Both Amyloid-β Peptide and Tau Protein Are Affected by an Anti-Amyloid-β Antibody Fragment in Elderly 3xTg-AD Mice

Alzheimer’s disease (AD) is the most common dementia worldwide. According to the amyloid hypothesis, the early accumulation of the Aβ-peptide triggers tau phosphorylation, synaptic dysfunction, and eventually neuronal death leading to cognitive impairment, as well as behavioral and psychological symptoms of dementia. ScFv-h3D6 is a single-chain variable fragment that has already shown its ability to diminish the amyloid burden in 5-month-old 3xTg-AD mice. However, tau pathology is not evident at this early stage of the disease in this mouse model. In this study, the effects of scFv-h3D6 on Aβ and tau pathologies have been assessed in 22-month-old 3xTg-AD mice. Briefly, 3xTg-AD female mice were treated for 2 weeks with scFv-h3D6 and compared with 3xTg-AD and non-transgenic (NTg) mice treated with PBS. The treatment with scFv-h3D6 was unequivocally effective in reducing the area of Aβ staining. Furthermore, a tendency for a reduction in tau levels was also observed after treatment that points to the interplay between Aβ and tau pathologies. The pro-inflammatory state observed in the 3xTg-AD mice did not progress after scFv-h3D6 treatment. In addition, the treatment did not alter the levels of apolipoprotein E or apolipoprotein J. Thus, a 2-week treatment with scFv-h3D6 was able to reduce AD-like pathology in elderly 3xTg-AD female mice.


Introduction
Alzheimer's disease (AD) is the most common cause of dementia worldwide. Alzheimer's International (ADI) estimated that more than 50 million people are suffering dementia worldwide nowadays and that the global economic cost of dementia for the year 2018 was above US$ 1 trillion [1]. Incidence rates of any dementia and AD are reported to be greater in women than men [2]. As an example, the incidence rate of AD in Europe was 11.08 per 1000 person-years; 13.25 per 1000 person-years in women, and 7.02 per 1000 person-years in men [3]. In consonance, the prevalence of AD was 7.13% in women and 3.31% in men (overall, 5.05%). Because the prevalence of AD is estimated to be tripled in 30 years, the finding of disease-modifying drugs is mandatory.
AD is a chronic and progressive neuropsychiatric disorder characterized by memory loss, cognitive impairment, and behavioral and psychological symptoms of dementia, including anxiety, phobia, and irritability, owing to synaptic loss and neuronal death. Although its etiology has not been completely unraveled yet, the main histopathological hallmarks of AD include the extracellular Aβ-peptide in glial cells in the late phase post-injection [25]. These findings showed that the mechanism underlying the Aβ-peptide withdrawal is the engulfment of the scFv-h3D6/Aβ complex by the glial cells.
Because AD is usually diagnosed after extracellular Aβ aggregates and tau pathologies become evident, the assessment of the therapeutic potential of any putative drug at the advanced stages of pathology is of capital relevance, particularly concerning the ability of an anti-Aβ antibody to ameliorate tau pathology. In this sense, tau immunotherapy has been shown to reduce amyloid deposition [26,27], which supports the hypothesis of the occurrence of a cross-talk between Aβ and tau. Thus, the searching for pharmacological approaches targeting both histopathological markers is mandatory. The aim of the present work is to evaluate the efficacy of scFv-h3D6 treatment in older animals with evident senile plaques and NFTs. For 2 weeks, 22-month-old 3xTg-AD female mice were treated with either scFv-h3D6 or PBS and, then, Aβ and tau levels were quantified, along with several other markers related to neuroinflammation.

ScFv-h3D6 Accumulates in the Hippocampus
Previous studies have shown that anti-Aβ antibodies were mainly accumulated in the central periventricular areas, especially in the hippocampus of AD mouse models, where Aβ load is high [28]. To assess whether scFv-h3D6 is homogenously distributed or accumulated in the hippocampus, immunohistostaining of brain sections from treated 3xTg-AD mice was performed. ScFv-h3D6 was accumulated in the hippocampus of the treated mice ( Figure 1A), whereas a very faint labeling was detected in the cortex ( Figure 1B). This very faint labeling was negligible and corresponds to a residual cross-reactivity of the polyclonal antibody used, as observed by us previously [22], because the same intensity of labeling was observed in samples from untreated animals ( Figure 1C). Moreover, scFv-h3D6 co-localized with Aβ, not only in the extracellular plaques of the hippocampus (tM1 = 0.342, max = 0.545, min = 0.192; tM2 = 0.187, max = 0.326, min = 0.098) but also, and more intensely, inside the cornus amonis 3 (CA3) neurons (tM1 = 0.597, max = 0.781, min = 0.349; tM2 = 0.476, max = 0.511, min = 0.382) ( Figure 1A, arrows).

