SEARCH

SEARCH BY CITATION

Keywords:

  • aging;
  • Alzheimer disease;
  • heme oxygenase-1;
  • iron;
  • oxidative stress;
  • Parkinson disease

Abstract

  1. Top of page
  2. Abstract
  3. Heme oxygenases
  4. Brain aging
  5. Glial senescence
  6. Alzheimer disease
  7. Parkinson disease
  8. Other neurological disorders
  9. Conclusions
  10. Acknowledgements
  11. References

The heme oxygenases (HOs), responsible for the degradation of heme to biliverdin/bilirubin, free iron and CO, have been heavily implicated in mammalian CNS aging and disease. In normal brain, the expression of HO-2 is constitutive, abundant and fairly ubiquitous, whereas HO-1 mRNA and protein are confined to small populations of scattered neurons and neuroglia. In contradistinction to HO-2, the ho-1 gene (Hmox1) is exquisitely sensitive to induction by a wide range of pro-oxidant and other stressors. In Alzheimer disease and mild cognitive impairment, immunoreactive HO-1 protein is over-expressed in neurons and astrocytes of the cerebral cortex and hippocampus relative to age-matched, cognitively intact controls and co-localizes to senile plaques, neurofibrillary tangles, and corpora amylacea. In Parkinson disease, HO-1 is markedly over-expressed in astrocytes of the substantia nigra and decorates Lewy bodies in affected dopaminergic neurons. HMOX1 is also up-regulated in glial cells surrounding human cerebral infarcts, hemorrhages and contusions, within multiple sclerosis plaques, and in other degenerative and inflammatory human CNS disorders. Heme-derived free ferrous iron, CO, and biliverdin/bilirubin are biologically active substances that have been shown to either ameliorate or exacerbate neural injury contingent upon specific disease models employed, the intensity and duration of HO-1 expression and the nature of the prevailing redox microenvironment. In ‘stressed’ astroglia, HO-1 hyperactivity promotes mitochondrial sequestration of non-transferrin iron and macroautophagy and may thereby contribute to the pathological iron deposition and bioenergetic failure amply documented in Alzheimer disease, Parkinson disease and other aging-related neurodegenerative disorders. Glial HO-1 expression may also impact cell survival and neuroplasticity in these conditions by modulating brain sterol metabolism and proteosomal degradation of neurotoxic protein aggregates.

Abbreviations used
AAT

α1-antitrypsin

AD

Alzheimer disease

apoE

apolipoprotein E

APP

amyloid precursor protein

CA

corpora amylacea

CO

carbon monoxide

CSH

cysteamine

DA

dopamine

DAB

diaminobenzidine

ER

endoplasmic reticulum

GFAP

glial fibrillary acidic protein

Hmox1/HMOX1

heme oxygenase-1 gene

HO

heme oxygenase

HSP

heat-shock protein

IL-1β

interleukin-1β

LXR

liver X activated receptor

MCI

mild cognitive impairment

Mrp1

multidrug resistance-associated protein 1

MS

multiple sclerosis

NFT

neurofibrillary tangle

NO

nitric oxide

OS

oxidative stress

PD

Parkinson disease

ROS

reactive oxygen species

SnMP

tin mesoporphyrin

TNFα

tumor necrosis factor-α

UPS

ubiquitin-proteosome system

The heme oxygenases (HOs) have been recognized as dynamic sensors of cellular oxidative stress (OS) and likely arbiters of tissue redox homeostasis across the phylogenetic spectrum. However, the HO reaction is a ‘double-edged’ sword! Akin to the activity of the nitric oxide (NO) synthases, superoxide dismutases, and other redox-active enzymes, the HOs may impart robust antioxidant defense in certain disease conditions while serving to amplify free radical damage in others. As illustrated in this article, the Janus faces of the HOs are particularly apparent in disorders of the mammalian CNS. The successful manipulation of tissue HO activity for clinical purposes, as considered by several laboratories (Dennery 2000; Scapagnini et al. 2002; Durante 2003; Otterbein et al. 2003; Abraham and Kappas 2008; Syapin 2008) including our own (Schipper 2004a,b; Schipper et al. 1995), pre-supposes a thorough understanding of the multiple and often disparate activities subserved by these enzymes under specific pathological conditions. This review focuses primarily on the roles of HO-1 in brain aging and the common human neurodegenerative disorders, Alzheimer disease (AD) and Parkinson disease (PD). Participation of HO-1 in other degenerative and neuroinflammatory conditions is also briefly discussed, with emphasis on pathophysiological processes shared with the more common aging-related neurodegenerations.

Heme oxygenases

  1. Top of page
  2. Abstract
  3. Heme oxygenases
  4. Brain aging
  5. Glial senescence
  6. Alzheimer disease
  7. Parkinson disease
  8. Other neurological disorders
  9. Conclusions
  10. Acknowledgements
  11. References

Regulation and physiology

In humans and other species, degradation of cellular heme is catalyzed by the HO family of enzymes (E.C. 1:14:99:3; heme-hydrogen donor:oxygen oxidoreductase). HOs are located within the endoplasmic reticulum (ER) where they act, in association with NADPH cytochrome P450 reductase, to oxidize heme to biliverdin, free ferrous iron and carbon monoxide (CO) (Fig. 1). Biliverdin reductase further catabolizes biliverdin to the bile pigment, bilirubin (Ryter and Tyrrell 2000). Mammalian cells express at least two isoforms of HO, HO-1 [a.k.a. heat-shock protein (HSP) 32] and HO-2. A third protein, HO-3 was determined to be a retrotransposition of the HO-2 gene (pseudogene) unique to rats (Scapagnini et al. 2002). HO-1 and HO-2 exhibit identical substrate and cofactor specificities. Yet, the isoforms are encoded by distinct genes, share only 43% amino acid sequence homology in humans, and exhibit significant differences with regard to molecular weight, electrophoretic mobility, tissue distribution, regulation, and antigenicity (Dennery 2000; Loboda et al. 2008). HO-1 contains a destabilizing proline-glutamic acid-serine-threonine sequence at the carboxy terminus that renders the peptide sensitive to rapid proteolysis (Dwyer et al. 1992), a characteristic not exhibited by HO-2. The half-lives of HO-1 mRNA and protein have been estimated to be 3 and 15–21 h, respectively (Dennery 2000).

image

Figure 1.  The heme catabolic pathway. The heme degradation products, ferrous iron (Fe II), carbon monoxide (CO), and biliverdin/bilirubin may behave as either pro-oxidants or antioxidants accounting for the disparate influences of heme oxygenase expression on cell function and survival (symbolized by Janus faces). M = methyl; V = vinyl; P = propionate [modified from Ryter and Tyrrell (2000) with permission].

Download figure to PowerPoint

In humans, the ho-1 gene (HMOX1) is located on chromosome 22q12 and contains four introns and five exons. A 500-bp promoter, a proximal enhancer, and two or more distal enhancers occur in the regulatory region of the mammalian Hmox1 gene. The latter exhibits activator protein 1, activator protein 2, nuclear factor kappa B, and hypoxia-inducible factor 1 binding sites, as well as heat shock consensus sequences, metal response elements, cadmium response elements, and stress response elements. These response elements render Hmox1 exquisitely sensitive to induction by a wide array of pro-oxidant and inflammatory stimuli including heme, β-amyloid, dopamine (DA), H2O2, hyperoxia, UV light, heavy metals, prostaglandins, NO, peroxynitrite, Th1 cytokines, oxidized lipid products, and lipopolysaccharide, as well as certain growth factors (Dennery 2000; Schipper 2000; Kinobe et al. 2006; Loboda et al. 2008). Curiously, under hypoxic conditions, HO-1 is induced in rodent, bovine, simian, and some human (D407 retinal pigment epithelium; dermal fibroblasts; and keratinocytes) cells, whereas in other human cells, HMOX1 is repressed (astrocytes; umbilical vein and coronary artery endothelial cells) or unaffected (ARPE19 retinal pigment epithelial cells; chorionic villus epithelium) (Loboda et al. 2008). In certain tissues (e.g. rat liver, lung), HO-1 expression appears to be developmentally-regulated at both transcriptional and post-transcriptional levels (Dennery 2000). Numerous transcription factors participate in the regulation of HO-1 expression, with tendencies for one or more signaling pathways to predominate in a species-specific manner. NF-E2-related factor-2 transcription factor binding to Maf response elements in the Hmox1 promoter and its repression by the heme-regulated protein, basic leucine zipper transcription factor 1 have emerged as a key control mechanism for HO-1 induction and homeostasis in stressed brain and other organs (Ogawa 2002; Sun et al. 2002; Kitamuro et al. 2003; Suzuki et al. 2003). Finally, a 56-bp sequence (Signal Transducer and Activator of Transcription binding site) within the Hmox1 promoter confers susceptibility to transcriptional suppression by glucocorticoids (Lavrovsky et al. 1996).

