Alzheimer’s Disease-associated Region-speciﬁc Decrease of Vesicular Glutamate Transporter Immunoreactivity in the Medial Temporal Lobe and Superior Temporal Gyrus

—Alzheimer’s disease (AD) is a progressive neurodegenerative disorder for which there are very limited treatment options. Dysfunction of the excitatory neurotransmitter system is thought to play a major role in the pathogenesis of this condition. Vesicular glutamate transporters (VGLUTs) are key to controlling the quantal release of glutamate. Thus, expressional changes in disease can have implications for aberrant neuronal activity, raising the possibility of a therapeutic target. There is no information regarding the expression of VGLUTs in the human medial temporal lobe in AD, one of the earliest and most severely aﬀected brain regions. This study aimed to quantify and compare the layer-speciﬁc expression of VGLUT1 and VGLUT2 between control and AD cases in the hippocampus, subiculum, entorhinal cortex, and superior temporal gyrus. Free-ﬂoating ﬂuorescent immunohistochemistry was used to label VGLUT1 and VGLUT2 in the hippocampus, subiculum, entorhinal cortex, and superior temporal gyrus. Sections were imaged using laser-scanning confocal microscopy and transporter den-sitometric analysis was performed. VGLUT1 density was not signiﬁcantly diﬀerent in AD tissue, except lower staining density observed in the dentate gyrus stratum moleculare ( p = 0.0051). VGLUT2 expression was not altered in the hippocampus and entorhinal cortex of AD cases but was signiﬁcantly lower in the subiculum ( p = 0.015) and superior temporal gyrus ( p = 0.0023). This study indicates a regionally speciﬁc vulnerability of VGLUT1 and VGLUT2 expression in the medial temporal lobe and superior temporal gyrus in AD. However, the causes and functional consequences of these disturbances need to be further explored to assess VGLUT1 and VGLUT2 as viable therapeutic targets. (cid:1) 2024 The Author(s). Published by Elsevier Inc. on behalf of IBRO. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).


INTRODUCTION
Alzheimer's disease (AD) is the most ubiquitous phenotype of dementia worldwide, and as the population of earth's average age increases, so does the prevalence and therefore burden of disease (Soria Lopez et al., 2019;Alzheimer's Disease International, 2023).AD is characterized by many clinical indications, primarily aberrant cognitive processing, resulting in behavioural abnormalities and impaired memory retention (DeTure and Dickson, 2019).The disease can be characterized by cortical volume loss, in turn affecting synaptic and neurovascular function (DeTure and Dickson, 2019).Historically, the causative underpinnings for AD have been primarily attributed to protein pathologies, aggregation of amyloid-beta (Aß) and neurofibrillary tangles (NFT) induced by tau hyperphosphorylation (Hardy and Higgins, 1992;Arnsten et al., 2021;Karran and De Strooper, 2022).Both have been robustly associated with AD progression, though neither has shown to be the keystone in resolving or developing a highly effective treatment for this neurological disease.Instead, AD is multifaceted; involving hypoperfusion, neurometabolic dysfunction, and disturbed homeostatic inhibitory-excitatory neuromodulation (Kaur et al., 2021;Hafizi and Rajji, 2023).
Glutamate is the primary excitatory neurotransmitter in the brain, and acts on ionotropic and metabotropic receptor families, propagating excitatory synaptic signaling (Danbolt, 2001;Zhou and Danbolt, 2014).Glutamate levels are finely controlled in the extracellular spaces across the central nervous system (CNS), a process managed by various neuronal and glial proteins (Danbolt, 2001).Intracellular control of glutamate synaptic vesicle filling for release is mediated by the vesicular glutamate transporter (VGLUT) family (Liguz-Lecznar and Skangiel-Kramska, 2007).There are three subtypes of VGLUT, displaying close homology and function at the synapse (Reimer and Voglmaier, 2014).VGLUT1 and VGLUT2 are the most common, mainly expressed in the cortex and cerebellum respectively (Ozkan and Ueda, 1998;Herzog et al., 2001;Liguz-Lecznar and Skangiel-Kramska, 2007;Reimer and Voglmaier, 2014).Although, both VGLUT1 and VGLUT2 have been found to be expressed in the human hippocampus (Vigneault et al., 2015).
VGLUT expression across the brain is influenced by the intracellular milieu, neuronal subtypes and the developmental age of the organism.VGLUT expression tends to be complementary, with neuronal populations either VGLUT1 or VGLUT2 positive, representing unique classes of excitatory synapses (Fremeau et al., 2001;Liguz-Lecznar and Skangiel-Kramska, 2007).The subtype is theorized to affect synaptic vesicle release probability, with VGLUT1 expression indicating low release probability and long-term potentiation, while VGLUT2 expression implies the inverse -long-term depression and high glutamate release probability (Liguz-Lecznar and Skangiel-Kramska, 2007;Weston et al., 2011).The latter occurs primarily in developmental phases as synaptic maturation occurs (Danbolt, 2001;Liguz-Lecznar and Skangiel-Kramska, 2007).Thus, there are likely to be subtle variations in physiological functions of the VGLUT isoforms, which are yet to be fully elucidated.
In AD, immunolocalization of Aß with VGLUT1 and VGLUT2 has been shown in the human parietal cortex, preferentially the co-expression with VGLUT1 resulting in intracellular Aß-associated glutamate transport dysfunction (Fremeau et al., 2004;Sokolow et al., 2012;Martineau et al., 2017).This raises intracellular glutamate concentration, inducing further synaptic dysfunction and also the release of Aß from glutamatergic synapses (Sokolow et al., 2012).A loss of VGLUT1 or VGLUT2 expression has been documented in the frontal, parietal and occipital cortices in AD, which appears to correlate with the cognitive presentation of patients (Kirvell et al., 2006;Kashani et al., 2008;Mi et al., 2023).However, the regional hippocampal-specific alterations have not yet been explored.The hippocampus and the surrounding anatomical areas, including the entorhinal cortex, subiculum, and superior temporal gyrus (STG), are typically severely affected in AD and are some of the first brain regions to show pathology and neuronal loss (Braak and Braak, 1995;Soria Lopez et al., 2019;Leng et al., 2021).This warrants further investigation to better understand VGLUT remodeling in AD, and potentially propose new therapeutic strategies.

