Effects of memantine treatment on hippocampal protein expression in AD are subtle but specific
In total, 9087 proteins were identified from which 7737 were quantified in all samples (Table S2). To first check the global features of the dataset, samples were compared by principal component analysis (PCA,Figure 1B) and hierarchical clustering (Figure S1). Both analyses confirmed dominant role of the disease for clustering of the data (AD/AD+M(memantine) vs controls) with only minor role of memantine treatment, suggesting that memantine intervention does not have profound effects on hippocampal proteome of AD patients (Figure 1A, S1). Interestingly, proteomic profile of one AD patient resembled rather those of healthy controls (labeled as AD 1, Figure S1). Although the patient was diagnosed by AD, reexamination of hippocampal FFPE tissue by pathologist was inconclusive. Nevertheless, the sample was kept in the dataset for further interpretation to follow the original clinical diagnosis. To gain deeper insight into the impact of memantine treatment, proteins quantified in all samples were tested by ANOVA (FDR <0.05) and those found significant were clustered based on their expression profiles in patient groups (Figure 1C). Memantine treatment had either no effect or slightly amplified the expression trends set by AD in the majority of regulated hippocampal proteins (clusters 1 and 3, C1 and C3, Figure 1C), which was expected based on global inspection of data (Figure 1B, Figure S1), further strengthening the finding that memantine does not induce dramatic changes in hippocampus of AD pateints. Notably however, relatively small group of proteins responded to memantine intervention in AD patients by increasing (C2) or decreasing (C4) their expression and thus counteracting the effect of AD pathology on their hippocampal levels (Figrue 1C).
Memantine specifically upregulates mitochondrially-encoded respiratory complex subunits
Clusters defined in Figure 1C were further characterized by Gene Ontology (GO) term enrichment (graphs in Figure 1C). Both mitochondrial and proteasomal proteins were downregulated in course of AD. The drop in expression remained relatively insensitive to memantine medication (C1, Figure 1C) despite several studies reporting memantine-mediated altering of mitochondrial functions (19-23). In the presented data, many proteins involved in oxidative mitochondrial metabolism such as TCA enzymes or components of respiratory complex I (RCI, Figure 2A), RCIII, RCIV, and ATP synthase (Figure S2A) were consistently downregulated in hippocampi of AD patients independently of memantine treatment. The notable exception were products of mtDNA-encoded genes such as subunits of RCI (MT-ND1-6, Figure 2A), MT-CYB, and MT-CO1 (Figure S2A). Nevertheless, such mematine-dependent upregulation of mitochondrial expression does not seem to improve global metabolic decline (Figure 2A, Figures S2A and S2B).
Memantine enhances expression of GABA pathway components
In contrast to mitochondrial proteome, proteins involved in formation of basement membrane (BM) were upregulated in AD patients (C3 in Figure 1C). Indeed, BM proteins like collagens, laminins, nidogens, or thrombospondins showed increased expression independently of memantine treatment (Figure 2B). Beside components of BM, proteins participating in GABA signaling were also enriched in C3 (Figure 1C). Notably, several subunits of GABAA (GABRA2-5 and GABRG1) and GABAB (GABBR1-2) receptors (Figure 2C), together with proteins involved in GABA metabolism (GAD1 and SLC32A1; Figure S3), were associated with the cluster. Interestingly, closer inspection revealed that all quantified GABA receptors were upregulated following memantine treatment when compared to controls and levels of GABA-synthesizing enzyme GAD1, GABA transporter SLC32A1 (Figure S3), and GABRA3-5 receptor subunits (Figure 2C) were significantly increased even when compared to unmedicated AD patients suggesting that memantine administration has impact on GABA signaling.
Memantine suppresses proteomic signature of phagocytic cells
Small fraction of proteins was upregulated in AD but their levels normalized to those in healthy subjects upon administration of memantine (C4, Figure 1C). While GO enrichment indicated these are predominantly associated with lysosomes or antigen presentation via MHC II (Figure 1C), closer inspection of the cluster revealed also proteins involved in immune cell signaling, migration, and apoptosis (Figure 3A) suggesting immune cell origin. In particular, increased presence of lysosomal enyzmes and components of MHC II (HLA) indicated accumulation of phagocytes in AD hippocampus (Figure 3B). Memantine administration in AD patients not only reverted this expression pattern but led to downregulation of several proteins related to phagocyte activation such as MAVS or IRF3 involved in type I IFN signaling (Figure 3B and Figure S4). Collectively, these data suggest that memantine treatment in AD has also a immunomodulating effect exerted on hippocampal phagocytes.
Memantine stimulates expression of proteins involved in glutamate-mediated neurotransmission
Memantine also specificaly enhanced expression of several hippocampal proteins related to glutamatergic synapse, which were otherwise downregulated in untreated AD patients (C2, Figure 1C). Several subunits of iGluRs including one from NMDA (GRIN2A, Figure 3C) and two from AMPA family (GRIA2 and GRIA3, Figure 3D) returned to levels observed in healthy controls after memantine administration. The positive effect of memantine on glutamate receptor expression seems unrelated to the disease since other iGluR subunits, apparently unaffected by AD, followed a similar trend. Indeed, two other members of NMDA receptor family (GRIN1 and GRIN2B, Figure 3C), two members of kainite receptor family (GRIK2 and GRIK5, Figure 3E) and even subunits of mGluRs (GRM1 and GRM2, Figure S5) were upregulated in memantine-treated AD patients when compared to controls. In parallel, enhanced GluR levels were mirrored by memantine-mediated increase in expression of glutamate transporters (SLC1A3, SLC17A6 and SLC17AA, Figure 3F). Such effect is limited only to proteins directly involved in glutamate sensing or metabolism. The levels of neuronal (Figure S6A) and both excitatory and inhibitory post-synaptic density (Figure S6B) markers appear memantine-insensitive suggesting that overall synaptic capacity does not change upon memantine treatment. Rather, these findings indicate that memantine administration in AD specifically stimulates the expression of synaptic components respofnsible for glutamate-mediated neurotransmission.