Polymorphisms in the lengths of GT repeats (11–40) within the HMOX1 promoter appear to be an important determinant of HO-1 expression and function in humans. Long GT sequences code for relatively unstable (Z-conformational) DNA with attenuated transcriptional activity and diminished baseline and stimulated HO-1 protein expression profiles. Robust HO-1 activity associated with the short-GT polymorphisms appears protective against atherosclerosis-linked conditions (e.g. coronary artery disease) whereas the malignant behavior of various neoplasms was fairly consistently enhanced (Exner et al. 2004; Jozkowicz et al. 2007; Loboda et al. 2008). For further information regarding the molecular biology and regulation of Hmox1, several recent reviews are recommended (Jozkowicz et al. 2007; Loboda et al. 2008; Syapin 2008).

All mammalian cells metabolize heme. OS may transiently augment the intracellular ‘free heme pool’ by stimulating the conformational modification and breakdown of hemoproteins such as cytochromes, myoglobin, peroxidises and respiratory burst enzymes (Ryter and Tyrrell 2000; Loboda et al. 2008). Some cellular heme is also reversibly bound to proteins such as neuronal PAS domain protein 2, the DiGeorge critical region-8 protein, basic leucine zipper transcription factor 1, and proteins of the N-end rule pathway (Syapin 2008). In stressed cells, the up-regulation of HO-1 may confer protection by accelerating the degradation of pro-oxidant heme to the radical-scavenging bile pigments, biliverdin and bilirubin (Stocker et al. 1987; Nakagami et al. 1993; Llesuy and Tomaro 1994; Dore et al. 1999; Baranano and Snyder 2001). Co-stimulation of apoferritin synthesis, a major iron sequestration pathway, may obviate potential toxicity related to the intracellular liberation of heme-derived ferrous iron (Dennery 2000; Ryter and Tyrrell 2000). However, in some cases, heme-derived iron and CO may exacerbate intracellular OS and substrate damage by promoting the generation of reactive oxygen species (ROS) within the mitochondrial and other subcellular compartments (Zhang and Piantadosi 1992; Frankel et al. 2000; Ryter and Tyrrell 2000; Desmard et al. 2007). The intensity and temporal pattern of HO-1 expression and the chemistry of the local redox microenvironment may determine whether free radical damage accruing from intracellular liberation of iron/CO or the antioxidant benefits of a diminished heme : bilirubin ratio predominate (Galbraith 1999; Suttner and Dennery 1999).

Basal HO expression in mammalian CNS

Heme oxygenase-1 expression in the normal, unstressed brain is minimal and confined to scattered neuroglia and sparse populations of neurons in the cerebellum (Purkinje cells), thalamus, hypothalamus, brain stem, hippocampal dentate gyrus, and cerebral cortex (Vincent et al. 1994; Matz et al. 1996; Bergeron et al. 1998; Nakaso et al. 2000; Baranano and Snyder 2001). In contrast, HO-2 mRNA and protein are widely distributed and strongly expressed in neurons of the mammalian brain and spinal cord, with highest concentrations in hippocampal pyramidal cells and dentate gyrus, olfactory epithelium and olfactory bulb, and cerebellar granule and Purkinje cell layers (Trakshel et al. 1988; Verma et al. 1993; Dwyer et al. 1995b). The distribution of HO-2 expression in rat brain exhibits considerable overlap with the topography of soluble guanylate cyclase. Heme-derived CO, like NO, may modulate neuronal activity by stimulating guanylate cyclase and raising intracellular levels of the second messenger, cyclic guanosine monophosphate (Verma et al. 1993).

Neuroprotective roles of HO-1

There exists ample evidence implicating a neuroprotective role for HO-1 both in intact animals as well as in tissue culture. Following exposure to a variety of noxious stimuli, HO-1 induction occurs in neuronal and non-neuronal brain cells, although it has been argued that the greater capacity of astrocytes than neurons to mount a robust HO-1 (and other HSP) response may partly account for the relative preservation of the former in the face of oxidative challenge (Dwyer et al. 1995a; Manganaro et al. 1995; Snyder et al. 1998). Cerebellar granule cells harvested from transgenic mice designed to over-express HO-1 in neurons (Maines et al. 1998) appear to be relatively resistant to glutamate- and H2O2-mediated oxidative damage in vitro (Chen et al. 2000). Similarly, neuroblastoma cell lines transfected with HO-1 cDNA were less prone than control cells to oxidative damage resulting from exposure to H2O2 (Le et al. 1999; Takeda et al. 2000) or β-amyloid1–40 (Le et al. 1999). Over-expression of HO-1 was also recently reported to increase brain-derived neurotrophic factor and glial cell-derived neurotrophic factor expression and protect neurons from MPP+ toxicity (Hung et al. 2008). Astrocytes harvested from HO-1 knockout mice exhibit enhanced vulnerability to hemin toxicity relative to wild type cells (Chen-Roetling et al. 2005), although HO-1 may be inimical to astroglia under other circumstances (see below). In an in vivo study, HO-1 transgenic mice subjected to cerebral ischemia manifested decreased tissue staining for lipid peroxidation end-products, enhanced expression of the anti-apoptotic factor, B-cell CLL/lymphoma 2 and smaller infarct volumes relative to normal littermates (Panahian et al. 1999). HO-1 may also confer neuroprotection in animal models of traumatic (Fukuda et al. 1996; Beschorner et al. 2000) and excitotoxic (Huang et al. 2005; Ahmad et al. 2006) brain damage and spinal cord injury (Lin et al. 2007b). As described above, rapid heme/hemoprotein degradation and the intracellular accumulation of antioxidant bile pigments may be responsible, at least in part, for the observed neuroprotection associated with the induction of HO-1. CO released in the course of heme catabolism may also mediate some of the cytoprotective benefits of HO-1. For example, by engendering smooth muscle relaxation (Verma et al. 1993), heme-derived CO has been postulated to ameliorate cerebral vasospasm, a cause of significant morbidity in patients with subarachnoid hemorrhage (Matz et al. 1996; Suzuki et al. 1999; Tanaka et al. 2000).

Neurodystrophic effects of HO-1

While HO-1 induction may provide neuroprotection under certain conditions (see above), the action of this enzyme in other models of CNS injury and disease has proven detrimental. It has long been known that excessive hyperbilirubinemia of untreated neonatal jaundice may engender irreversible neurological injury (kernicterus). The latter can be prevented in these children by administration of synthetic metalloporphyrins, competitive inhibitors of HO activity, or photodegradation of circulating bilirubin (Qato and Maines 1985). Moreover, metalloporphyrin suppression of HO activity has been shown to diminish tissue necrosis and edema formation following focal cerebral ischemia in intact rats (Kadoya et al. 1995), confer neuroprotection in an experimental model of intracerebral hemorrhage (Koeppen and Dickson 1999; Wang and Dore 2007) and alleviate traumatic cornu ammonis 1 insults in rat hippocampal slices (Panizzon et al. 1996). Suppression of hippocampal HO-1 expression has also been posited as a mechanism by which nimodipine treatment ameliorates aluminum neurotoxicity in mice (Yuan et al. 2008). Of possible relevance to PD, HO-1 also facilitates dopaminergic cell injury following exposure to polychlorinated biphenyls (Lee et al. 2006). Germane to several human neurodegenerative disorders, and as discussed at length below, HO-1 over-expression in the astroglial compartment may transduce a host of noxious stimuli into altered patterns of cellular iron mobilization and attendant mitochondrial damage (Schipper 2004b). Differences in species, experimental models and therapeutic protocols may account for the disparate data regarding the roles of HO-1 induction in brain aging and disease.

