Skip to content

comp-020 — Per-Node Findings (Brief-Scrubbed Verification Re-Run)

Independence statement: This re-run was conducted without consulting comp-018 / comp-019 outputs. The compound list emerged organically from target-anchored literature mining via Paperclip MCP (PMC + bioRxiv full-text + abstracts), supplemented by targeted WebSearch for primary-literature claims (rosmarinic acid PubMed records, ChEMBL coverage spot-checks).

Methodology: for each upstream complement node, anchor queries → top results → grep verification of load-bearing IC50/CH50/AP50 values from primary-paper full text → flag assay format → record source language, evidence tier, and primary citation.

Caveat: ALL hemolytic-assay IC50 values are highly assay-format-dependent. CH50 (sheep erythrocyte, classical pathway), AP50 (rabbit erythrocyte, alternative pathway), Wieslab ELISA (terminal MAC deposition), C4-deposition assay (MASP-2 specific), C3c ELISA (lectin-pathway specific) report different operative values for the same compound. See cross-cutting heterogeneity log below.


A. Initiation pathway — C1q (classical) / MBL-MASP (lectin) / AP tickover

A.1 Direct C1q-cascade modulators (CP via C1q + downstream C2/C9)

Compound Class CH50 (mM unless noted) AP50 Targets in cascade Evidence tier Source language Primary citation
(7S,8R)-Dihydrodehydrodiconiferyl alcohol (Helicteres compound 5) Plant lignan (benzofuran) 0.009 ± 0.002 0.021 ± 0.003 C1q, C2, C3, C9 In vitro (sheep + rabbit erythrocyte hemolysis) English (China-based authors) Yin X, Lu Y, Cheng ZH, Chen DF. Anti-Complementary Components of Helicteres angustifolia. Molecules 21(11):1506 (2016). PMC6273495 L18
Machicendonal (Helicteres compound 4) Plant lignan (benzofuran) 0.040 ± 0.009 0.105 ± 0.015 C1q, C2, C3, C4, C5, C9 In vitro (sheep + rabbit) English (China-based) Yin et al. 2016, PMC6273495 L18
Luteolin Plant flavone (3',4'-OH) 0.19 ± 0.02 0.17 ± 0.04 CP + AP (mechanism not target-mapped in this paper) In vitro English (China-based) Zhang T, Chen DF. Anticomplementary principles of a Chinese multiherb remedy for SARS. J Ethnopharmacol 117(2):351-61 (2008). PMC7126446 L10 + Table 1
Quercitrin Plant flavonol glycoside 0.53 ± 0.02 0.32 ± 0.04 CP + AP In vitro English (China-based) Zhang & Chen 2008, PMC7126446
Quercetin Plant flavonol 0.50 ± 0.02 1.02 ± 0.03 CP > AP In vitro English Zhang & Chen 2008, PMC7126446
Rutin Plant flavonol glycoside 0.58 ± 0.01 0.42 ± 0.04 CP + AP In vitro English Zhang & Chen 2008, PMC7126446
Hyperoside Plant flavonol glycoside 1.72 ± 0.01 0.25 ± 0.02 AP-selective In vitro English Zhang & Chen 2008, PMC7126446
Apigenin Plant flavone 3.40 ± 0.06 2.02 ± 0.21 CP + AP, weak In vitro English Zhang & Chen 2008, PMC7126446
Ginsenoside Rg3 Plant triterpene saponin not in matched-format paradigm C1q reduction (mouse brain, depression model) Animal model — chronic restraint depression mouse English (China-based) Yang D et al., PMC12594608 (2025)

Reference: Heparin sodium salt in the same Zhang & Chen 2008 paradigm: CH50 38.5 ± 1.78 μg/mL ≈ 2.6 μM (assuming MW 15 kDa). On a μM basis the Helicteres lignan compound 5 (CH50 9 μM) is comparable to heparin and 20× more potent than luteolin.

Top-tier within ~20% on lead metric (CH50 ≤ 0.20 mM = 200 μM): the Helicteres benzofuran lignans (compound 5 + compound 4) and luteolin are the only three sub-200-μM compounds; compounds 4 and 5 are ~5–20× more potent than luteolin and stand alone at the top. Per the no-headline rule, both Helicteres benzofuran lignans tie for top tier; luteolin sits at the next tier (~5× weaker on CH50, ~1× on AP50).

A.2 Lectin pathway — MASP-2 / MBL-MASP

Compound Class IC50 Assay Target Evidence tier Source Primary citation
Heparin (unfractionated) Sulfated GAG (FDA-approved drug) 2 μg/mL (LP) / 39 μg/mL (CP) / 76 μg/mL (AP) WieLISA (LP/CP/AP) at 1:100 serum dilution MASP-2 (LP) + C1q (CP) + AP serine proteases In vitro English (Netherlands group) Talsma DT et al. MASP-2 Is a Heparin-Binding Protease. Front Immunol 11:732 (2020). PMC7212410 L45
Heparin tetrasaccharide (4-mer) Sulfated GAG fragment 21 μg/mL (LP), no CP/AP inhibition at tested concentrations WieLISA MASP-2-selective In vitro English Talsma 2020, PMC7212410
Heparin hexasaccharide Sulfated GAG fragment 4 μg/mL (LP), LP-selective WieLISA MASP-2 In vitro English Talsma 2020
Heparin octasaccharide Sulfated GAG fragment 3 μg/mL (LP), LP-selective WieLISA MASP-2 In vitro English Talsma 2020
Bupleurum smithii crude polysaccharide (BPs) Plant pectic polysaccharide LP IC50 1.057 mg/mL (C3c ELISA), CH50 0.34 mg/mL, AP50 0.081 mg/mL C3c ELISA + hemolysis C1s, C3, C4 (per prior work cited in PMC4629277 L46) In vitro English (China-based) Wu M, Li H, Zhang YY, Chen DF. C3c-based ELISA for Bupleurum polysaccharides. Acta Pharm Sin B (2015). PMC4629277
Bupleurum chinense crude polysaccharide (BCPs) Plant pectic polysaccharide LP IC50 0.098 mg/mL, CH50 0.35 mg/mL, AP50 0.337 mg/mL C3c ELISA + hemolysis C1q, C2, C5, C9 (per PMC4629277 L46) In vitro English (China-based) Wu et al. 2015, PMC4629277
Suramin Synthetic polysulfonated naphthylurea (FDA-approved trypanosomiasis drug) LP IC50 not stated as μM but operates on LP per PMC4629277 Hemolysis + C3c ELISA CP and LP Pre-clinical (FDA-approved for off-label complement use) English PMC4629277 — used as positive control
TFPI1-derived peptide (engineered MASP-2 inhibitor) Protein engineering High potency vs MASP-2 (Ki sub-nM range) Direct enzymatic + ischemia-reperfusion model MASP-2-selective Animal model English Szakács D et al. JBC 2019, PMC6527154

