Antifungal Therapy: Drug–Fungus Mechanism Design Pattern

Status: Phase 1 — Antifungal Mechanism Module Set Built

This project extends the ANTIMICROBIAL drug–bug mechanism design pattern to antifungal therapy. Same machinery — a Treatment links via target_mechanisms to the specific pathophysiology node (a fungal biosynthetic target or a gating principle) that makes the drug work, and recurrent fungus-property × drug-class interactions are captured once as kb/modules/ that disease entries conforms_to.

All five antifungal mechanism modules are now built and validated (schema + term + independent snippet-substring verification of every evidence quote): four drug-target modules (ergosterol synthesis, polyene membrane binding, echinocandin β-glucan synthesis, flucytosine antimetabolite) plus the species-level intrinsic-resistance gating module. Invasive candidiasis is the first target-only echinocandin conformer, Coccidioidomycosis is the first full ergosterol-synthesis conformer, and Otomycosis is the first intrinsic-resistance conformer. The ergosterol-synthesis tranche now wires all four named existing candidates; the remaining missing flagship is cryptococcal meningitis, tracked with the other follow-up modules in §7.

Antifungal pharmacology has a tighter, more conserved target set than antivirals because fungi are eukaryotes — selective toxicity hinges on a handful of fungal-specific structures (ergosterol instead of cholesterol; a β-1,3-glucan cell wall mammalian cells lack).

Multi-module conformers (the payoff — one disease constrained by several independent antifungal mechanisms, mirroring Leprosy/Whipple on the antibacterial side): - Cryptococcal meningitis (entry to be created): ergosterol binding (fungal_membrane_ergosterol_binding, amphotericin B) + antimetabolite (fungal_nucleic_acid_antimetabolite, flucytosine) for induction, then azole (fungal_ergosterol_synthesis_inhibition, fluconazole) consolidation — and antifungal_intrinsic_resistance_gating because Cryptococcus is intrinsically echinocandin-resistant (β-glucan synthase is present but not a viable target), explaining why the cell-wall module does not apply. - Invasive candidiasis: the target-only first tranche now models anidulafungin inhibition of Fks glucan synthase via fungal_cell_wall_glucan_synthesis_inhibition. Azole step-down and species/C. auris resistance gating remain future multi-module extensions.

0. Scope and Positioning (what this is and is NOT)

This is an explanatory mechanism layer, not a drug–indication database and not a clinical decision support system (DSS). Identical positioning to ANTIMICROBIAL §0:

• Not duplicating drug-indication curation. DrugBank, DrugCentral, ChEMBL, and antifungal labels already enumerate which antifungals treat which mycoses. dismech's contribution is to explain those associations mechanistically — linking a treatment, via target_mechanisms, to the fungal biosynthetic step it blocks — and to make that explanation queryable and consistency-checkable across diseases via shared modules.
• Not a DSS. IDSA antifungal guidelines, CLSI/EUCAST breakpoints, and stewardship pathways already give point-of-care guidance. These modules are at most a mechanistic substrate such tools could cite, not a prescribing engine.
• Open question worth a short survey: how DrugMechDB and antifungal resistance/MIC databases represent fungal drug-target mechanism paths — borrow vocabulary, avoid reinventing, sharpen where mechanistic explanation is additive. Tracked as a follow-up, not a blocker.

1. The Problem

Naively, the treatment block for every mycosis collapses to Pharmacotherapy → some antifungal → disease. That discards the knowledge clinicians actually use: the small antifungal armamentarium is sharply gated by organism. Whether a class works depends on conserved fungal-cell properties — does the organism have an azole-druggable CYP51, an echinocandin-druggable β-glucan wall, a polyene-bindable ergosterol membrane, and does intrinsic resistance exclude a class outright? We want to encode that depth.