Brain aging

  1. Top of page
  2. Abstract
  3. Heme oxygenases
  4. Brain aging
  5. Glial senescence
  6. Alzheimer disease
  7. Parkinson disease
  8. Other neurological disorders
  9. Conclusions
  10. Acknowledgements
  11. References

CNS senescence has been implicated as the single most important risk factor for a host of human neurodegenerative conditions, including sporadic (non-familial) AD, idiopathic PD, and amyotrophic lateral sclerosis (Calne 1994). In accord with the prevailing ‘Free Radical-Mitochondrial’ theory of aging (Finkel and Holbrook 2000; Sohal et al. 2002; Hekimi and Guarente 2003), OS, altered patterns of iron mobilization and mitochondrial insufficiency have been recognized as fairly constant features of senescent and degenerating neural tissues. Based on extensive evidence adduced from neuropathological surveys, animal studies and cell culture experiments, we have argued that these ubiquitous pathological changes are causally interrelated phenomena that forge a vital link between normal brain aging and neurodegeneration (Schipper 2004a). Moreover, as discussed below, the consolidation of this pathological ‘triad’ may depend, at least in part, on the antecedent up-regulation of HO-1 by the astrocytic compartment.

HO-1 expression in normal aging brain

Heme oxygenase-1 expression in human brain has not been rigorously mapped or quantified. However, a Japanese study showed that numbers of neurons and neuroglia immunoreactive for HO-1 in the normal human cerebral cortex and hippocampus (procured from 31 autopsied brains) increased progressively between 3 and 84 years of age (Hirose et al. 2003). In 1995, we surveyed HO-1 immunoreactivity in post-mortem brain specimens derived from five AD (discussed below) and five neurohistologically normal human subjects aged 79.4 ± 3.4 and 79.8 ± 3.0 years, respectively (Schipper et al. 1995). In the non-demented subjects, neurons in the temporal cortex and hippocampus generally exhibited faint or no detectable HO-1 immunoreactivity. When present, HO-1 staining appeared as diffuse precipitates largely confined to the perikarya and apical dendrites. Stronger HO-1 immunostaining was observed in rare neurofibrillary tangles (NFT) encountered in the control specimens. The percentages of glial fibrillary acidic protein (GFAP)-positive astrocytes co-expressing immunoreactive HO-1 protein were approximately 6.8% in the normal hippocampus and 12.8% in the substantia nigra. In normal senescent human brain, robust HO-1 staining was noted in the cytoplasm of ependymocytes and choroid plexus epithelial cells, in cerebrovascular endothelial cells, and within many glial and extracellular corpora amylacea (CA) (Schipper et al. 1995; Anthony et al. 2003). In a subsequent study (Schipper et al. 1998a), moderate HO-1 immunoreactivity was consistently observed in the cytoplasm of neuromelanin-containing (dopaminergic) neurons of the normal subtantia nigra pars compacta (n = 7; mean age = 69.6 ± 9.6 years). In contrast, HO-1 staining of other neurons within this brain region was faint or undetectable. The latter findings suggest that, during the course of normal aging, dopaminergic neurons in the human substantia nigra are subjected to increased levels of OS (and possibly additional noxious influences) relative to other neuronal populations. This formulation is consistent with the high levels of ROS generated in these cells by the enzymatic deamination and iron-mediated oxidation of DA (Stokes et al. 1999), and reports that aging-associated loss of nigral dopaminergic neurons may occur at an accelerated rate in comparison with other brain cell populations (Mann 1994). Other work revealed substantial HO-1 immunoreactivity in olfactory neuroepithelium of normal elderly human subjects (Perry et al. 2003). HO-1 expression also increases markedly in optic nerve astrocytes and in retinal ganglion cells and photoreceptors of the human eye between the second and seventh decade of life (Mydlarski et al. 2003). As HO-1 transcription and translation are widely accepted as sensitive and reliable reporters of tissue OS, these HO-1 expression patterns may denote regions of the human CNS that are particularly prone to OS in the course of normal aging. Moreover, as described in the following section, stress-related induction of glial HO-1 may de-regulate iron homeostasis and contribute to the decay of mitochondrial function in the senescent human CNS.

Glial senescence

  1. Top of page
  2. Abstract
  3. Heme oxygenases
  4. Brain aging
  5. Glial senescence
  6. Alzheimer disease
  7. Parkinson disease
  8. Other neurological disorders
  9. Conclusions
  10. Acknowledgements
  11. References

The peroxidase-positive subcortical glial system

In aging rats, humans, and other vertebrates, a sub-population of subcortical astrocytes progressively accumulates cytoplasmic inclusions that exhibit an affinity for Gomori stains, orange-red autofluorescence, and non-enzymatic (pseudo-) peroxidase activity mediated by ferrous iron (Schipper et al. 1981, 1998b; Schipper 1996, 1999). In the rat substantia nigra, the iron-laden glial inclusions increase in abundance by a factor of four between 4 and 14 months of age (Schipper et al. 1998b). As described below, sustained or repeated induction of the Hmox1 gene may be necessary and sufficient for the development of this senescent glial phenotype.

The cysteamine model

Exposure of dissociated fetal or neonatal rat brain cell cultures to the sulfhydryl agent, cysteamine (CSH; 2-mercaptoethylamine) induces a marked accumulation of peroxidase-positive astrocytic inclusions that are structurally and histochemically identical to those that naturally accumulate in subcortical astroglia of the intact aging brain (Schipper et al. 1990). Elemental iron is readily detected in the inclusions by electron microprobe analysis, and the presence and concentration of the metal correlates closely with the presence and intensity of diaminobenzidine (peroxidase) staining (McLaren et al. 1992). Within 24–72 h of CSH exposure, many astroglial mitochondria exhibit progressive swelling, rearrangement or dissolution of their cristae, subcompartmental sequestration of redox-active iron and fusion with lysosomes or cisternae of the ER (Brawer et al. 1994a; Chopra et al. 1997). In young adult rats, subcutaneous CSH injections (150–300 mg/kg twice weekly for 3 weeks) induce two- to threefold increases in numbers of peroxidase-positive astrocyte granules in the basal ganglia, hippocampus and other brain regions (Schipper et al. 1993). As in the case of the CSH-treated cultures, peroxidase-positive glial granules in the intact rat and human brain invariably exhibit mitochondrial epitopes (as well as identical profiles of HSP expression) in immunohistochemical preparations (Brawer et al. 1994b; Schipper and Cissé 1995; Schipper et al. 1998b). Further studies indicated that intracellular OS may be responsible for the transformation of normal astrocyte mitochondria to peroxidase-positive granules and CA, glycoproteinaceous inclusions characteristic of aging and degenerating neural (and other) tissues (Cavanagh 1999), in vitro and in the intact aging brain (Srebro 1971; Manganaro et al. 1995; Sahlas et al. 2002; Schipper 2004a).

