Phase 4 — Ranked Candidate List (v2)¶
Date: 2026-05-06 Total compound × chokepoint pairs evaluated: 42 Excluded by toxicity filter: 2 Ranked candidates (after dedup): 36
Top tier — in vivo / clinical / multi-chokepoint¶
| Rank | Compound | Chokepoint(s) | Producer | Evidence | Finding / Potency | Score |
|---|---|---|---|---|---|---|
| 1 | LOVASTATIN | HDAC6 (2 chokepoints) | Aspergillus terreus / Pleurotus ost | in_vitro_biochem | IC50 16285 nM | 6 |
| 2 | LOVASTATIN | PPARgamma (2 chokepoints) | Aspergillus terreus | in_vitro_biochem | IC50 50118.7 nM | 6 |
| 3 | 2,4-dihydroxybenzoic acid methyl ester (DAE) | XO (2 chokepoints) | Ganoderma applanatum | animal | DAE (Ganoderma applanatum) reduces serum uric acid in mice from 407 to 134 µmol/L via | 6 |
| 4 | 2,4-dihydroxybenzoic acid methyl ester (DAE) | URAT1 (2 chokepoints) | Ganoderma applanatum | animal | Same DAE entry — also down-regulates URAT1 (this is the load-bearing dual-chokepoint | 6 |
| 5 | Cordyceps sinensis preparations | urate_axis_general | Cordyceps sinensis | CLINICAL — Cochrane systematic review of RCTs | CLINICAL EVIDENCE: 5 RCTs (n=447) showed Cordyceps reduces hyperuricemia, hyperlipide | 6 |
| 6 | Cordyceps sinensis (natural + fermented) | urate_axis_general | Cordyceps sinensis | CLINICAL — meta-analysis of 9 RCTs | CLINICAL: 9 RCTs (mostly Chinese, including CNKI sources). Cordyceps + cyclosporine A | 6 |
| 7 | bois d,arc | URAT1 (3 chokepoints) | [query-hint] Ganoderma, [query-hint | in_vitro_biochem | IC50 2000 nM (pChEMBL 5.7) | 6 |
| 8 | 5,8-dihydroxy-10-methoxy-2-methylbenzo[h]chromen-4 | ABCG2 | [query-hint] Aspergillus niger | in_vitro_biochem | IC50 12800 nM (pChEMBL 4.89) | 5 |
| 9 | bois d,arc | ABCG2 (3 chokepoints) | [query-hint] Ganoderma, [query-hint | in_vitro_biochem | IC50 21000 nM (pChEMBL 4.68) | 5 |
| 10 | bois d,arc | XO (3 chokepoints) | [query-hint] Ganoderma, [query-hint | in_vitro_biochem | IC50 10100 nM (pChEMBL 5) | 5 |
| 11 | davallialactone | XO | Sanghuangporus vaninii (Phellinus / | in vitro + animal | Davallialactone isolated from S. vaninii at 97.7% purity exhibits mixed XO inhibition | 4 |
| 12 | 2,5-dihydroxyacetophenone (DHAP) | XO | Cordyceps spp. | animal | DHAP from Cordyceps reduces SUA, inhibits hepatic XOD, computationally screened again | 4 |
| 13 | cordycepin | URAT1 | Cordyceps militaris (cordycepin = 3 | animal | Cordycepin lowers SUA from 337 to 203 µmol/L via URAT1 modulation — direct URAT1 chok | 4 |
| 14 | ethanol extract (mixed) | URAT1 | Sanghuangporus vaninii | animal | S. vaninii ethanol extract regulates uric acid transporters (URAT1/GLUT9/OAT1/ABCG2 f | 4 |
| 15 | AMC-BFE (mixed extract) | URAT1 | Cordyceps militaris (co-fermented w | animal | AMC-BFE downregulates URAT1 + GLUT9, alters ABCG2, activates PPARα — multi-chokepoint | 4 |
| 16 | FZ formula (mixed); Poria cocos contribution unspe | ABCG2 | Poria cocos (component of FZ formul | animal | FZ formula modulates ABCG2/GLUT9/OAT1 expression and suppresses NLRP3 inflammasome in | 4 |
| 17 | FZ formula | NLRP3 | Poria cocos (in FZ formula) | animal | Suppresses NLRP3 + ASC + cleaved-caspase-1 in HUA-nephropathy mouse model. | 4 |
| 18 | exopolysaccharides (EPS) | NLRP3 | Ganoderma lucidum | animal + in vitro | DIRECTIONALITY CAVEAT: G. lucidum EPS ACTIVATE (not inhibit) NLRP3 via dectin-1/Syk/N | 4 |
