AD is a progressive, degenerative, dementing disorder and the sixth leading cause of the death in the US [1]. It is estimated that by the year 2050 the number of patients living with AD in the US may rise to 13-16 million while the worldwide prevalence could reach a staggering 1 in 85 or 75-100 million cases [2-5]. Pathological hallmarks of AD include extracellular deposits of β-amyloid (plaques), neurites containing hyperphosphorylated tau (neurofibrillary tangles) and loss of neurons in discrete regions of the basal forebrain, hippocampus and association cortices [6-11]. A largely overlooked pathological feature of AD, originally described over a century ago by Alois Alzheimer, is the accumulation of “adipose inclusions” or “lipoid granules” suggesting aberrant lipid homeostasis in this degenerative state [12]. Although medications (cholinesterase inhibitors) capable of conferring transient symptomatic relief are available for the management of AD, there is currently no disease-modifying intervention which unequivocally slows arrests or reverses the degenerative process.

Cholesterol (CH), its precursors and oxysterols have been consistently implicated in brain aging and neurodegeneration [2,13- 15]. The mechanisms underlying this relationship, however, remain ill-defined. In the CNS, sterols including CH subserves numerous biological functions essential for cell survival and homeostasis [16]. Mammalian CNS development is highly dependent on CH [17]; it is therefore not surprising that altered CH homeostasis occurs in a host of neurological conditions [18,19]. Glial CH can be enzymatically (via CH 7a-, 24- and 25-hydroxylase) or non-enzymatically (via oxidative stress) metabolized to oxysterols. On the one hand, augmented oxysterol concentrations may protect neural tissues by attenuating β-amyloiddeposition and by activation of the liver X receptor (LXR). The latter mediates up-regulation of apolipoprotein E (apoE) which, in turn, promotes glial CH efflux that facilitates neuronal membrane repair. However, LXR activation may also inhibit glial 3-hydroxy- 3-methylglutaryl-CoA (HMG-CoA) reductase and de novo CH biosynthesis. The latter establishes a negative servomechanism that limits the extent of oxysterol production and CH efflux by these cells. Moreover, primary neuronal cultures have been shown to be highly sensitive to 25-OH cytotoxicity, with lesser vulnerability to other common oxysterols such as 7-keto, 22-OH and 7β-OH [20]. Oxysterols may also promote cytotoxicity in primary astrocytes and apoptosis in microglial cells.

Several critical lines of evidence have implicated sterol (dys) regulation and the formation of oxysterols in the pathogenesis of AD. For example, genetic association studies have linkedpolymorphisms ofAPOE, SORL1, CLU andABCA7 genes, which directly or indirectlyimpact lipid homeostasis, with the development of AD [21,22]. In addition, dyslipidemia has been identified in several distinct populations as an important midlife risk factor for sporadic AD [23]. Although the mechanisms responsible for this relationship remain poorly understood, investigations conducted in our laboratory suggest that over-expression of heme oxygenase-1 in AD astrocytes may transduce a broad spectrum of noxious stimuli into altered, and potentially deleterious, patterns of brain sterol homeostasis.

Heme oxygenase-1 belongs to a family of enzymes widely recognized as dynamic sensors of cellular oxidative stress. HO-1 is a 32 kDa stress protein that mediates the catabolism of heme to biliverdin, free iron, and carbon monoxide (CO). Activating protein-1, activating protein-2, nuclear factor kappa B, heat shock factor and hypoxia inducible factor-1 binding sites, metal response elements (MtRE, CdRE) and antioxidant response elements (ARE) render the mammalian Hmox1 gene highly susceptible to up-regulation by a wide array of pro-oxidant and inflammatory stimuli. TheHmox1 gene is inducible by β-amyloid, hydrogen peroxideand Th1 cytokines and accumulates in CNS tissues affected by AD, Parkinson disease and other neurological conditions [24,25]. The up-regulation of HO-1 may confer cytoprotection by enhancing the breakdown of pro-oxidant heme to the radical scavenging bile pigments, biliverdin and bilirubin [26-29]. Under certain conditions, however, the heme-derived iron and CO may exacerbate intracellular oxidative stress and substrate damage by provoking free radical generation within mitochondria and other subcellular organelles [30]. Germane to the current discussion,HO-1 promotes mitochondrial iron sequestration, oxidative membrane injury and macroautophagy in rodent astroglia both in vitro and in the intact brain [25,31].