ScFv-h3D6 Reduces the Area of 6E10 Staining
We have previously reported that scFv-h3D6 reduces extracellular Aβ oligomers [21] and intracellular Aβ levels [22] after a single intraperitoneal dose to 5-month-old 3xTg-AD mice. We could also observe these effects after treatment for 6 weeks to a similar group of animals [23]. However, the effect of the scFv-h3D6 in elderly animals with an advanced stage of AD-like pathology, where plaques are present, has not yet been evaluated.
Many senile plaques were observed in the hippocampus of 22-month-old 3xTg-AD mice, especially in the subiculum (Figure 2A, arrows). After scFv-h3D6 treatment, the percentage of 6E10-immunoreactive area ( Figure 2B) was significantly reduced (U-value = 0; p = 0.0079; r = 1). However, in spite of the median for the number of Aβ extracellular aggregates was lower after treatment ( Figure 2C), significance was not reached (U-value = 5; p = 0.1508; r = 0.6). Quantification of Aβ 40 (Figure 2D-F) and Aβ 42 ( Figure 2G-I) levels in the different hippocampal protein extracts by ELISA showed that the levels of both Aβ 40 and Aβ 42 were negligible in NTg mice. Even though both Aβ 40 and Aβ 42 levels were reduced after scFv-h3D6 immunization to 3xTg-AD mice in all the protein fractions (TBS-soluble, SDS-soluble, and FA-soluble), differences were not significant, except for Aβ 40 in the SDS-soluble fraction, where the difference was marginally significant, and effect size indicated this difference is real (U-value = 3; p = 0.056; r = 0.76). It is also worth noting that the Aβ load was higher in the SDS-and FA-soluble fractions than in the TBS-soluble fraction, as expected at this advanced stage of pathology.

ScFv-h3D6 Reduces the Area of 6E10 Staining
We have previously reported that scFv-h3D6 reduces extracellular Aβ oligomers [21] and intracellular Aβ levels [22] after a single intraperitoneal dose to 5-month-old 3xTg-AD mice. We could also observe these effects after treatment for 6 weeks to a similar group of animals [23]. However, the effect of the scFv-h3D6 in elderly animals with an advanced stage of AD-like pathology, where plaques are present, has not yet been evaluated.

ScFv-h3D6 Decreases Total Tau Levels
Because 3xTg-AD mice develop both histological hallmarks of AD, senile plaques composed of Aβ, and NFT of tau protein [29], and considering that bapineuzumab was able to reduce tau levels in the CSF from patients with mild-to-moderate AD [30], the effect of scFv-h3D6 in tau pathology was evaluated.
Localization of Aβ and total tau in the hippocampus was examined ( Figure 3). Aβ was detected extra-and intracellularly whereas NTFs were located intracellularly within Aβ-containing neurons in the CA1 of the hippocampus (tM2 = 0.932, max = 0.982, min = 0.893). Since amyloid pathology precedes NFTs formation, not all 6E10 reactive neurons were NFTs positive (tM1 = 0.087, max = 0.101, min = 0.079). soluble fraction. (H) Aβ42 levels in the SDS-soluble fraction. (I) Aβ42 levels in the FA-soluble fraction. Three replicates for each experiment. Data are expressed as medians in box plots, and whiskers represent the minimum and maximum values. Statistical differences were assessed with the nonparametrical Mann-Whitney U-test. * Significant differences p ≤0.05; # Marginal significance p ≤ 0.1; § r ≥ 0.68.

