Skip to content

atopic-dermatitis

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

atopic-dermatitis

README.md

Atopic Dermatitis (AD) — Quantitative Systems Pharmacology Model

AD QSP Map

Click the image above to open the full-resolution SVG mechanistic map.

Disease Overview

Atopic dermatitis (AD) is the most prevalent chronic inflammatory skin disease, affecting ~15–20% of children and ~5–10% of adults worldwide. It is characterized by:

  • Epidermal barrier dysfunction (FLG mutation → TEWL↑, pH↑, allergen penetration)
  • Th2-skewed immune activation (IL-4, IL-13, IL-31, IgE, TARC/CCL17)
  • Chronic pruritus (itch-scratch cycle via IL-31Ra, TRPV1, DRG neurons)
  • Comorbidities (atopic march: asthma, allergic rhinitis, food allergy; anxiety/depression)

The pathology involves innate alarmins (TSLP, IL-33, IL-25) activating ILC2 and dendritic cells, which drive Th2 differentiation via GATA-3/STAT6 signaling. IL-4/IL-13 co-signal through the shared IL-4Rα chain — the therapeutic target of dupilumab.

Model Files

File Description
ad_qsp_model.dot Graphviz mechanistic map source (130+ nodes, 13 clusters)
ad_qsp_model.svg Rendered vector diagram (interactive)
ad_qsp_model.png Rendered raster image (150 dpi)
ad_mrgsolve_model.R mrgsolve ODE model (R) — PK/PD/disease dynamics
ad_shiny_app.R Shiny interactive dashboard (6 tabs)
ad_references.md 52 PubMed references, categorized

Mechanistic Map Structure (13 Clusters)

Cluster Contents Nodes
1. Environmental & Genetic Triggers Allergens, microbiome, FLG/SPINK5 mutations, UV 8
2. Epidermal Barrier FLG, ceramides, CLDN1, TEWL, pH, KLK5 11
3. Innate Immunity & Alarmins TSLP, IL-25, IL-33, ILC2, mast cells, DC 14
4. Th2 Adaptive Immunity IL-4, IL-13, IL-31, STAT6, JAK1/2, IgE, TARC 20
5. Eosinophil / Mast Cell Effectors IL-5, ECP, MBP, PGD2, leukotrienes 10
6. Th1 / Th17 / Th22 Counter-Regulation IFN-γ, IL-17A, IL-22, Treg, TGF-β 10
7. Pruritus Neural Pathway IL-31Ra, TRPV1, TRPA1, DRG neuron, NRS 13
8. Keratinocyte Inflammatory Responses TSLP-KC, CCL27, CXCL10, IL-36 10
9. Biologic PK Dupilumab, tralokinumab, lebrikizumab, nemolizumab 10
10. JAK Inhibitor PK Upadacitinib, abrocitinib, baricitinib, deucravacitinib 10
11. Drug PD Effects RO, pSTAT6 inhibition, TARC/eosinophil reduction 11
12. Clinical Endpoints EASI, IGA, SCORAD, NRS, DLQI, POEM 14
13. Comorbidities Atopic march, anxiety/depression, CardioRisk 10

Total: 130+ nodes, 200+ directed edges

mrgsolve ODE Model (20 Compartments)

Drug PK Compartments

Drug Model Key Parameters
Dupilumab 2-cmt SC mAb CL=0.21 L/day, V1=3.5 L, t½≈21d, F_SC=64%
Upadacitinib 1-cmt oral CL/F=38 L/day, Vd=166 L, t½≈8h
Nemolizumab 1-cmt SC mAb CL=0.15 L/day, V1=3.8 L, t½≈22d

Disease Dynamics Compartments

State Variable Description Calibration Source
Rfree / RC Free/bound IL-4Rα receptor (nM) Kovalenko et al. 2016 (PK TMDD)
pSTAT6 Phosphorylated STAT6 (AU) Hamilton et al. 2014
Th2 Th2 cell index (skin AU) Guttman-Yassky et al. 2019
ILC2 ILC2 activation index Kim et al. 2013
TARC TARC/CCL17 plasma (pg/mL) Normal <450, AD 3000-8000
Eos_blood Blood eosinophil count (/µL) Normal 100-500, AD 350+
IgE Total IgE (IU/mL) AD baseline: 1000-5000
IL31 IL-31 skin/plasma (pg/mL) Baseline 45 pg/mL
FLG Filaggrin expression (relative) Normal=1.0, AD≈0.40
TEWL Trans-epidermal water loss Normal <10, AD 20-40 g/m²/h
SkinInfl Composite inflammation index Maps to EASI
EASI Eczema Area & Severity Index Calibrated to SOLO-1/2 data

