The Dysregulation of OGT/OGA Cycle Mediates Tau and APP Neuropathology in Down Syndrome

Protein O-GlcNAcylation is a nutrient-related post-translational modification that, since its discovery some 30 years ago, has been associated with the development of neurodegenerative diseases. As reported in Alzheimer’s disease (AD), flaws in the cerebral glucose uptake translate into reduced hexosamine biosynthetic pathway flux and subsequently lead to aberrant protein O-GlcNAcylation. Notably, the reduction of O-GlcNAcylated proteins involves also tau and APP, thus promoting their aberrant phosphorylation in AD brain and the onset of AD pathological markers. Down syndrome (DS) individuals are characterized by the early development of AD by the age of 60 and, although the two conditions present the same pathological hallmarks and share the alteration of many molecular mechanisms driving brain degeneration, no evidence has been sought on the implication of O-GlcNAcylation in DS pathology. Our study aimed to unravel for the first time the role of protein O-GlcNacylation in DS brain alterations positing the attention of potential trisomy-related mechanisms triggering the aberrant regulation of OGT/OGA cycle. We demonstrate the disruption of O-GlcNAcylation homeostasis, as an effect of altered OGT and OGA regulatory mechanism, and confirm the relevance of O-GlcNAcylation in the appearance of AD hallmarks in the brain of a murine model of DS. Furthermore, we provide evidence for the neuroprotective effects of brain-targeted OGA inhibition. Indeed, the rescue of OGA activity was able to restore protein O-GlcNAcylation, and reduce AD-related hallmarks and decreased protein nitration, possibly as effect of induced autophagy.


Introduction
Down syndrome (DS; Trisomy 21) is the most common chromosomal disorder and the most frequent genetic cause of intellectual disability affecting about 6 million people worldwide [1,2]. Because of the advances in health care and management of co-occurring illnesses, the life expectancy of people with DS has largely improved [3,4]. The triplication of genes on chromosome 21 and of their products can alter diverse pathways, including those involved with brain development, metabolism, and neuronal networks [5,6]. Individuals with DS are also more likely to develop certain pathological conditions, including hypothyroidism, autoimmune diseases, epilepsy, hematological disorders, and Alzheimer-like dementia [7]. The clinical manifestation of Alzheimer-like dementia in DS resembles that occurring in the general population [8,9], with slight differences in early presentation [10]. Nearly all individuals with full trisomy 21 aged 40 and older are found to have typical signs of Alzheimer's disease (AD) neuropathology, including extracellular amyloid plaques and intracellular neurofibrillary tangles [11]. The extra copy of amyloid precursor protein (APP) gene on chromosome 21 is associated with a 4-to 5-fold overexpression of APP that leads to an early onset and rapid accumulation of β-amyloid protein (Aβ) with age [9,12]. Cortical deposits of Aβ 1-42 have even been discovered as early as at 12 years of age [13]. Triplication of specific kinases (e.g., DYRK-1) interacting with APP and tau represents a further link between gene imbalance and neuropathological features DS [14]. Furthermore, brain hypoglycemia and insulin resistance are emerging as common mechanisms of neurodegeneration in DS and AD [15][16][17]. Several studies using FDG PET at different stages of AD revealed the potential role of reduced glucose uptake in driving neurodegeneration [18][19][20]. Brain samples from DS and AD demonstrated defects in insulin receptor signaling, decreased levels of brain glucose transporters, decreased activity of enzymes from the Krebs cycle, and a decline in mitochondrial respiratory chain complexes in the brain [5,16,[21][22][23][24].
Recent lines of evidence suggested that the nutrient-related dynamic changes of O-GlcNAcylation might underlie the neuropathological mechanisms observed in AD and DS [25][26][27][28]. O-GlcNAcylation is the non-canonical glycosylation of nucleocytoplasmic proteins with a single O-linked Nacetylglucosamine (O-GlcNAc) moiety. The dynamic cycling of O-GlcNAc on proteins is regulated by the concerted actions of two enzymes: the O-GlcNAc transferase (OGT) and a neutral β-hexosaminidase known as O-GlcNAcase (OGA). O-GlcNAcylation occurs widely in all tissues, but heavily in the brain, where OGT and OGA levels were found highly expressed [29][30][31][32]. Because O-GlcNAcylation depends on the availability of UDP-GlcNAc, the product of the hexosamine biosynthetic pathway (HBP), and in turn intracellular UDP-GlcNAc level determines OGT activity, O-GlcNAcylation is considered a valuable intracellular sensor of cell metabolic status that can be directly regulated in a nutrient-responsive manner [33]. Notably, O-GlcNAcylation and phosphorylation, by occurring on serine and threonine residues of proteins, are mutually related, allowing cells to modulate a variety of signaling pathways and transcription factors in response to nutrients or stress [34,35]. Many of the O-GlcNAcylated proteins are directly related to neurodegenerative diseases, including APP, α-synuclein, neurofilaments, tau, and synapsin I [26,[36][37][38][39][40]. Several studies have recently correlated brain hypoglycemia with decreased O-GlcNAcylation supporting their convergence to AD neurodegeneration [23,26,28,41]. The accepted hypothesis is that the general decrease of protein O-GlcNAcylation associated with the specific reduction of APP and tau GlcNAc levels may promote Aβ plaque formation and tau aggregation in the brain, thus contributing to cognitive decline [37,[42][43][44][45]. Noteworthy, several studies reported the increase of O-GlcNAc levels in the brain of AD subjects suggesting a more complex crosstalk between O-GlcNAcylation and neurodegenerative processes [42,46,47].
Given that DS and AD share similar metabolic alterations and common pathological markers within the brain [8,16,48], it is conceivable to suppose a role for aberrant O-GlcNAcylation in driving DS neurodegeneration. The present work discloses, for the first time, the reduction of GlcNAc levels in the hippocampus of a murine model of DS supporting the notion that this phenomenon, by heavily affecting APP and tau post-translational modifications, might have an evident role in the pathological mechanisms mediating the progression of AD-like dementia in DS individuals.

