*2.5. LGF Modulates the Expression of Proteins Involved in Cell Survival*

Our previous studies showed how in an experimental model of Parkinson's disease, LGF potentiates cell survival through the regulation of Bcl2 and Bax protein expression [9]. In the hippocampus of APPswe mice the Bcl2/Bax ratio was significantly reduced, and LGF treatment increased this ratio to control values (Figure 6A). This effect was due to a reduction in the expression of the pro-apoptotic protein Bax which levels were significantly up-regulated in the hippocampus of APPswe mice (Figure 6B).

Akt is another protein involved in cell survival modulated by LGF in neurodegeneration [9]. As shown in Figure 6C, phospho-Akt/Akt ratio was up-regulated in the hippocampus of APPswe mice where phospho-Akt levels were significantly increased in comparison with WT mice (Figure 6D). The APP-LGF group of mice also showed an increased phospho-Akt/Akt ratio and higher levels of phospho-Akt than control mice (Figure 6C,D). Neither Bcl2/Bax nor phospho-Akt/Akt ratio were affected in the cerebral cortex at any of the experimental conditions used in this study (Figure 6A,B).

**Figure 5.** LGF reduces the inflammatory response in APPswe mice. Panel (**A**) shows how LGF significantly reduces GFAP protein expression in the cerebral cortex and hippocampus of APPswe. LGF administration also decreases the expression of ASC (**B**), which is a protein involved in the production of pro-inflammatory cytokines. Panel (**C**) shows quantitative Western blot analysis of the receptor scavenger CD36 that is over-expressed in the hippocampus of APP mice. Panel (**D**) represents the quantitative analysis of the transcription factor Nrf2 as an oxidative stress regulator. Representative Western blots of GFAP, Nrf2, ASC, CD36 and their respective β-actin as charge are shown in panel (**E**). Results represent the mean ± SEM of 4 to 10 independent mice. The statistical analysis was performed by one-way ANOVA followed by Newman–Keuls test. \* *p* ≤ 0.05, \*\* *p* ≤ 0.01, \*\*\* *p* ≤ 0.001 vs. WT. + *p* ≤ 0.05, ++ *p* ≤ 0.01, +++ *p* ≤ 0.001 vs. APPswe mice.

**Figure 6.** Effect of LGF on the expression of proteins involved in cell survival. Protein expression was analyzed by Western blot using specific antibodies against Bcl2, Bax, p-Akt and Akt. Panel (**A**), shows that LGF increases Bcl-2/Bax ratio in the hippocampus. This effect is mainly due to a reduction in Bax protein levels (**B**). Panels (**C**,**D**) show that phospho-Akt/Akt ratio (**C**) and phospho-Akt protein levels (**D**) are significantly increased in the hippocampus of APP and APP-LGF treated mice. Results represent the mean ± SEM of 4 to 8 independent mice. The statistical analysis was performed by one-way ANOVA followed by Newman–Keuls test. \* *p* ≤ 0.05, \*\* *p* ≤ 0.01, vs. WT.+ *p* ≤ 0.05, ++ *p* ≤ 0.01 vs. APPswe mice.

#### **3. Discussion**

In the experimental model of AD used in this study (APPswe mouse), LGF treatment reduced Aβ and phospho-Tau protein levels as well as the number of Aβ plaques with a diameter higher than 25 μm. LGF administration also modulated protein ubiquitination and HSP70 protein levels, reduced glial reactivity and inflammation, and up-regulated the expression of proteins involved in cell survival. These LGF beneficial effects correlate with a better cognitive outcome in APPswe mice.

