Methionine Sulfoxide Reductase-B3 (MsrB3) Protein Associates with Synaptic Vesicles and its Expression Changes in the Hippocampi of Alzheimer’s Disease Patients

Human hippocampi

Human formalin-fixed paraffin-embedded (FFPE) tissue blocks of hippocampi of 23 subjects were acquired through the Framingham Heart Study Brain Donation Program, Framingham, Massachusetts, the Netherlands Brain Bank, Amsterdam, Netherlands, and Boston Medical Center as described in Table 1. Boston University Medical Center’s Institutional Review Board approved this study and the authors state adherence to these standards. The analyzed subjects were stratified into three groups based on Clinical Dementia Rating (CDR) score [9–11] and Braak and Braak (BB) stage [12]. The CDR was assigned months prior to death based on antemortem assessment and confirmed with a postmortem retrospective CDR based on a family interview with one or more family members [13]. Group 1 consisted of true control individuals defined as cognitively healthy subjects without any presence of NFTs in the CA fields (CDR0, BB0; n = 6, age mean 66.33±5.82 y, 2 female (F)/4 male (M)); Group 2 included subjects either cognitively intact or with minimal cognitive dysfunction (CDR0-0.5) in the limbic BB stages (CDR0-0.5, BBI-III; n = 11, age mean 91.27±7.35 y, 5F/6M); and Group 3 consisted of subjects with definite AD by NINCDS-ADRDA criteria and in the isocortical BB stages (CDR1-2, BBIV-V; n = 6, age mean 89.50±4.64 y, 3F/3M) (Table 1). Group 1 subjects (true controls without any evidence of NFTs in the hippocampus) are inherently younger than Group 2 and 3 subjects, as NFTs increase with age even in cognitively intact subjects. No Lewy body pathology was reported in any of the subjects.

All subjects were de-identified and authors were blinded to subjects’ CDR score and BB stage during data acquisition. We performed tau immunohistochemistry (IHC) on all obtained hippocampi to corroborate reported BB stages in the accompanying neuropathological reports and to perform quantification of NFTs in CA1 field. In this study, we deliberately omitted the analysis of the ultimate BB stage (BBVI) because the associated pervasive but fluctuating neuronal loss in CA1 regions of end-stage AD patients precludes reliable analysis. No neuritic plaques were present in cognitively intact subjects of Group 1 (CDR0, BB0; Table 1) while amyloid-β pathology was reported in all Group 3 subjects—patients diagnosed with AD clinically and neuropathologically (CDR1-2, BBIV-V; Table 1). In between Groups 1 and 3 were subjects with only hippocampal/limbic NFTs (BBI-III) and heterogeneous in terms of both amyloid-β pathology and cognitive status (CDR0-0.5; Table 1). Neuropathology reports for four subjects (two cognitively intact individuals, CDR0, and two with minimal cognitive deficits, CDR 0.5; Group 2, Table 1) reported absence of any amyloid-β pathology. It is possible that some or all four of these individuals (ages 89–97) could have met criteria for Primary Age-Related Tauopathy (PART) [14] and not develop AD.

Rodent hippocampi

Eight-month-old C57BL/6J mice and three-month-old Wistar rats (strain 003, Charles River Laboratories International, Massachusetts) were anesthetized with 10 mg/kg xylazine and 80 mg/kg ketamine hydrochloride given intraperitoneally and subsequently perfused transcardially with 0.1 M phosphate buffer (PB) followed by 4% paraformaldehyde with 0.01% glutaraldehyde in 0.1 M PB, pH 7.4, for use in electron microscopy (EM) and immunohistofluorescence (IF) studies. All animal procedures were approved by the Institutional Animal Care and Use Committee of Boston University and the authors state adherence to these standards.

MsrB3-Immunoreactivity is found in the human hippocampal pyramidal layer, vascular walls, choroid plexus epithelium, and ependymal ventricular lining

We observed MsrB3-immunoreactivity (MsrB3-IR) in multiple tissue compartments (Fig. 1A). Cortical blood vessel walls were MsrB3-immunoreactive (MsrB3-ir) throughout while CA1 and CA3 pyramidal layers exhibited specific patterns of MsrB3 signal (Fig. 1B, C). In CA1, MsrB3 IHC highlighted distinct somatic puncta in pyramidal neurons (Fig. 1B). In contrast, CA3 field was characterized by fine, diffuse MsrB3 signal in pyramidal somata and punctate signal in the surrounding neuropil (Fig. 1C). In addition to blood vessel luminal linings, epithelium in the choroid plexus showed strong cytoplasmic signal (Fig. 1D). Finally, ependymal cells were MsrB3-ir, in particular at the apical surface (Fig. 1E).