Materials and Methods
Animal Model E x p e r i m e n t s w e r e c o n d u c t e d o n T s 2 C j e m i c e (Rb(12.Ts171665Dn)2Cje) and correspondent euploid animals (B6EiC3SnF1). Ts2Cje are a well-established murine model of DS characterized by a triple copy of a Robertsonian fusion chromosome carrying the distal end of Chr16 and Chr12. Parental generations were purchased from Jackson Laboratories (Bar Harbor, ME, USA). Mouse colony was raised by repeated crossbreed of Ts2Cje (Ts2) trisomic females with euploid (Eu) males. Since these breeding pairs produce litters containing both trisomic and euploid offspring, resultant progeny was genotyped to determine the presence of the trisomic segment using quantitative PCR, as previously described by Reinoldth et al. [49]. Mice were housed in clear Plexiglas cages (20 × 22 × 20 cm) under standard laboratory conditions with a temperature of 22 ± 2°C and 70% humidity, a 12-h light/dark cycle and free access to food and water. For the initial longitudinal study, both Ts2 and Eu male and female mice were sacrificed by cervical dislocation at different time points (3,6,9,12, and 18 months of age) and brain areas were collected for preliminary western blot analysis. In order to perform the 6-month-focused analysis, Ts2 and Eu males and females at the selected age (n = 6/group) were perfused with saline solution through intracardiac injection. Brain was dissected in halves: one hemisphere was fixed for immunofluorescence analysis while the other section was used for remaining biochemical evaluations. A 2-way ANOVA analysis was performed to exclude the influence of sex (Sup. Tab. 1). All experiments were performed in strict compliance with the Italian National Laws (DL 116/92), and the European Communities Council Directives (86/609/EEC). Experimental protocol was approved by the Italian Ministry of Health (#1183/2016-PR). All efforts were made to minimize the number of animals used in the study and their suffering. All samples were flash-frozen and stored at − 80°C until utilization.

Thiamet-G Intranasal Treatment
A pilot study was performed to identify the effective Thiamet-G (TMG; HY-12588, MedChemExpress) dose to use for the intranasal treatment. This preliminary dose-response study was performed on a restricted number of Eu animals (n = 3/ group; Eu = 3 m) which were respectively administered with vehicle solution (Veh; PBS 1X solution) or 1 μg, 5 μg, 10 μg, 25 μg, and 50 μg of TMG solution. Animals were treated with a single intranasal delivery of 10 μL to each nostril and sacrificed 8 h later through cervical dislocation. Brain areas were collected, and western blot analysis was performed to evaluate treatment efficacy. Our data demonstrated that the intranasal administration of 25 μg of TMG was able to significantly increase the global levels of O-GlcNAcylated proteins (Sup. Fig. 7). Once the effective dose was identified, 6-monthold Ts2Cje and euploid mice were divided into four experimental groups (n = 6/group) according to genotype and intranasal treatment received. Based on previously published data on TMG treatment in rodents, which demonstrated that the administration of the drug achieved the peak of increased levels of O-GlcNAc proteins after 8-10 h [50], we opted to treat our animals 2 × day with the effective dose of 25 μg for 5 days, with the aim of achieving a stable elevation of O-GlcNAc proteins for a short period of time. Treatment was well tolerated although a physiological loss of weight was noticed due to prolonged animal manipulation. Indeed, an equal weight loss was observed in TMG-treated mice as well as animals treated with vehicle solution (Table 1). No change in food intake or drinking water consumption was observed. At the end of treatment, animals were euthanized and perfused with sterile PBS through an intracardiac puncture. After sacrifice, brain areas were saved for subsequent analysis.

Immunofluorescence
Entire brains from 6-month-old Ts2Cje mice and corresponding euploids were fixed in a 4% formaldehyde aqueous solution for 24 h at 4°C. Fixed brains were then cryoprotected for the next 48 h at 4°C with a solution containing 20% of sucrose and 0.02% of NaN 3 .
Brains were frozen on a temperature-controlled freezing stage, coronal sectioned (20 μm) on a sliding cryostat (Leica Biosystems, Wetzlar, Germany), and stored in a solution of PBS containing 0.02% NaN 3 at 4°C until utilization. Brain sections were mounted on glass slide. Once dried, brain sections underwent to a heat-induced antigen retrieval step in a 10 mM EDTA solution, pH = 6.0, for 20 min at 55°C [51]. After 4 washes with filtered PBS, sections were blocked with a solution containing 10% normal goat serum and 0.2% Triton X-100 in filtered PBS. Slides were then incubated 24 h at 4°C with following antibodies: GFAP (1:500; anti-rabbit; 840,001, BioLegend), IBA1 (1:250; anti-rabbit; GTX100042, GeneTex), NeuN-1 (1:500; anti-rabbit, 702,022, Invitrogen Thermo Fisher Scientific), O-GlcNAc CTD110.6 (1:100; anti-mouse, SC-59623, Santa Cruz Biotechnology), O-GlcNAc RL2 (1:50; anti-mouse, #MABS157, Sigma-Aldrich). Slides were then washed with filtered PBS and incubated with Alexa Fluor − 488 nm and − 594 nm secondary antibodies (1:1500; A11029, A11034, Invitrogen Thermo Fisher Scientific) for 1.5 h at room temperature. Tissues were then stained with Sudan black (0.1% Sudan Black B in 70% ethanol; 199,664, Sigma-Aldrich) to block auto fluorescence inherent to the sample. Slides were then washed, incubated with DAPI (10 mg/mL; IS-7712, Immunological Sciences) for 1 min and washed again. One slide per group was stained without primary antibodies to establish nonspecific background signal. At the end, cover slip glasses were placed using a drop of Fluoromount aqueous mounting medium (F4680, Sigma-Aldrich) and glasses were kept at room temperature to dry. All slides were imaged using Zeiss AXio (Carl Zeiss, Oberkochen, Germany). All immunolabeling acquisition intensities, field sizes, and microscopy settings were kept consistent across images. Images were analyzed using ImageJ. Image montages for figures were collated in Illustrator and Photoshop Cs6 (Adobe System) software programs and were based upon brain images that most closely approximated the group means.

Western Blot
All mice samples used for longitudinal and coming from intranasal TMG treatment were homogenized following the same procedure.

