Alzheimer's disease as an autoimmune disorder of innate immunity endogenously modulated by tryptophan metabolites

Abstract

Introduction: Alzheimer's disease (AD) is characterized by neurotoxic immuno-inflammation concomitant with cytotoxic oligomerization of amyloid beta (Aβ) and tau, culminating in concurrent, interdependent immunopathic and proteopathic pathogeneses.

Methods: We performed a comprehensive series of in silico, in vitro, and in vivo studies explicitly evaluating the atomistic-molecular mechanisms of cytokine-mediated and Aβ-mediated neurotoxicities in AD. Next, 471 new chemical entities were designed and synthesized to probe the pathways identified by these molecular mechanism studies and to provide prototypic starting points in the development of small-molecule therapeutics for AD.

Results: In response to various stimuli (e.g., infection, trauma, ischemia, air pollution, depression), Aβ is released as an early responder immunopeptide triggering an innate immunity cascade in which Aβ exhibits both immunomodulatory and antimicrobial properties (whether bacteria are present, or not), resulting in a misdirected attack upon "self" neurons, arising from analogous electronegative surface topologies between neurons and bacteria, and rendering them similarly susceptible to membrane-penetrating attack by antimicrobial peptides (AMPs) such as Aβ. After this self-attack, the resulting necrotic (but not apoptotic) neuronal breakdown products diffuse to adjacent neurons eliciting further release of Aβ, leading to a chronic self-perpetuating autoimmune cycle. AD thus emerges as a brain-centric autoimmune disorder of innate immunity. Based upon the hypothesis that autoimmune processes are susceptible to endogenous regulatory processes, a subsequent comprehensive screening program of 1137 small molecules normally present in human brain identified tryptophan metabolism as a regulator of brain innate immunity and a source of potential endogenous anti-AD molecules capable of chemical modification into multi-site therapeutic modulators targeting AD's complex immunopathic-proteopathic pathogenesis.

Discussion: Conceptualizing AD as an autoimmune disease, identifying endogenous regulators of this autoimmunity, and designing small molecule drug-like analogues of these endogenous regulators represents a novel therapeutic approach for AD.

Keywords: Alzheimer's disease; amyloid beta; antimicrobial peptide; arginine; autoimmune; cytokine; tryptophan.

FIGURE 1

Combined proteopathic–immunopathic conception of AD. (A) Summary of cellular and protein elements implicated in AD pathogenesis. (B) Summary of synaptotoxic pathologies and signaling in AD. Conceptualizing Aβ as an immunopeptide enables a unifying harmonization of the multiple postulated pathogeneses for AD. Aβ oligomers are directly neurotoxic damaging the neuronal membrane and eliciting excitotoxicity via neurotransmitter receptors such at the NMDA receptor. Concomitantly, Aβ oligomers are directly immunotoxic interacting with microglia and astrocytes (with enhanced microglia‐astrocyte cross‐talk mediated in part by MCP‐1, ICAM‐1), skewing microglial activation towards the proinflammatory M1 phenotype, promoting release of pro‐inflammatory cytokines (IL‐1β, IL‐6, TNFα), and suppressing the release of anti‐inflammatory cytokines (IL‐3, IL‐4, IL‐10, IL‐13). In turn this promotes neurotoxic tau aggregation (with microtubule instability) and mitochondrial damage leading to increased membranotoxcic reactive oxygen species (peroxides, superoxide, hydroxyl radical, singlet oxygen). Finally, this culminates in cellular death by both necrosis and apoptosis with associated symptom expression in the form of disordered cognition and short term memory information processing.

FIGURE 2

In silico calculations of amyloid beta (Aβ) and membrane dynamics. (A) Simulation of a single Aβ peptide inserting into a model neuronal membrane after 450 ps. Insertion occurred only in the presence of cholesterol, irrespective of variations in computational parameters. Aβ favorably inserted adjacent to cholesterol with the C‐terminus (residues 28–42) inserting first. B) Proposed model of Aβ membrane insertion in which HHQK motif facilitates anchoring (aided by glycosaminoglycans [GAGs] and metal ions), with insertion occurring at the C‐terminus via twisting about the hinge region. (C) Simulation of Aβ oligomer (3–5 peptides) inserting into membrane after 450 ps, upon pre‐equilibration outside the membrane for 30 ns. Oligomeric insertion formed a loosely organized membrane‐disrupting aggregate. (D) Schiffer‐Edmundson helical wheel conformation of Aβ. Negatively charged residues (red) align on a single face of the helix, compatible with the proposed HHQK motif anchored membrane insertion model (blue: positive; black: hydrophobic; unmarked: polar, uncharged). (E) Molecular modeling demonstrates notable overlap between GM1 and lipopolysaccharide, and (F) lipoteichoic acid, upon energy minimization calculations, suggesting molecular mimicry among markers of infection and necrosis toward Aβ production