Glial HO-1 expression and mitochondrial iron sequestration

Cysteamine (880 μM), β-amyloid40/42 (3–15 μM), DA (0.1–1.0 μM), tumor necrosis factor-α (TNFα) (20 ng/mL) or interleukin-1β (IL-1β) (20 ng/mL) significantly augments the incorporation of 59Fe (or55Fe) into astroglial mitochondria within 3–6 days of treatment without affecting transfer of the metal into whole-cell and lysosomal compartments (Wang et al. 1995; Schipper et al. 1999; Ham and Schipper 2000; Mehindate et al. 2001). These effects were only discernible when inorganic 59FeCl3, but not 59Fe-diferric transferrin, served as the metal donor (ibid). Within 3–12 h of exposure to these stimuli, and preceding mitochondrial iron sequestration, augmented HO-1 mRNA, protein and/or activity levels were observed in these cells (Schipper 1999; Mehindate et al. 2001). Pharmacological studies revealed that OS is a likely common mechanism mediating glial Hmox1 induction under these experimental conditions (Mydlarski et al. 1993; Schipper et al. 1999). Co-administration of tin mesoporphyrin (SnMP) (1 μM), a competitive inhibitor of HO activity, or dexamethasone (50 μg/mL), a transcriptional suppressor of Hmox1, significantly attenuated mitochondrial iron sequestration in cultured astrocytes exposed to DA, TNFα, or IL-1β (Schipper et al. 1999; Mehindate et al. 2001). Similarly, administration of SnMP or dexamethasone abolished the pathological accumulation of mitochondrial 55Fe observed in rat astroglia engineered to over-express human HMOX1 by transient transfection (Schipper et al. 1999). These findings indicate that up-regulation of HO-1 is a key event in the cascade leading to excessive mitochondrial iron deposition in oxidatively challenged astroglia.

HO-1, intracellular oxidative stress, and the mitochondrial permeability transition pore

Treatment with SnMP or antioxidants (ascorbate, melatonin, or resveratrol) blocks the compensatory induction of the manganese superoxide dismutase (Mnsod) gene in astrocytes challenged with DA or transiently transfected with HMOX1 cDNA (Frankel et al. 2000). Furthermore, levels of protein carbonyls (protein oxidation), 8-epi-prostaglandinF2α (lipid peroxidation), 8-hydroxy-2-deoxyguanosine (nucleic acid oxidation) and a synthetic redox reporter molecule were significantly increased in glial mitochondrial fractions after 3–4 days of HMOX1 transfection relative to sham-transfected controls and HO-1-transfected cells receiving SnMP (Song et al. 2006; Vaya et al. 2007). These convergent data clearly indicate that HO-1 over-expression in astroglia exacerbates intracellular OS that targets the mitochondrial compartment. In 2004, Gennuso et al. (2004) reported that nanomolar concentrations of unconjugated bilirubin up-regulate multidrug resistance-associated protein 1 (Mrp1) expression and translocation of Mrp1 from the Golgi apparatus to the cell membrane in cultured mouse astrocytes. As Mrp1 promotes egress of bilirubin from astrocytes (ibid), resultant increments in cellular (mitochondrial?) ratios of Fe/CO : bilirubin may pre-dispose to OS in astrocytes over-expressing HO-1.

Treatment with cyclosporin A or trifluoperazine, potent inhibitors of the mitochondrial permeability transition pore, curtails mitochondrial iron trapping in hHO-1 transfected glia and cells exposed to DA, TNFα, or IL-1β (Schipper et al. 1999; Mehindate et al. 2001). Conceivably, intracellular OS accruing from HO-1 activity promotes pore opening (Petronilli et al. 1993; Bernardi 1996) and influx of cytosolic iron to the mitochondrial matrix.

HO-1 and glial macroautophagy

Recent ultrastructural studies in combination with dynamic secondary ion mass spectrometry disclosed profound alterations in organellar morphology and the topography of elemental iron in rat astrocytes transiently transfected with HMOX1 cDNA (Zukor et al. 2009). On post-transfection day 3, transmission electron microscopic analysis revealed distended mitochondria with disrupted cristae, cytoplasmic dense inclusions and small multilamellar bodies consistent with membrane damage and possible lysosomal engagement, and spherical concretions bearing ultrastructural resemblance to human CA. The HMOX1-associated cytopathology was more elaborate on day 6 post-transfection and featured advanced mitochondrial degeneration, huge multilamellar bodies, multivesicular bodies, pleomorphic osmiophilic inclusions, and CA-like structures. These morphological changes were abrogated in HMOX1-transfected cultures exposed to SnMP, indicating that they devolve from HO enzymatic activity rather than some signaling or other non-canonical function recently attributed to the HO-1 protein (Li Volti et al. 2004; Lin et al. 2007a, 2008). The cytopathological changes represent classical ultrastructural features of macroautophagy and demonstrate that over-expression of the HMOX1 gene is capable of driving this process in astroglia. Using dynamic secondary ion mass spectrometry, intense foci of elemental iron were observed within HMOX1-transfected cells which were not seen in control preparations (Zukor et al. 2009). Within the former, there was discrete co-localization of iron with small elliptical, vermiform, and distended organelles identified as normal and pathological mitochondria on the basis of shape, size, number, investing membranes, oxygen-labeling, and peri-nuclear deployment. Many of the large, heterogeneous autophagic vacuoles within the cytoplasm of HMOX1-transfected cells displayed prominent iron signals. The latter also occurred within the outer rims of the CA-like inclusions, akin to reports of histochemically detectable iron within human and experimental CA (Cisse and Schipper 1995; Schipper 1998a). HMOX1 transfection had no significant effects on immunoreactive transferrin receptor, ferritin, ferroportin, or iron response protein 2 levels or iron response protein 1 binding, suggesting that HO-1 activity promotes mitochondrial macroautophagy and deposition of redox-active iron in cultured astroglia independently of classical iron mobilization pathways (Zukor et al. 2009). As HO-1 is primarily an ER enzyme, the subcellular iron pathway involved may entail (active or passive) delivery of iron from the ER to the mitochondria via iron-carrier proteins in the cytosol. Of note, ultrastructural studies of the biogenesis of iron-laden astroglial inclusions have revealed occasional close apposition and possibly membrane fusion of ER strands with osmiophilic cytoplasmic granules derived from damaged mitochondria (Brawer and Sonnenschein 1975; Brawer et al. 1978, 1994b). Thus, in stressed astroglia, putative organellar fusions or transient ‘kiss-and-run’ relationships of the ER with mitochondria may facilitate delivery of HO-1 enzyme, and hence heme-derived iron, to the mitochondrial compartment. Another intriguing possibility is that HO-1 enzyme native to mitochondria may contribute to the sequestration of heme-derived iron in this organelle. HO-1 protein has been detected in mitochondria of rat pulmonary endothelial cells (Kim et al. 2004) and liver cells (Converso et al. 2006) where it appears to modulate heme content and metabolism. Whether astroglial mitochondria similarly contain endogenous HO-1 remains unknown.

Alzheimer disease

  1. Top of page
  2. Abstract
  3. Heme oxygenases
  4. Brain aging
  5. Glial senescence
  6. Alzheimer disease
  7. Parkinson disease
  8. Other neurological disorders
  9. Conclusions
  10. Acknowledgements
  11. References

AD pathology: role of iron deposition, oxidative stress and mitochondrial injury

Alzheimer disease is an aging-associated dementia featuring progressive neuronal degeneration, gliosis, and the accumulation of intracellular inclusions (NFTs) and extracellular deposits of amyloid (senile plaques) in the basal forebrain, hippocampus, and association cortices (Selkoe 1991). OS and mitochondrial deficits have been consistently implicated in the pathogenesis of this disorder (Reichmann and Riederer 1994; Beal 1995; Mattson 2002). OS in the AD brain may be secondary to ROS production by effete mitochondria, the accumulation of β-amyloid (Butterfield 2002), pro-inflammatory cytokines (TNFα, IL-1β) and NO released from activated microglia (McGeer and McGeer 1995), and pathological deposition of redox-active iron in the basal forebrain and association cortices (Schipper 1998b; Sayre et al. 2001). In both the AD and PD brain, enhanced expression of tissue ferritin, the major intracellular iron storage protein, parallels the topography of the excess iron and largely implicates non-neuronal (glial, endothelial) cellular compartments (Schipper 1998b). Regional concentrations of transferrin binding sites remain unchanged or vary inversely with the elevated iron stores, suggesting that the transferrin pathway of iron mobilization, important for normal iron delivery to most peripheral tissues, plays a minimal role in the pathological sequestration of this redox-active metal in AD and PD brain (Schipper 1999).