OGA Assay
OGA enzymatic activity was measured using the synthetic substrate p-nitrophenyl N-acetyl-β-D-glucosaminide (pNP-GlcNAc) as described by Zachara and colleagues [52]. All samples used for 6-month-focused analysis and intranasal TMG treatment were processed using the same procedure. Briefly were incubated with activity assay buffer containing 2 mM pNP-GlcNAc, 50 mM sodium cacodylate (pH = 6.4), 50 mM Nacetylgalactosamine, and 0.3% BSA at 37°C for 2 h. Reaction was stopped by the addition of 500 mM Na 2 CO 3 and absorbance was measured at 405 nm (Multiskan EX, Thermo Labsystems). OGA activity is reported as enzyme activity units where 1 U catalyzes the release of 1 μmol pNP/min from pNP-GlcNAc. OGA activity for each group was normalized on corresponding protein expression levels.

GFAT1 Assay
GFAT1 enzymatic activity was performed adapting a procedure developed by McClain and colleagues [53]. GFAT1 activity for 6-month-focused analysis was assessed through the measuring of its enzymatic product glucosamine 6-phosphate (GlcN6P). Briefly, 15 mg of hippocampus (n = 6/group) was thawed in 80 μL lysis buffer (pH = 7.5) containing 100 mM KCl, 1 mM EDTA, 50 mM Na 3 PO 4 , protease inhibitor cocktail (1:100; 539132, Millipore), and phosphatase inhibitor cocktail (1:100; P5726, Sigma-Aldrich). Brains were homogenized by 20 strokes of a Wheaton tissue homogenizer, sonicated and centrifuged at 14000 rpm for 40 min at 4°C to remove debris. Supernatant was collected and total protein concentration was determined by the BCA method (Pierce™ BCA Protein Assay Kit, 23227, Thermo Fisher Scientific) according to the manufacturer's instructions. Samples (240 μg of proteins) were incubated with activity assay buffer containing 1 mM EDTA, 1 mM DTT, 40 mM NaHPO 4 (pH = 7.4), 12 mM fructose 6-phosphate, and 12 mM L-glutamine at 37°C for 45 min. Reaction was stopped by the addition of PCA 1 M (1:2) to induce protein precipitation. Samples were then incubated 10 min on ice and centrifuged at 16000g 4°C for 10 min. Supernatant was extracted with chloroform (1:2) and 100 μL of the aqueous phase was collected for HPLC analysis. GlcN6P generated during the reaction was detected by derivatization of the sample with 2 volumes of ophthalaldehyde (OPA) reagent (100 μL of 10 mg/mL OPA in EtOH, 900 μL sodium borate 100 mM pH = 9.7, and 2 μL 3-mercaptopropionic acid). Reaction was incubated for 10 min at room temperature protected from light, and sample was diluted 1:1000 in the mobile phase for HPLC detection. Chromatographic separation was performed using an isocratic elution; the mobile phase was composed by Na 3 PO 4 15 mM, pH = 7.2 (phase A) and acetonitrile (phase B) (90:10). A Symmetry C18 column (300 Å, 5 μm, 4.6 mm × 250 mm, 1/pk, Waters Corporation) was used for separation. Fluorescence of the sample eluent (λ = 340/450) was analyzed using a fluorescent detector (RF-551, Shimadzu) and the peak area was integrated using dedicated software (Empower 2, Waters Corporation). OPA-derivatized GlcN6P standards (G5509, Sigma-Aldrich) were run separately to determine the retention time (1.8 s) and to generate a standard curve to correlate area to activity. The correlation coefficient between the concentration of GlcN6P standards and the area under the GlcN6P peak was 0.999. Activity is expressed as units/ milligram of protein where 1 U represents the generation of 1 pmol of GlcN6P/min. GFAT1 activity for each group was normalized on corresponding protein expression levels.

RNA Extraction and Quantitative Real-Time RT PCR
RNA was extracted from the hippocampus of Ts2Cje and Eu treated both with TMG and vehicle (n = 6/group) using Tissue Total RNA Kit according to the manufacturer's instructions (Abcam). RNA was quantified using the Biospec Nano spectrophotometer (Shimadzu, Columbia, MD, USA), and RNA was reverse transcribed using the cDNA High-Capacity kit (Applied Biosystems, Foster City, CA, USA), including reverse transcriptase, random primers, and buffer according to the manufacturer's instructions. The cDNA was produced through a series of heating and annealing cycles in the MultiGene OPTIMAX 96-well thermocycler (LabNet International, Edison, NJ, USA). Real-time PCR (Q-PCR) was carried out using the following cycling conditions: 35 cycles of denaturation at 95°C for 20 s; annealing and extension at 60°C for 20 s, using the SensiFAST™ SYBR® No-ROX Kit (Bioline, London, UK). PCR reactions were carried out in a 20-μl reaction volume in a CFX Connect Real Time PCR machine (Bio-Rad Laboratories). Primers used for the evaluation of gene expression were designed as follows: GADPH (Fw: ACAGTCCATGCCATCACTGCC; Rv:

Statistical Analysis
Statistical analyses were performed using Student t test for the evaluation of differences between two groups and a non-parametric 1-way ANOVA with post hoc Bonferroni t test for the evaluation of differences between more than two groups. To determine the influence of genotype (Ts2Cje and euploids), treatment (TMG and vehicle), or sex (male and female), we performed a 2-way ANOVA analysis. Data are expressed as mean ± SEM per group. All statistical analyses were performed using Graph Pad Prism 8.0 software (GraphPad, La Jolla, CA, USA).