ScFv-h3D6 Decreases Total Tau Levels
Because 3xTg-AD mice develop both histological hallmarks of AD, senile plaques composed of Aβ, and NFT of tau protein [29], and considering that bapineuzumab was able to reduce tau levels in the CSF from patients with mild-to-moderate AD [30], the effect of scFv-h3D6 in tau pathology was evaluated.
Localization of Aβ and total tau in the hippocampus was examined ( Figure 3). Aβ was detected extra-and intracellularly whereas NTFs were located intracellularly within Aβ-containing neurons in the CA1 of the hippocampus (tM2 = 0.932, max = 0.982, min = 0.893). Since amyloid pathology precedes NFTs formation, not all 6E10 reactive neurons were NFTs positive (tM1 = 0.087, max = 0.101, min = 0.079). When focusing on tau and the effect of scFv-h3D6, the 3xTg-AD, but not NTg mice, showed strong labeling for mAb HT7 (U-value = 0; p = 0.0079; r = 1), indicative of high levels of total tau, which were partially decreased after treatment (U-value = 4; p = 0.0952; r = 0.68) ( Figure 4A,B). When using mAb AT8 directed against Ser202 and Thr205 phosphorylated tau, the labeling before the treatment When focusing on tau and the effect of scFv-h3D6, the 3xTg-AD, but not NTg mice, showed strong labeling for mAb HT7 (U-value = 0; p = 0.0079; r = 1), indicative of high levels of total tau, which were partially decreased after treatment (U-value = 4; p = 0.0952; r = 0.68) ( Figure 4A,B). When using mAb AT8 directed against Ser202 and Thr205 phosphorylated tau, the labeling before the treatment was not strong and the effect of treatment was not so evident ( Figure 4C). Phosphorylated tau was also increased in 3xTg-AD compared to NTg mice (U-value = 0; p = 0.0079; r = 1)) but no effect of scFv-h3D6 treatment was observed (U-value = 10; p = 0.8016; r = 0.12) ( Figure 4D). Therefore, total and phosphorylated tau levels in the SDS-soluble fraction of the hippocampus were determined by WB ( Figure 4E-G). Total tau levels were negligible in the NTg and elevated in the untreated 3xTg-AD group (U-value = 0; p = 0.0079; r = 1) ( Figure 4E,G). After the treatment, a tendency for a reduction in the total tau levels was found and the effect size showed these reductions as real (U-value = 4; p = 0.0952; r = 0.68). However, non-significant differences in phosphorylated tau levels, as quantified with the mAb AT8, were observed after the treatment (U-value = 8; p = 0.5317; r = 0.28) ( Figure 4F,G). It should be noted that a large range in tau levels is inherent to the 3xTg-AD model for both total and phosphorylated tau and that scFv-h3D6 treatment reduced this dispersion ( Figure 4E-G). This means that treatment with an anti-Aβ scFv reduced overall tau load.
h3D6 treatment was observed (U-value = 10; p = 0.8016; r = 0.12) ( Figure 4D). Therefore, total and phosphorylated tau levels in the SDS-soluble fraction of the hippocampus were determined by WB ( Figure 4E-G). Total tau levels were negligible in the NTg and elevated in the untreated 3xTg-AD group (U-value = 0; p = 0.0079; r = 1) ( Figure 4E,G). After the treatment, a tendency for a reduction in the total tau levels was found and the effect size showed these reductions as real (U-value = 4; p = 0.0952; r = 0.68). However, non-significant differences in phosphorylated tau levels, as quantified with the mAb AT8, were observed after the treatment (U-value = 8; p = 0.5317; r = 0.28) ( Figure 4F,G). It should be noted that a large range in tau levels is inherent to the 3xTg-AD model for both total and phosphorylated tau and that scFv-h3D6 treatment reduced this dispersion ( Figure 4E-G). This means that treatment with an anti-Aβ scFv reduced overall tau load.

ScFv-h3D6 Slightly Ameliorates the Pro-Inflammatory Status
Several studies support the role of the immune system in neurodegeneration, as reviewed by reference [31]. There are some proteins involved in the immunological processes, such as the triggering receptor expressed on myeloid cells 2 (TREM2), that are associated with an increased risk of AD onset [32]. Besides, the administration of full-length antibodies promotes astroglia and microglia reactivity and an overall inflammatory process [33]. Considering the importance of the immune system and the propensity of its over-activation after Aβ-immunotherapy, the effect of scFv-h3D6 treatment on inflammation in 22-month-old 3xTg-AD female mice has been studied here.
In Figure 5, co-localization of different combinations of astroglia or microglia and Aβ peptide or tau protein in the hippocampus of 3xTg-AD mice are shown. In the first row ( Figure 5A), reactive astrocytes (green) are found surrounding Aβ plaques (red). This concurs with wrapping the soma of some Aβ-containing neurons by astrocytes and microglia reported in a previous deep study of neuronal populations in the 3xTg-AD model [10]. In the second row ( Figure 5B), amoeboid microglia (green) are found close to Aβ plaques (red). Similarly, in the third ( Figure 5C) and fourth ( Figure 5D) rows, astroglia and microglia, respectively, surround tau positive neurons in the CA1. Thus, the interplay among Aβ, tau, and glial cells becomes apparent.
To further analyze the inflammatory state of treated and untreated 3xTg-AD mice, the levels of TNFα, IL-33, IL-1β, and IL-6 were measured in the TBS-soluble fraction ( Figure 6H-K). TNFα levels tended to be increased in the untreated 3xTg-AD group, and effect size revealed that the difference is real (U-value = 4; p = 0.0952; r = 0.68), but not in the treated 3xTg-AD animals ( Figure 6H). IL-33 was also increased in untreated 3xTg-AD mice (U-value = 1; p = 0.0159; r = 0.92) with respect to NTg mice, and no differences were found between treated and untreated 3xTg-AD mice ( Figure 6I). However, a tendency to be equal was found between treated 3xTg-AD and NTg, and effect size indicated they are actually equal (U-value = 4; p = 0.0952; r = 0.68, not in Table 1). Similarly, IL-1β was elevated in untreated 3xTg-AD mice (U-value = 0; p = 0.0079; r = 1), and in treated 3xTg-AD mice, there were no effects, although it should be noted that there was an outlier point masking putative differences ( Figure 6J). Finally, no differences in IL-6 levels were found among the groups ( Figure 6K).