OPEN IN VIEWER

Fig.1

MsrB3 protein signal in hippocampal pyramidal layer, vascular walls, choroid epithelium, and ependymal ventricular lining. A) MsrB3-immunoreactivity in CA1 (inset b), CA3 (inset c), choroid plexus (inset d), and the ependymal ventricular lining (inset e). In pyramidal neurons of healthy human (CDR0, BB0) CA1, MsrB3 was visualized as few somatic, sharply demarcated, 1–3 μm puncta (B, arrows); cortical glial cells do not give rise to MsrB3 signal (B, arrowheads). Asterisk indicates MsrB3-immunoreactive (-ir) cortical blood vessel wall. CA3 neuropil displays punctate MsrB3 signal (C, arrowheads) while CA3 pyramidal somata show diffuse signal (C, arrows). D) The choroid plexus epithelium is MsrB3-ir (arrows). The thicker vascular wall of an artery (asterisk) and a thinner wall of a small vein (double asterisk) are MsrB3-ir. E) Ependymal cells are MsrB3-ir at the apical surface (arrows) near an MsrB3-ir blood vessel wall (asterisk). # indicates (direction of) lateral ventricle. Scale bar A = 200 μm, B-E = 20 μm.

MsrB3-immunoreactivity patterns in CA1 and CA3 have similar features in human and rat hippocampus

To determine the subcellular localization of MsrB3 in hippocampal cell populations, we performed studies on freshly fixed transcardially-perfused rat brain— a biological material superior to the post-fixed human tissue obtained a significant period of time after death. In comparative analysis of human and rat CA1 and CA3, we found similar MsrB3 signal patterns, while the overall signal was stronger in rat (Fig. 2). Somatic MsrB3-ir puncta were found in human and rat CA1 pyramidal neurons (Fig. 2A, B), while the punctate neuropil signal in human CA3 (Fig. 2C) was more robust in rat CA3, where prominent MsrB3-IR puncta frequently outlined pyramidal somata and their apical dendrites (Fig. 2D). The similarity of the staining patterns in humans and rats indicated that the rat may be a good model to study MsrB3 expression at the ultrastructural levels.

OPEN IN VIEWER

Fig.2

MsrB3-immunoreactive pattern in CA1 and CA3 is similar in human and rat hippocampus. MsrB3 punctate signal in human CA1 pyramidal neuron somata (A, arrowheads) was also observed in the rat (B, arrowheads). Punctate MsrB3 neuropil signal surrounding the pyramidal neuron apex (C, arrows) and diffuse MsrB3 signal in the pyramidal somata (C, asterisks) in human CA3 are similar in distribution to the signals in rat CA3 (D). Rat CA3 shows particularly prominent MsrB3-immunoreactive puncta along apical dendrites of pyramidal neurons (D, arrows); diffuse somatic signal is also strong (D, asterisks). # indicates direction of lateral ventricle. Scale bar A, C, D = 20 μm, B = 15 μm.

Some synaptic vesicles of rodent hippocampi associate with MsrB3

To identify the cellular organelle that might harbor MsrB3 signal we examined sections from mouse and rat hippocampus using EM, as human postmortem brain tissue is seldom suitable for ultrastructural analysis. Human CA1 and CA3 field–specific patterns of MsrB3 protein expression were not reflected in markers for ER and mitochondria (Supplementary Fig. 3), the sites of MsrB3 localization in primate epithelial cell lines [7]. Considering that rodents had a single MsrB3 isoform (MsrB3A) that was found to associate with ER in epithelial cell lines, we carefully looked for ER-associated MsrB3 signal in mouse and rat hippocampal sections. However, we found MsrB3 signal labeling at multiple SV in CA3 (Fig. 3). Immunogold labeling was suboptimal in CA1 of mouse hippocampus, and we repeated the experiment using Wistar rat hippocampus, which is heavily MsrB3-immunoreactive (see Fig. 2).

OPEN IN VIEWER

Fig.3

Hippocampal MsrB3 signal associates with some synaptic vesicles (SV) in mouse CA3. MsrB3 immuno-gold labeling highlights some SV (A-C, arrowheads) in the presynaptic terminal; asymmetric, excitatory synapses (A and B, arrows) in CA3 of mouse hippocampi. Scale bars = 300 nm.

