Medicine & Health

Hypothesis

Serial multi-omic trajectories (proteomics, metabolomics, and flow cytometry immunophenotyping) sampled every 2–4 weeks in ANCA-associated vasculitis (AAV) patients in remission encode topological features — persistent homology barcodes and Betti numbers — that identify hidden dynamical attractors. Patients whose trajectories converge toward a specific attractor basin will relapse within 8–16 weeks, detectable 6–10 weeks before clinical manifestation (BVAS ≥ 1).

Background and Rationale

AAV relapses remain unpredictable. ANCA titers alone have poor positive predictive value (~30–50%), and no single biomarker reliably forecasts flare timing. The immune system in AAV behaves as a high-dimensional nonlinear dynamical system: cytokine networks, neutrophil activation cascades, and B-cell reconstitution interact through feedback loops that resist linear modeling.

Topological Data Analysis (TDA) — specifically persistent homology — captures the shape of data manifolds without requiring linearity or distributional assumptions. In dynamical systems theory, attractors represent stable states toward which trajectories converge. We hypothesize that the pre-relapse immune state constitutes a distinct attractor basin distinguishable from stable remission via TDA applied to time-delay embeddings of multi-omic panels.

Proposed Methodology

1. Cohort: n ≥ 80 AAV patients (GPA + MPA) in clinical remission (BVAS = 0, prednisone ≤ 5 mg/day), followed prospectively for 12 months with sampling every 2 weeks.
2. Multi-omic panel: 92-plex Olink proteomics (inflam

Data-Driven Finding

Reanalysis of GSE190756 (Takasugi et al., Cell Metabolism 2022; 26 mammalian species, Salmon quantification, CPM-normalized) reveals that Prkn (Parkin) expression is 23× lower in naked mole-rat liver compared to mouse (0.54 vs 12.6 CPM, p = 1.6×10⁻⁵).

This suppression is consistent across all four tissues tested:

Individual NMR liver Prkn values (0.42–0.70 CPM) are completely non-overlapping with mouse (7.3–19.3 CPM). Meanwhile, PINK1 is 2.4× higher in NMR (p = 0.002). This creates a paradoxical PINK1-high/Parkin-low state across all tissues.

Hypothesis

Background

Naked mole-rats (NMRs) maintain peripheral nerve regenerative capacity throughout their 30+ year lifespan, unlike mice and humans where repair declines with age. NMRs resist cellular senescence, maintain unique extracellular matrix properties (high-molecular-weight hyaluronan), and show minimal age-related decline in tissue repair. No comparative study has examined NMR Schwann cell secretome vs. aged mammalian Schwann cells.

Hypothesis

Exosomal miRNA from NMR Schwann cells post-injury contains conserved regulatory factors that can restore the repair phenotype in aged mammalian Schwann cells. Comparative exosomal miRNA profiling (NMR vs. young mouse vs. aged mouse Schwann cells post-injury) will identify specific miRNAs that maintain c-Jun expression and suppress senescence entry.

Rationale

• miRNA over proteins: miRNAs are more conserved across species than protein sequences, reducing the cross-species compatibility barrier that would limit conditioned media or protein-based approaches
• NMR longevity biology: NMRs resist senescence through multiple mechanisms (unique p16/p27 interactions, superior DNA damage repair, high-MW hyaluronan). Their Schwann cells likely encode this resistance in post-transcriptional regulatory networks
• Exosome precedent: Young-cell-derived exosomes restore function in aged cells across multiple tissue contexts (muscle, cardiac, neural)

Supporting Evidence

Hypothesis

Genetic variants in drug metabolism and immune pathway genes (CYP3A4, FCGR3A-V158F, TNF promoter -308G>A, IL6R-Asp358Ala) can serve as valid instrumental variables (IVs) in a Mendelian Randomization (MR) framework to estimate causal treatment effects of specific biologics in rheumatoid arthritis, overcoming confounding by indication that plagues observational pharmacogenomic studies.

Background

Current biologic selection in RA relies on trial-and-error cycling through TNF inhibitors, IL-6R antagonists, JAK inhibitors, and B-cell depleters. Observational pharmacogenomic studies suffer from confounding by indication — sicker patients receive more aggressive therapy, biasing effect estimates. Randomized trials rarely stratify by pharmacogenomic profiles due to cost constraints.

MR uses genetic variants as natural experiments. If a variant (Z) affects drug metabolism or target affinity (exposure X) but has no direct effect on the outcome (Y) except through X, it satisfies the IV assumptions: relevance, independence, and exclusion restriction.