Effect of ScFv-h3D6 on Apolipoproteins' Levels
Increasing evidence supports the role of apoE, apoJ, and their common receptor LRP1, in promoting Aβ clearance [34]. Considering this, ELISA was used to assess whether the scFv-h3D6 treatment modulates apolipoproteins levels and WB was used to check the effect of the treatment on LRP1 expression. Differences among experimental groups were not detected for apoJ levels or for the LRP1 expression ( Figure 7B,C). However, increased levels of apoE were detected ( Figure 7A) in 3xTg-AD mice in comparison with NTg mice (U-value = 2; p = 0.0317; r = 0.84). Although apoE reduction was previously reported after scFv-h3D6 treatment in 5-month-old 3xTg-AD mice [21], there was a small decrease in the median value at 22 months of age that never tended to be close to the values of the NTg group (U-value = 10; p = 0.6667; r = 0.2).
However, a tendency to be equal was found between treated 3xTg-AD and NTg, and effect size indicated they are actually equal (U-value = 4; p = 0.0952; r = 0.68, not in Table 1). Similarly, IL-1β was elevated in untreated 3xTg-AD mice (U-value = 0; p = 0.0079; r = 1), and in treated 3xTg-AD mice, there were no effects, although it should be noted that there was an outlier point masking putative differences ( Figure 6J). Finally, no differences in IL-6 levels were found among the groups ( Figure  6K).

Effect of ScFv-h3D6 on Apolipoproteins' Levels
Increasing evidence supports the role of apoE, apoJ, and their common receptor LRP1, in promoting Aβ clearance [34]. Considering this, ELISA was used to assess whether the scFv-h3D6 treatment modulates apolipoproteins levels and WB was used to check the effect of the treatment on LRP1 expression. Differences among experimental groups were not detected for apoJ levels or for the LRP1 expression ( Figure 7B,C). However, increased levels of apoE were detected ( Figure 7A) in 3xTg-AD mice in comparison with NTg mice (U-value = 2; p = 0.0317; r = 0.84). Although apoE reduction was previously reported after scFv-h3D6 treatment in 5-month-old 3xTg-AD mice [21], there was a small decrease in the median value at 22 months of age that never tended to be close to the values of the NTg group (U-value = 10; p = 0.6667; r = 0.2).

