**3. Discussion**

Regardless of the fact that "type 3 diabetes mellitus" has received much attention in the past decade, it is still unclear whether an overall failure of brain glucose and insulin metabolism regulations is associated with AD pathogenesis. It is generally accepted that altered cerebral glucose uptake and insulin resistance are some of the hallmarks of the progression of neurodegenerative processes [40]. Investigators from Thiambisetty laboratory were one of the first to find, in 2017, that lower rates of glycolysis and higher brain glucose levels were correlated to more severe amyloid plaques and neurofibrillary tangles found in the brains of people with AD [41]. Current evidence from the AD-T2DM animal model, where scientists combined the APP/PS1 mouse model with the genetic db/db model

of type 2 diabetes, highlights neuroinflammatory processes manifested by a significant increase of microglia burden in the cortex and general brain atrophy in animals of 14 weeks of age when T2DM was started, but no AD pathology was observed [42]. Moreover, recent studies have found that AD-T2DM mice are characterized by the upregulation of a broad profile of pro-inflammatory cytokines, such as IL-1*α*, IFN-*γ* and IL-3 in the brain [43].

In our study, we demonstrated that environments with high levels of glucose or insulin have a negative impact on neuronal metabolism by elevating levels of oxidative stress markers. Our present data support the view that hyperglycemia and/or insulinemia causes activation of ROS and NO levels in neurons, which also affects double-stranded DNA breaks. Furthermore, there was a significant positive correlation between those parameters after 1 h from glucose administration. The statistical analysis did not confirm any significant differences between oxidative stress parameters DBSs after insulin treatment with two exceptions for ROS vs. DBSs after 1 and 24 h of incubation. It is suggested that, over the years, these small changes in the neuronal tissue may lead to the deactivation or shift of crucial cell cycle pathways such as the MEK-ERK1/2-NF-κB pathway as well as to the upregulation of pro-apoptotic factors, which leads to amyloid *β* accumulation. Although caution must be exerted in extrapolating in vitro findings to the in vivo situation, our data are in line with the suggestion that the fast-shifting concentrations of glucose and/or insulin in neuronal tissue may be one of the first pathological changes connected to AD and begin several years prior to the onset of clinical symptoms. Some clinical studies have suggested a significant role for S100B in neurodegeneration processes, pointing to its increased levels in the body fluids of patients with AD [44–46]. Christl et al. [47] estimated that S100B protein levels in cerebrospinal fluid might have a diagnostic value, particularly at the early stages of the disease, as it declines to normal levels in more advanced stages. Animal studies provide additional data showing the importance of S100B in neurodegenerative processes. The performance of S100B-overexpressing Tg2576 mice was inferior in the spatial learning study, dependent on the functioning of the hippocampus. Animals showed higher levels of brain parenchymal *β*-amyloid depositions and cerebral amyloid angiopathy, enhanced amy-loidogenic APP metabolism, augmented reactive astrocytosis and microgliosis and increased levels of pro-inflammatory cytokines (TNF-*α*, IL-1*β* and IL-6) as early as at 7–9 months of age. Mice whose gene encoding S100B was inactivated showed increased spatial memory and memorization under the influence of anxiety and increased long-term synaptic enhancement in the CA1 sector of the hippocampus. These results indicate that despite playing an important role in encouraging brain inflammatory responses, S100B has a role in directly promoting amyloidogenic APP processing [48].