Ts2Cje Mice Show an Aberrant and Tissue-Specific O-GlcNAcylation Profile at 6 Months of Age
In the last few years, strong interest has been focused on understanding the role of protein O-GlcNAcylation in the development of AD-associated pathological features [26,28,36,38,[55][56][57]. We took advantage of the Ts2Cje (Ts2) mice to investigate the O-GlcNAcylation profile and its relevance in DS neurodegeneration. To begin, we conducted an agedependent study with the aim of assessing putative changes of the total levels of O-GlcNAcylated proteins in the hippocampus from 3-, 6-, 9-, and 12-month-old Ts2 compared with aged-matched euploids (Eu). We observed, in Ts2 mice, a premature reduction of O-GlcNAcylated proteins as early as at 6 months of age. Indeed, we found a trend of increase of total O-GlcNAc levels in 3-month-old Ts2 hippocampus, compared to aged-matched euploids, but a sudden significant switch is shown at 6 months of age ( Fig. 1(A); **p < 0.01, Eu vs Ts2: − 18%). A trend of reduced O-GlcNAc levels persisted also in Ts2 hippocampus at 9 months of age and 12 months of age compared to respective euploids, suggesting the reduction of O-GlcNAc levels during aging. Accordingly, we observed by immunofluorescence microscopy a diffuse reduction of O-GlcNAcylated proteins in the entire hippocampus area of Ts2 mice at 6 months of age compared to respective euploids ( Fig. 1(E-F); **p < 0.01, Eu vs Ts2: − 10%).
Considering the significant alteration of protein O-GlcNAcylation in 6-month-old DS mice, we focused our following experiments on this age group. To investigate the interplay between O-GlcNAcylation and O-phosphorylation on serine-threonine residues [35,38,45,58,59], we evaluated the levels of total protein phosphorylation on these residues only. As postulated, we detected a significant increase of total protein phosphorylation in Ts2 hippocampus compared to respective euploids at 6 months of age ( Fig. 1(C-D); *p < 0.05, Eu vs Ts2: + 44%). Several studies supported that in AD and metabolic diseases, the reduction of protein O-GlcNAcylation is a brain-specific effect associated with reduced glucose uptake, altered HBP flux, and/or aberrant phosphorylation process [23,26,41,60,61]. In contrast, peripheral organs often  [29,31,55,66]. Furthermore, O-GlcNAcylation plays a role in regulating hippocampal synaptic transmission and plasticity, thus influencing learning and memory processes [67,68]. Considering the relevance of protein O-GlcNAcylation in this brain area, we further examined the distribution of O-GlcNAcylated proteins in different subregions of Ts2 hippocampus at 6 months of age. As expected, Ts2 mice showed a general impairment of protein O-GlcNAcylation compared to respective euploids in each of the hippocampal subregions analyzed (Fig. 2(A1-2)). In detail, a relevant reduction of O-GlcNAc fluorescent signal was observed in the CA1 area ( Fig. 2(B1-6, E); **p < 0.01, Eu vs Ts2: − 20%), in the CA3 area ( Fig. 2(C1-6, E); *p < 0.05, Eu vs Ts2: − 70%), and in the dentate gyrus (DG) region ( Fig. 2(D1-6, E); *p < 0.05, Eu vs Ts2: − 30%) of the hippocampus from Ts2 mice compared to the respective euploid mice. In addition, Ts2 mice showed a different distribution of O-GlcNAcylated protein reductions, with a higher decrease O-GlcNAc levels in the CA3 subregion compared to both the DG area and the CA1 area ( Fig. 2(E)). In the second instance, we evaluated cell-type distribution of O-GlcNAcylated proteins in the CA3 subregion, which has proved to be the area most affected by the reduction of O-GlcNAc fluorescence intensity in Ts2 mice. We determined that protein O-GlcNAcylation occurs primarily in neurons in Ts2Cje (Fig. 3(A5-8)). Indeed, co-localization analysis showed significant O-GlcNAc signal (green) that overlaps with the neuronal marker NeuN-1 (red) (Fig. 3(A5-8)), while partial or no overlap between O-GlcNAc and microglia (IBA-1) or astrocytes (GFAP) occurs ( Fig. 3(B5-8, C5-8) respectively) in the CA3 of 6-month-old Ts2 mice. Costes' and Manders' coefficient (M1 and M2) analyses of colocalization showed that 90% of the O-GlcNAc signal colocalized with neurons [69]. In Eu mice, co-localization analysis demonstrated a strong correlation as well with NeuN-1 and O-GlcNAc that, although with lower values, persists also in microglia and astrocytes. These results suggest a cell-type specific impairment of O-GlcNAcylated proteins in Ts2Cje at 6 months of age and a consistent reduction of protein O-GlcNAcylation in neurons and astrocytes. We note that the analysis of GFAP and IBA-1 markers by immunochemical methods demonstrated no significant changes in Ts2 compared to Eu mice, thus supporting no astrocytosis or microgliosis at 6 months of age (Sup. Fig 2) [70].
Reduced O-GlcNAcylation Results from the Dysregulation of OGT/OGA Cycle Considering the relevance of O-GlcNAcylation cycle homeostasis, many regulatory mechanisms exist to balance OGA/OGT activity, thus calibrating protein O-GlcNAc levels according to cellular status [71]. We investigated OGA/OGT functionality in the hippocampus of Ts2 mice to test whether a dysfunctional cycling could be responsible for the observed reduction of global O-GlcNAcylation. Ts2 mice showed no difference in both OGT protein expression (Fig. 4(A-B)) and transcript ( Fig. 4(C)) compared to the aged-matched Eu group. Since OGT itself undergoes to O-GlcNAcylation and phosphorylation on different sites, we performed an immunoprecipitation assay to analyze OGT PTMs that could potentially affect its ability to transfer O-GlcNAc moiety [61,66]. We noted a significant reduction in O-GlcNAc OGT/ OGT levels in Ts2 mice compared to the Eu group (Fig. 4(D-E); *p < 0.05, Eu vs Ts2: − 67%) and a trend on increase in p Ser/ Thr OGT/OGT levels ( Fig. 4(D-E); Eu vs Ts2: + 20%). The reduction in the O-GlcNAc/phosphorylation ratio of OGT ( Fig.  4(D-E); **p < 0.01, Eu vs Ts2: − 68%) in Ts2 hippocampus suggests that the global reduction of O-GlcNAcylation might result from OGT-altered functionality. Subsequently, we analyzed the removal process of the O-GlcNAc moiety, measuring both OGA levels and enzymatic activity. OGA resulted as significantly more expressed in the hippocampus of 6-month-old Ts2 mice (Fig. 4(F-G); **p < 0.01. Eu vs Ts2: + 15%) in comparison with the control group. Furthermore, OGA protein levels also reflect the difference in mRNA transcript, which is significantly increased in Ts2 mice ( Fig. 4(H); *p < 0.05; Eu Vs Ts2: + 1.20-fold change), confirming an upregulation of OGA in Ts2 mice at 6 months of age. The analysis of OGA activity demonstrated an upregulation of the global removal process of O-GlcNAc moiety (Fig. 4(I), left panel; *p < 0.01. Eu vs Ts2: + 29%) that appeared to be strictly associated with its increased expression; indeed, no changes in enzyme activity were detected after normalization on protein levels (specific activity) (Fig. 4(I), right panel). In detail, our data suggests that O-GlcNAc removal process is markedly increased in Ts2 hippocampus at 6 months of age compared to euploids as effect of increased OGA protein levels but not of OGA-specific hydrolytic activity increase. Therefore, we support that increased OGA levels promote the aberrant subtraction of O-GlcNAc moiety from serine and threonine residues, thus leading to the global reduction of O-GlcNAcylated proteins in Ts2Cje mice.