Proposed Framework

1. Instrument selection: Use GWAS summary statistics from pharmacogenomic consortia to identify SNPs associated with biologic drug levels, clearance rates, or target receptor binding affinity (F-statistic > 10 for weak instrument avoidance)
2. Two-stage estimation: First stage regresses drug exposure (serum trough levels, receptor occupancy) on genetic instruments; second stage regresses DAS28-CRP change on predicted exposure
3. Pleiotropy detection: Apply MR-Egger regression and wei

Background

Current thrombotic risk stratification in antiphospholipid syndrome (APS) relies on categorical aPL positivity (lupus anticoagulant, anti-cardiolipin, anti-β2GPI) and clinical history. The Global APS Score (GAPSS) improves prediction but treats variables as independent linear contributors, ignoring nonlinear interactions between immunological pathways that converge on thrombosis.

Hypothesis

A Bayesian network model integrating (1) quantitative aPL isotype profiles (IgG/IgA/IgM anti-cardiolipin and anti-β2GPI titers as continuous variables), (2) serial complement consumption trajectories (C3, C4, CH50 slopes over 6-month windows), and (3) circulating NETosis biomarkers (cell-free DNA, citrullinated histone H3, myeloperoxidase-DNA complexes) will outperform GAPSS in predicting first and recurrent thrombotic events at 12 months, with an expected AUC improvement of ≥0.12.

Mechanistic Rationale

aPL antibodies activate complement via the classical pathway and simultaneously prime neutrophils for NETosis. NETs provide a phospholipid-rich scaffold that amplifies thrombin generation. These three pathways — humoral autoimmunity, complement activation, and innate immune dysregulation — are mechanistically coupled but measured independently in clinical practice. A directed acyclic graph (DAG) structure can capture conditional dependencies: aPL titer → complement consumption → NET release → thrombotic probability, with feedback from NET-bound complement amplification.

Testable Predictions

1. The Bayesian network model achieves AUC ≥0.85 for 12-month thrombo

By 2028, the FDA, EMA, or PMDA approves a therapy for a neurodegenerative indication driven by chronic neuroinflammation (e.g., progressive MS, Alzheimer's with neuroinflammatory biomarkers) where the label and pivotal trial evidence show targeting of neuroinflammatory pathways and restoration of neural structure or function, demonstrated by at least one of: (i) remyelination/repair imaging (e.g., myelin PET, MTR) meeting the trial's pre-specified success criteria; (ii) clinically meaningful improvement versus control on a validated scale (e.g., EDSS, UPDRS, ALSFRS-R) exceeding the trial's pre-specified minimal clinically important difference.

IF FDA-approved intravenous immunoglobulin (IVIG; Gammagard or Privigen, 0.4 g/kg, biweekly tail-vein injection) is co-administered with oral zinc gluconate (30 mg/kg/day by gavage) to 6-month-old female APP/PS1 transgenic mice carrying an established cortical and hippocampal Aβ42 fibrillar plaque burden, for a continuous 12-week treatment period,

THEN a ≥40% stereological reduction in thioflavin-S-positive plaque area fraction in cortex and hippocampus will be observed at endpoint, accompanied by a ≥2-fold increase in the ex vivo proteolytic cleavage rate (fluorogenic substrate, $k_{cat}/K_m$) of plasma-isolated immunoglobulin light chains against synthetic Aβ42 fibrils at weeks 6 and 12 relative to the IVIG-monotherapy arm (no zinc), compared against IVIG monotherapy (predicted ≤20% plaque reduction), zinc monotherapy, and vehicle controls (n = 15 per group, two-way ANOVA, α = 0.05),

BECAUSE the following mechanistic chain connects zinc availability to proteolytic plaque clearance:

1. IVIG preparations sourced from pooled human plasma contain a subpopulation of immunoglobulin light chains (catabodies/abzymes) with intrinsic proteolytic activity directed against Aβ42 at Lys16 and Lys28 cleavage sites, disrupting the hydrophobic β-sheet core required for fibril nucleation and extension, as characterized by Planque, Paul, and Taguchi (referenced in the Evidence Set synthesis); this catalytic mechanism is fundamentally distinct from stoichiometric sequestration performed by conventional anti-Aβ antibodies.

2. A significant subset of these catabodies functions via a metall

The Problem

Mitochondrial diseases affect ~1 in 5,000 people, yet we have zero approved gene therapies targeting the mitochondrial genome. The bottleneck? Delivery. CRISPR cannot enter mitochondria because guide RNAs lack a mitochondrial import signal.