FIGURE 3

Antibacterial and antiviral functionality of amyloid beta (Aβ). (A) Fluorescence microscopy of Vero cells infected with Vesicular Stomatitis Virus tagged with green fluorescent protein (VSV‐GFP) pretreated overnight with varying Aβ_1‐42 concentrations. Control (CTL) samples were diluted and infected immediately, without overnight treatment; cells imaged 24 hours post‐infection and reveal a dose‐dependent antiviral response. (B) Typhoon scans of culture plates, where plaques are visible as dark‐gray spots, show the same. (C) Dose‐response curve of Aβ against herpes simplex virus 1 (HSV‐1), derived from fluorescence experiments, demonstrating clear antiviral functionality. Calculated EC₅₀ = 1 μg/ml or 0.2 μM. (D) Toxicity/antimicrobial peptides (AMP) activity of Aβ_1‐42 against E.coli, in the presence of cholesterol and metal ions (Cu²⁺ and Zn²⁺), as determined by relative 12‐hour optical density (OD_12H); and calculated IC₅₀ of Aβ_1‐42 with/out cholesterol and metal ions from fitted nonlinear regressions (sigmoidal) to data. Cholesterol and Zn²⁺ enhance Aβ_1‐42 toxicity against bacteria. (E) Aβ production in lysate and supernatant of SK‐N‐AS incubated with liopopolysaccharide (LPS), lipoteichoic acid, and bacterial fragments. Aβ production quantified by enzyme‐linked immunosorbent assay; significant increases observed in cell lysates only, exposed to E.coli (*P = 0.03), K. pneumoniae (**P < 0.01), Methicillin‐susceptible S. aureus (MSSA, *** P = 0.02), LPS (100 ng/ml, **** P < 0.01). MRSA denotes Methicillin‐resistant S. aureus. (F) Time‐dependent production of Aβ in lysates of SK‐N‐AS showing upregulation within 1 hour of initial exposure, suggesting innate (as opposed to adaptive) immune response; P < 0.01 (compared to control) in all cases. (G) Relative neurotoxicity of Aβ (1‐40 and 1‐42) and AMPs (Ceropin and LL‐37) against SK‐N‐AS cells, assayed by 3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazolium bromide (MTT). (H) Relative neurotoxicity of Aβ and AMPs against primary cultured rat neurons, assayed by MTT

FIGURE 4

Amyloid beta (Aβ) production in response to innate immune activation. (A) To specify the signaling leading to upregulation of Aβ production, cells were treated with mechanical agitation (scraping) to induce necrosis and ultraviolet light (10 minutes with 254 nm light) to induce apoptosis. Evaluation of caspase‐3/7 activity, by relative fluorescence per minute (ΔRFU/min), verified which treatments led to a specified mode of cell death, with only the apoptosis treatment significantly elevating caspase activity (P < 0.01, in lysate). (B) Aβ production in response to incubation with fluorescently labeled Aβ (10 μM), and GM1 (molecular marker of neural necrosis, 5 μM). Significant elevation in Aβ production followed exposure to Aβ_1‐40. Aβ_1‐42 induces less Aβ production, though addition of GM1 significantly elevates Aβ levels (P > 0.01)—indicating synergy between neural necrosis and self‐perpetuated release of Aβ. (C) Scheme to quantify origin of elevated Aβ signaling, exposing control/necrotic/apoptotic cells’ supernatant and lysates to fresh SK‐N‐AS cells, with subsequent quantification of Aβ production in supernatant and lysates. (D) Aβ production in supernatant of cells exposed to supernatant of noxious cells; necrotic supernatant significantly (P = 0.03) elevated Aβ production. E, Aβ production in lysates of cells exposed to supernatant of noxious cells; no significant difference. (F) Aβ production in supernatants of cells exposed to noxious lysates; no significant difference. (G) Aβ production in lysates of cells exposed to noxious lysates; no significant difference