HO-1 expression in AD brain

In AD brain, HO-1 protein co-localizes to neurons, GFAP-positive astrocytes, choroid plexus epithelial cells, ependymocytes, NFTs, senile plaques, CA and some vascular smooth muscle and endothelial cells (Smith et al. 1994; Yan et al. 1994; Schipper et al. 1995). In one study, 86% of GFAP-positive astrocytes residing within the AD hippocampus exhibited HO-1 immunoreactivity, whereas the fraction of hippocampal astroglia expressing HO-1 in age-matched normal tissue was in the range of only 6–7%. Similarly, western blots of protein extracts derived from AD hippocampus and temporal cortex revealed strong HO-1 bands, whereas the latter were faint or undetectable in normal control preparations (Schipper et al. 1995). Importantly, significant augmentation of astroglial HO-1 expression was also documented in post-mortem brain specimens procured from individuals with mild cognitive impairment (MCI), a frequent harbinger of incipient AD (Schipper et al. 2006). In the latter, glial HO-1 immunoreactivity in temporal cortex correlated with the burden of neurofibrillary pathology. Furthermore, astroglial HO-1 expression in the temporal cortex was associated with decreased scores for global cognition, episodic memory, semantic memory, and working memory; hippocampal astroglial HO-1 expression was associated with lower scores for global cognition, semantic memory, and perceptual speed. The results of the MCI study strongly suggest that glial HO-1 induction is a relatively early event in the pathogenesis of sporadic AD (Schipper et al. 2006). OS provoked by the aforementioned stimuli may be responsible for the elaboration of HO-1 in the Alzheimer-diseased and MCI cerebral cortex and hippocampus. On the basis of data adduced previously (vide supra), up-regulation of HO-1 in AD/MCI-affected astroglia may perpetuate intracellular OS and contribute to pathological iron trapping, bioenergetic (mitochondrial) failure and CA formation characteristic of Alzheimer-diseased neural tissues.

Heme oxygenase-1 suppressor factor in AD plasma

In contradistinction to the enhanced HO-1 expression documented in AD brain parenchyma, HO-1 protein or mRNA levels are suppressed in AD CSF, choroid plexus epithelial cells, blood mononuclear cells, and plasma compared to control values (Schipper et al. 2000, 2001; Anthony et al. 2003). Several laboratories (Kimpara et al. 1997; Shibata et al. 2009) failed to disclose over-representation of specific HMOX1 promoter polymorphisms in patients with sporadic AD (vs. normal elderly or PD subjects), arguing that primary genetic determinants of HO-1 expression are not responsible for the sub-normal levels of HO-1 protein or mRNA in the blood, CSF and choroid plexus of patients with sporadic AD. The diminished CSF HO-1 concentrations observed in AD may reflect decreased synthesis and release of the protein by the choroid plexus or, alternatively, curtailed delivery (transudation) from the systemic circulation. Using serial affinity chromatography and an astroglial bioassay (HO-1 mRNA response to CSH challenge), we demonstrated the presence of augmented HO-1 mRNA suppressor (HOS) activity in plasma of patients with sporadic AD relative to normal elderly and neurological controls; mean plasma HOS activity of the MCI patients was intermediate between normal elderly control and AD values (Maes et al. 2006; Schipper 2007). The HOS factor was determined to be a 50–100 kDa heat-labile, heparin-binding glycoprotein that was unrelated to antioxidant ingestion, plasma total antioxidant capacity, circulating cortisol levels or apolipoprotein E (apoE) ε4 carrier status. Proteins found to be specifically enriched in AD plasma fractions by 2D gel electrophoresis were identified by matrix-assisted laser desorption time of flight mass spectrometry as isoforms of human α1-antitrypsin (AAT), hemopexin, and transferrin (Yu et al. 2003). Of these HOS candidates, only exogenous AAT recapitulated HOS bioactivity in our astroglial assay. Moreover, AAT levels were significantly elevated in AD plasma and correlated with HOS activity and mini-mental state exam scores. The HOS activity of AD plasma was significantly attenuated by AAT immunodepletion. In AD brain, AAT immunoreactivity was augmented relatively to control tissues and co-distributed with HO-1. In addition, the PI*M3 allele of AAT was reported in a Polish study to occur with increased frequency in patients with sporadic AD relative to the general population (Kowalska et al. 1996). We conjectured that the HOS activity of AAT, derived from the circulation or elaborated within the CNS (Gollin et al. 1992), may limit HO-1-dependent derangement of cerebral iron homeostasis and account for diminished HO-1 expression in AD choroid plexus and peripheral tissues (Maes et al. 2006). Conceivably, net HO-1 mRNA and protein levels remain elevated in AD brain because the HOS activity of the AAT is overwhelmed by the pro-oxidant effects of β-amyloid, Th1 cytokines and other endogenous inducers of the HMOX1 gene. In the absence of similar provocation outside the affected brain parenchyma, AAT suppresses HMOX1 gene expression in AD blood and choroid plexus to sub-normal values. The biomarker potential of circulating HO-1/HOS for the diagnosis of MCI and AD is considered elsewhere (Schipper 2007). Finally, AAT may not be the only factor participating in the negative regulation of HO-1 in AD brain: Takahashi and co-workers demonstrated protein–protein interactions between amyloid precursor protein (APP) and both HO-1 and HO-2 that, in cortical neurons, attenuated enzymatic activity by 25–35% (Takahashi et al. 2000).

AD, HO-1, and sterol homeostasis

CNS cholesterol metabolism is pivotal for normal neurodevelopmental processes as well as for membrane repair (neuroplasticity) in AD and other neurological disorders (Bogdanovic et al. 2001; Refolo et al. 2001; Shobab et al. 2005). In AD, astrocytes and other neuroglia synthesize cholesterol de novo and recycle sterols released from degenerating neurons (Jurevics and Morell 1995; Pfrieger 2003). Pro-oxidant conditions prevailing in AD-affected neural tissues promote the formation of oxysterols and other cholesterol oxidation products. The latter are effective ligands of liver X activated receptor (LXR) nuclear receptors, major regulators of genes subserving cholesterol homeostasis. Oxysterol-mediated LXR activation may impact AD pathogenesis by (i) induction of glial apoE biosynthesis, further enhancing cholesterol re-distribution and removal; (ii) up-regulation of other genes governing key sterol elimination pathways; and (iii) impacting APP processing which may serve to attenuate β-amyloid toxicity (Vaya and Schipper 2007).

Heme oxygenase-1 over-expression in rat astroglia by transient transfection of HMOX1 cDNA for 3 days (enhancing HO activity three- to fourfold) significantly decreases intracellular cholesterol concentrations and increases the levels of at least four oxysterol species (measured by gas chromatography/mass spectrometry) compared to untreated control cultures and HMOX1-transfected cells exposed to SnMP. OS was shown to be augmented in mitochondria and culture media derived from the HMOX1-transfected astrocytes, as evidenced by enhanced oxidation of the synthetic reporter molecules, linoleoyl tyrosine, linoleoyl tyrosine cholesterol ester, and linoleoyl tyrosine deoxyguanosyl ester (Vaya et al. 2007). Furthermore, HMOX1 transfection in these cells promoted de novo cholesterol biosynthesis and stimulated cholesterol efflux, effects abrogated by SnMP. CO released from the synthetic CO donor molecule, CO releasing molecule-3 mimicked the effects of HMOX1 transfection on glial cholestrerol biosynthesis; and the combination of exogenous CO and iron stimulated cholesterol efflux. Treatment with pharmacological LXR antagonists implicated LXR activation in the modulation of cholesterol homeostasis by these heme degradation products. In these cells, cholesterol efflux exceeded biosynthesis such that net glial cholesterol content was diminished (supplying further stimulus for de novo cholesterol biosynthesis) (Hascalovici et al. 2009). Thus, in diseased neural tisssues, glial HO-1 induction may transduce ambient OS and other noxious stimuli into altered patterns of cholesterol homeostasis in a manner favoring enhanced production and delivery of the sterol for neuronal membrane repair.

