AD is the most common cause of dementia. Among the many other pathological hallmarks, neuronal loss is considered as one of the main causative events of dementia in this catastrophic disease. The neuronal death in AD may result directly and/or indirectly from the triggering insults caused by Aβ toxicity, glutamate excitotoxicity, long-lasting oxidative stress, DNA damage, and elevation of intracellular calcium levels [3
]. Thus, the mode of cell death in AD remains a matter of controversy [21
], and it is possible that both apoptotic and non-apoptotic cell death coexist in the brains of affected patients.
Apoptosis, a type programmed cell death (PCD), is not always harmful. It is widely involved in normal regulatory life events and to some extent participates in pathologic events as well. Since the duration of the apoptotic process is relatively short, only a few cells that are undergoing apoptosis can be detected at a single moment in the course of a chronic neurodegenerative disease [22
]. Previously, it was generally considered that apoptotic neuronal death in chronic neurodegenerative disease, e.g., AD, Parkinson's disease, etc., is associated with classical caspase mediated cell death [23
]. However, in part, it was suggested that the caspase-independent pathway might also participate in the pathogenesis of the disease [27
]. According to Yang et al. [29
], occasional neurons in brain sections of older PS/APP mice displayed abnormal morphological changes such as cell shrinkage, condensed nucleus and cytoplasmic organelles, and evidence of plasmalemmal blebbing. These features and the appearance of corkscrew-like dendrites in some neurons correspond to so-called "dark neurons." On sections processed with anti-activated caspase-3 antibodies, few, if any, silver-gold particles were detected in dark neurons, indicating that caspase 3 is not activated in the dark neurons. Thus, the existence of at least two modes of neurodegeneration in the same mouse model underscores the complexity of cell death patterns in chronic neurodegenerative disease. Cross talk is extensive between different cell death pathways, which include multiple types of caspase-dependent and caspase-independent programmed cell death. AIF appears to play an important role in acute neural tissue damage induced by trauma, hypoglycemia, transient ischemia, and chronic neurodegenerative diseases. Bahi et al. [30
] demonstrated that experimental ischemia in embryonic cardiomyocytes triggers caspase activation, but the mode of cell death switches into caspase-independent PCD in differentiated cardiomyocytes. Although it is the result of in vitro
procedures, it indicates that the cell death mechanisms are different depending on the cell's status. Unlike in vivo
, in vitro
study indicates that the forms of cell death are the same and that the translocation of AIF is not caused by nuclear shrinkage [31
], because AIF is clearly seen in the nucleus at early points in time when the nucleus is clearly distinguishable from cytoplasmic structures. For many years, there has been some debate about whether there is apoptotic cell death in AD and whether it is in anyway related to cognitive impairment and dementia. While the study of apoptosis and apoptotic mechanisms is fairly easy to perform in cell culture systems and animal models, it is much more difficult in human organs including the brain. A few years ago, it was demonstrated that Aβ-induced neuronal apoptosis lead to AIF translocation from mitochondria to the nucleus in embryonic rat cortical cultures [32
]. Interestingly, both mitochondrial and nuclear effects of AIF have been observed in neuronal death associated with rodent aging and acute traumatic injuries [11
]. What was shown is consistent with a possible involvement of AIF in neuronal cell death on AD pathology. Interestingly, a study on AIF protein levels in the cortical areas of human brains at various ages and in AD was reported [14
]. Recently, a study by the same group provided the first demonstrations of increased nuclear translocation of AIF and its colocalization with NFT in the AD brain in midto late stages, but not in early stages [15
]. Studies on defining the plausible mechanism and role of AIF on cell death in AD human brains have not been widely performed thus far. In this article, the focus has been centered on the possibility of early involvement of AIF in neurodegeneration and cell death in AD. This study proposes that the dysfunction of neurons precedes the obvious pathological abnormalities and that the intraneuronal molecular changes caused by this stress-induced injury are subtle. The present study was noteworthy in documenting AIF protein expression in the hippocampal pyramidal neurons in human AD brains. Our study confirmed caspase 3 immunoreactivity in AD (data not shown). Our in vivo
results suggest that the AIF-mediated caspase-independent apoptotic pathway associated with dark neurons may be involved in the hippocampal pyramidal neuron death from early stages of AD.
The basal forebrain contains a population of large cholinergic neurons and is subdivided into four groups, Ch1-Ch4 [34
]. Ch4, which is very prevalent in human brains, most closely corresponds to the nucleus basalis of Meynert which is embedded in the substantia innominata, and at least 90% of the neurons in the nucleus basalis are cholinergic. It is well known in AD that the cholinergic neurons in the nucleus basalis of Meynert undergo a profound and selective degeneration from the early stages of the disease [36
The amygdala is a gray mass situated in the dorsomedial portion of the temporal lobe. It has been known that the amygdala is involved from in early AD. The amygdala nuclear complex is divided into two main nuclear masses: a corticomedial nuclear group and a basolateral nuclear group. The cholinergic innervation of the amygdala in human is severely affected in case of AD [38
]. It is still debated whether it is the reduction in cholinergic innervations or other mechanisms that directly influence on the neuronal death in the amygdala. Early studies reported that the neurons in the corticomedial nuclear group are extensively degenerated from the early stage, but the neurons in the basolateral nuclear group are not [39
]. In summary, in order to identify the involvement of AIF-mediated caspase-independent cell death in the pathology of AD, we investigated the AIF protein expression in the hippocampus, amygdala, and BFCN in accordance with AD progression. The major findings are that 1) AIF immunoreactivity is increased in the nucleus of hippocampal pyramidal neurons in the AD brain, and the expression of this molecule is altered in accordance with the progression of AD. 2) From the early stage of AD, AIF is translocated into the nucleus of apoptotic pyramidal neurons. 3) AIF immunoreactivity was increased in the neurons of corticomedial nucleus of the amygdala and BFCN in early AD and the neurons showing nuclear translocation of AIF were also increased in early AD in comparison to the age-matched control in those areas. Unfortunately, we do not have proper specimens for biochemical study which would be helpful in identifying the precise molecular mechanism involved in AIF induced cell death in AD, and thus our study is inevitably limited only to morphological study. Although our information is not enough to solve the exact molecular events involved, this study gives the first morphological evidence that from early stages of AD, caspase-independent neuronal death is also involved in the pathogenesis of AD. Our results may provide a novel concept for developing new therapeutic strategies against chronic neurodegenerative diseases. To identify, however, the precise molecular mechanism involved in AIF-mediated neuronal death in AD, further investigations will be required.