FIGURE 5

Identification and optimization of endogenous anti‐immunopathic/anti‐proteopathic molecules. (A) In silico screen of endogenous brain molecules (see Table S1) identified numerous metabolites of tryptophan as potential multifunctional agents against Alzheimer's disease (AD). Molecules representing intermediates in tryptophan's major metabolic pathways (shown: tryptamine, 5‐OH tryptamine, 3‐OH anthranilic acid) were selected for in vitro screening. (B) Thioflavin T (ThT) measured aggregation of amyloid beta (Aβ; in which ThT fluorescence increases when bound to β‐sheets of fibrils, rather than α‐helices of monomers) revealed multiple viable inhibitors of Aβ aggregation. (C) Chemical synthesis of tryptophan‐based analogue: NCE217. (D) ThT measured aggregation of Aβ in the presence of varying concentrations of NCE217, demonstrating dose‐dependent inhibition of aggregation from 0.4 μM to 10 μM. (E) ThS measured aggregation of tau observed considerable inhibition of fibrillization with 50μM of NCE217—evidenced by decreased fluorescence of ThS in the treatment group over time. (F) Transmission electron microscopy (TEM) of Aβ_1‐42 in the absence and presence of NCE217 (8μM and 40μM), demonstrating substantial inhibition of Aβ fibril formation, and preponderance of smaller species upon exposure to compound. (G) TEM at 6000X of tau (control and treated with 10μM of NCE217) showing disruption of large tangles, and preponderance of smaller species after 72‐hour exposure. (H) Circular dichroism (CD) spectra of Aβ₄₀ in the absence (left) and presence (right) of NCE217 (100μM), demonstrating preservation of α‐helical conformation past 144 hours, in contrast to uninhibited Aβ. (I) SH‐SY5Y neuroblastoma cell viability (assayed by 3‐[4,5‐dimethylthiazol‐2‐yl]‐2,5‐diphenyltetrazolium bromide [MTT]) upon exposure to Aβ (10 μM). NCE217 significantly increases cell viability (*P = 0.05 at 20μM; ** P = 0.01 at 50 μM) after Aβ exposure, suggesting it may alleviate Aβ‐induced toxicity

FIGURE 6

In vivo efficacy of tryptophan‐based anti‐Alzheimer's disease (AD) therapeutic compound. (A) Long‐term potentiation is strengthened in APP/PS1 transgenic murine hippocampal slices (400 μm) upon exposure to 50 μM of NCE217 (indicated by arrow)—as evidenced by heightened field excitatory postsynaptic potential (fEPSP) slope upon θ‐burst stimulation (P < 0.01). (B) In radial arm water maze testing, APP/PS1 transgenic mice treated with NCE217 (20mg/kg/day) made statistically fewer errors than mice treated with a vehicle (P = 0.04). Performance during retention (R) proved comparable to wild‐type (WT) mice, suggesting possible preservation of memory in APP/PS1 murine models. (C) In Morris water maze testing, APP/PS1 murine models treated with NCE217 (20mg/kg/day) exhibited a lower latency (time to locate a hidden platform) than vehicle‐treated models. Performance was again comparable to WT mice, suggesting possible preservation of memory (P < 0.01). (D) Amyloid beta (Aβ) plaque burden in brains of APP/PS1 mice treated with vehicle (top) and NCE217 (bottom). Plaque burden is visibly diminished with administration of NCE217, across rostral coronal sections (left), coronal sections near bregma (center), and in the cerebral cortex (right). E, Computational plaque quantification (by ImageJ software) observed significant reduction (*P = 0.024) of amyloid plaques in the cortex. F, Immunohistochemistry of Aβ in the brain, assayed by dot blots using the A11 anti‐oligomeric Aβ antibody, revealed a lower preponderance of neurotoxic oligomers in the brain upon NCE217 treatment. G, Quantification (by ImageJ) revealed a statistically significant decrease in the prevalence of oligomers (P = 0.04)