EXPERIMENTAL PROCEDURES
Tissue preparation and neuropathological analysis Donated post-mortem brain tissue was obtained from the Neurological Foundation Human Brain Bank at the University of Auckland.Experimental procedures were approved by the University of Auckland Human Participant's Ethics Committee (Approval number: 011654).Processing of the donated brain tissue was conducted as previously described (Waldvogel et al., 2006).In short, the right hemispheres of donated brains were perfused with 15% formalin solution and dissected into anatomical blocks.These blocks were further processed in sucrose for cryoprotection and then stored at 80 C until experimental analysis.A neuropathological examination was also conducted in the middle frontal gyrus, middle temporal gyrus, cingulate gyrus, hippocampus, caudate nucleus, substantia nigra, locus coeruleus, and cerebellum, staining for A and tau immunohistochemically.This information was used to determine The Consortium to Establish a Registry for Alzheimer's Disease (CERAD) classification and Braak and Braak stages for AD cases.Notes from the pathological reports for cases are reported in Tables 1 and 2. Brain tissue from 7 control (Table 1) and 7 AD cases (Table 2) was used in this study with a mean age of 78.6 years and a maximum PM delay of 33 h.All the AD cases used in this study had clinical dementia, and control cases had no history of any neurodegenerative or neuropsychiatric disorders.

