Serum levels of S100B and NSE proteins in Alzheimer's disease patients

Alzheimer's disease (AD) is a progressive brain disorder that results in memory impairment, personality alterations, global cognitive dysfunction, and functional impairments [1]. It is the most common dementia in the elderly, accounting for 60-80% of cases, and it is estimated to affect more than 4 million of USA citizens [2]. The lifespan of individuals diagnosed with AD is reduced by about 50% as compared with those of similar age without disease, and the survival expectancy is negatively associated with the severity of the disease at the time of diagnosis [3]. Furthermore, there is no definitive ante-mortem diagnostic test for AD, and when the clinical diagnosis is made, is difficult to access and following the course of neural cells loss [4].

The inherent hurdles of studying brain tissues in human populations, especially in vivo, are the permanent stimulus for finding peripheral markers of central nervous system (CNS) alterations. In this context, a number of proteins have been proposed as peripheral biochemical markers of neuronal damage and glial injury/activation, which peripheral assessment may represent a relevant step forward in the diagnostic and monitoring of CNS diseases [5‐9]. For this reason, the clinical usefulness of peripheral biochemical markers as complementary tool in the neuropsychiatric evaluation has claimed further attention.

S100B and neuron specific enolase (NSE) are brain derived proteins extensively studied as peripheral biochemical markers for brain injury [8‐11]. S100B is a calcium binding protein physiologically produced and released predominantly by astrocytes, whereas NSE is a cytoplasmatic glycolytic pathway enzyme, being the γγ isoform mainly neuronal [12, 13]. Since their levels may increase in CSF and/or blood in several brain pathologies, both proteins are considered to be markers of astrocytic damage/reaction (S100B) and neuronal damage (NSE) [14‐16]. Considering the prominent neural death observed in the course of AD, some studies have also attempted to clinically evaluate the levels of these proteins, resulting in contradictory findings [17‐21], but alternatively, experimental and human studies have strengthened the belief that S100B is implicated in the mechanisms underlying neurodegeneration in AD [22‐24]. Accordingly, it was reported an association between the deposition of cerebral amyloid beta protein and the presence of activated astrocytes over expressing S100B.

Furthermore, life-long over expression of S100B in Down syndrome patients and transgenic mice cause neuronal and glial morphological alterations similar to those found in AD patients, as well as behavioral deficits in animals [17‐19]. These conjectures have been the rational for studying CSF/serum S100B and NSE levels in AD as markers of neurodegeneration and severity of the disease. Also, it is important to take into account that, due to the insufficiency of professional, methodological and background conditions, the diagnosis of AD in non specialized centers may not be accurately performed, which encourage identifying potential peripheral biomarkers for AD, aiming an easier, accurate and widespread diagnosis [4].

The major aim of our study was to evaluate serum S100B and NSE levels (more readily assessed than CSF) in AD patients and control elderly individuals without pathological cognitive impairment. In addition, we searched for correlations among their levels and the severity of dementia, cognitive status and brain morphological changes accessed by MRI.

The main goal of this work was to investigate the course of brain neurodegenerative processes in AD patients through biochemical brain markers, neuropsychological and neuroimaging evaluations. Our main findings were: i) although serum NSE levels were not different between AD patients and control elderly individuals, serum of S100B levels were significantly lower in DA patients; ii) there were a positive correlation between S100B levels and AD severity (evaluated by CDR and MMSE) as well as a negative correlation between NSE levels and the severity of morphological brain alterations, evaluated by MRI.

In the last years the possibility of evaluating brain damage/activity through quantification of neuronal and glial derived proteins (such as S100B and NSE) in peripheral samples has gained appropriate attention in clinical and experimental settings [5, 14, 20, 32]. However, S100B and NSE proteins are also expressed in other non neural cell types under physiological and pathological conditions. Thus, the brain specificity of these proteins has been questionable by some works, including from our laboratory [33], when assessing serum samples. Moreover, recent epidemiological and clinico-pathologic data suggest overlaps between AD and cerebrovascular lesions that may magnify the effect of mild AD pathology and promote progression of cognitive decline or even may precede neuronal damage and dementia [34]. So, we cannot rule out that vascular lesions also could account for increase S100B levels.