Treatment Scenarios (6 Scenarios)

# Scenario Drug(s) Key PD Mechanism
1 No Treatment — Natural Th2 progression
2 TCS Only Topical CS STAT6 partial inhibition (+35%)
3 Dupilumab Q2W 600mg load → 300mg Q2W IL-4Rα blockade, RO >85%
4 Upadacitinib 30mg QD 30mg daily JAK1 inhibition >90%, broad cytokine block
5 Nemolizumab Q4W 60mg Q4W SC IL-31Ra blockade, rapid itch
6 Dupilumab + TCS 300mg Q2W + TCS Additive STAT6 inhibition

Simulated Response Endpoints at Week 16 (Model)

Scenario EASI NRS EASI-75 IGA 0/1 TARC Δ%
No Treatment ~28 ~9.5 0% 0% ~0%
TCS Only ~21 ~7 5% 10% ~-20%
Dupilumab Q2W ~10 ~5.5 45% 35% ~-70%
Upadacitinib 30mg ~8 ~4.0 55% 42% ~-65%
Nemolizumab Q4W ~22 ~4.5 8% 12% ~-15%
Dupilumab + TCS ~7 ~5.0 60% 50% ~-75%

Model-predicted values; clinical trial values (SOLO-1/2, Rising Up) used for calibration.

Shiny App — 6 Interactive Tabs

Tab Content
1. Patient Profile Disease severity sliders (EASI, IgE, Eos, TARC, BSA), AD subtype, comorbidities
2. Drug PK Profiles Dupilumab Cp, IL-4Rα RO%, Upadacitinib Cp, JAK1 inhibition %
3. PD Biomarkers TARC, eosinophil, IL-31, IgE, pSTAT6 time courses
4. Clinical Endpoints EASI, IGA, NRS itch, EASI-75/90 response bars, data table
5. Scenario Comparison Multi-arm overlay: EASI, NRS, TARC with summary table
6. Barrier & Pruritus FLG expression, TEWL, itch-inflammation feedback loop

Key Biological Insights

Dupilumab Mechanism

  • Blocks shared IL-4Rα chain → inhibits both IL-4 and IL-13 signaling simultaneously
  • Downstream effects: pSTAT6↓ → TARC↓, FLG↑, IgE↓ (slow), eosinophil normalization
  • Transient eosinophilia early (~4-8 weeks) before long-term reduction
  • Ocular comorbidity (conjunctivitis ~15%) — mechanism under investigation (IL-13 suppression in conjunctiva)

Upadacitinib (JAK1 Inhibitor)

  • Broadest cytokine inhibition profile: TSLP, IL-4, IL-13, IL-31, IFN-γ all use JAK1
  • Fastest itch improvement in head-to-head trials (Heads Up: upadacitinib > dupilumab at Wk 16)
  • Concerns: herpes zoster reactivation, platelet changes, lipid effects (JAK2 off-target)

Nemolizumab (Anti-IL-31Ra)

  • Itch-selective biologic — does not significantly impact EASI at standard doses
  • Rapid NRS improvement (within 2 weeks) without broad anti-inflammatory effect
  • Best used in itch-predominant AD phenotype

Epidermal Barrier Recovery

  • FLG suppression by STAT6 creates a vicious cycle: barrier defect → more Th2 sensitization
  • Dupilumab restores FLG expression via STAT6 inhibition within 4-8 weeks
  • TEWL normalization is a delayed but objective marker of barrier recovery

References Summary

Category Count
Epidemiology & burden 4
Epidermal barrier & FLG 5
Innate immunity & alarmins 7
IL-4/IL-13/STAT6 signaling 4
IL-31 & pruritus 4
Dupilumab clinical trials 4
Dupilumab PK/PD modeling 2
Tralokinumab 2
Lebrikizumab 2
Nemolizumab 2
JAK inhibitors 4
TARC & biomarkers 3
QSP & mathematical modeling 4
mrgsolve tools 2
Scoring systems 3
Total 52

Installation & Usage

# Required R packages
install.packages(c("mrgsolve", "dplyr", "ggplot2", "tidyr", "purrr",
                   "shiny", "shinydashboard", "DT", "plotly"))

# Run mrgsolve simulation
source("ad_mrgsolve_model.R")

# Launch Shiny app
shiny::runApp("ad_shiny_app.R")

Model developed as part of the QSP Disease Model Library. Parameters calibrated from published clinical trial data (SOLO-1/2, ECZTRA-1/2/3, Rising Up, JADE MONO-1/2).