Imaging and analysis
Imaging of immunolabeled tissue sections was performed using a Zeiss-710 confocal microscope (Carl Zeiss, Jena, Germany).Anatomical brain regions were differentiated based on relative location, cell morphologies, and Neu-N and Hoechst staining.An argon laser (488 nm wavelength) was used to activate Neu-N-stained neurons, and a helium laser (647 nm wavelength) was used to excite either VGLUT1 or VGLUT2 stained transporters depending on the experiment.A blue diode laser (405 nm wavelength) was used to excite the Hoechst-stained nuclei.The gain threshold for the 647 channels was set at a constant value before images were taken for quantitative purposes.Two hippocampal and STG sections per case were stained and imaged.Images were taken from the dentate gyrus (DG), cornu ammonis 1 (CA1), CA2, CA3, subiculum, entorhinal cortex, and STG using a 20 objective lens (Zeiss 20X (0.8NA) Plan-Apochromat).Tile scanning functionality on the Zeiss-710 microscope was used to image sections through the cell layers of each region, where one tile represented a tissue area of 182,000 m 2 .The number of tiles imaged for each anatomical area was as follows: DG (1), CA3 (4), CA2 (3), CA1 (4), subiculum (6), entorhinal cortex (10),and STG (11).
Using ImageJ software (U. S. National Institutes of Health, Bethesda, Maryland, USA), transporter integrated density measurements were conducted as previously described (Yeung et al., 2021).Integrated density (referred to as density hereafter) is the sum of the number of pixels in the region of interest, a set area for each region.Background subtraction and grayscale threshold determination were performed before the transporter densities were obtained.Density measurements were performed from a 31,000 m 2 area in each analyzed layer in the DG (str.granulosum, str.moleculare and hilus), CA1, CA2, and CA3 (str.oriens, str.pyramidale, str.radiatum).For CA1, CA2, CA3, and the DG, density measurements were taken from the middle of the str.pyramidale and str.granulosum.For the str.oriens, str.radiatum, str.moleculare and hilus, measurements were recorded directly adjacent to the pyramidal or granular layers.This layer distinction was determined based on Neu-N and Hoechst staining.Areas of 432,000 m 2 , 605,000 m 2 , and 692,000 m 2 were used for density measurements in the subiculum, entorhinal cortex, and STG, respectively, through all cortical layers.Integrated density measurements from the two sections from each case

RESULTS
Expression of VGLUT1 in the hippocampus, subiculum, entorhinal cortex, and superior temporal gyrus VGLUT1 immunoreactivity appeared strong throughout most of the regions analyzed in the medial temporal lobe (MTL), outlining and surrounding neuronal cell bodies (Figs. 1 and 2).VGLUT1 is expressed in presynaptic terminals, (Wilson et al., 2005;Du et al., 2020) so this immunoreactivity reflects strong labeling in synaptic boutons across all regions.However, the one exception was that the DG str.granulosum showing minimal VGLUT1 immunoreactivity which was restricted to thread-like processes piercing the granular cell layer (Fig. 1(D)).These staining patterns are in agreement with previous research conducted in the human hippocampus (Vigneault et al., 2015;2021).Quantification revealed that VGLUT1 density was not significantly altered in the CA regions of AD cases relative to control cases, and the staining patterns generally showed high similarity to control tissue (Fig. 3(A-C)).However, two AD cases in the CA1 str.pyramidale displayed minimal VGLUT1 immunoreactivity (Fig. 3(A)).AD cases showed significantly lower density (p = 0.0051) of VGLUT1 density in the str.moleculare of the DG and no changes in the hilus region and the str.granulosum (Fig. 3(D)).VGLUT1 integrated density measurements in AD cases showed high variability in the subiculum, STG, and entorhinal cortex, overall showing a slight trend towards lower density but not reaching significance (Fig. 3(E-G)).VGLUT1 density showed a significant correlation with case age, only in the control DG str.granulosum (positive) and control subiculum (negative) (both p = 0.0167).In addition, significant negative correlations were found between VGLUT1 density and post-mortem delay in the CA1 str.oriens (p = 0.0040), CA1 str.pyramidale (p = 0.0302) and the subiculum (p = 0.0333) of AD cases.No other correlations reached significance.Expression of VGLUT2 in the hippocampus, subiculum, entorhinal cortex, and superior temporal gyrus VGLUT2 immunolabeling was visible in both the control and AD cases in the neuropil and neuronal cell where immunoreactivity appeared higher on the membrane.These patterns were visible across the hippocampus, subiculum, entorhinal cortex and STG (Figs. 4  and 5).Quantification of VGLUT2 density revealed no significant differences between control and AD cases in all layers of the CA regions and DG (Fig. 6(A-D)).We detected significantly lower (p = 0.015) VGLUT2 density in the subiculum of AD cases relative to control, as well as in the STG (p = 0.0023) (Fig. 6  (E,G)).A trend toward lower mean integrated density was observed in AD cases in the entorhinal cortex, but this did not reach significance (Fig. 6(F)).We also observed vascular VGLUT2 labeling, particularly in the subiculum, entorhinal cortex, and STG, that appeared weaker in AD cases (Fig. 5(A-C)).VGLUT2 neuronal soma staining was also visible in the CA3 region of control cases in a recent study (Woelfle and Boeckers, 2021).There was a significant negative correlation between VGLUT2 density and post-mortem delay in str.oriens of the CA1 region in the control hippocampus (p = 0.0167).Additionally, in AD cases, a statistically significant negative correlation was observed between VGLUT2 density and A load in the str.radiatum of the CA1 region (p = 0.0333).No other correlations reached significance.