Consistently, ultrastructural evaluation of MsrB3 immuno-gold signal in CA3 (Fig. 4A) and CA1 (Fig. 4B) in rat hippocampus revealed MsrB3-IR in many SVs. We next sought to identify SV- and neurotransmitter-associated proteins in the hippocampal subregions that might co-localize with MsrB3 (Fig. 4C-I). Consistent with ultrastructural findings in mouse and rat hippocampus in CA3, we found MsrB3 signal co-localized with vesicle-associated membrane protein 2 (VAMP2) (Fig. 4D-F). Consistent with strong MsrB3 punctate signal along apical dendrites in rat CA3 pyramidal neurons (as shown in Fig. 2D), the site of glutamatergic mossy fibers synapses, we observed co-localization of MsrB3 and vesicular glutamate transporter 1, VGLUT1 (Fig. 4G-I). However, the rare MsrB3-ir puncta at CA1 cell bodies did not co-localize with VAMP2 or VGLUT1 (Supplementary Figure 4).

OPEN IN VIEWER

Fig.4

MsrB3 co-localizes with synaptic vesicle proteins VAMP2 and VGLUT1 in CA3. MsrB3 immuno-gold labeled synaptic vesicles in rat CA3 (A) and CA1 (B). In rat CA3 immunohistofluorescence (C) MsrB3 punctate signal outlined pyramidal somata and co-localized with vesicle associated membrane protein 2 (VAMP2) (D-F) and with vesicular glutamate transporter type 1 (VGLUT1) (G-I). Scale bar A, B = 300 nm, C = 800 μm, F, I = 30 μm.

MsrB3 expression pattern in CA1 pyramidal layer changes in AD

Because MsrB3 signal associated with SVs and demonstrated localization consistent with intrinsic connections of pyramidal layers of CA1 and CA3, we examined whether MsrB3-IR patterns in these hippocampal fields might change with onset and progression of AD. As shown earlier in Figs. 1B and 2A, the MsrB3-IR pattern in CA1 pyramidal neurons of a healthy human hippocampus is apparent as distinct but rare somatic puncta (Fig. 5A). Across the three groups of subjects (see Table 1), we observed that as NFTs accumulated in progressive BB stages (I-V) and cognitive status declined (CDR0.5-2), MsrB3-ir puncta appeared smaller and became more numerous (Fig. 5B). Consistently, NFT-containing neurons often contained frequent, smaller MsrB3-ir puncta (Fig. 5C, D).

OPEN IN VIEWER

Fig.5

MsrB3 signal in CA1 changes during AD progression. In CA1 neurons in cognitively intact subjects without hippocampal neurofibrillary tangles (NFTs) (Group 1– CDR0, BB0), MsrB3 signal appears as rare somatic puncta (A) that become more frequent in AD patients (Group 3– CDR1, BBIV) (B). IHC for MsrB3 (C) and tau (D) in adjacent CA1 sections from an AD patient show NFT-containing neurons with frequent, smaller MsrB3-immunoreactive puncta in the soma (arrows), while non-tangled neurons have rare, larger MsrB3-immunoreactive somatic puncta (arrowheads). # indicates direction of lateral ventricle. Scale bar A, B = 20 μm, C, D = 30 μm.

Quantitative analysis confirmed a significant decrease in the percentages of CA1 pyramidal neurons with rare MsrB3-ir somatic puncta (“healthy phenotype”) in subjects with early AD-associated pathology (Group 2– CDR0-0.5, BBI-III) and in AD patients (Group 3– CDR1-2, BBIV-V) compared to controls (Fig. 6A, Supplementary Table 1.1). Simultaneously, compared to control subjects, the percentage of neurons with frequent somatic (≥5) MsrB3-ir puncta (“degenerative phenotype”) was increased significantly in AD patients (Fig. 6B, Supplementary Table 1.2). The MsrB3-IR degenerative phenotype correlated positively with the accumulation of NFTs in CA1 while the healthy MsrB3-IR phenotype did not (Fig. 6C, Supplementary Table 1.3 and 1.4).

OPEN IN VIEWER

Fig.6

MsrB3-immunoreactive phenotypes in CA1 pyramidal neurons during Alzheimer’s disease (AD) progression correlate with the accumulation of AD pathological hallmarks. The percentage of CA1 pyramidal neurons with rare somatic (≤2) MsrB3-immunoreactive (-ir) puncta decreased in AD patients (CDR1-2, BBIV-V) and subjects with early AD pathology (CDR0-0.5, BBI-III) compared to controls (CDR0, BB0) (A), while the percentage of neurons with frequent somatic (≥5) MsrB3-ir puncta increased in AD patients compared to controls (B). The percentage of pyramidal neurons with frequent MsrB3-ir puncta correlated positively with the percentage of tangled neurons in CA1 (C). ^*p < 0.05, ^**p < 0.01, by one-way ANOVA and Pearson correlation.

Considering CA3 subfield-specific MsrB3 neuropil signal and intrinsic connections between CA3 and CA1, we examined the relationship between MsrB3-IR phenotypes in CA1 pyramidal neurons and CA3 MsrB3 neuropil signal intensity. Indeed, the intensity of MsrB3 signal in CA3 neuropil correlated positively with the percentage of CA1 pyramidal neurons exhibiting the MsrB3-IR healthy phenotype, i.e., rare (≤2) somatic puncta (Fig. 7, Supplementary Table 1.5).