| 19 | S-GLSP (mixed: 42 compounds — flavonoids, alkaloid | NLRP3 | Ganoderma lucidum (spore powder, sp | animal + in vitro | S-GLSP INHIBITS NLRP3 in AD model — opposite directionality from PMID 40450627. Spore | 4 |
| 20 | GLE extract (mixed) | NLRP3 | Ganoderma lucidum | animal | GLE represses microglial NLRP3 in Parkinson model. Confirms G. lucidum extracts inhib | 4 |
| 21 | GLP (β-glucan) | NLRP3 | Ganoderma lucidum | animal | GLP inhibits NLRP3 in cognitive-dysfunction model. Note: this β-glucan INHIBITS NLRP3 | 4 |
| 22 | GLPS polysaccharides | NLRP3 | Ganoderma lucidum | animal | GLPS suppresses NLRP3/ASC/caspase-1 in CCl4 liver injury. | 4 |
| 23 | GM (mycelium powder), GP (extracellular polysaccha | urate_axis_general | Ganoderma lucidum (mycelium + EPS) | animal | GM at 200-400 mg/kg lowers serum UA from 5.56 to 4.45 mg/dL, downregulates hepatic XO | 4 |
| 24 | GLPP (polysaccharide-peptide) | urate_axis_general | Ganoderma lucidum | animal + in vitro | GLPP reduces blood UA up to 40.6% in HUA mice via ADA inhibition + GLUT9 down + OAT1 | 4 |
| 25 | fruiting body extract (FBCC) | urate_axis_general | Cordyceps chanhua | animal (acute HUA mice + chronic gout rats) | FBCC reduces SUA + suppresses XOD + reduces renal tubulointerstitial damage in chroni | 4 |
Multi-chokepoint hits (mechanism diversity ≥ 2)¶
These are the highest-leverage candidates because they hit multiple chokepoints simultaneously, suggesting platform-level rather than single-target mechanism.
- LOVASTATIN — hits HDAC6, PPARgamma (2 chokepoints). Producer: Aspergillus terreus / Pleurotus ostreatus / Monascus ruber. Source: ChEMBL (Phase 2a).
- 2,4-dihydroxybenzoic acid methyl ester (DAE) — hits URAT1, XO (2 chokepoints). Producer: Ganoderma applanatum. Source: PubMed 35750011 (2022).
- bois d,arc — hits ABCG2, URAT1, XO (3 chokepoints). Producer: [query-hint] Ganoderma, [query-hint] Ganoderma lucidum. Source: LOTUS × ChEMBL (Phase 3a).
- quercetin — hits ABCG2, XO (2 chokepoints). Producer: [query-hint] Agaricus. Source: LOTUS × ChEMBL (Phase 3a).
Excluded by toxicity filter¶
| Compound | Chokepoint | Reason | Source |
|---|---|---|---|
| FUMITREMORGIN (C) | ABCG2 | EXCLUDED (Aspergillus fumigatus) | ChEMBL (Phase 2a) |
| CYCLOTRYPROSTATIN A (12,13-dihydroxyfumitremorgin C) | ABCG2 | EXCLUDED (Aspergillus fumigatus) | ChEMBL (Phase 2a) |
PubMed high-signal findings (verbatim)¶
- DUAL-CHOKEPOINT HIT: Ganoderma applanatum 2,4-DAE (PMID 35750011) inhibits BOTH XOD and URAT1 in mouse model — structurally allopurinol-like, in vivo SUA reduction from 407 to 134 µmol/L. This is the strongest single-compound finding from the breadth scan.
- CORDYCEPIN AS DIRECT URAT1 MODULATOR: Cordyceps militaris cordycepin (PMID 29422889) — dose-response in mice, SUA 337→203 µmol/L, mechanism proven via ELISA/RT-PCR/WB. Already a well-characterized molecule (3'-deoxyadenosine) with established pharmacology.
- CLINICAL-LEVEL EVIDENCE: Cordyceps sinensis has TWO meta-analyses (PMIDs 26457607 Cochrane + 28137532) of RCTs in transplant patients showing hyperuricemia reduction. Highest evidence-tier finding in the entire scan — clinical, not preclinical.
- MULTI-CHOKEPOINT EXTRACT: AMC-BFE (Cordyceps militaris × Astragalus biotransformation, PMID 41905012, 2026) hits URAT1 + GLUT9 + ABCG2 + PPARα simultaneously. Bidirectional fermentation enriches isoflavone aglycones — koji-relevant biotransformation strategy comparator.