DISCUSSION
This study investigated the expression of VGLUT1 and VGLUT2 in the hippocampus, subiculum, entorhinal cortex, and STG in AD.We observed strong VGLUT1 immunolabeling throughout most of the MTL, outlining, but not staining the cytoplasm of neurons.VGLUT2 staining intensity was weaker, and neuronal cell bodies were labeled.These differences indicate a likely divergent function of the two transporter isoforms in the MTL which is not yet fully understood.We report significantly lower VGLUT1 and VGLUT2 protein expression in AD which was region-specific; in the DG str.moleculare for VGLUT1, and in the subiculum and STG for VGLUT2.These alterations may impair neuronal communication in AD.

Expressional alterations to VGLUT1 in Alzheimer's disease
VGLUT1 was mostly preserved in AD post-mortem tissue compared to age-matched control cases, with no significant changes detected in the CA subregions of the hippocampus, the subiculum or entorhinal cortex.However, significantly lower density was observed in the DG str.moleculare in AD.Our image analysis workflow measured VGLUT1 density in the str.moleculare directly adjacent to the str.granulosum, i.e., the inner molecular layer (IML).Mossy cells, which are present in the hilus of the DG, are innervated by and synapse back to granule cells, with these projections terminating in the IML of the DG in an excitatory feedback circuit (Seress, 2007).Therefore, lower VGLUT1 density in the IML may represent a loss in these mossy cell and granule cell connections.Mossy cells are known to be especially vulnerable to excitotoxic cell death (Buckmaster and Schwartzkroin, 1994;Scharfman and Myers, 2013), and their degeneration or death in AD could be causative of the lower VGLUT1 staining observed.However, the functional consequences of reduced density on excitatory feedback to granule cells are not yet known.Interestingly, AD cases showed considerable variation in VGLUT1 staining across the DG hilus, while little variation was seen in control cases.This may lead to differential activation of mossy and other granule cell feedback-providing cells in the hilus in AD, such as the inhibitory hilar perforant path-associated cell.This heterogeneity could indicate complex neurodegenerative processes at different stages of AD progression, with implications on overall network function.
A study on presynaptic vesicle fusion protein components from post-mortem tissue in the hippocampus, revealed lower expression in AD, restricted to the DG outer molecular layer (Haytural et al., 2021).Interestingly, these components were spared in the IML, but our results indicate that VGLUT1 density was significantly lower in this sublayer.It could indicate that vesicle fusion protein components and VGLUTs are differentially regulated in AD, but ultimately these distinct changes may both lead to decreased glutamate release.The decrease in vesicle fusion machinery staining in the AD DG may represent a homeostatic response to dampen excitatory transmission and spare the hippocampus proper from excitotoxic damage, given that neither the presynaptic vesicle fusion proteins (Haytural et al., 2021) nor VGLUT1 appears to be significantly affected in the CA regions.
Previous studies examining VGLUT1 expression in the human brain found significantly lower expression in frontal, parietal and occipital, but not temporal cortices (Kirvell et al., 2006;Kashani et al., 2008;Mi et al., 2023).Interestingly, although reporting lower VGLUT1 density in the AD frontal cortex, Mi et al.'s results show higher VGLUT1 intensity, suggesting potentially higher levels in remaining synaptic terminals (Mi et al., 2023).However, on the whole, our results support this earlier evidence, suggesting the temporal lobe is less susceptible to VGLUT1 alterations in AD (Kirvell et al., 2006).The reasons for this apparent preservation remain uncertain, but may be related to temporal areas accumulating A plaques at a later stage than the frontal and parietal cortices (Thal et al., 2002).However, we saw no correlation between A load and VGLUT1 density, suggesting other factors and pathological mechanisms might also be involved.Interestingly though, we indicate that a specific neuronal population that innervates the DG str.moleculare (most likely mossy cells), is vulnerable to AD-associated changes in VGLUT1 expression, demonstrating its expressional modulation in AD is regionally specific.This may have significant functional implications for normal neuronal networking in the hippocampus, which warrants further investigation.