Amyloid-β aggregation and deposition of Aβ plaques are considered as causative agents in AD, so many therapies have been directed to modify Aβ levels (reviewed in [21]). Under our experimental conditions, LGF modulated Aβ by reducing its levels to WT values in the hippocampus and cerebral cortex. This reduction was observed in the expression of those Aβ peptides with a molecular weight lower than 60 kDa. This fraction includes the Aβ1-42 peptide that has been extensively studied due to its abundance in patients predisposed to AD [22] and mouse models of AD [19]. Here, we also show that LGF treatment reduced the accumulation of polyubiquitinated proteins, and up-regulated HSP70 protein expression in both structures. The high levels of ubiquitinated proteins observed in APPswe mice indicate that proteins are not correctly proteolyzed, probably due to the generation of free radicals and/or an inflammatory environment promoted by the accumulation of Aβ HSP70 is a chaperone which overexpression could suppress the progression of AD by promoting Aβ clearance (reviewed in [23]). HSP70-mediated Aβ clearance from the brain has been associated with the up-regulation of insulin-degrading enzyme (IDE) and transforming growth factor β1 (TGF-β1) [24]. This study did not evaluate IDE and TGF-β1 expression but LGF administration significantly reduced plasmatic levels of insulin in an experimental model of atherosclerosis in mice [25], and modulated the synthesis of TGF-β1 in bile duct-ligated rats [11].

Another major histopathological characteristic in AD is neurofibrillary tangles composed of aggregates of hyperphosphorylated forms of the protein associated with Tau microtubules [26,27]. As shown here, LGF treatment reduced phospho-Tau/Tau ratio in the hippocampus of APPswe mice. Interestingly, Tau homeostasis is controlled through the action of molecular chaperones, such as HSP70 [28,29] that, as noted above, was up-regulated in this structure. In the case of the cerebral cortex, phosphoTau/Tau ratio in APP mice was not significantly different from untreated WT mice but, in this brain structure, LGF treatment reduced phospho-Tau/Tau ratio in both, WT and APP mice. At present we do not have a possible explanation for this event. We could suggest that this is a potential specific effect of LGF not associated to the accumulation of Aβ plaques and inflammation observed in APP mice. However, according to our results it does not seem the case because similar results were not observed in the hippocampus of WT mice where, as already mentioned, LGF did not reduced p-Tau/Tau ratio. Besides HSP70, which was also up-regulated in the cerebral cortex of APP-LGF mice, LGF could modulate Tau phosphorylation in this brain structure by different mechanisms not analyzed in this study. Further studies must be conducted in order to elucidate these discrepancies.

Activated microglia and reactive astrocytes are two pathological markers observed in AD [30–32]. Both cell types are considered cell targets of LGF in the CNS [6,7,9,20]. According to the present results, LGF significantly reduced proliferation (as measured by BrdU incorporation and PCNA protein expression) and Iba1 protein levels, which are two features of activated microglia. These results are in agreement with previous studies that show how chronic administration of LGF reduced microglia cell reactivity in an experimental model of Parkinson's disease in rats [9] and in ataxic rats [15]. Our immunohistochemical studies also suggest that the administration of LGF could prime microglia to acquire a phagocytic phenotype. Thus, although the number of Iba1-positive cells was not reduced in APP-LGF treated mice, this treatment increased by 3-fold the percentage of Iba1-positive cells that invaded Aβ plaques and reduced the number of Aβ large plaques while increasing the number of smaller plaques, which would explain the slight, but not significant increase observed in this parameter. How LGF modulates microglia activity in APPswe mice is unknown but, as mentioned above, LGF could stimulate TGFβ1expression that is a cytokine that stimulates Aβ clearance through activation of phagocytic microglia [33]. On the other hand, activation of phagocytic microglia has been associated with increased levels of interleukin 6 and TNF-alpha [34]. This latter cytokine seems to play an important role in the mitogenic cascade of LGF in the liver, and in LGF-induced neuroregeneration (reviewed in [6]) and neuroprotection [20].

Reactive astrogliosis is a well-known hallmark of AD that is identified, among others, by an increased expression of GFAP. We also found that LGF reduced GFAP over-expression in the hippocampus and cerebral cortex of APPswe mice, suggesting that astrocytes are potential targets of LGF in this experimental model of AD. Microglia cell reactivity to Aβ and phagocytic activity are also under the regulation of astrocytes [35] so, their stimulation with LGF could lead to the release of factors with antioxidant and/or anti-inflammatory effects that could modulate both events. In this regard, we should mention that LGF modulates the expression of TNF-alpha in cultures of mesencephalic glia [20], and significantly reduced by 36 ± 11% the expression of ASC in these cultures (*p* ≤ 0.05 vs. untreated cultures).