Regarding the role of S100B protein in T2DM, Kheirouri et al. [49] measured its concentration in the blood serum of patients with metabolic syndromes characterized by intermittent fasting, central obesity, dyslipidemia and arterial hypertension. Moreover, the participants of the study had elevated insulin levels, showed high values on the HOMA-IR index of insulin resistance, and the serum level of S100B protein was significantly elevated compared to healthy volunteers. The extracellular action of the S100B protein is primarily based on its interaction with RAGE—the best-known class of advanced glycation end receptors [50]. S100B binds to RAGE in the extracellular space and activates a number of intracellular biochemical pathways such as MAPKs in microglia and neurons. These receptors mediate both the trophic and the toxic effects of the S100B protein on microglia and endothelial cells. RAGE is also one of the known receptors for A*β* peptide [51,52]. In the physiological state, RAGE expression in cells is kept at a low level. Increasing the amount of RAGE ligands observed in inflammation, oxidative stress, diabetes and AD results in the induction of the expression of this receptor [53]. Literature data show a relationship between glucose concentration and S100B protein secretion. Cultures of astrocytes obtained from rat brains under conditions of metabolic stress: glucose, oxygen and serum (FBS) deprivation increased the secretion of S100B protein. However, after 12 and 24 h of exposure to metabolic stress, mRNA expression for the S100B protein decreased considerably, and a significant reduction in S100B protein secretion was observed after 48 h

of incubation. Based on these observations, it is hypothesized that the S100B protein may be actively secreted in the microglia in the early stages of metabolic stress [54].

The aim of our work was to broaden current knowledge on the role of S100B protein in oxidative stress instigated by local variations in glucose or insulin concentrations in neurons. We analyzed extracellular and intracellular S100B protein levels in order to evaluate the potential role of S100 protein in metabolic disturbances occurring in neuronal cells. The single marked observation to emerge from the data comparison was that neither hyperglycemic nor insulinemic conditions provoked S100B protein efflux from the neurons, and moreover, they even decreased secretion. Interestingly, a small but still statistically significant rise in S100B concentrations was observed in cell lysates. No substantial statistical correlation between intracellular and extracellular S100B levels and analyzed markers of oxidative stress and DNA damage was found after glucose administration, excluding levels in supernatants vs. ROS, where significance was observed. In contradiction, hyperinsulinemic conditions provoked a strong correlation between S100B concentration vs. NO or double-stranded DNA break levels. Collectively, these data support the notion that cellular S100B protein could behave as a neuroprotective factor against extracellular glucose/insulin fluctuations. Our observations are consistent with the results carried out on the primary cortical astrocytes of rats by Nardin et al. [55]. After 24 h, the extracellular S100B levels were reduced by around 45% in astrocytes cultured in a high-glucose medium. A decrease in glutathione content but not glutamate uptake activity was also observed. It is important to emphasize the fact that S100 protein family members exert a dual effect (neurotrophic or neuroprotective) on neurons and astrocytes, depending on the concentration attained in the brain's extracellular space. Our findings are consistent with the previous results presented by Alhemeyer et al. [56] who showed that neurotoxicity caused by glutamate and staurosporine can be counteracted by the S100B protein. Another study performed on PC12 cells found that the S100B protein could inhibit NGF-induced cell differentiation, but increased expression of S100B did not reverse the effects of differentiation in PC12 already differentiated with NGF cells. This may suggest an effect of the S100B protein on the inhibition of cell differentiation in the early stages of cell development [57].

In conclusion, our data support the view that there is a putative relationship between S100B and metabolic disturbances in AD-like pathology and that the S100B protein acts probably as a cytoprotective factor and may protect neurons against the toxicity of local high levels of glucose or insulin during the initial phases of the pathophysiology of AD. This study is the first step towards enhancing our understanding of the role of S100 proteins in neuronal metabolism in the specific condition of "type 3 diabetes mellitus". Further research is necessary to clarify the mechanism of action of S100B protein as well as to explain the reasons for the significant differences in their effect on neuronal properties. As the brain's energetic regulation is entirely insulin-dependent, it would be especially interesting to evaluate the interrelation between S100 protein family members and the glucose transporters: GLUT1 and GLUT3 in the human brain as well as their relation to the receptor for insulin (IR), and these results may represent an excellent initial step toward using S100B levels as diagnostic markers in the early stages of neuropathological disorders.