Expressional alterations to VGLUT2 in Alzheimer's disease VGLUT2 expression was preserved in the hippocampus of AD cases, showing no alteration in expression compared to control cases.However, we found significantly lower density in the subiculum and STG.Recent evidence has found that VGLUT2 is a marker for subicular output neurons in the hippocampus, specifically the burstfiring neurons (Wozny et al., 2018).The subiculum contains two main types of pyramidal neurons; burstfiring, and regular-firing, which describe their differential inherent firing patterns (Taube, 1993;Staff et al., 2000).Burst neurons typically fire multiple action potentials when they become depolarised, while regular spiking neurons discharge a single charge when stimulated (Staff et al., 2000;Matsumoto et al., 2019).Thus, the lower VGLUT2 density in the AD subiculum may reflect lower expression in burst-firing neurons.
The two classes of subicular neurons also have subtle variations in anatomical location, affecting where they primarily project.Regular firing neurons are localized mainly to the proximal subiculum and primarily (approximately 80 percent) project to the amygdala and lateral entorhinal cortex.Burstfiring neurons are reported to be more highly expressed in the distal subiculum and mainly project to the medial entorhinal cortex, retrosplenial cortex, and hypothalamus (Kim and Spruston, 2012).The differential locations and projection areas of regular and burst-firing neurons suggest that these cells may convey separate information to these anatomical areas.However, the role of these two neuron classes in normal physiology is still uncertain and confirmation is needed to establish the localization of VGLUT1 and VGLUT2 to distinct neuronal populations (Wozny et al., 2018).Based on the data from the literature, the lower VGLUT2 density seen in the subiculum in AD may reflect a decreased glutamate output of burst-firing neurons.The significantly lower density seen in the STG is also interesting, potentially pointing to disturbances to the secondary auditory cortical processing and could also reflect lower VGLUT2 in burst-firing neurons in AD.
A recent ex vivo study provided further evidence that VGLUTs and vesicle fusion proteins may be differentially regulated in AD, backing up the previously mentioned study by Haytural and colleagues.The vesicle fusion protein synaptophysin was lower in slices from the rat entorhinal cortex acutely exposed to A 1-42 , while interestingly, VGLUT2 protein expression was higher when compared to controls (Olajide and Chapman, 2021).These two alterations seem contradictory, as the increase in VGLUT2 is likely to facilitate higher vesicular concentrations of glutamate and thus a higher quantal release.Comparatively, the reduction in synaptophysin would be expected to reduce vesicle fusion, so the net effects on overall glutamate release are not clear (Olajide and Chapman, 2021).More studies are needed to fully understand this apparent differential regulation of both VGLUT1 and VGLUT2 in AD.
The high expression around neuronal bodies and blood vessels relative to VGLUT1 could indicate higher VGLUT2 expression in presynaptic terminals that contact these neurons or potentially astrocytes that are associated with blood vessels.The notion of VGLUT2 expression in terminals that connect to cell bodies appears consistent with our results reported in mice, where VGLUT2 but not VGLUT1 is expressed in the str.pyramidale and str.granulosum, outlining cell bodies (Yeung et al., 2020a(Yeung et al., , 2020b)).Another possibility is that this apparent pyramidal neuron staining may reflect actual cell body or cytoplasmic labeling.Previous evidence in mice has shown that VGLUT2 was evident in neuronal cell bodies, but only in early development (Real et al., 2006).This expression was suggested to occur due to a higher rate of protein translation than the rate of axonal transport during early development, increasing the likelihood that the protein is detected where it is translated in the cytoplasm.Thus, our results could suggest that pyramidal neurons in the human hippocampus produce VGLUT2 at a higher rate than its axonal transport.It is uncertain which of the two scenarios regarding the exact location of VGLUT2 staining is correct but both suggest a critical role for VGLUT2 in the human brain, which is likely functionally distinct from VGLUT1.The implications for lower density VGLUT2 in AD may thus be different to VGLUT1, a theory that requires further examination.
Previous research on VGLUT2 expression in the human brain is limited and contradictory.One study found reduced VGLUT2 in the prefrontal cortex in AD via Western blot analysis (Kashani et al., 2008).However, another study found no change to VGLUT2 protein levels in AD, despite using similar methodology and investigat-ing the same brain region (Poirel et al., 2018).In addition, the frontal cortex showed no difference to VGLUT2 density or intensity between control and AD subjects (Mi et al., 2023).Our results show a regionally specific susceptibility of VGLUT2 expression in AD, with spared expression in the hippocampal formation but a lower VGLUT2 density in the AD subiculum and STG.Importantly, a better understanding of the exact physiological roles and divergent functions of VGLUT1 and VGLUT2 will be critical to determine what these changes mean in the context of pathology and symptoms of AD.