In a related investigation, HO-1 protein levels and various sterols were measured (by enzyme-linked immunosorbent assay and gas chromatography/mass spectrometry, respectively) in post-mortem human frontal cortex derived from subjects with sporadic AD, MCI and no cognitive impairment enrolled in the Religious Orders Study (Hascalovici et al. 2008). In the AD and some MCI (but not in cognitively normal) specimens, HO-1 levels correlated significantly with decreased cholesterol levels, increased cholesterol precursor levels and increased oxysterol concentrations. In addition, total cholesterol content, cholesterol precursors and specific oxysterols correlated with disease state, increasing neuropathological burden, neuropsychological deficits, age, apoE ε4 genotype and gender. These observations, in conjunction with our in vitro data (vide supra), suggested a model for altered sterol homeostasis in AD brain (Hascalovici et al. 2008): (i) in the normal aging brain, sterol homeostasis is maintained by pathways governing baseline cholesterol biosynthesis, cholesterol efflux and oxysterol formation; glial HO-1 expression and its influence on sterol metabolic pathways are minimal. In MCI and early (compensated) AD, enhanced HO-1 expression stimulates cholesterol biosynthesis, oxysterol formation and cholesterol efflux by the astroglial compartment. Glial cholesterol efflux (to sites of neuronal repair and for egress across the blood–brain barrier) exceeds biosynthesis and total cholesterol levels in affected brain are normal or diminished. In advanced (de-compensated) AD, a massive increase in the free cholesterol pool (derived from widespread neuronal degeneration) saturates sterol efflux mechanisms resulting in increased brain cholesterol levels which, in turn, exacerbate amyloid deposition and neurodegeneration in this disease (Vaya and Schipper 2007). HO-1 may also interact with the presence of the apoE ε4 allele, a well-established and robust genetic risk factor of sporadic AD (Saunders et al. 1993; Strittmatter et al. 1993), as Jofre-Monseny et al. (2007) reported greater basal and lipopolysaccharide-stimulated HO-1 expression in apoE ε4-transfected murine macrophages in comparison with apoE ε3-transfected cells.

HO-1 and tau degradation

The ubiquitin-proteosome system (UPS) mediates turnover of normal, mis-folded and chemically modified (e.g. oxidized) intracellular proteins. Tau protein is degraded by the UPS in vitro and in vivo, and ubquitinated tau is present in NFTs. Proteosomal activity is reduced in AD brain and β-amyloid peptide inhibits the UPS in cultured cells (Ciechanover and Brundin 2003; Shastry 2003). Expression of HMOX1 by transient transfection triggers proteosomal tau degradation in M17 human neuroblastoma cells, an effect that can be blocked with SnMP or the proteosomal inhibitor, lactacystin (Song et al. 2009). Up-regulation of HO-1 in AD brain (Smith et al. 1994; Premkumar et al. 1995; Schipper et al. 1995) may therefore serve to stimulate a failing UPS in an attempt to limit the accumulation of toxic tau aggregates. Interestingly, HO-1 may keep tau levels in check by more than one mechanism as Smith and co-workers (Takeda et al. 2000) reported transcriptional suppression of tau in neuroblastoma cells transfected with human HO-1 cDNA. Finally, the impact of HO-1 may not be restricted to native tau. The P301L tau mutation, found in kindreds with heritable frontotemporal dementia-parkinsonism linked to chromosome 17, promotes tau aggregation in vitro and in transgenic mouse models of the disease (Lewis et al. 2000; Rizzu et al. 2000; Vogelsberg-Ragaglia et al. 2000). HO-1 over-expression in M17 cells elicits proteosomal catabolism of P301L tau protein to the extent observed with native tau (Song et al. 2009), countering the assumption that the P301L mutation fosters tau pathology by rendering the protein less susceptible to degradation (Lee et al. 2001). A very different situation prevails concerning HO-1-stimulated degradation of native vs. mutant α-synuclein, as discussed below.

Parkinson disease

  1. Top of page
  2. Abstract
  3. Heme oxygenases
  4. Brain aging
  5. Glial senescence
  6. Alzheimer disease
  7. Parkinson disease
  8. Other neurological disorders
  9. Conclusions
  10. Acknowledgements
  11. References

Neuropathology of PD

Idiopathic PD is a late-onset movement disorder of uncertain etiology characterized pathologically by progressive degeneration of dopaminergic neurons in the substantia nigra pars compacta, formation of intraneuronal fibrillar inclusions (Lewy bodies) in this cell population, and variable depletion of noradrenaline and serotonin in other brain stem nuclei. As in the case of AD, there is abundant evidence of oxidative tissue damage, bioenergy deficits and transferrin-independant iron sequestration in PD-affected brain regions (Schipper 1998b, 1999).

HO-1 expression in PD brain

Moderate HO-1 immunoreactivity is observed in dopaminergic neurons of both the normal and Parkinson-diseased substantia nigra (Schipper et al. 1998a). In PD specimens, cytoplasmic Lewy bodies within affected dopaminergic neurons are prominently decorated with immunoreactive HO-1 (Castellani et al. 1996; Schipper et al. 1998a). The proportion of GFAP-positive astroglia that express HO-1 in the PD nigra was found to be significantly increased (77.1%) relative to age-matched control specimens (18.7%), whereas percentages of GFAP-positive astroglia co-expressing HO-1 in other subcortical nuclei, such as the caudate, putamen and globus pallidus, were relatively low in both groups (Schipper et al. 1998a). DA-derived pro-oxidant intermediates, MPTP-like neurotoxins generated endogenously or stemming from the environment, and microglia-derived cytokines and NO are plausible inducers of astroglial HMOX1 gene expression in the PD nigra (Rieder et al. 2004; Schipper 2004a). On the basis of the aforecited in vitro data, the enhanced glial HO-1 activity may promote the transferrin receptor-independent accumulation of iron and mitochondrial electron transport (complex I) deficits consistently reported in the basal ganglia of PD subjects (Beal 1996; Schipper 2001).

Pro-toxin bioactivation and PD

Iron sequestered in astroglial mitochondria as a result of HO-1 activity behaves as a pseudo-peroxidase that promotes the oxidation of catechol-containing compounds (DA, 2-hydroxyestradiol) to potentially neurotoxic ortho-semiquinone radicals (Schipper et al. 1991; Schipper 2001). The redox-active glial iron also mediates bioactivation of the pro-toxin, MPTP, to the dopaminergic neurotoxin, MPP+, in the face of monoamine oxidase blockade (Di Monte et al. 1995). Neuron-like PC12 cells grown on a substratum of astrocytes replete with mitochondrial iron induced by CSH pre-treatment (Frankel and Schipper 1999) or HMOX1 transfection (Song et al. 2007) were noted to be far more susceptible to DA/H2O2-related killing than PC12 cells co-cultured with control, ‘iron-poor’ astroglia. Oxidative PC12 cytotoxicity was significantly attenuated by the administration of SnMP, iron chelators (deferroxamine, phenanthroline) or natural antioxidants (ascorbate, trans-resveratrol, and melatonin) attesting to the canonical role of HO-1 and iron-mediated redox reactions, respectively, in this paradigm (Frankel and Schipper 1999; Song et al. 2007). Incubation of PC12 cells in astrocyte-conditioned media resulted in similar patterns of PC12 cell death observed in the co-culture experiments except on a smaller scale (Song et al. 2007). Thus, in addition to direct neuronal–glial interactions, HMOX1-transfected astroglia exposed to DA/H2O2 may release neurotoxins (e.g. CO, Fe, and quinones) to the extracellular space and/or deprive nearby PC12 cells of glia-derived trophic substances (e.g. glial-derived neurotrophic factor). The glial mitochondriopathy may also pre-dispose to neural injury by diminishing ATP production and thereby compromising important energy-dependent processes such as uptake of excitotoxic neurotransmitters (glutamate) from the synaptic cleft and de novo biosynthesis of glutathione destined for neuronal delivery (Aschner 2000).