The Aberrant O-GlcNAc/Phosphorylation Ratio of Tau and APP Drives AD-Like Neurodegeneration in Ts2Cje Mice
The abnormal hyperphosphorylation of tau, on specific serine and threonine residues, induces protein self-assembly and gives rise to toxic NFTs, a well-established hallmark of ADlike pathology [44,58]. Recent studies highlighted that tau phosphorylation is inversely regulated by O-GlcNAc and that tau O-GlcNAcylation plays a key role in hindering its aggregation [25,37,41,57,72]. Our data confirmed the aberrant phosphorylation of tau protein in 6-month-old Ts2 mice compared to Eu group on both Ser202-Thr205 residues (Fig. 5(A-B), AT8/tau; *p < 0.05; Eu vs Ts2: + 75%) and Ser404 (Fig.  5(A-B); *p < 0.05; Eu vs Ts2: + 14%). Furthermore, through immunoprecipitation analysis, we found that increased phosphorylation of tau is associated with a significant reduction of its O-GlcNAcylation (Fig. 5(C-D); *p < 0.05; Eu vs Ts2: − 20%). These results suggest a role for O-GlcNAc levels in the early disturbance of tau PTMs and confirm, in Ts2 neuropathology, the mutual inverse relationship between tau-reduced O-GlcNAcylated levels and its increased phosphorylation.
It was largely proven that APP undergoes O-GlcNAcylation [43] and recent advances demonstrated that increased APP O-GlcNAc levels could switch its processing from the on DAPI signal is reported for each hippocampal region from Ts2 and Eu animals. Number of animals for each condition was as follows: n = 3/group for immunofluorescence staining. All bar charts reported in (E) show mean ± SEM. *p < 0.05, **p < 0.01, using Student's t test amyloidogenic pathway to the non-amyloidogenic via, thus reducing the production of Aβ plaques [73]. Since APP is encoded on chromosome 21, the role of O-GlcNAcylation in APP processing could be further exacerbated in DS neuropathology. As previously demonstrated [74], APP protein levels were significantly higher in Ts2 mice compared to respective Eu controls (Fig. 5(E, G); **p < 0.01, Eu vs Ts2: + 60%). Subsequently, O-GlcNAc APP/APP levels were demonstrated to be significantly reduced in 6-month-old Ts2 mice compared to Eu group (Fig. 5(F,  H); **p < 0.01, Eu vs Ts2: − 35%) and in parallel a significant increase in p Ser/Thr APP/APP levels ( Fig. 5(F, H); *p < 0.05, Eu vs Ts2: + 23%) was observed.

Ts2Cje Mice Show Alteration of the HBP and the Induction of the Insulin Cascade
Since DS is characterized by a significant altered metabolic profile, with a prevalence of less efficient fermentative metabolism [75,76], we decided to analyze the activation state of AMP-activated protein kinase (AMPK) in our Ts2 model and Co-localization graphs are reported for both Eu (B4) and Ts2 (B8) mice; O-GlcNAc (green); IBA-1 (red); co-localization (yellow). (C1-8) O-GlcNAc colocalization with the astrocytic marker GFAP in the CA3 hippocampal area from Ts2 and respective euploid mice. Consistent co-localization of O-GlcNAc signal and GFAP was detected in the CA3 area of euploid animals (C1-4), while apparently no signal has been identified in the same area of Ts2 hippocampus (C5-8). Co-localization graphs are reported for both Eu (C4) and Ts2 (C8) mice; O-GlcNAc (green); GFAP (red); co-localization (yellow). Number of animals for each condition was as follows: n = 3/group for immunofluorescence staining its relevance in the control of HBP at early stage of disease. A significant increase in AMPK protein levels was observed in the hippocampus of 6-month-old Ts2 compared to the Eu group ( Fig. 6(A-B); **p < 0.01, Eu vs Ts2: + 78%), together with a significant reduction of AMPK activation measured by its p Thr172 AMPK levels normalized on respective protein levels ( Fig. 6(A-B); **p < 0.01, Eu vs Ts2: + 40%). The analysis of GFAT1 protein levels showed a significant difference among the two groups ( Fig. 6(C-D); *p < 0.05, Eu vs Ts2: + 50%), while GFAT1 phosphorylation on Ser 243, which is controlled by AMPK and regulates the inhibition of its catalytic activity, showed a significant reduction in Ts2 mice compared to euploids (Fig. 6(C-D); *p < 0.05; Eu Vs Ts2: + 34%). In order to evaluate GFAT1 activation state, we took advantage of an HPLC-based method to measure its direct enzymatic product. In accordance with the lack of AMPK inhibitory effect, a trend of increase was observed in GFAT1 global ability to synthetize glucosamine-6-phosphate in Ts2 mice compared to the Eu group (Fig. 6(E-G)). However, since GFAT1 protein levels change between Ts2 and Eu animals, the activity was also normalized on respective protein levels demonstrating no significant alterations. To fully clarify if Ts2 shows metabolic alterations that can impact the HBP and the O-GlcNAcylation process, we evaluated the insulin cascade and/or glucose uptake. At first, we analyzed in 6-month-old animals the phosphorylated levels of IR on Tyr1146-1150-1151 to assess its activation state: a significant increase in the phosphorylated levels of IR/IR was observed in Ts2 hippocampus compared to respective euploids (Fig. 6(H-I); *p < 0.05, Eu vs Ts2: + 130%), while no significant changes were noticed on IR protein levels in the two groups. The evaluation of IR's direct substrate, IRS-1, demonstrated a significant increase in p Tyr632 IRS-1 (activatory site) compared to p Ser636 IRS (inhibitory site) in Ts2 mice compared to the Eu group ( Fig. 6(H, L); *p < 0.05, Eu vs Ts2: + 60%). No relevant differences were observed in IRS-1 protein levels between the two groups. According to our results, Ts2 mice at 6 months of age suggest the increased activation of the insulin cascade compared to respective euploids. The subsequent analysis of GSK3β demonstrated the inhibition of the protein (Sup. Fig. 5) supporting, as reported in human brain, an uncoupling of the signal [22]. Furthermore, the observed Ts2 mice compared to respective euploids. Both OGA protein and mRNA levels were found significantly increased Ts2 mice in comparison to the respective control group. Representative blot is reported in (F). (I) OGA activity assay. Global OGA activity is significantly increased in Ts2 mice compared to the respective control group. However, the enzyme-specific hydrolytic activity of OGA obtained through normalization on respective protein levels does not show relevant changes in the two groups. Number of animals for each condition was as follows: n = 6/group for western blot and RT-qPCR, n = 4/group for immunoprecipitation analysis, and n = 5/group for OGA activity assay. All bar charts reported in (B), (C), (D), (G), (H), and (I) show mean ± SEM. *p < 0.05, **p < 0.01, using Student's t test activation of the insulin cascade in Ts2 mice was associated with unaltered levels of brain glucose transporters and GLUT4 translocation to the membrane as suggested by unaltered p Thr142 As160/AS160 ratio in Ts2 mice hippocampus (Sup. Fig. 6).