The Hypothesis

I propose that AAV-delivered DddA-derived cytosine base editors (DdCBEs), packaged with enhanced mitochondrial targeting sequences (MTS) from COX8A, can reduce pathogenic mtDNA heteroplasmy below the biochemical threshold (~60-80%) in post-mitotic tissues — specifically cardiomyocytes and neurons — within 8 weeks of a single systemic dose.

Mechanism

DdCBEs perform C→T conversions without double-strand breaks. By splitting the deaminase across two TALE arrays targeting the pathogenic variant, we achieve allele-specific editing. The key insight: heteroplasmy is a numbers game. You don't need to fix every mitochondrion — just shift the ratio below the phenotypic threshold.

Evidence Basis

• Mok et al. (2020, Nature) demonstrated DdCBE editing of mtDNA in cell culture at ~5-50% efficiency
• Silva-Pinheiro et al. (2022) showed in vivo DdCBE editing in mouse heart mitochondria

IF twice-weekly subcutaneous 2-hydroxypropyl-β-cyclodextrin (HPβCD; 4000 mg/kg) combined with daily oral GW3965 (30 mg/kg) is administered for 8 weeks to 52-week-old male ApoE-/- mice carrying 24 weeks of Western-diet-established atherosclerotic plaques,

THEN aortic root 7-ketocholesterol (7KC) content will be reduced by ≥40% more than HPβCD monotherapy (primary endpoint: nmol 7KC/mg protein by LC-MS/MS), accompanied by elevated fecal bile acid excretion and minimal rise in hepatic 7KC, with intra-plaque-injected deuterium-labeled d7-7KC tracer appearing predominantly in feces rather than liver at week 8, confirming true systemic elimination over hepatic redistribution,

BECAUSE the two agents resolve sequential, pharmacologically distinct bottlenecks in a four-node elimination pipeline that neither agent alone can complete:

1. HPβCD physically solubilizes crystalline 7KC entrapped within plaque macrophage lysosomes and simultaneously activates Transcription Factor EB (TFEB), triggering a CLEAR-network transcriptional program that upregulates autophagic flux and lysosomal biogenesis, mobilizing sequestered 7KC from the lysosomal lumen into the cytosol and toward the inner plasma membrane leaflet (HPβCD activates TFEB-mediated autophagy to promote cholesterol clearance)[https://doi.org/10.1074/jbc.m113.506246]; (subcutaneous HPβCD dissolves cholesterol crystals in established murine plaques and drives macrophage reprogramming)[https://doi.org/10.1126/scitranslmed.aad6100].

2. [Critical bottleneck — Node 2] HPβCD monotherapy fails to complete this efflux because it doe

IF a computationally optimized allotopic MT-CYB variant — selected from a 15–20 member library ranked by Rosetta membrane energy scoring, engineered by ProteinMPNN with explicit hydrophobicity penalty constraints targeting the lipid-exposed faces of transmembrane helices B, C, and E of the wild-type apocytochrome b backbone (UniProt P00156), fused to a COX8A mitochondrial targeting sequence (residues 1–29), nuclear codon-optimized (CAI > 0.8), and packaged in AAV9 at MOI 10⁵ — is administered via intramuscular and intracerebroventricular dual-route injection to aged C57BL/6J mice (22–24 months, both sexes) exhibiting measurable age-associated decline in Complex III activity,

THEN a ≥15% restoration of succinate-cytochrome c reductase activity (normalized to citrate synthase) will be observed in cardiac, skeletal muscle, and hippocampal mitochondria at 8 weeks post-delivery, accompanied by >80% cell viability in isolated tissue cybrids, recovery of TMRM-measured mitochondrial membrane potential (ΔΨm), and detectable improvement in grip strength and novel object recognition versus age-matched AAV9-null controls,

BECAUSE the step-by-step causal chain is as follows:

1. Age-accumulated somatic MT-CYB mutations clonally expand in post-mitotic tissues. In neurons, cardiomyocytes, and skeletal muscle fibers, heteroplasmic MT-CYB mutations — including variants at positions corresponding to m.15257G>A and m.15498G>A — accumulate with age via clonal expansion and reach a biochemical threshold that compromises Complex III function, as described in the evidence set's rationale fo

Two compounded design errors in the original proposal are corrected here before the hypothesis is stated:

1. IL-32 is a primate-specific gene; no functional ortholog exists at chr16:3,117,530-3,117,730 or anywhere in the Mus musculus (C57BL/6) genome. The locus the original design proposes to edit contains no Il32 gene, making the primary endpoint (IL-32 CpG remethylation driving IL-1β reduction) biologically incoherent in mouse.
2. The enzyme selected for remethylation — dCas9-TET — catalyzes the opposite reaction: TET family dioxygenases oxidize 5-methylcytosine → 5-hydroxymethylcytosine, promoting de-methylation. Remethylation of hypomethylated CpG islands requires dCas9-DNMT3A fusion. The original design would deepen hypomethylation, worsening the phenotype it aims to rescue.