In contrast to the vulnerability of the PC12 cells, DA/H2O2 injury to either the CSH pre-treated, HMOX1-transfected or control astroglia themselves was negligible. The capacity of stressed astroglia to up-regulate robust antioxidant defenses, elaborate cytoprotective HSPs, and successfully convert to anaerobic metabolism may account for the resiliency of these cells in the face of oxidative challenge (Schipper 2004a). The data summarized herein suggest that progressive, HO-1-mediated trapping of glial mitochondrial iron within subcortical brain regions may be an important mechanism predisposing the senescent nervous system to PD and other free radical-related neurodegenerative disorders (Schipper 2001, 2004a). This conjecture is consistent with reports of enhanced neurotoxicity and parkinsonism in old animals exposed to DA (Cantuti-Castelvetri et al. 2003), MPTP (Jarvis and Wagner 1990; Ali et al. 1993), amphetamines (Vasilev et al. 2003) and manganese (Desole et al. 1995) compared to respective lesions incurred in younger counterparts.

HO-1 and α-synuclein degradation

Dysfunction of the UPS contributes to the formation of toxic protein aggregates and cellular inclusion bodies in PD and other aging-related human neurodegenerations (Rubinsztein 2006). In particular, aberrant parkin activity and other identified UPS deficiencies may foster protofibrillization and aggregation of α-synuclein, a major component of the hallmark Lewy body (Ciechanover and Brundin 2003; McNaught and Olanow 2003; McNaught et al. 2003). Transfection of HMOX1 in M17 cells promotes intracellular degradation of endogenous and co-transfected wild-type α-synuclein which is abrogated by treatment with the proteasomal inhibitor, lactacystin (Song et al. 2009). Blockade of HO-1-related α-synuclein degradation by SnMP, methylene blue or deferroxamine (but not bilirubin) provided evidence that heme-derived iron and CO are responsible for proteasomal activation and attendant α-synuclein degradation in these cells. These data also indicate that the catabolism of α-synuclein is dependant on HO enzymatic activity rather than direct HO-1/α-synuclein protein-protein interactions or some other HO-1 function. These findings suggest that up-regulation of HO-1 in the PD nigra (Schipper et al. 1998a) may represent an adaptive response to enhance the proteasomal degradation, and thereby pre-empt the toxic aggregation, of α-synuclein. In contrast to wild-type α-synuclein, the A30P mutant α-synuclein, implicated in some families with autosomal-dominant PD (Polymeropoulos et al. 1997; Kruger et al. 1998), exhibited marked resistance to proteosomal degradation when co-expressed with HMOX1 in M17 cells (Song et al. 2009). As non-degradable, aggregated forms of α-synuclein are cytotoxic (Volles and Lansbury 2002), the resistance of the A30P protein to HO-1-mediated proteasomal catabolism may be a factor promoting Lewy body formation and parkinsonism in individuals bearing this mutation.

Other neurological disorders

  1. Top of page
  2. Abstract
  3. Heme oxygenases
  4. Brain aging
  5. Glial senescence
  6. Alzheimer disease
  7. Parkinson disease
  8. Other neurological disorders
  9. Conclusions
  10. Acknowledgements
  11. References

In addition to AD and PD, HO-1 has been implicated in the pathobiology of numerous other degenerative and non-degenerative CNS conditions. Among the neurodegenerative disorders, the focus of this review, immunoreactive HO-1 protein localizes to diseased motor neurons in amyotrophic lateral sclerosis (Calingasan et al. 2005), Pick bodies in subjects with frontotemporal dementia, NFT in cases of progressive supranuclear palsy, and ballooned neurons in corticobasal degeneration (Castellani et al. 1995). On the basis of evidence discussed herein, it is conceivable that induction of HO-1 may contribute to the pathological brain iron deposition, oxidative substrate damage and mitochondrial lesions that have been documented in these late-onset human neurodegenerations (Reichmann and Riederer 1994; Chang et al. 1995; Connor 1997; Delisle et al. 2000). The exquisite inducibility of the Hmox1 gene and the highly pleotropic bioactivities of all three heme degradation products have also suggested important roles for HO-1 in the pathogenesis of various neuroinflammatory [multiple sclerosis (MS), falciparum malaria], cerebrovascular (ischemic and hemorrhagic stroke), traumatic (cerebral contusions) and neuro-oncological (malignant glioma) disorders. These conditions are not primarily neurodegenerative in nature and interested readers are referred to a previous publication addressing these topics (Schipper 2004c). However, a brief consideration of MS, an autoimmune demyelinating disorder of central white matter, is germane to the current discussion given recent recognition of the significant post-inflammatory, ‘degenerative’ phases of the illness (Trapp and Nave 2008). In a neuropathological survey (Mehindate et al. 2001), the fraction of GFAP-positive astrocytes expressing HO-1 (57.3%) in spinal cord plaques derived from patients with MS was noted to be substantially higher than that computed in the spinal white matter of normal subjects (15.4%). Glial HMOX1 induction in MS may be secondary to the enhanced release of IL-1β, TNFα (Mehindate et al. 2001) or myelin basic protein (Businaro et al. 2002) within the diseased tissues. In primary astrocyte cultures, TNFα (20 ng/mL) or IL-1β (20 ng/mL) stimulated the accumulation of non-transferrin 55Fe by astroglial mitochondria, effects that were abrogated by co-administration of HO inhibitors, mitochondrial pore anatagonists or antioxidants (Mehindate et al. 2001). Thus, up-regulation of HO-1 in astrocytes may give rise to the aberrant iron deposits and electron transport chain defects reported in the vicinity of MS plaques (Lu et al. 2000; Mehindate et al. 2001; Levine and Chakrabarty 2004) and in experimental autoimmune encephalomyelitis, a rodent model of the disease (Levine and Chakrabarty 2004). Stahnke et al. (2007) reported increased levels of HO-1 protein in oligodendrocytes (myelin-producing cells) in early MS lesions and in astrocytes, microglia, and macrophages in acute disseminated encephalomyelitis. These investigators also provided evidence that acute HO-1 expression in cultured OLN-93 oligodendroglia confers cytoprotection in the face of H2O2 stress, whereas longer-term induction of the enzyme in these cells exacerbates oxidative injury to mitochondria and the microtubular network.