Intranasal Thiamet G Rescued Aberrant Protein O-GlcNAcylation and OGA Activity in 6-Month-Old Ts2Cje Mice
Our results demonstrated that aberrant protein O-GlcNAcylation occurs in Ts2 mice as a result of increased OGA expression, which plays a significant role in the onset of AD-related markers. Hence, to question the possible neuroprotective effects of rescuing protein O-GlcNAcylation in Ts2 neuropathology, we performed a short-term intranasal treatment with Thiamet G (TMG), a potent and selective OGA inhibitor. Both 6-month-old Ts2 and respective Eu were administered by intranasal route with vehicle solution (Veh) or 25 μg of TMG (Fig. 7(A)), with the aim of targeting the brain directly and avoiding effects on peripheral organs [77] that might behave differently in terms of O-GlcNAcylation. After 5 days of treatment, we observed that TMG treatment was able to rescue lower levels of O-GlcNAcylated proteins in Ts2 hippocampus, restoring euploid-physiological levels ( Fig.  7(B, D); **p < 0.01, Ts2 Veh vs Ts2 TMG: + 46%), affecting neither OGT protein levels ( Fig. 7(C, E)) nor transcript (Fig.  7(F)). Accordingly, 2-way ANOVA analysis supported the effect of TMG treatment in rising O-GlcNAcylated protein levels ( Table 2; F (1,21) = 21.25, ***p < 0.001). In agreement with our hypothesis, the analysis of OGA enzymatic activity demonstrated the efficacy of the intranasal TMG administration in inhibiting the aberrant removal of O-GlcNAc moiety from serine and threonine residues of hippocampal proteins. Indeed, TMG treatment induced a significant reduction of OGA activity normalized on protein levels (specific activity) in Ts2-treated mice compared to respective Ts2 treated with vehicle solution (Fig. 7(I), right panel; *p < 0.05: Ts2 Veh vs Ts2 TMG: − 24%). In any event, due to increased OGA expression levels in treated-Ts2Cje mice, we did not observe reduction on global OGA enzyme activity. Furthermore, we confirmed a genotype-dependent upregulation of both OGA protein levels ( Fig. 7(C, G)) and transcript (Fig. 7(H)) in Ts2 mice compared to euploids treated with vehicle (Table 2; OGA protein: F (1,22) = 3.7, *p < 0.05; OGA mRNA: F (1,20) = 5.3, *p < 0.05). Intriguingly, as described elsewhere [78,79], TMG treatment was able to trigger a significant increase of OGA protein levels in both Eu (Fig. 7(C, G); *p < 0.05, Eu Veh vs Eu TMG: + 27%) and Ts2 animals ( Fig. 7(C, G); ***p < 0.001, Ts2 Veh vs Ts2 TMG: + 20%). Furthermore, an effect of TMG treatment in increasing OGA protein levels was also assessed by 2-way ANOVA analysis ( Table 2; F (1,22) = 16.8, ***p < 0.001). Surprisingly, upregulated OGA protein levels did not significantly reflect OGA mRNA levels both in Eu and Ts2 mice treated with TMG ( Fig.  7(H)). The effect of TMG on OGA supports the occurrence of Analysis of AMPK activation status in Ts2 mice compared to respective euploids. A significant increase in the AMPK protein levels was observed in Ts2 mice compared to Eu, together with a significant impairment in p Thr172 AMPK/AMPK levels, thus resulting in reduced AMPK activation. Representative blots are reported in (A). (C-D) Analysis of GFAT1 activation status in Ts2 mice compared to respective euploids. A significant increase in GFAT1 levels was observed in Ts2 mice compared to the control group, together with an impairment of p Ser243 GFAT1/ GFAT1 ratio, resulting in reduced GFAT1 inhibition. Representative blots are reported in (D). (E-G) GFAT1 activity assay. GFAT1 global activity showed a trend of increase in Ts2 hippocampus compared to respective euploids, while GFAT1 enzymatic normalized on corresponding protein expression levels showed no alteration.
Representative spectra of GFAT1-synthetized glucosamine-6-phosphate for both Ts2 and euploid animals are reported in (E) and (F) and respective bar graph of global and normalized activity (G). (H-L) Analysis of the insulin cascade in Ts2 mice compared to respective euploids. A significant increase in the phosphorylated levels of insulin receptor (Tyr1146-1150-1151)/IR was observed in Ts2 mice compared to Eu (I). Ts2 mice also showed an increase in the activation of the insulin receptor substrate (IRS-1), with increased ratio between phosphorylated levels on activatory site (Tyr632) and inhibitory site (Ser636) (L). Representative blots are reported in (H). Number of animals for each condition were as follow: n = 6/group for both western blot analysis and n = 5/group for GFAT1 activity assay. All bar charts reported in (B), (C), (G), (I), and (L) show mean ± SEM. *p < 0.05, **p < 0.01, using Student's t test a TMG-induced compensatory mechanisms as a result of the inhibition of OGA catalytic activity [80].