2. The Core Principle: Target the Fungal Biosynthetic Node, Not the Disease

A Treatment links to specific pathophysiology nodes via target_mechanisms (TreatmentMechanismTarget). For antifungals, the edge should point at the fungus's druggable biology, which lives as a pathophysiology node:

• Otomycosis now keeps its existing biofilm and Aspergillus-specific fluconazole-resistance nodes separate from a complete Cyp51 → ergosterol → membrane-organization slice. Topical clotrimazole and systemic itraconazole carry agent-specific INHIBITS edges to Cyp51, while clotrimazole separately BYPASSES only the fluconazole exclusion; this does not assert universal isolate susceptibility.
• Coccidioidomycosis now carries a complete Cyp51 → ergosterol → membrane organization slice, with agent-specific target_mechanisms edges from fluconazole and itraconazole. Its rich host-immunity pathophysiology (CLEC7A/ dectin-1 β-glucan recognition, CARD9, Th1/Th17) remains distinct. Note the elegant duality: the same β-glucan that dectin-1 recognizes for host defense is the polymer echinocandins block the synthesis of — a candidate cross-link between the host-immunity nodes and a future cell-wall drug-target node.

A single target_mechanisms edge encodes why this drug, and it predicts failure modes: an echinocandin has a viable target only where β-1,3-glucan synthase is essential and accessible (Candida, Aspergillus) — not in Cryptococcus or the Mucorales, which it cannot treat.

3. Three-Tier Encoding

Tier 1 — crude baseline. treatment_term: NCIT:C15986 (Pharmacotherapy/antimicrobial agent therapy as appropriate) or NCIT:C15986 (Pharmacotherapy) / NCIT:C15986 (Pharmacotherapy) + therapeutic_agent (CHEBI for the drug, e.g. CHEBI:3764 clotrimazole, plus fluconazole/itraconazole/amphotericin B/caspofungin/flucytosine) + the modality appropriate to that agent (usually SMALL_MOLECULE; echinocandins such as anidulafungin are PEPTIDE). Says "clotrimazole is used for otomycosis."

Tier 2 — the mechanistic edge (the depth). Add a pathophysiology node for the targeted fungal step and link target_mechanisms to it: an azole → "Lanosterol 14-alpha-Demethylation by CYP51 (ERG11)"; terbinafine → "Squalene Epoxidation by Squalene Epoxidase (ERG1)"; amphotericin B → "Ergosterol Membrane Integrity"; an echinocandin → "β-1,3-Glucan Cell-Wall Synthesis by Fks glucan synthase"; flucytosine → "Fungal DNA/RNA Synthesis (cytosine-deaminase activation)". Use target_phenotypes for adjuncts (surgical debridement, immune reconstitution).

Tier 3 — conserved generalizations as modules. The reasons "some classes are useless against some fungi" are conserved fungus-property × drug-class interactions — kb/modules/ + conforms_to. See the proposed module table above. Because antifungal selectivity rests on a few fungal-specific structures, the module set is small and high-coverage.

4. The Axes That Make Specific Drugs Better for Specific Fungi

5. Recommended Encoding Pattern (per fungal entry)

1. Populate infectious_agent (InfectiousAgent, NCBITaxon term) — the fungal identity anchor (already present in Coccidioidomycosis, Chromoblastomycosis, Otomycosis, Mycetoma).
2. In pathophysiology, add fungal drug-target nodes (ergosterol synthesis, ergosterol membrane, β-glucan wall synthesis, nucleic-acid antimetabolite) and gating-property nodes (intrinsic resistance, biofilm/dimorphic phase, CNS sanctuary) as appropriate.
3. Each Treatment: Tier-1 fields plus target_mechanisms → the specific node(s) it inhibits, and target_phenotypes for adjuncts (debridement, immune reconstitution).
4. Where the mechanism recurs, conforms_to a shared antifungal-mechanism module.
5. Consider cross-linking host β-glucan recognition (dectin-1/CLEC7A) nodes to the β-glucan-synthesis drug-target node where both are modeled — the same polymer seen from the host-defense and drug-target sides.

6. Open Decision for the Register

Surface to docs/explanation/design-decisions.md: the same per-disease-vs-module decision recorded for antibacterials applies. Antifungal-specific question: whether the two ergosterol-biosynthesis targets (CYP51 demethylase for azoles, squalene epoxidase for allylamines) are one module with two target nodes (recommended — both normal dependencies converge on ergosterol production and membrane organization) or two modules. Recommendation: one fungal_ergosterol_synthesis_inhibition module with distinct, parallel CYP51 and squalene-epoxidase target nodes, analogous to the bacterial folate module holding distinct DHPS and DHFR targets.