Conclusions

  1. Top of page
  2. Abstract
  3. Heme oxygenases
  4. Brain aging
  5. Glial senescence
  6. Alzheimer disease
  7. Parkinson disease
  8. Other neurological disorders
  9. Conclusions
  10. Acknowledgements
  11. References

Delineation of the diverse roles of HO-1 in brain senescence, aging-related human neurodegenerative disorders and other CNS conditions (summarized in Fig. 2) has progressed substantially since we last reviewed this topic in 2004 (Schipper 2004a,b,c). Whereas the acute induction of this enzyme in neural and other tissues is predominantly cytoprotective in nature, protracted or repeated up-regulation of the Hmox1 gene in astrocytes, oligodendroglia and possibly neurons may perpetuate cellular dysfunction and demise in many chronic degenerative and neuroinflammatory conditions long after provocative stimuli have dissipated. In the case of normal brain aging, AD, PD, and other late-life neurodegenerations, there is converging evidence that the associated tissue iron sequestration, intracellular OS and mitochondrial insufficiency may constitute a single cytopathological ‘lesion’ that forms downstream from the sustained action of HO-1 within the astrocytic compartment. The fact that induction of the Hmox1 gene is both a cause and consequence of intra-glial OS in our models led to the prediction that iron homeostasis will invariably be disturbed in aging and diseased CNS tissues harboring mitochondrial lesions (a key source of endogenous ROS); neural tissues exhibiting primary or secondary iron overload will exhibit obligatory bioenergetic failure; and at least some portion of the sequestered metal will be confined to the mitochondrial matrix (Schipper 2004a). The fairly ubiquitous co-registration of these pathological features over a wide range of human and experimental neurological afflictions is consistent with this formulation (Schipper 2004a,b).

image

Figure 2.  Putative roles of HO-1 in aging-related human neurodegenerative disorders. Implications for AD and PD are emphasized, although neural HMOX1 induction may exert homologous influences on brain structure and function in other degenerative and neuroinflammatory conditions. Sustained or repeated up-regulation of HO-1 may exacerbate certain degenerative processes (red) while concomitantly activating various neuroprotective responses (green), a situation possibly reflecting antagonistic pleiotropy. Dystrophic effects of HO-1 are represented by mitochondrial damage and ferrous iron sequestration in astrocytes which may, in turn, pre-dispose to bioenergetic failure, pro-toxin bioactivation, excitotoxicity, macroautophagy, and corpora amylacea formation. Neuronal degeneration stimulates microglial activation resulting in the release of ROS, NO, and proinflammatory cytokines (TNFα, IL-1β). The latter may further induce HMOX1 in indigent astroglia, completing a self-sustaining loop of pathological cellular interactions that may perpetuate oxidative damage and mitochondrial insufficiency within senescent and degenerating neural tissues. Genetic and environmental risk factors may confer disease specificity by superimposing unique pathological signatures on this core lesion. In AD, AAT and APP may suppress glial HO-1 transcription and activity, respectively, in an attempt to restore central heme/iron homeostasis. Potential benefits conferred by the up-regulation of HO-1 include (i) stimulation of neuronal tau and α-synuclein degradation by the UPS to curtail protein aggregate toxicity (in AD and PD, respectively) and (ii) modulation of sterol homeostasis in favor of neuronal repair/neuroplasticity. See text and references for full details. Solid lines (arrows) denote causation or stimulation; dashed lines with spheres denote inhibition. AAT, α1-antitrypsin; AD, Alzheimer disease; APP, amyloid precursor protein; CHOL, cholesterol; CO, carbon monoxide; DA, dopamine; Fe2, ferrous iron; GSH, glutathione; HO-1, heme oxygenase-1; IL-1β, interleukin-1β; MPP+, methyl-4-phenylpyridinium; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; NO, nitric oxide; OS, oxidative stress; PD, Parkinson disease; TNFα, tumor necrosis factor-α; UPS, ubiquitin-proteosome system. Electron micrographs are reproduced from Zukor et al. (2009) with permission.

Download figure to PowerPoint

The aforementioned, potentially dystrophic effects of HO-1 in aging and degenerating neural tissues may inform the development of neuroprotective and other disease-modifying strategies. To the extent that pathological brain iron mobilization in these conditions is HO-1-dependant, attempts to minimize iron-related neurotoxicity by targeting components of classical iron-regulatory pathways (transferrin receptor, iron response protein 1/2, and ferritin) may prove futile. On the other hand, an enzymatic (HO-1) basis for oxidative mitochondrial injury in degenerating or inflamed neural tissues may be considerably more amenable to pharmacological management than free radical damage accruing from unregulated chemical processes (Droge 2002). Several metalloporphyrin inhibitors of HO activity have been employed in the clinic for the control of neonatal hyperbilirubinemia (jaundice) and certain adult liver conditions (Kappas et al. 1993; Drummond and Kappas 2004) and could potentially be adapted for neurotherapeutic use. Along similar lines, we are currently investigating a novel class of CNS-permeable, non-peptidic small molecule inhibitors which are highly selective for the HO-1 isoenzyme. As previously demonstrated with SnMP (above), a lead compound attenuated glial HO-1 activity and mitochondrial protein carbonyl formation resulting from HMOX1 transfection in a dose-dependent manner (Schipper et al. 2009). The data reviewed in this article support our earlier contention (Schipper 2004a,b) that targeted suppression of glial HO-1 expression or activity may ameliorate iron-mediated toxicity and bioenergy deficits in the brains of subjects with AD, PD, and other human neurodegenerative disorders. That HO-1 and other proteins may acquire nefarious functions in the senescent mammalian CNS should not be surprising as there is little evolutionary pressure in post-reproductive life to thwart such eventualities, a possible example of antagonistic pleiotropy (Williams 1957; Kirkwood and Rose 1991).

Several important caveats and further pre-clinical research demand consideration before advocating central HO-1 inhibition as a potential therapeutic modality for these neurodegenerative conditions. As described above, HO-1 exhibits Janus-faced behavior as a determinant of neuronal survival in various disease models. On theoretical grounds, there is concern that suppression of its activity in the AD or PD brain may render the latter increasingly vulnerable to the impact of intercurrent ischemic stroke or other acute injury. In the past, we viewed such risk as potentially acceptable were the intervention to mitigate substantially the chronic neurodegenerative process and inexorable clinical decline. However, the situation has since grown more complex given recent evidence suggesting that, within a given neurodegenerative disorder, the up-regulation of HO-1 may ameliorate certain aspects of of the disease process while concomitantly aggravating others. Thus, the effects of HO-1 on brain sterol/oxysterol homeostasis and on the proteosomal degradation of α-synuclein and tau may represent adaptive responses in AD and PD, while the accruing iron deposition and mitochondrial damage feed into the neurodegenerative cascade (Fig. 2). As such, the inhibitory effects of AAT on HO-1 transcription (Maes et al. 2006) and APP/HO-1 interactions on HO activity (Takahashi et al. 2000) may prove detrimental in AD when implicating affected neurons, but beneficial when operating within the astrocytic compartment. Addressing this potential therapeutic conundrum, Syapin (2008) suggested that selective short interfering RNA delivery could be exploited to achieve HO-1 knock-down within specific CNS cell populations (e.g. astrocytes), guided by recent successes with targetted short interfering RNA delivery to neurons (Kumar et al. 2007) and hepatocytes (Rozema et al. 2007).

Intriguingly, reduced locomotion, exploratory behaviour (Maines et al. 1998), and impaired spatial navigation (Morgan et al. 1998) have been documented in transgenic mice over-expressing rat HO-1 under the control of a neuron-specific enolase promoter. Studies have been initiated in our laboratory to evaluate cognitive/locomotor behavior and neuropathology in (i) transgenic mice engineered to over-express HMOX1 selectively within the mature astroglial compartment and (ii) an APP mouse model of AD in the presence and absence of pharmacological HO-1 blockade (Schipper et al. 2009). The latter paradigms may help resolve some of the unsettled issues raised in this review and determine the effects of sustained glial HO-1 induction on brain structure and neurological health in intact organisms.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Heme oxygenases
  4. Brain aging
  5. Glial senescence
  6. Alzheimer disease
  7. Parkinson disease
  8. Other neurological disorders
  9. Conclusions
  10. Acknowledgements
  11. References

The author thanks Adrienne Liberman for assistance with the preparation of this manuscript. Contributions by the host laboratory were enabled by grants to HMS from the Canadian Institutes for Health Research, the Fonds de la Recherche en Santé du Québec, the Alzheimer’s Association (US), the Parkinson Foundation of Canada, and Osta Biotechnologies (Montreal).

References

  1. Top of page
  2. Abstract
  3. Heme oxygenases
  4. Brain aging
  5. Glial senescence
  6. Alzheimer disease
  7. Parkinson disease
  8. Other neurological disorders
  9. Conclusions
  10. Acknowledgements
  11. References