Discussion
The knowledge about proteins and pathways involved in AD has exponentially grown in the last few years. However, the translation of therapeutic strategies from preclinical studies to clinical assays has not been successful [35]. The reason for this lack of positive results in clinical trials is not well understood, but several factors that could influence the outcome have been discussed. Among them, targeting Aβ or tau pathologies, instead of targeting both simultaneously, and the advanced stage of the disease when the treatment started during the trials, could explain this apparent standstill.
Bapineuzumab, the humanized form of the murine mAb 3D6, directed to the N-terminal Aβ region, is able to bind both fibrillar and soluble Aβ species [36]. Even though a reduction in Aβ and tau levels was achieved after bapineuzumab administration, severe side effects, together with the lack of positive outcomes for cognition, forced the discontinuation of the trials [14]. Therefore, in order to avoid the adverse effects associated with Fc-dependent microglia over-activation, scFv-h3D6 lacking the Fc region was constructed [16]. A single administration to 5-month-old 3xTg-AD female mice, when pathology was in an early stage, was able to reduce the first hallmarks of AD, and improve the cognitive impairment associated with an overload of intraneuronal Aβ [21,22]. ScFv-h3D6 also decreased Aβ uptake by the human primary astrocytes that should be beneficial because sustained Aβ uptake by astrocytes may impair their normal functions [37,38]. However, nowadays, AD is usually diagnosed when Aβ burden is high, and symptomatology is evident [4]. Thus, studying the effect of scFv-h3D6 administration in elderly animals, where Aβ and tau pathologies develop, is of capital relevance for assessing the therapeutic potential of this antibody fragment.
At this initial point in the discussion, an important methodological issue is worth mentioning. As we [18,[21][22][23][24][25] and other authors [39][40][41][42] have extensively argued, although 6E10 mAb is able to recognize Aβ, β-CTD, and APP, what we are measuring in the 3xTg-AD mouse with this mAb is nowadays agreed to be the Aβ peptide. The present work constitutes another demonstration of this fact since scFv-h3D6, which exclusively recognizes the Aβ peptide, promotes a comparable decrease in both 6E10-labeling and Aβ40-and Aβ42-specific signals by ELISA.
Concerning mAbs pharmacokinetics, it has been reported that peripherally administered antibodies appear in the CSF [30], which would indicate that the mAbs may enter into the brain by crossing the blood-CSF barrier, rather than the blood-brain-barrier (BBB) [43]. If this were true, scFvs levels would be higher in the central periventricular areas and almost non-detectable in the cortex. ScFv-h3D6 labeling was detected in the hippocampus of treated 3xTg-AD but not in the cortex, supporting this hypothesis. Therefore, the use of bispecific antibodies combining an anti-Aβ mAb with an anti-transferrin R antibody that has been used as a BBB transporter, which could improve the distribution of therapeutic mAbs in the brain, spreading its activity and increasing the efficacy of treatment [28]. As recently reported, APP can be relocated into transferrin-positive recycling endosomes. Thus, the use of anti-transferrin R antibodies is able not only to promote transport through the BBB but also to drive the therapy directly to an APP-containing organelle [44].
Since scFv-h3D6 was mainly found in the hippocampus, we would refer hereinafter to this brain region. After a relatively short period of treatment, a reduction in Aβ burden was achieved as quantified by immunohistochemistry. Although the median value for the number of senile plaques was reduced in the 3xTg-AD brains after scFv-h3D6 treatment, there was not a significant difference. Importantly, the total 6E10-immunoreactive area was significantly decreased after treatment. Therefore, scFv-h3D6 clears amyloid aggregates not only from the intracellular space [22] but also from the extracellular space. This is in agreement with the reduction in Aβ 40 oligomers, i.e., trimers, hexamers, nonamers, and dodecamers, previously reported in 5-month-old 3xTg-AD mice [21]. In this work, Aβ 40 and Aβ 42 levels were also reduced in TBS-soluble, SDS-soluble, and FA-soluble fractions and, although the differences were not significant, the trend observed for the decrease in Aβ 40 in the SDS-fraction was revealed as real after considering the effect size. In addition, the trend observed in all the fractions concurs with immunohistochemical analyses, which indicate that scFv-h3D6 reduces Aβ burden, even at this late stage of pathology. It could be possible that anti-Aβ antibodies were retained in senile plaques [28], hindering the removal of oligomers. This could also partially explain why Aβ immunotherapy is not as effective as expected; if antibodies were trapped in Aβ fibrils, oligomers, the most toxic species, could not be targeted. Thus, a higher doses, an increased time-period of treatment, or a redesign of the antibodies to avoid fibril binding and enhance removal of oligomers, could improve the efficacy of the treatment.
Importantly, not only Aβ but mainly hyperphosphorylated tau protein correlates with cognitive impairment associated with AD progression [45,46]. Thus, reducing tau levels is critical for ameliorating dementia-related symptoms. The interplay between Aβ and tau is well documented [47]. On the one hand, Aβ and APP intracellular domain (AICD) can promote tau hyperphosphorylation [48]. On the other hand, some pathogenic characteristics of transgenic models are rescued by ablating the tau gene [49]. Considering that bapineuzumab was effective in removing tau protein from the CSF [30], we investigated whether scFv-h3D6 can reduce tau burden in the hippocampus. A tendency to reduce total tau levels was achieved after scFv-h3D6 treatment (p ≤ 0.1), and effect size indicated this reduction is real (r ≥ 0.68). It is important to note that 2 weeks of treatment was effective at a very advanced stage of the disease when tau pathology was well established. Even though a reduction in Ser202 and Thr205 phosphorylated tau was not observed after scFv-h3D6 treatment, future studies analyzing other phospho-tau epitopes could shed light on the effects of anti-Aβ immunotherapy on tau pathology. In a recent study, active full-length DNA Aβ 42 immunization to 3xTg-AD mice was able to reduce tau levels [50]. However, there are two main differences with our study; 1) they treated animals for 20 months whereas we did so for 2 weeks and 2) our rout of administration (i.p.) was safer than active DNA immunization. How anti-Aβ antibodies could influence the clearance of tau is not well understood, but it is likely a consequence of Aβ reduction. Because both pathologies are linked, it is not surprising that reducing one of them could result in the amelioration of the other [51].