VGLUT modulation by amyloid-beta and tau pathology
There are several studies which indicate A and tau may be involved in VGLUT1 and VGLUT2 expressional modulations in AD.In hippocampal cultures, a decrease in VGLUT1 expression was observed after incubation with A (Rodriguez-Perdigon et al., 2016).However, we previously showed that VGLUT1 and VGLUT2 hippocampal expression is generally well preserved after acute A 1- 42 exposure via injection in mice, although we did detect lower density of both isoforms in the DG, suggesting these alterations are highly localized (Yeung et al., 2020a(Yeung et al., , 2020b)).In line with this, in the human MTL a negative correlation between VGLUT2 density and A load was only found in the str.radiatum of the CA1 region, possibly due to false positive error, while no correlations were detected for VGLUT1.Another study also reported no alteration to VGLUT1 expression in cultured rat neuronal/glial primary cells after A exposure (Buntup et al., 2008).For tau protein, its translocation to the nucleus has been shown to modulate VGLUT1 in cultured mouse hippocampal neurons, increasing its expression (Siano et al., 2019;Siano et al., 2020).Siano and colleagues point out that alteration to tau in the AD prefrontal cortex appears to correlate to VGLUT1 expressional changes.VGLUT1 expression is increased early in the disease corresponding to tau detachment from microtubules, while later timepoints of AD are denoted by lower expression with increasing aggregation of tau (Siano et al., 2020).This is in keeping with observed changes to glutamatergic transmission in AD with an increase in glutamatergic components and hyperexcitability seen at the early stages of AD, and subsequent hypo-excitation with disease progression (Bell et al., 2007;Targa Dias Anastacio et al., 2022).
Overall, our study reports significantly lower VGLUT1 and VGLUT2 protein expression in the MTL of AD cases, confined to specific regions and cell layers.Understanding the disease mechanisms behind reduced density in AD, functional consequences, and the timeline in relation to AD pathology and symptomatology, are the next questions to investigate.Additionally, understanding the differences in isoform functionality in normal physiology will also give more clues about their expressional changes in AD.Overall, this future research could allow the exploration of therapeutic strategies aiming to modify VGLUT expression and function, potentially changing the disrupted excitatory-inhibitory balance in AD and altering the course of the disease.

SUMMARY FOR SOCIAL MEDIA, IF PUBLISHED
The expression of vesicular glutamate transporters (VGLUTs) is critical to controlling the release of glutamate, the brain's primary excitatory neurotransmitter.Expressional alterations previously reported in Alzheimer's disease (AD) tissue and animal models, likely alter neural excitation and networking, and attempting to modulate such changes may have potential as a therapeutic target.However, there are no studies that have examined the expression of VGLUTs in the post-mortem human hippocampal formation in AD, one of the most early and severely affected brain regions in this devastating disease.Thus, we aimed to investigate and compare the expression of VGLUT1 and VGLUT2 in the control and AD hippocampus, subiculum, entorhinal cortex, and superior temporal gyrus.Overall, our results indicate that VGLUT1 and VGLUT2 are generally preserved in AD across medial temporal lobe regions.However, we report significantly lower immunoreactivity in several highly specific cell layers and regions, with lower density of VGLUT1 in the dentate gyrus stratum moleculare and VGLUT2 in the subiculum and superior temporal gyrus.Future studies investigating VGLUTs or the glutamatergic system as therapeutic targets in AD should take these findings into account.

POTENTIAL CONFLICT OF INTERESTS
Nothing to report.