- ABCG2 + NLRP3 SIMULTANEOUS HIT: Poria cocos in Fuling-Zexie formula (PMID 37788785) modulates ABCG2/GLUT9/OAT1 + suppresses NLRP3/ASC/cleaved-caspase-1. Poria is the only fungal ingredient — needs decomposition to identify which compound class drives which axis.
- POLYSACCHARIDE DIRECTIONALITY PROBLEM: G. lucidum exopolysaccharides ACTIVATE NLRP3 in cryptococcosis (PMID 40450627, beneficial) but G. lucidum spore powder INHIBITS NLRP3 in Alzheimer (PMID 41378217). β-Glucan structure (MW, branching, sulfation, triple-helix conformation) determines directionality (PMID 40592243 review). Critical for Phase 5: spore powder ≠ EPS ≠ fruiting-body β-glucan. Cell-wall vs cytoplasmic vs lipid-fraction extracts have OPPOSITE NLRP3 effects.
- C5aR1 PLATFORM GAP CONFIRMED (per validation-experiments.md §1.21): NO direct fungal antagonist of C5aR1 surfaced in PubMed scan. β-glucans ACTIVATE complement (CR3/iC3b opsonization, PMID 17895634) — opposite of inhibition. Complement-NLRP3 crosstalk review (PMID 41983141) confirms the mechanistic synergy rationale but identifies no natural fungal small-molecule antagonist. Phase 5 C5aR1 work will need synthetic-derivative or non-fungal-source approaches; whole-mushroom polysaccharide programs will not fill this gap.
- NOVEL CHOKEPOINT — MITOPHAGY: Cordyceps cicadae regulates PINK1-mediated mitophagy in HUA nephropathy (PMID 40334761). Not in current Open Enzyme chokepoint map — worth surfacing to validation-experiments.md as a new candidate.
- NOVEL CHOKEPOINT — ADENOSINE DEAMINASE (ADA): Ganoderma lucidum GLPP polysaccharide-peptide inhibits ADA (PMID 36385640). ADA upstream of XO — earlier intervention point. Worth adding to chokepoint map.
- Sanghuangporus vaninii davallialactone (PMID 36801789): XO IC50 = 90 µM. Fruiting-body XO inhibitor at characterized purity (97.7%). Mid-tier potency (allopurinol IC50 ~7 µM) but novel scaffold.
- MULTILINGUAL CATCH-NET: PubMed indexed extensive Chinese-author primary research (Shandong U, Peking U, Northwest A&F, Shanghai Academy, Guangdong CM University). The TCM × hyperuricemia search (PMID 38907332, 39176260, etc.) confirms the 'Chinese-language clinical evidence is real' framing — do not gate on Western evidence alone.
- Lp-PLA2 (15 hits total) and HDAC6 (155 hits total): essentially zero direct fungal hits on these chokepoints. Phase 5 will need ChEMBL/LOTUS querying or structural prediction work — PubMed primary literature is empty.
- Trametes versicolor / PSK: only 2 hits total, neither NLRP3-focused. PSK is well-characterized as immune adjuvant in Japan, but PubMed primary research on PSK × NLRP3 is essentially absent. Phase 5 may need direct screening.
- Pleurotus ostreatus / oyster mushroom: 189 hits dominated by lovastatin/cholesterol. Statin axis is real (Pleurotus produces lovastatin natively) but does not map to our XO/URAT1/NLRP3/C5aR1 chokepoints. Off-target for current scope.
Caveats¶
- Plant-origin compounds (quercetin, genistein, bois d'arc) are flagged and discounted but appear in LOTUS mushroom records due to substrate accumulation.
- LOTUS organism attribution is via [query-hint] (search-term name) for many records — not per-record biosynthetic provenance. Phase 4b should re-attribute via LOTUS per-compound endpoint.
- ChEMBL coverage of canonical mushroom compounds is severely Western-pharma-biased (G. applanatum 2,4-DAE, GLPP, davallialactone, AMC-BFE — none in ChEMBL).
- HIT (TCM-target), KNApSAcK, NPASS, TCMSP, BATMAN-TCM not yet pulled. SwissTargetPrediction for orphan compounds not run. Phase 3 is incomplete on the long-tail side.
- Two new chokepoint candidates surfaced (ADA, PINK1/mitophagy) — not yet in canonical wiki chokepoint inventory. Phase 5 decision.