In AD-affected brain regions (temporal cortex, hippocampus, etc.), HO-1 protein is increased several-fold and co-localizes to neurons, astrocytes, neurofibrillary tangles, senile plaques, corpora amylacea and vascular smooth muscle and endothelial cells [32-34]. Potential inducers of HMOX1 in AD brain include reactive oxygen species generated by senescent mitochondria, β-amyloid, pro-inflammatory cytokines (TNFα, IL-1β) and nitric oxide (NO) released from activated microglia [35],and redox-active transition metals [25,31].

We hypothesized that HO-1 over-expression may impact sterol homeostasis in AD neural tissues by amplifying intracellular oxidative stress within the astrocytic compartment [25]. We determined that HO-1 over-expression in cultured rat astroglia significantly decreases intracellular CH concentrations and increases the levels of several oxysterol species. We identified CO and ironas mediators of the observed net decrease in total CH and increase in oxysterol concentrations [36]. In a subsequent study [37], we observed that HO-1 over-expression (2–3-fold) for 3 days resulted in a 30% rise in CH biosynthesisand a two-fold enhancement of CH efflux. Both effects were abrogatedby the competitive HO inhibitor, tin mesoporphyrin. CO, releasedfrom exogenous CORM-3, significantly augmented CH biosynthesis; a combination of CO and iron stimulated CH efflux; and heme-derived iron likely fostered oxysterol formation in our model. In addition, co-treatmentwith LXR antagonists implicated LXR activation in the modulation of CH homeostasis by heme degradation products (Figure 1). We then proceeded to measure HO-1 protein levels and various sterols/oxysterols (by GC-MS) in post-mortem human brain specimens ofpersons with sporadic AD, mild cognitive impairment (MCI) and no cognitive impairment (NCI) [38]. In AD and some MCIsamples, but not in NCI specimens, HO-1 levels correlated significantly with decreased total CH, increased CH precursors (together suggesting augmented CH efflux) and increased oxysterol concentrations [39], commensurate with our in vitro data. We also explored HO-1/sterol interactions in two in vivo animal models: a novel GFAP.HMOX1transgenic mouse which conditionally and selectively over-expresses the human HMOX1 gene in astrocytes [40,41]; and an established triple transgenic (3xTg-AD)mouse model of AD [42,43]. Total oxysterols significantly decreased as HO-1 expression increased in GFAP.HMOX1 mice (‘antioxidant’ HO-1 behavior), whereas total oxysterols increased as HO-1 expression increased in aged 3xTg-AD mice harboring AD-like pathology (a ‘pro-oxidant’ HO-1 effect). We concluded that in diseased (but not healthy aging) neural tissues, glial HO-1 induction transducesdystrophicstimuli,including many that have been linked to known or suspected AD risk factors [23], into altered patterns of sterol/oxysterol homeostasis. Together, our findings underscore a differential impact of HO-1 on the metabolism of brain sterols/oxysterols contingent on the level of HO-1 expression and the presence or absence of AD-like neuropathology[43].

Figure 1: A model depicting the impact of HO-1 over-expression on astroglial sterol homeostasis. Republished from Hascalovici et al. J Neurochem (2009) 108, 72-81 with permission.

We adduced evidence implicating glial HO-1 expression as a likely transducer of altered sterol homeostasis in human ADand across several experimental models of neurodegeneration. Our data indicate that HO-1 over-expression and the heme catabolic products, CO and Fe significantly impact CNS sterol homeostasis by influencing brainregulatory mechanisms governing sterol biosynthesis, efflux and oxidation. Further investigations along these lines should disclose whether aberrant HO-1/sterol dynamics akin to those described here constitute a ubiquitous pathway of pathological brain aging and degeneration and, hence,a robust target for therapeutic intervention. Several metalloporphyrin inhibitors of hemeoxygenase activity are already in clinical use for the management of neonatal hyperbilirubinemia (jaundice) and certain adult liver conditions and could be adapted for the treatment of CNS disorders. Recently, novel imidazole-based inhibitors that are relatively selective for the HO-1 isoform and display favorable blood-brain barrier penetration were shown to ameliorate cognitive dysfunction in the APPs we/PS1_ΔE9 mouse model of AD [43]. It will be important to ascertain whether this salutary effect on behavior is due, in whole or in part, to normalization of brain sterol/oxysterol homeostasis in these animals.