in vivo substrates of cathepsin B, L. An assumption can be made that a significant number of proteins should be degraded for the processing of neuronal necrosis. But, there had been no evidence concerning the in vivo cathepsin substrates during the occurrence of necrosis, because of the very short duration and complexity of the degradation process by cathepsin enzymes. At acidic pH, cathepsins are optimally active and can degrade their substrates quickly. When cathepsins are released into the cytosol by lysosomal rupture, they face unfavorable conditions because of the higher pH. Therefore, after the release of cathepsins into the cytosol, not every proteins that are degraded within lysosomes at acidic pH can become their substrate in vivo (Turk and Turk 2009). Nevertheless, acidification of the cytosol, which is observed during hypoxia (Syntichaki et al. 2005; Artal-Sanz et al. 2006), may help stabilizing and activating cathepsins to digest molecules that are normally not exposed to these proteases.

Degenerative neuronal death

Activation of μ-calpain, being as much as sevenfold up-regulation compared with the age-matched brains, was demonstrated also in the Alzheimer’s brains on the Western blot analysis, using the same anti-activated μ-calpain antibody used for Fig. 4 (Taniguchi et al. 2001; Yamashima 2004). Accordingly, activated μ-calpain might have an important implication to consider the mechanism of neuronal death associated with the Alzheimer’s disease. The author speculates that the calpain- and HNE-mediated lysosomal rupture of ischemic neuronal death, being emerged from the analysis of cerebral ischemia in monkeys, might be applicable to elucidating mechanisms of degenerative neuronal death in the Alzheimer’s disease. This speculation emerged not only from calpain activation in the Alzheimer’s brain (Taniguchi et al. 2001) but also from such observation that at the high-power light microscopic magnification, the degenerated CA1 neurons in the Alzheimer’s brain show similar eosinophilic coagulation necrosis with many cytoplasmic tiny vacuoles (generally called granulo-vacuolar degeneration, data not shown here) as seen in the monkey CA1 neurons (Fig. 5f, g, h, and i). It is interesting to imagine how the calpain–cathepsin cascade, although should be very mild and gradual compared with cerebral ischemia, can participate in the development of degenerative neuronal necrosis, because elevations of the free HNE concentrations were detected in the cerebrospinal fluid of Alzheimer’s disease (Lovell et al. 1997).

Although the detailed mechanism still remains unknown, the oxidative stress associated with the membrane lipid peroxidation is currently accepted to play an important role in the pathogenesis of Alzheimer’s disease (Jomova et al. 2010). As described previously, ROS may lead to destabilization of the lysosomal membrane. Intralysosomal irons, derived from degraded iron-containing proteins, such as metallo proteins, react with H2O2 and generate highly reactive hydroxyl radicals that can trigger lipid peroxidation of the lysosomal membrane and contribute to the generation of HNE. HNE is a cytotoxic aldehyde byproduct generated by the peroxidation of membrane fatty acids such as linoleic and arachidonic acids (Esterbauer et al. 1991). As HNE is a diffusible factor, it can modify proteins of cell membranes in the surrounding area of its generated sites (Pedersen et al. 2000). Immunohistochemical analysis of the postmortem brains suffering Alzheimer’s disease revealed increased protein modification by HNE (Sayre et al. 1997). As carbonylation and cleavage of Hsp70.1 compromise integrity of the lysosomal membrane in the post-ischemic CA1 neurons (Fig. 1), it is tempting to speculate that the same HNE-mediated lysosomal injury might occur in the neurons involved by Alzheimer’s disease, but this is extremely difficult to demonstrate in the living brain.

As distinguishing between necrosis and apoptosis, particularly in vivo, is very difficult, it is unclear that Hsp70-mediated lysosomal necrosis is very relevant in neurodegenerative diseases that are relatively slowly progressing. Although the Alzheimer’s disease displays many elements of cell stress such as ROS production, these are at a relatively low level. Accordingly, one would rather stipulate that the term ‘apoptosis’ would still better describe the nature of neuronal death in this disease rather than ‘necrosis’. Hsp70 has been shown to play a direct role in the regulation of apoptosis (Nylandsted et al. 2000; Bivik et al. 2007). For instance, the Hsp70.1-KO mice brain was more vulnerable to the focal ischemia compared with the wild-type brain (Lee et al. 2004). On the contrary, Hsp70.1 can rescue various cell types from apoptosis after exposure to heat shock, tumor necrosis factor-α, oxidative stress, irradiation, or anti-cancer drugs (Mosser and Morimoto 2004). Interestingly, Bivik et al. (2007) showed that Hsp70.1 can rescue melanocytes from ultraviolet irradiation-induced apoptosis by preventing the release of cathepsins from lysosomes as well as Bax translocation and cytochrome c release from mitochondria. Hsp70 protein is a potent cell death-inhibitory factor, and its protective and anti-apoptotic effects are associated with the reduction of nuclear translocation of apoptosis-inducing factor (AIF) as a result of an increased binding of Hsp70 to AIF (Matsumori et al. 2005). By binding directly to AIF, Hsp70 may prevent AIF-induced chromatin condensation in a caspase-independent apoptotic pathway (Ravagnan et al. 2001). On the contrary, deletion of Hsp70 exaggerates ischemia/reperfusion-induced cardiac dysfunction, which is associated with increased cardiac apoptosis, decreased cardiac output, and larger infarct size mediated by AIF activation. It is suggested from these data that calpain-mediated cleavage of carbonylated Hsp70.1 can also induce neuronal apoptosis being associated with neurodegenerative diseases.