Discussion
In the last decade, a great deal of effort has been made to investigate the role of protein O-GlcNAcylation in neurodegenerative diseases; however, no evidence is available on its possible implications in DS neuropathology. Our analysis in Ts2Cje mice showed a persistent reduction of total GlcNAc bound to protein in each area of the hippocampal region, suggesting the alteration of O-GlcNAcylation homeostasis as an early molecular event that could increase susceptibility to neurodegenerative phenotypes. Previous reports demonstrated that decreased protein O-GlcNAcylation in the hippocampal region drives synaptic and cognitive decline in the aging brain, facilitating later onset of dementia [97]. Among the proposed mechanisms through which loss of O-GlcNAcylation promotes neurodegenerative processes, the extensive interplay of this PTM with protein phosphorylation has special relevance [38,45,58]. In agreement, our analysis showed increased global protein phosphorylation on Ser/Thr residues and the imbalance of the O-GlcNAcylation/phosphorylation equilibrium in the hippocampal area, recapitulating the alterations observed in the brain of both AD humans and murine models [23,26,28,38,41,55,98]. Most interesting, the reciprocal interplay between protein O-GlcNAcylation and phosphorylation is not only related to the competitive modification of the same residues but also to the ability of each modification to regulate the other's enzymatic machinery. Recently, insulin stimulation of 3T3-L1 adipocytes was proved to increase both OGT tyrosine phosphorylation and catalytic activity [99]. Furthermore, Kaasik et al. found that GSK3β can enhance OGT activity by phosphorylation on Ser residues [61]. However, the decreased O-GlcNAcylation of OGT, observed in our model, may alter the sites available for phosphorylation independently of GSK3β, thus affecting its activity. Among control mechanisms of O-GlcNAc cycling, fluctuations in protein O-GlcNAcylation are known to affect both OGT and OGA transcription in physiological context, in the direction of compensating the imbalance [71,80]. Recent evidence has   [27,40,42,97,102]. In Ts2 mice, the relevant reduction in protein O-GlcNAcylation was concomitant with a significant increase of both OGA transcript and protein levels, hinting at a lack of compensation for O-GlcNAc imbalance. Taken together, our data suggest that the overall removal process of O-GlcNAc moiety is markedly increased in young DS animals as an effect of OGA overexpression. We cannot exclude that increased gene dosage, occurring in DS both human and mice, might directly or indirectly drive the overexpression of OGA transcript and protein. However, no interaction between triplicated genes in DS and OGA gene sequence or product has yet been observed, and further studies in that sense are needed. Molecular pathways regulating the O-GlcNAcylation of proteins include the HBP, an offshoot of the glycolytic flux that integrates several major metabolic pathways into the synthesis of UDP-GlcNAc. Zibrova et al. demonstrated that GFAT1, the first rate-limiting enzyme of the HBP, undergoes negative regulation through increased phosphorylation on Ser243, by the action of AMPK, one of the master sensors of cellular energy. Moreover, AMPK itself was found to be O-GlcNAcylated on several residues, allowing to theorize further feedback mechanisms between these systems [103,104]. In line with the reduced AMPK activation (Thr172 phosphorylation), GFAT1 inhibitory phosphorylation decreases and a trend of increase is observed for GFAT1 enzyme activity. In addition, HBP is finely regulated by nutrient availability and brain metabolic changes, and many studies demonstrated that DS brain is characterized by the early presence of insulin resistance markers, which precede and contribute the development of AD-like brain damage [16]. We demonstrate in Ts2 mice at 6 months of age that reduced GlcNAc levels couple with the overactivation of insulin signal with no massive defects in glucose uptake. These events suggest not considering the alteration of glucose utilization as a possible negative regulator of the HBP flux. Despite that, prolonged IR stimulation is known to result in IRS-1 inhibition through negative feedback mechanisms [105], though the early alteration of O-GlcNAcylation cycling may interfere with correct insulin signaling in young Ts2 mice, as observed in AD mice [55], paving the way for the onset of insulin resistance c o n d i t i o n o b s e r v e d i n D S b r a i n [ 1 6 , 2 2 , 7 7 ] . Accordingly, streptozotocin-treated rats showed decreased global O-GlcNAcylation before the appearance of commonly recognized markers of insulin resistance, confirming the relevance of O-GlcNAcylation disturbances in the onset of insulin signaling defects [56].
In the last decade, OGA inhibitors have been proposed as a promising approach to recover the pathological implications of reduced O-GlcNAcylation in neurodegenerative diseases. Growing evidence in murine models of AD have shown that rescuing brain GlcNAc levels, by OGA inhibition, reduces the levels of pathological tau [36,[106][107][108], limits APP amyloidogenic cleavage and Aβ accumulation [84,85], boosts mitochondrial activity [28], and promotes the removal of toxic aggregates through macro-autophagy [90]. Data collected in Ts2 mice support a role for dysregulated OGT/OGA cycle in promoting unbalanced O-GlcNAcylation/phosphorylation ratio of tau and APP, thus proposing a role for their aberrant PTMs in favoring the accumulation of their resulting toxic aggregates. Further, previous data in Ts65Dn mice, which share the same genetic background with Ts2, demonstrate altered mitochondrial function and impaired autophagy [77,109]. In this scenario, Ts2 mice describe a pathological context that strongly encouraged the administration of OGA inhibitors to rescue GlcNAc levels and protect the brain from AD-like neurodegeneration. Yet, our analysis of liver samples in Ts2 mice showed an opposite profile in terms of O-GlcNAcylated/phosphorylated ratio, as observed in diabetes [21,102,110], advising for a brain-specific targeting of OGA with the aim to directly subject the brain and avoid possible disturbances in other organs. The intranasal administration of TMG provided positive outcomes in rescuing protein O-GlcNAcylation levels through the inhibition of OGA activity in Ts2 hippocampus. Such effect was followed by the recovery of O-GlcNAc tau deficiency and site-specific reduction of tau phosphorylation. Indeed, upon TMG administration, a reduction on Ser404 phosphorylation was observed while Ser202 and Thr205 were not affected in our model. However, the sitespecific effect on tau phosphorylation should not be considered a surprise since compelling evidence indicated that the neuroprotective effect of increased O-GlcNAc tau levels is not necessarily associated with the reduction of each tau