The anti-oxidative and anti-inflammatory activities of LGF have also been observed in vivo. Thus, chronic LGF decreased oxidized glutathione concentration in ataxic mice [14], and reduced TNF-alpha protein levels in 6-OHDA-lesioned rats [9]. As presented here, LGF modulates the expression of several proteins that regulate oxidative stress and inflammatory responses in AD. ASC is an adaptor protein that participates in the production of pro-inflammatory cytokines, and contributes to the inflammatory response and cognitive impairment in AD pathology [36–39]. Chronic administration of LGF down-regulated ASC protein levels in the hippocampus and cerebral cortex of APPswe mice. In these mice, ASC is mainly expressed by microglia, which is associated with Aβ plaques (Reimers et al. manuscript in preparation). Because ASC released by microglia forms an aggregate complex with Aβ that promotes inflammation [37,39], we may argue that ASC reduction is associated with a potential anti-inflammatory effect of LGF in this experimental model of AD. Although we were unable to detect any significant change in TNF-alpha and Il-1beta protein expression at the experimental time analyzed in this study, we cannot exclude the possibility that LGF modulates the expression of these cytokines

at earlier times of treatment, as observed for TNF-alpha in 6-OHDA-lesioned rats [9]. In addition, small changes in the levels of both proteins may not be detected by Western blot analysis, so that it may be necessary to use other methods with higher sensitivity as ELISA. Other studies have reported that reduced levels of ASC stimulate phagocytosis in glial cells in a mouse model of AD [40], so LGF-induced down-regulation of ASC could also mediate the increase in phagocytic activity proposed in our study.

CD36 is a class B scavenger receptor involved in phagocytosis which expression in microglia plays a key role in the Aβ-mediated activation of this cell type (reviewed in [41]). Moreover, CD36 over-expression in microglia seems to mediate the inflammatory processes associated to Aβ pathology in transgenic mice [42], and the production of ROS in response to Aβ in AD human brains [43,44]. Similarly, the increased levels of CD36 observed in the hippocampus of APPswe mice could contribute to creating an inflammatory environment in this structure. According to our results, LGF treatment could modulate inflammation and/or ROS production in the APPswe hippocampus by decreasing CD36 protein expression and, in consequence, its interaction with Aβ in this structure. In fact, the inhibition of Aβ interaction with CD36 has been proposed as a potential therapeutic strategy in AD [45]. An interesting observation of our study is that LGF up-regulated CD36 levels in the hippocampus of WT mice. These high levels of CD36 could promote the phagocytic activity of microglia to improve the clearance of debris that is accumulated in the normal brain during the processes of aging [46,47]. Although similar results were observed in the cerebral cortex of APPswe and APP-LGF treated mice, no statistically significant differences were detected. Perhaps CD36 levels are not sufficient to be detected through Western blot in this structure, or we need to study a higher number of mice, or mice at different ages. De facto, we were unable to detect CD36 up-regulation in the hippocampus of thirteen months old APPswe mice (100 ± 7 (*n* = 4) and 104 ± 6 (*n* = 4) % of CD36 expression in WT and APPswe mice, respectively).

Nrf2 is a transcription factor that regulates oxidative stress [48,49] and anti-inflammatory responses [50,51]. Nrf2 mRNA and protein deficiency have been observed in AD patients [52,53] and AD models in mice [54,55]. As shown here, LGF up-regulated Nrf2 protein levels that were reduced in the cerebral cortex and the hippocampus of APPswe mice. Nrf2 up-regulation ameliorated cognitive impairment in APP Knock-in AD mice [56], and in APPswe/PSEN1dE9 hypoxic mice, the intranasal administration of Nrf2 modulated CD36 protein expression, reduced Aβ deposition, and improved spatial memory [57]. In both studies, these beneficial effects were associated to a reduction in oxidative stress and neuroinflammation promoted by the increased levels of Nrf2. Since in our mice LGF significantly ameliorates cognitive impairment, we may consider that up-regulation of Nrf2 could underlie this behavioral improvement through the activation of anti-oxidant and anti-inflammatory responses, as has been reported in other experimental pathologies[58].