Despite the controversies on their brain specificity, S100B and NSE have been investigated in different brain diseases as peripheral markers of therapeutic interventions, as well as of neurological and neuropsychological outcome [14, 16, 22‐24]. Additionally, a recent work suggests that blood-brain barrier permeability may also be damaged even at an early stage of AD indicating different blood-brain-CSF compartmental kinetics [35]. Thus, the leakage of proteins from brain to blood could be facilitated.

There is a considerable number of clinical studies regarding CSF levels of S100B in AD, which from the perspective of cerebral protein release provides more sensitivity and specificity than serum samples, however, the availability of CSF samples is somewhat limited for routine clinical use [20, 21, 36]. These previous studies showed discrepant results with respect to the differences in S100B levels between AD and control groups and also to clinical associations. While some studies have reported increased CSF levels of S100B in AD patients compared to controls [21, 36], others did not find any difference [20]. There are also reports suggesting a positive correlation between CSF S100B levels and MMSE, and a negative correlation with severity of dementia by CDR scale [20] while in others CSF S100B levels did not correlate with impaired cognitive status evaluated by MMSE [21, 36].

In contrast to these cited works, here we demonstrated AD patients a positive correlation between serum S100B levels and brain function (evaluated by MMSE and CDR). Ethnic differences, education, prior life style and the predominance of female in our study, compared to others, could account for the differences observed among the previous published results. Considering that in our work MMSE and CDR scales identified patients with impaired cerebral functionality affected by AD and that the S100B levels were higher in more severely affected patients, we preliminarily suggest that serum S100B levels could help to distinguish severity or to follow up the progression of dementia in AD disease, even though the diagnosis is based on clinical investigations. This result may reflect the participation of this protein in the pathogenesis of AD. Indeed, glial cells particularly microglia and astrocytes are able to modulate cerebral plasticity and to protect brain from insults [13]; thus, in this context the increase in S100B could indicate astrocytic reaction to neuronal injury (reactive astrogliosis) in AD patients. Once activated, these cells increase the expression of substances that can participate in the excitotoxicity and inflammatory processes that occur during the evolution of AD [37].

Further, dysfunction of glial cells may promote neurodegeneration and, eventually, the retraction of neuronal synapses, which leads to cognitive deficits [38]. However, our major concern is regarding the decreasing levels of S100B in total AD group compared to controls. This finding was confirmed by an additional experiment with a new set of patients. Interestingly, the severity of CDR score was associated with elevations in serum S100B levels, contrasting to our previous findings with schizophrenic patients [39], when the levels were more elevated in the early onset of disease. In this context, increased astrocytic expression of S100B could be involved in the progression of brain neuropathological changes and behavioral deficits observed in patients and animal models [20, 13, 40‐42]. While in this study S100B levels in AD patients did not correlate with brain morphological changes evaluated by MRI, lower serum levels of NSE were related to higher degree of atrophy and brain macroscopic alterations. Furthermore, this loss was not associated with dementia severity in AD patients. The lack of a correlation with atrophy and cognition reported here is somewhat curious as many recent reports of the literature suggest the opposite [31].

Levels of NSE have been shown to be elevated in acute phase of several disorders of the CNS [43‐47]. This data could preliminary suggest that AD patients, prior the onset of symptoms, could have altered NSE levels, which is now reflected by brain atrophy and ventricle enlargement. However, the reports of CSF NSE vary from decreased levels [17], no difference [19] to increased levels [48]. Interestingly, Palumbo et al. 2008 [48] showed that CSF NSE level has the same behavior as the other accepted markers of AD, being correlated with Abeta42 and total protein tau (h-tau).

Taking into consideration, the strong association of the prognosis of AD patients with the severity of dementia at the diagnosis [3], and the difficulty of evaluating the degree of neural cells loss in AD [4], further experimental and clinical studies regarding these serum markers in AD disease should be encouraged.

In conclusion, we showed that in AD patients the serum S100B levels increased with the severity of the disease whereas decreased serum levels of NSE were associated with increased brain morphological damage.