Membrane lipid peroxidation is strongly implicated in the pathogenesis of Alzheimer’s disease, and the disease susceptibility is increased in individuals with an apolipoprotein Eε4 allele, compared to those with only apolipoprotein Eε2 or Eε3 alleles. Inheritance of the apolipoprotein Eε4 allele is well known to influence the pathogenesis of Alzheimer’s disease (Smith et al. 1998). These are presumably because apolipoprotein Eε2, 3 alleles can play a major role in detoxifying HNE, which can protect Hsp70.1 from carbonylation, or neurons from lysosomal rupture/permeabilization, whereas apolipoprotein Eε4 allele cannot play a role in protecting against the neurotoxic effects of HNE (Pedersen et al. 2000). Apolipoprotein E proteins differ in only two amino acids: E2 has two cysteine residues, E3 has one cysteine residue, whereas E4 has none. Although there is only a minor variation in the cysteine content among each isoforms, this is sufficient to confer a differential ability to protect against HNE toxicity. Each isoforms differ in the amount of HNE they can bind, with the order of E2 > E3 > E4, which correlates with the differential ability of each isoforms to protect against HNE (Pedersen et al. 2000). This can explain why the isoform-specific neuroprotective activity is weakest in E4 (Jordán et al. 1998; Buttini et al. 1999). Unraveling the cascade of calpain- and HNE-mediated lysosomal destabilization may be important also for understanding the pathogenesis of degenerative neuronal death.

Summary

  1. Top of page
  2. Abstract
  3. Background: apoptosis versus necrosis
  4. Lysosomal stabilization
  5. Cleavage of carbonylated Hsp70.1 by calpain
  6. Lysosomal lipidosis underlying its rupture
  7. What remains unclarified in the ‘calpain–cathepsin hypothesis’?
  8. Summary
  9. Acknowledgements
  10. References

An important unanswered question about necrosis has been whether this cell death process is programmed, involving dedicated effector proteins. Activated μ-calpain translocated to the lysosomal membrane after ischemia/reperfusion compromises its integrity, induces spillage of cathepsins into the cytoplasm and neuropil, and eventually dismantles hippocampal CA1 neurons a few days later (Fig. 1). This observation in the monkey experimental paradigm led to the formulation of the ‘calpain–cathepsin hypothesis’ in 1998. However, the molecular mechanisms of lysosomal rupture had been incompletely understood, because nobody could explain whether and how μ-calpain and ROS work together. The common substrate of μ-calpain and ROS was recently identified to be Hsp70.1, and this discovery led to unraveling the whole calpain–cathepsin cascade at last. Accordingly, the present review has attempted to outline the whole cascade of programmed neuronal necrosis by focusing on the mechanism of lysosomal stabilization and destabilization that are mediated by Hsp70.1 and its related factors such as ASM, BMP, and NF-κB (Fig. 6). As Hsp70.1 has a pivotal role in regulating survival and death of neurons via stabilization or destabilization of lysosomal membranes, interfering with its carbonylation and cleavage could have clinical relevance in the treatment of not only ischemic but also degenerative neuronal death. The study of lysosomal rupture in the programmed neuronal necrosis is still young, but the therapeutic approaches based on the ‘calpain–cathepsin hypothesis’ would promise a powerful strategy for various brain diseases including Alzheimer’s disease. Not only calpain and cathepsin but also Hsp70.1 and NF-κB would represent attractive targets for the avoidance of unwarranted neuronal death in humans.

image

Figure 6.  Role of Hsp70.1 and related molecules in the calpain–cathepsin cascades. Whatever the case may be, lysosomal rupture occurs by two pathways that are fairly synergistic: (i) calpain-mediated cleavage of carbonylated Hsp70.1, and (ii) inactivity of ASM because of either genetic disorder or the experimental impairment of Hsp70.1-BMP binding. Original by the author.

Download figure to PowerPoint

References

  1. Top of page
  2. Abstract
  3. Background: apoptosis versus necrosis
  4. Lysosomal stabilization
  5. Cleavage of carbonylated Hsp70.1 by calpain
  6. Lysosomal lipidosis underlying its rupture
  7. What remains unclarified in the ‘calpain–cathepsin hypothesis’?
  8. Summary
  9. Acknowledgements
  10. References