Several mechanisms for explaining how anti-Aβ antibodies reduce amyloid pathology have been proposed. The peripheral sink hypothesis proposes that circulating mAbs could reduce Aβ from blood and other tissues, and thus promote its efflux from the brain to the blood. In support of the peripheral sink hypothesis, Aβ 40 levels in blood were increased after bapineuzumab administration [52]. Although this mechanism could be shared between bapineuzumab and the scFv-h3D6 derived fragment, it has been demonstrated that the antibody fragment could enter the brain and localize in both the extracellular and intracellular compartments. Importantly, the lack of the Fc domain prevents glial activation via Fc receptors and consequently would avoid the increased inflammatory response observed with full-length mAbs [53].
Inflammatory response and neurodegeneration are linked in AD [54]. The exact role that the immune system plays in triggering AD and progression is not completely understood. However, it is well known that the immune system plays a key role in AD pathogenesis. It has been reported that both complement and microglia mediate synaptic loss [55], that some variants of TREM2 increase the risk for developing AD [56], and that there are astroglial and microglial reactive cells surrounding Aβ plaques [57]. Moreover, this pro-inflammatory state is enhanced after some anti-Aβ treatments as a consequence of glia over-activation [58,59].
In this work, we report the effect of scFv-h3D6 on the inflammatory state of 22-month-old 3xTg-AD female mice brains. Astroglia and microglia were found surrounding senile plaques and NFTs in the hippocampus, as expected. Moreover, an increased GFAP, but not an Iba1 signal, was detected in 3xTg-AD mice, which indicates the occurrence of astrogliosis in this mouse model, but not microgliosis. IL-1β and TNF-α, both secreted by astrocytes [60], were increased in old 3xTg-AD mice. IL-33 was also increased in this model. IL-33 administration to APP/PS1 double-transgenic mouse model of AD could ameliorate cognitive impairment and reduce both Aβ 40 and Aβ 42 in the cortex [61]. Besides, it has been reported that IL-33 ablation causes tau accumulation and late-onset neurodegeneration in the cortex and hippocampus [62]. These data suggest that IL-33 is protective against AD. Importantly, elevated expression of IL-33 has been found in the vicinity of Aβ plaques and tau NFTs [63]. Thus, the increased levels of IL-33 in the hippocampus of 22-month-old 3xTg-AD female mice could be a natural response to counteract the progression of the disease. After repeated scFv-h3D6 injections, interleukin levels were not altered with respect to those in the untreated 3xTg-AD. Neither Iba1 nor GFAP were increased, which means that no over-activation of the immune system after scFv-h3D6 administration occurred. Moreover, levels of GFAP tended to decrease (p ≤ 0.1) after treatment, and effect size pointed to a real decrease (r ≥ 0.68), and although differences in interleukin levels were not significant, their values were closer to the NTg interleukin levels than to the untreated 3xTg-AD animals' levels.
Finally, considering that apoE and apoJ are involved in AD pathology and that we previously found apolipoprotein levels increased in 5-month-old 3xTg-AD female mice [21], the same study has been performed at 22-months of age. In this case, only apoE levels were found elevated in the 3xTg-AD group, and no effect of the treatment was observed. In fact, it is interesting to note that the role of apoE in AD is not well understood, and it has been suggested that apoE could be detrimental at early stages and neuroprotective when the pathology is advanced [64]. Thus, in response to oxidative stress, aging, brain damage, or Aβ deposition, neurons synthesize increasing amounts of apoE, which, in turn, undergoes proteolytic processing, generating fragments that cause mitochondrial dysfunction, NFTs formation, and eventually neurodegeneration [65,66]. Therefore, scFv-h3D6 presents many advantages in comparison with the full-length mAb from which it derives such a greater penetrance into the CNS [22] and no over-activation of the immune system [18,22], even at the late stage of the disease, as demonstrated in this work. Thus, the use of antibody fragments could be a safe strategy to be translated to the treatment of other neurodegenerative diseases. However, the fact that Aβ-immunotherapy has not proved an effect in phase 3 could be related to the additional involvement of other non-amyloidogenic or -taugenic mechanisms in AD pathophysiology. Therefore, combined therapies against the amyloid peptide and/or tau with immune system modulators and/or other alternatives such as apolipoprotein mimetic peptides [23,37,38], could constitute the solution for an effective AD treatment.
As a general conclusion, scFv-h3D6 ameliorates the histological hallmarks of AD in 22-month-old 3xTg-AD female mice, without eliciting any detectable inflammatory response. It is important to remark that although other antibodies improve performance in behavioral testing, they do not affect overall amyloid plaque and tau burden in 22-month-old 3xTg-AD [67]. Future studies, targeting both Aβ and tau pathologies at the same time or in combination with other approaches, would improve the efficacy of Aβ-immunotherapy and could constitute an improved pharmacological approach.