Fig. 1 .
Fig. 1.VGLUT1 expression in control and Alzheimer's disease cases in the hippocampus.Panels show representative images displaying VGLUT1 labeling from control and Alzheimer's disease cases for CA1 (A), CA2 (B), CA3 (C) and the dentate gyrus (DG) (D).VGLUT1 labeling is shown in red (a,c), and VGLUT1 overlayed with Neu-N in green (b,d).Scale bars = A-C = 100 m; D = 50 m.

Fig. 2 .
Fig. 2. VGLUT1 expression in control and Alzheimer's disease cases in the subiculum, entorhinal cortex and superior temporal gyrus.Panels show representative images displaying VGLUT1 labeling from control and Alzheimer's disease cases from the subiculum (A), entorhinal cortex (B), and superior temporal gyrus (STG) (C).VGLUT1 labeling is shown in red (a,c).VGLUT1 overlayed with Neu-N labeling in green is displayed in (b,d).Scale bars = A-C = 100.

Fig. 3 .
Fig. 3. Quantification of VGLUT1 integrated density in control and Alzheimer's disease cases in the hippocampus, subiculum, entorhinal cortex, and superior temporal gyrus.(A-G) show the quantification of VGLUT1 integrated density in regions and layers of the medial temporal lobe.Data shown represent CA1 (A), CA2 (B), CA3 (C), dentate gyrus (DG) (D), subiculum (E), entorhinal cortex (F) and superior temporal gyrus (STG) (G).Control cases are shown as closed circles and Alzheimer's disease cases as open circles.Data are expressed as the mean ± SEM. ** = p < 0.01 (Unpaired Mann-Whitney test).

Fig. 4 .
Fig. 4. VGLUT2 expression in control and Alzheimer's disease cases in the hippocampus.Panels show representative images displaying VGLUT2 labeling from control and Alzheimer's disease cases for CA1 (A), CA2 (B), CA3 (C) and the dentate gyrus (DG) (D).VGLUT2 labeling is shown in red (a,c) and VGLUT1 overlayed with Neu-N in green (b,d).Scale bars = A-C = 100 m; D = 50 m.

Fig. 5 .
Fig. 5. VGLUT2 expression in control and Alzheimer's disease cases in the subiculum, entorhinal cortex, and superior temporal gyrus.Panels show representative images displaying VGLUT2 labeling from control and Alzheimer's disease cases from the subiculum (A), entorhinal cortex (B), and superior temporal gyrus (STG) (C).VGLUT2 staining is shown in red (a,c).VGLUT2 overlayed with Neu-N labeling in green is displayed in (b,d).Scale bars = A-C = 100 m.

Table 1 .
Details of control cases used for immunohistochemical analysis *case not used for VGluT1 quantification; NA -Not available.

Table 2 .
Details of Alzheimer's disease cases used for immunohistochemical analysisWood et al. / Neuroscience 546 (2024)[75][76][77][78][79][80][81][82][83][84][85][86][87]were combined to obtain an average value for each anatomical region in each case.Tissue incubated without the primary antibody was included in each quantification experiment, which did not result in any staining apart from minor labeling attributed to lipofuscin or cellular debris.onefrom the AD entorhinal cortex.For the VGLUT2 experimental run (6 control and 6 AD cases), two AD cases were lost, again due to poor tissue quality in all layers of CA1 and CA2, but no data points were excluded as outliers from the ROUT test.The Spearman rank correlation test was used to assess correlations between the density of VGLUT1 and VGLUT2 in each of the analysed regions and layers in the hippocampal formation, and density levels in the STG, respectively with case age, post-mortem delay, Braak and Braak stages, as well as tau and A pathology scores reported in Table 2.All experimental data are expressed as the mean ± Standard Error of Mean (SEM).All statistical analyses were conducted using GraphPad Prism software version 9 (GraphPad software; RRID:SCR_002798) with a value of p 0.05 considered significant.Adobe Photoshop CC 2018 (Adobe Systems Software, San Jose, CA, USA) was utilized to make the figures.
# case not used for VGLUT2 quantification; NA -Not available.O. W. G. for VGLUT1 density measurements.A few data points were also lost to poor tissue quality in the VGLUT1 experimental run (6 control and 7 AD cases).This included one control case for CA1 str.oriens, CA1 str.pyramidale, all DG layers, subiculum, and entorhinal cortex.Two control cases were lost from the CA1 str.radiatum and all layers of CA2 and CA3, and two cases were also lost to poor tissue quality from the AD subiculum and