OPEN IN VIEWER

Fig.7

MsrB3 signal in CA3 neuropil correlates positively with the percentage of CA1 pyramidal neurons exhibiting “healthy” MsrB3-immunoreactive phenotype (rare ≤2, somatic puncta). A) The intensity of MsrB3 signal in CA3 neuropil positively correlated with the percentage of CA1 pyramidal neurons with rare (≤2) MsrB3-immunoreactive (-ir) puncta. B) MsrB3 signal in CA3 neuropil in a healthy control (CDR0, BB0, arrowheads) compared to a hardly discernible neuropil signal in an AD subject (CDR2, BBV) (C, arrowheads). Scale bars = 20 μm. ns = nonsignificant, ^**p < 0.01, by Pearson correlation.

MsrB3 signal intensity decreases in white matter arteriolar walls of AD patients’ hippocampi

As indicated in Fig. 1, MsrB3 signal was observed in the vessel walls of choroid plexus and hippocampal cortex and white matter. MsrB3 IHC strongly highlighted arteriolar smooth muscle layer (Fig. 8A). Endothelial/pericyte cell layer appeared MsrB3-ir as well (Fig. 8B). We observed variation in the intensity of MsrB3-IR in the hippocampal arteriolar walls in our subjects, often depending on the BB stage and CDR score (Fig. 8C, D). The presence of cerebral amyloid angiopathy (CAA) did not explain the variation in MsrB3 signal within the arteriolar walls in our sample, as CAA was present in the hippocampi of only four of our subjects and often restricted to leptomeningeal vessels (see Table 1, Supplementary Table 2.1). Due to the presence of vascular comorbidity in our subjects, we evaluated MsrB3 signal in the arteriolar vascular walls in the hippocampal cortex and white matter in relationship to various parameters of neurologic vascular health (Supplementary Table 2.1). The level of arteriolosclerosis in the hippocampal white matter (Supplementary Table 2.2) and cortex (Supplementary Table 2.3) did not correlate with MsrB3-IR measures or AD pathological hallmarks. There was no significant difference in most MsrB3-IR measures when subjects were grouped according to the history of hypertension (Supplementary Table 2.4), history of stroke or transient ischemic attack (TIA) (Supplementary Table 2.5), microscopic infarcts present in the hippocampus (Supplementary Table 2.6), or hippocampal vascular white matter disease (Supplementary Table 2.7). Only in subjects with a history of stroke or TIA, the MsrB3-IR signal in arteriolar walls in white matter, but not in cortex, was significantly increased compared to subjects without stroke or TIA (Supplementary Table 2.5). Also, subjects with hippocampal microinfarcts had significantly lower MsrB3-IR in CA3 neuropil compared to subjects without hippocampal microinfarcts (Supplementary Table 2.6). To examine the significance of variation in MsrB3 signal intensity in arteriolar walls further, we tested the vascular MsrB3-IR in correlation with the MsrB3-IR in CA pyramidal neurons and neuropil. MsrB3 signal intensity in the arteriolar walls in hippocampal white matter and cortex correlated positively with the percentage of CA1 neurons with the MsrB3-ir healthy phenotype (i.e., rare (≤2) somatic puncta) (Supplementary Table 2.8). Further, MsrB3 signal intensity in arteriolar walls in hippocampal white matter (but not cortex) was significantly decreased in AD patients compared to controls (Group 1) and subjects with early AD-associated pathology (Group 2) (Fig. 8E, Supplementary Table 2.9 and 2.10).

OPEN IN VIEWER

Fig.8

MsrB3-immunoreactivity in the hippocampal white matter arteriolar walls changes in Alzheimer’s disease. MsrB3 was strongly expressed in smooth muscle of vessel walls (arrows), highlighted in adjacent section with muscle-specific actin (MSA) IHC, in hippocampal cortex (A) and white matter (WM) (B). B) An arteriole in hippocampal WM shows MsrB3 immunoreactivity in smooth muscle (arrow) and endothelial/pericyte (arrowheads) cell layers. Hippocampal WM arteriolar walls were strongly MsrB3-immunoreactive in healthy controls (CDR0, BB0) (C), while weak signal was often found in AD subjects (CDR1, BBIV) (D). E) The intensity of MsrB3 signal in the arteriolar walls of hippocampal white matter was significantly reduced in AD patients compared to controls and subjects with early AD-associated pathology. Scale bars = 20 μm. ^*p < 0.05, ^**p < 0.01, by one-way ANOVA.