ScFv-h3D6 Production
ScFv-h3D6 was produced as previously described [18]. Briefly, the scFv-h3D6 gene was cloned into the pPicZαA vector (Invitrogen, Carlsbad, CA, USA) and expressed in KM71H P. pastoris cells. Cells were grown until an OD 600 of 2.0 was reached. Then, methanol was supplemented every 24 h to induce protein expression. After 48 h of expression, cells were centrifuged at 3000× g and 4 • C for 10 min. ScFv-h3D6 was precipitated with ammonium sulfate. After centrifugation at 100,000× g and 4 • C for 1 h, the pellet was re-suspended in 10 mM Na 2 HPO 4 (pH 6.5) and double dialyzed. Finally, cationic exchange chromatography was performed in a Resource S column coupled to a UP10 AKTA (GE Healthcare, Chicago, IL, USA). Pure scFv-h3D6 was dialyzed to PBS (pH 7.4) and stored at −20 • C until further use.

Animals
The triple-transgenic mouse model of AD (3xTg-AD), harboring PS1 M146V , APP Swe , and tau P301L transgenes, was initially engineered at the University of California, Irvine [29]. Both the 3xTg-AD colony and their corresponding non-transgenic (NTg) littermates with the same genetic background (B6129SF2) were purchased from the Jackson Laboratory (Bar Harbour, ME, USA) and bred at the Animal Facility of our university (Autonomous University of Barcelona, UAB, Barcelona, Spain). Female mice were selected because, in contrast to male mice, they conserve the phenotype as originally described [68] and exhibit an exacerbated AD-like pathology [69]. In addition, AD incidence was reported to be greater in women than men [2]. 3xTg-AD mice progressively displayed Aβ deposition, with intraneuronal immunoreactivity at 3-month-old and extracellular deposition at 6-month-old, and tau aggregation at 12-month-old. Cognitive impairment was first detected at 4-month-old as long-term memory impairment [29,70]. No physical alterations were detected in this mouse model with a mean lifespan of 386.9 ± 161.2 days for male mice and 517.1 ± 197.1 days for female mice [71]. Five animals (with the same genotype and sex) were maintained in Makrolon home cages 35 cm × 35 cm × 25 cm at a temperature of 22 ± 2 • C, with a relative humidity of 55 ± 5%, 12 h light:dark cycle starting at 08:00 am, and food and water ad libitum.

Experimental Design and Statistical Analysis
Ten 22-month-old 3xTg-AD female mice were randomly distributed into 2 experimental groups (n = 5): Untreated triple-transgenic mice (3xTg/−) and treated triple-transgenic mice (3xTg/+). Simultaneously, the untreated non-transgenic mice (NTg/−) group constituted of 5 age-and gender-matched NTg mice. A treated non-transgenic mice (NTg/+) group was not included because of 3 reasons derived from observations explained in the introduction Section. First, we already demonstrated that the treatment to NTg mice was safe (as well as to 3xTg-AD mice). Second, the benefit observed in the NTg/+ groups in different works never reached significance, and thus no effect was expected in older animals. Third, we focused on the interplay between Aβ and tau pathologies in this work, and the NTg group did not develop tau pathology.
Animals were intraperitoneally administered every 3 days with 200 µg of scFv-h3D6 (~6.6 mg/kg) diluted in 200 µL of vehicle (PBS, pH 7.4), or with 200 µL of vehicle (referred as untreated for the sake of clarity), for 15 days. This was in contrast to previous studies, where a single dose of 3.3 mg/kg was administered to 5-month-old mice [18,[20][21][22]25] or repeatedly for 6 weeks since mice were 4.5 month-old [23].
All the experiments were performed in accordance with the requirements of the Committee on Ethics in Animal and Human Experimentation at UAB (CEEAH 0661). All experiments were approved by the UAB Animal Research Committee and the Government of Catalonia on May 15th 2020.
The degree of overlapping between Aβ and scFv-h3D6 signals in co-localization experiments was calculated using Mander's coefficient [72] from 3 sections per animal. tM1 means the proportion of 6E10 signal that overlaps with anti-scFv-h3D6 signal, and tM2 means the proportion of anti-scFv-h3D6 signal that overlaps with the 6E10 signal. Similarly, the degree of overlapping between Aβ and tau was shown as tM1, meaning the proportion of 6E10 signal that overlaps with the HT7 signal, and tM2, meaning the proportion of HT7 signal that overlaps with the 6E10 signal. Values were expressed as medians, maximum, and minimum. Mander's coefficient = 1 implied total overlapping, and Mander's coefficient = 0 implied no overlapping.
For the rest of the experiments, statistical differences among groups were evaluated with the non-parametric Kruskal-Wallis test. Next, a two-to-two comparison was performed using the Mann-Whitney U-test. Values were expressed by medians, maximum, and minimum, and depicted as boxplots. A p-value ≤ 0.05 was considered statistically significant. A p-value between 0.051 and 0.100 was considered a tendency [73]. The effect size was measured by computing Rank-biserial correlation coefficients (Table 1) with the use of the Wendt formula: r = 1 − (2U/n 1 n 2 ); where U is the Mann-Whitney -statistics, and n 1 and n 2 the sizes of the samples being compared. The correlation r expresses the difference between the proportions of pairs that support the hypothesis, minus the proportion of pairs that do not. For n 1 = 5 and n 2 = 5 the minimum value of r significant at p-value ≤ 0.05 is 0.68 [74]. It is important to note that retaining alive 10 22-month-old 3xTg-AD female mice born the same week was rather challenging since this mice model begins to die at 15-month-old.
Graphical and statistical analyses were performed using Graphpad Prism v6 (GraphPad Software, San Diego, CA, USA). Table 1. Comparison between the p-values (Mann-Whitney U-test) and the r-coefficients (using the obtained U-values in the Wendt formula [74]). Considering the size of the groups in this study (n = 5), r-coefficients ≥0.68 are significant at p-value ≤ 0.05, whereas values ≥0.92 (more restrictive because the r-coefficient ranges from 0 to 1) are significant at p-value ≤ 0.05 to p ≤ 0.01 [74].