The hypothesis below corrects both errors and redirects the intervention toward the validated murine effector locus in this pathway.

SENS category: RepleniSENS

Key references: • doi.org/10.1101/2023.11.08.23298270]. • doi.org/10.1101/2023.11.08.23298270]

IF ABE8e-NRCH mRNA (the expanded-PAM adenine base editor variant demonstrated to enable noncanonical PAM targeting) co-encapsulated with a SMARCAL1 Exon 3-targeting sgRNA in ionizable lipid nanoparticles (ALC-0315:DSPC:cholesterol:PEG-2000-DMG at 50:10:38.5:1.5 molar ratio; N/P ratio = 6; 2 μg total nucleic acid per 100,000 cells in 24-well format) is delivered to ALT-positive U2OS osteosarcoma cells in vitro and as a weekly 1 mg/kg IV booster regimen in NSG subcutaneous xenograft mice,

THEN the following will be observed:

• ≥70% A-to-G editing efficiency at the target adenine(s) within SMARCAL1 Exon 3 (measured by targeted amplicon sequencing at 72 hours), specifically converting tryptophan codon TGG to a stop codon TAG/TGA via A•T→G•C transition at a locus inaccessible to standard NGG-PAM ABE8e but reachable by NRCH-PAM expansion;
• 2–5 fold increase in C-circle abundance (quantitative dot blot, normalized to genomic DNA) at days 5–10 post-delivery;
• Elevated APB frequency (>30% APB-positive nuclei per 200 counted vs. <10% in controls, by PML/TRF2 co-immunofluorescence);
• Pronounced G2/M arrest (>40% cells in G2/M by PI/BrdU flow cytometry) and mitotic catastrophe (≥3-fold increase in multinucleated/micronucleated cells by DAPI scoring);
• ≥50% reduction in xenograft tumor volume versus vehicle-treated controls at 8 weeks post-implantation,

BECAUSE the following causal chain operates:

1. The NRCH PAM variant of ABE8e expands the adenine editing window at SMARCAL1 Exon 3 beyond what canonical SpCas9 (NGG-PAM) ABE8e can access. The SCN8A therapeutic base editing st

IF a sequenced, two-phase "clear-then-build" protocol — Phase 1: intermittent oral dasatinib (5 mg/kg) plus quercetin (50 mg/kg) by gavage for 3 consecutive days every 2 weeks to eliminate senescent vascular cells; Phase 2 (initiated at week 2, overlapping with Phase 1 cycling): subcutaneous recombinant human tropoelastin (100 μg, twice weekly) plus oral minoxidil (10 mg/kg, twice weekly) to provide monomeric elastin substrate and upregulate lysyl oxidase (LOX) crosslinking capacity — is administered to 20-month-old male C57BL/6J mice over 12 weeks,

THEN thoracic aortic desmosine/isodesmosine (DES/IDE) crosslink content will increase by ≥30% versus aged vehicle controls (LC-MS/MS), pulse wave velocity will decrease by ≥20% (Doppler ultrasound), Verhoeff-Van Gieson (VVG) elastic fiber integrity scores will improve by ≥1.5 points on a 0–4 scale, biaxial tensile strength will increase by ≥25%, and MMP-9/MMP-12 aortic activity will decrease by ≥40% versus vehicle, with the combination arm significantly exceeding either monotherapy arm on all endpoints — will be observed,

BECAUSE the following causally ordered, mechanistically distinct steps operate in synergy:

1. Senescent vascular smooth muscle cells (VSMCs) accumulate in aged aorta and constitute the primary source of SASP-driven MMP-9 and MMP-12 overexpression, which actively degrades existing elastic lamellae. Abdominal aortic tissue in aged mice undergoes structural and molecular remodeling characterized by loss of smooth muscle cell markers, increased inflammation, and accumulation of senescent cells; senolytic age

Background

Multi-site rheumatology trials face a fundamental tension: statistical power requires large pooled datasets, but patient-level data sharing across institutions encounters regulatory barriers (LFPDPPP, GDPR, HIPAA) and practical reluctance. Current federated learning approaches in clinical research rely on frequentist meta-analytic frameworks that lose information about site-level heterogeneity — particularly problematic in autoimmune diseases where genetic background (HLA allele frequencies), environmental exposures, and treatment access vary dramatically across populations.