phosphorylation sites [85]. In agreement, a previous study by Yang Yu et al., regarding the effects of acute TMG treatment, observed a reduction in tau phosphorylation on Ser404, while an increase in Ser202 was reported together with a timedependent effect on Thr205 [108]. As a matter of fact, a sitespecific reciprocity between O-GlcNAcylation and phosphorylation was described for Ser404 [58,111]. As regarding APP, TMG has proved to compensate for reduced O-GlcNAc APP levels that characterized DS mice and was able to further reduce its aberrant hyper-phosphorylation. TMGdriven modulation of APP PTMs acquires special relevance in DS neuropathology. Indeed, DS-affected subjects have increased APP gene dosage and overexpress APP, showing early signs of Aβ buildup [13,112]. Among APP residues that can be O-GlcNAcylated, Thr576 is known to regulate APP trafficking and processing, attenuating Aβ generation [39]. Conversely, increased levels of phosphorylated APP on Thr668 are found in AD brains and seem to facilitate BACE1 cleavage of APP that results in higher Aβ production [81,82]. In line with this neuroprotective role of APP O-GlcNAcylation, our TMG treatment indicates a redirection of APP toward a non-amyloidogenic processing as suggested by the trend of reduction of β-CTF/α-CTF ratio and of soluble Aβ 1-42. Despite the absence of amyloid plaques, elevation of APP, β-CTF, and soluble Aβ was demonstrated during aging in the hippocampus of the Ts65Dn model [86]. Since increased soluble Aβ oligomers characterized middle-aged Ts65Dn mice [87] only, the absence of increased soluble Aβ 1-42 peptide at 6 months of age in our model was not surprising. However, the modulation of O-GlcNAc APP levels by OGA inhibition suggested a role in the reduction of soluble Aβ 1-42 levels in both Eu and Ts2 mice, as reported elsewhere [73]. In line with this neuroprotective effect, TMG treatment also proved to boost PSD95, Syntaxin 1A, and BDNF expression, whose rescue is commonly associated with improved synaptic plasticity and amelioration of cognitive performances [77,89]. Although we have shown the positive effects of increasing O-GlcNAcylated levels of both tau and APP PTMs, this is not the only mechanism through which TMG exerts its neuroprotective purposes. A recent study by Zhu et al. strongly indicates that TMG acts in the brain to induce the disposal of toxic aggregates through the enhancement of autophagy [90]. The ability of TMG to modulate autophagy holds particular importance in DS, where defective degradative systems are known to exacerbate neurodegenerative phenotype by preventing proper aggregate clearance [22,77,95,113]. Despite this, Ts2 mice did not show signs of massive autophagy impairment at 6 months of age, and an increase of SQSTM1 was observed, possibly indicating insufficient autolysosomal degradation [114]. TMG treatment demonstrated boosting autophagic flux in our DS mice at both initial and later steps. Indeed, increased levels of Atg7 and Beclin-1 support an effect of TMG treatment at the level of autophagic induction, while increased LC3-II/I ratio and the reduction of SQSTM1 levels suggest a TMG-driven increase of autophagosome maturation and lysosomal degradation in DS mice. The valuable interplay between O-GlcNAcylation and autophagy was also observed in a number of different in vitro and in vivo models, where authors showed how reducing protein O-GlcNAcylation by genetic and pharmacological manipulation of OGT or OGA resulted in increased Fig. 10 Role of distupted O-GlcNAcylation homeostasis in DS neuropathology. The hexosamine biosynthetic pathway (HBP) is a minor branch of the glycolitic pathway that results in the production of UDP-GlcNAc, the activated substrate for protein O-GlcNAcylation. As the HBP flux integrates molecules from carbohydrate (fructose-6phosphate), amino acid (glutamine/glucosamine), neucleotide (UTP), and lipid (Acetyl-CoA) metabolism, the production of UDP-GlcNAc is considered a valuable intracellular sensor of cell metabolic status. An early upregulation of the insulin cascade (IR-IRS) is present in our DS murine model, which could reasonably imply an increase in glucose availability. In line with that, the HBP rate-limiting enzyme GFAT1 lack the inhibitory action of the metabolic-sensor kinase AMPK. Therefore, the altered OGT functionality and, mostly, aberrant increase of OGA-driven hydrolysis of O-GlcNAc seem to be the main cause for reduced protein O-GlcNAcylation in our DS model. The loss of protein O-GlcNAcylation is kown to give rise to an aberrant increase of protein phosphorylation, because of the mutual inverse relationship between these two modifications. This process aquires particular relevance when it comes to the post-transaltional modifications of proteins implicated in the DS neurodegenerative process. Indeed, the unbalanced O-GlcNAcylation/phosphorylation ratio of tau is known to promote its aggregated forms (NTFs, neurofibrillary tangles) while the aberrant increase of phosphorylated APP favors its amyloidogenic cleavage that results in the formation of β-CTF and thus, to β-amyloid accumulation. In this scenario, data collected on our model confirm the relevance of O-GlcNAcylation disruption in the appearance of AD-related hallmarks. Furthermore, TMG-mediated inhibition of OGA proved to restore protein O-GlcNAcylation and further exert its neuroprotective effects by boosting autophagic clearance of toxic aggregates and inducing the expression of proteins related to synaptic transmission autophagy [90,[115][116][117]. In this scenario, our results support that recovering protein O-GlcNAcylation in an altered phenotype such as DS could be an effective strategy to activate autophagic flux.
Notably, TMG treatment was also able to ameliorate the increased 3-NT levels which characterize the pathological alterations occurring in the brain of DS human and mice. Several lines of evidence demonstrated a role for autophagy in the removal of oxidized proteins [77,96,[118][119][120] and our data from TMG treatment support this hypothesis. However, the decreased burden of nitrated proteins might also be associated with the reduced formation of toxic tau and APP aggregates and/or with the induction/modulation of antioxidant pathways regulated by the O-GlcNAcylation process.
Overall, our study supports a pathological role for reduced O-GlcNAcylation in DS mice and poses the dysregulation of OGT/OGA cycle as a central contributor to tau and APP hyperphosphorylation. Inhibiting brain OGA activity, by intranasal TMG administration, recovers total and specific GlcNAc levels of tau and APP, and protects the hippocampus from increased protein nitration in a mechanism involving the induction of autophagy (Fig. 10). In this scenario, the brain-targeted rescue of protein O-GlcNAcylation may represent a valuable therapeutic strategy to ameliorate the early development of AD-like pathology in DS subjects.