7. Next Steps

• [x] Draft fungal_ergosterol_synthesis_inhibition as the proof-of-concept module. Built at kb/modules/fungal_ergosterol_synthesis_inhibition.yaml: five nodes: parallel CYP51/ERG11-demethylase and ERG1-squalene-epoxidase targets converge on normal ergosterol production and plasma-membrane organization; target alteration/overexpression and efflux are recorded as a standalone resistance state. Evidence: 33374996 and 31643715 (CYP51), 1543672 (ERG1, ergosterol, and membrane dependency). Key target: #Lanosterol 14-alpha-Demethylation by CYP51 (ERG11).
• [x] Draft fungal_membrane_ergosterol_binding (polyenes). Built: three nodes (membrane ergosterol as the polyene binding target → permeabilization / fungicidal killing → rare resistance via reduced ergosterol). Evidence: 31643715 (LiverTox), Anderson 24681535 (sterol-sponge), Czajka 37998390.
• [x] Draft fungal_cell_wall_glucan_synthesis_inhibition (echinocandins, Fks glucan synthase). Built: five nodes (β-1,3-glucan synthase target → normal cell-wall assembly, plus intervention-conditional integrity failure / osmotic lysis, FKS-mediated resistance, and intrinsic resistance in Cryptococcus). Evidence: Perlin 26190298 / 26567278, Emri 23463246, Aruanno 31138565, Cappelletty 21694887, Iyer 33558691.
• [x] Draft fungal_nucleic_acid_antimetabolite (flucytosine). Built: three nodes (fungal cytosine-deaminase activation of 5-FC to 5-FU → disruption of fungal RNA/DNA synthesis → rapid monotherapy resistance mandating combination). Evidence: Vermes 10933638, Houšť 32178468, Noël 12654658.
• [x] Draft antifungal_intrinsic_resistance_gating (the antifungal analog of intracellular_pathogen_persistence — species-level class exclusion). Built: three nodes forming parallel intrinsic and acquired resistance arms that converge on resistance-gated narrowing of antifungal options; neither resistance arm causes the other. Otomycosis now conforms on the Aspergillus/fluconazole intrinsic arm. Evidence: 33558691, 38445857, 31159914, 33091071, and 28911043.
• [x] Wire additional ergosterol-synthesis conformers. Coccidioidomycosis is the first full conformer, with Cyp51 → ergosterol → membrane nodes and fluconazole/itraconazole target edges. Otomycosis duplicates the same Cyp51 slice for clotrimazole and persistent-disease itraconazole while preserving its separate Aspergillus/fluconazole intrinsic-resistance gate. A later, separate module tranche can add Coccidioidomycosis's amphotericin-B ergosterol-membrane target without conflating the two modules.
• [ ] Create Cryptococcal_Meningitis, the remaining missing flagship conformer (AmB + flucytosine induction → fluconazole consolidation; echinocandin-resistant).
• [x] Create Invasive_Candidiasis with Candidemia as a subtype and the minimal Fks glucan-synthase target-only conformer: anidulafungin → INHIBITS → fungal beta-1,3-glucan synthase. Azole step-down and acquired-resistance wiring remain later extensions.
• [x] Wire Chromoblastomycosis and Mycetoma (Eumycetoma subtype) on the high-precision Cyp51 branch. Chromoblastomycosis models the itraconazole component of documented combination therapy without asserting monotherapy; Mycetoma limits the fungal target chain and itraconazole edge to Eumycetoma and anchors it to Madurella mycetomatis. Terbinafine/Erg1 branches and surgical-adjunct targeting remain optional extensions.
• [ ] Short survey of antifungal drug-mechanism KBs / resistance databases (DrugMechDB, EUCAST/CLSI breakpoint rationales) — see §0; borrow vocabulary, sharpen the niche.
• [x] Record the ergosterol-one-module-two-target-branches decision in the design register, including branch-specific conformance.