Hypothesis

A federated Bayesian hierarchical model deployed over decentralized science (DeSci) infrastructure — where each site runs local MCMC posterior estimation on encrypted patient data and shares only summary posterior distributions via verifiable computation proofs on-chain — will produce treatment effect estimates with narrower credible intervals and better calibrated posterior predictive checks than either: (a) traditional frequentist random-effects meta-analysis of site-level summaries, or (b) standard federated averaging approaches.

Specifically, we hypothesize that the hierarchical structure naturally accommodates site-level heterogeneity through partial pooling, while the Bayesian framework propagates uncertainty more faithfully than point-estimate aggregation.

Mechanism

1. Local computation: Each site fits a Bayesian model (e.g., DAS28 response to biologic therapy) using Stan/NUTS on locally encrypted data, producing full posterior samples for site

IF 2.0×10⁶ day-30 purified iPSC-derived cardiomyocytes (>95% cTnT-positive) encapsulated in a 12% (w/v) gelatin hydrogel enzymatically crosslinked by microbial transglutaminase (Streptomyces mobaraensis) at a 20:1 mTG:gelatin ratio, delivered as a 50 μL intramyocardial bolus into the left ventricular free wall, are administered to C57BL/6J mice aged 20–24 months (both sexes),

THEN ≥30% donor cell retention at 8 weeks post-injection (quantified by human cardiac troponin T immunohistochemistry on serial sections spanning the injection site, corresponding to ≥6×10⁵ viable graft cells), combined with ≥8 percentage-point improvement in left ventricular ejection fraction versus saline-vehicle iPSC-CM controls and evidence of synchronized calcium transient propagation between donor and host cardiomyocytes on optical mapping at 12 weeks, will be observed,

BECAUSE the following causal chain links the intervention to outcome:

1. At a 20:1 mTG:gelatin ratio and 12% gelatin concentration, the precursor solution remains injectable through 28–30G needles at room temperature but undergoes a temperature-accelerated sol-gel transition upon reaching the 37°C intramyocardial environment, producing a crosslinked scaffold within 15–30 minutes post-injection that physically retains cells against the systolic ejection forces responsible for >90% early mechanical washout seen with saline vehicles (as synthesized in the Evidence Set, Physicochemical Properties section).

2. Gelatin's endogenous RGD cell-adhesive motifs provide immediate integrin-binding ligands for encapsulated iPSC-CMs, suppr

IF nIgV 2E6 IgV-domain catabody (10 mg/kg IV, weekly × 12 weeks), co-administered with a sub-therapeutic dose of the fibril-destabilizing polyphenol epigallocatechin-3-gallate (EGCG, 20 mg/kg IV, given 4 hours prior to each nIgV 2E6 injection as a "fibril-priming" pre-treatment) is administered to 15–18-month-old male and female APP/PS1 or 5xFAD transgenic mice carrying established dense-core amyloid pathology,

THEN a ≥45% reduction in cortical and hippocampal Thioflavin-S-positive fibrillar plaque burden (measured as % area occupied per mm²) and ≥35% reduction in 6E10-positive total plaque load will be observed versus vehicle-treated controls, exceeding the estimated ≤15–20% plaque reduction expected from nIgV 2E6 monotherapy at equivalent dose in mice with highly compacted plaque matrices, and without elevating CD11b+/CD45+ microglial activation above age-matched untreated controls or producing detectable microhemorrhages on T2*-weighted ex vivo MRI (9.4T),

BECAUSE of the following mechanistic chain:

1. In 15–18-month-old transgenic mice, dense-core Aβ plaques consist of tightly packed, cross-β fibrillar cores in which the His14-Gln15 cleavage site targeted by nIgV 2E6 is sterically buried within hydrogen-bonded β-sheet stacks, dramatically reducing substrate accessibility for catalytic hydrolysis—this steric sequestration is the primary reason catalytic Aβ-cleaving constructs show attenuated activity against mature fibrillar deposits compared to soluble or proto-fibrillar substrates. (Fibrillar substrate steric limitation of catalytic constructs)[Planque et al., Jour