Sample Collection and Processing
One hour after the last administration, animals were anesthetized using 1% isoflurane (Esteve, Barcelona, Spain) and euthanized by decapitation. This time period was based on the pharmacokinetic knowledge of scFv-h3D6 in this mouse model [25]. Brains were immediately removed from the skulls, rinsed in cold PBS, weighed, and dissected on ice. One hemisphere was rapidly fixed by immersion in 4% paraformaldehyde (PFA) diluted in PBS for 48 h. Samples were embedded in paraffin following common procedures, serially sectioned in the coronal plane (10-µm thick), and mounted on SuperfrostTM Plus microscope slides (Thermo Fisher Scientific, Waltham, MA, USA). Dissected areas (cerebral cortex and hippocampus) of the other hemisphere were mechanically homogenized using a tissue homogenizer (Sigma-Aldrich Química SL, Saint Louis, MO, USA) in cold TBS buffer (pH 7.6) supplemented with a protease inhibitor cocktail (Roche). After a brief sonication step (1 cycle of 35 s, at 35% duty cycle, and output 4, in a Dynatech Sonic Dismembrator ARTEK 300 (Biologics Inc., Manassas, VA, USA), the samples were centrifuged at 100,000× g at 4 • C for 1 h. The supernatant was labeled as the TBS-soluble fraction. The pellet was re-suspended in cold TBS, supplemented with an inhibitor cocktail and 2% SDS. Sonication and centrifugation steps were repeated, and the supernatant was labeled as the SDS-soluble fraction. Finally, the pellet was re-suspended in 70% formic acid and, after sonication and centrifugation, the supernatant was recovered and dried overnight in a vacuum concentrator (Savant SpeedVac, Thermo Fischer Scientific). The dried formic extract was re-suspended in DMSO and labeled as the FA-soluble fraction. All fractions were immediately stored at −80 • C until further use.

Immunohistochemistry
Sections were deparaffinized by immersion in xylene and rehydrated in serial dilutions of alcohol. Endogenous peroxidase activity was blocked by immersion in 3% H 2 O 2 in pure methanol for 10 min.

Image Capture and Processing
Bright-field images were captured using a Leica DMRB microscope and a Leica DFC 500 camera with a Leica PL Fluotar lens. Confocal imaging was performed with a spectral Fluoview-1000 microscope (Olympus, Shinjuku, Japan) by capturing 10 optical sections with the UPLSAO0 objective lens. The cerebral sections corresponded to the range of coordinates from Figure 43 (interaural 2.34 mm and Bregma −1.46 mm) to Figure 48 (interaural 1.74 mm and Bregma −2.06 mm) in The Mouse Brain in Stereotaxic Coordinates [75].
The number of plaques and percentage of 6E10-immunoreactive area were semi-quantified using the ImageJ software (NIH, Bethesda, MD, USA).