Our previous results indicate a potential neuroprotective role of LGF in the brainstem and cerebellum of ataxic rats [15], and in the substantia nigra and striatum of 6-hydroxydopamine-lesioned rats [9]. In both experimental models of neurodegeneration, LGF raised the ratio Bcl2/Bax due to the up-regulation of the anti-apoptotic protein Bcl2, which is involved in the survival of neural cells [59]. In the experimental model of AD used in this study, LGF also restored Bcl2/Bax ratio to control values, but this effect was due to the down-regulation of the pro-apoptotic protein Bax that was over-expressed in the hippocampus of APPswe mice. Up-regulation of Bax has been observed in neurons that exhibit neurofibrillary tangle pathology [60], and APP transgenic mice [61]. Besides, a neuroblastoma cell line co-transfected with Swedish mutant APP showed a higher Bax/Bcl2 ratio [62]. LGF-induced down-regulation of Bax could contribute to decreasing the excessive neuronal apoptosis reported in AD [63]. The increase in Bax levels observed in APPswe could result in neuronal death. However, APPswe mice did not show any reduction in the expression of the specific nuclear neuronal marker NeuN in the hippocampus (100 ± 8 (*n* = 6) and 99 ± 10 (*n* = 5) % of NeuN expression in WT and APP mice, respectively), and cerebral cortex (100 ± 7 (*n* = 6) and 99 ± 12 (*n* = 5) % of NeuN expression in WT and APP mice, respectively). These results agree with the lack of neuronal death previously reported in this experimental model of AD [64,65]. Akt, is a serine/threonine kinase which phosphorylated

form plays a critical role in the regulation of neuronal survival [66,67] that has been involved in the pathogenesis of AD [68]. Our results showed an increased phospho-Akt/Akt ratio in the APPswe hippocampus due to the higher levels of phospho-Akt observed in this brain structure. Phospho-Akt may protect neuronal cell death through different mechanisms including the phosphorylation and inhibition of cytosolic pro-apoptotic proteins (reviewed in [69]). Because the phosphorylation of Bax by active Akt affects its translocation to the mitochondria [70], we may consider the up-regulation of phospho-Akt in the hippocampus as an intrinsic mechanism to protect neurons of Bax-mediated apoptotic cell death in the APPswe mice.

From our study, we cannot define if the observed effects of LGF are due to a direct or indirect action of the factor on the brain of APPswe mice. A potential direct action would imply that LGF could cross the blood–brain barrier by binding to specific receptors found in the endothelium known as RAGE receptors (receptors for advanced glycation end-products) [71]. In fact, LGF show some structural similarities with the advanced glycation end products (AGE) which structure is mainly based on albumins bound to glucose with the ability to change albumin conformation [72]. RAGE receptors are also expressed by astrocytes and microglia [73], and their activation elicits TNF-alpha release from these cells [73].

Present results show how LGF treatment also improves memory performance impairment in APPswe mice. The APPswe mice in comparison with the WT animals, had a reduced ambulatory behavior in the Y-maze, as shown by the total number of entrances and this pattern was unchanged by LGF treatment, so we may consider that motor activity and anxiety were affected in this experimental model of AD. However, the alternance index in the Y-maze, which is an indicator of mnemonic function, was significantly reduced in these mice. LGF treatment recovered this important parameter of working memory to similar values to those observed in WT mice. Other studies have reported a reduction in the exploratory behavior from 10 months old APPswe mice that are coincident with the presence of Aβ plaque depositions in the hippocampus and cerebral cortex (reviewed in [74]). Thus, LGF-induced reduction in Aβ plaques size and levels could contribute to the cognitive improvement observed in our study. In addition, LGF could ameliorate the exploratory behavior in these mice through the reduction of glial reactivity, and its anti-inflammatory and anti-oxidative responses referred to above.

The marble-burying test index was lower in the APPswe mice than in WT mice, and LGF rescued this behavioral parameter. Other studies have shown reduced scrabble behavior and buried marbles in mice with cortical and hippocampal lesions [75] and in Tg-APP/PS1 mice [76]. In the latter study, the behavioral deficits were attributed to the accumulation of Aβ in the brain, so we may consider that the beneficial effects observed in APP-LGF treated mice could be in part due to the reduction in Aβ levels promoted by the factor.