Top-tier on LP-selective metric: heparin octa-/hexasaccharides (3-4 μg/mL) are the most potent natural-derived LP-selective inhibitors; Bupleurum chinense polysaccharide (98 μg/mL i.e. 0.098 mg/mL) is the most potent plant-natural LP-active polysaccharide in the literature surveyed.

A.3 Complement-fixation polysaccharides (immunomodulatory; effect direction context-dependent)

Compound Class ICH50 (μg/mL) Note Citation
Ligusticum chuanxiong LCP-I-I Pectic polysaccharide (HG + RG-I + AG-I/II) 26.3 ± 2.2 Comparable to BP-II positive control 25.5 Zou YF et al. PMC6155779 (2017)

These cause complement fixation (consumption of complement at tissue sites away from inflammatory targets) — interpretive caution: in a gout context, fixation away from MSU sites could reduce or amplify C5a generation depending on tissue distribution. Flagged for further mechanism work.


B. Convertases — C3 convertase / C5 convertase

B.1 C3 convertase

Compound Class IC50 Assay Target Evidence tier Source Primary citation
Rosmarinic acid Plant phenolic acid (Lamiaceae/Boraginaceae) 34 μM (C3b covalent attachment to cells); 180 μM CP hemolysis; 160 μM AP hemolysis Cell-based C3b deposition + classical/alternative hemolytic Covalently modifies activated C3b → blocks C5 convertase formation; primarily C5 convertase In vitro + 3 in vivo models English Sahu A et al. Inhibition of complement by covalent attachment of rosmarinic acid to activated C3b. Biochem Pharmacol 57(12):1439-46 (1999). PMID 10353266. Also Englberger W et al. Int J Immunopharmacol 10(7):729-37 (1988). PMID 3198307. Also Peake PW et al. Int J Immunopharmacol 13(7):853-7 (1991). PMID 1761351
C5 convertase regulation by rosmarinic acid 1500 μM for C5 convertase direct inhibition Cell-based C5 convertase In vitro English Sahu 1999, PMID 10353266
Compstatin family (C3-binding peptide) Engineered cyclic peptide derived from phage display, NOT a natural product but FDA-approved adjacent (pegcetacoplan = compstatin derivative) sub-μM (peptide engineering) Direct C3 binding C3 (blocks both C3 and C5 convertase formation) Clinical (pegcetacoplan = Empaveli FDA 2021) English Mohan RR et al. PMC5082644; Gorham RD et al. PMC4306506

B.2 C5 convertase

Rosmarinic acid is the only natural compound with directly-named C5 convertase inhibition activity in the surveyed corpus (1500 μM — weak direct inhibition, but compound primarily acts upstream at C3b).


C. Soluble factors

C.1 Factor B / Factor D (alternative pathway)

Compound Class IC50 / Ki Target Evidence tier Citation
Iptacopan (LNP023) Synthetic small-molecule; approved for PNH/IgAN as Fabhalta sub-nM Factor B Ki Factor B Clinical (FDA 2023) Schubart A et al. PMC6475383 (2019); Tang Z et al. PMC11124358 (2024)
Danicopan (ACH-4471) Synthetic Factor D inhibitor sub-μM Factor D Clinical Risitano AM et al. PMC8634185 (2020)
MY008211A Synthetic small-molecule clinical Factor B/D class Clinical Ye L et al. medRxiv 2026
Vemircopan Synthetic Factor D inhibitor sub-μM Factor D Phase trial Schubart 2022 review PMC10092480

Natural-product Factor B/D direct-inhibition class: corpus surveyed surfaced no characterized natural-product factor B or factor D direct enzymatic inhibitor with in vivo evidence. The closest hits are sulfated polysaccharides (heparin, fucoidan, dextran sulfate) which broadly inhibit AP including factor B/D function via electrostatic competition with substrate, not via direct active-site binding.

C.2 Factor H upregulation (positive-direction intervention)

This was scoped in the brief as "upregulating Factor H is positive intervention." Surveyed corpus surfaced:

Compound Mechanism Effect Evidence tier Citation
Factor H mini-construct (mHDM-FH) Engineered protein, NOT a natural product Functional substitution Animal model Kamala O et al. PMC8696033 (2021)
Thrombospondin-1 (TSP-1) Endogenous human protein; not a "compound" but acts synergistically with Factor H Synergistic AP inhibition Animal model + in vitro Konwar S et al. bioRxiv 2024-07-31
Native Factor H protein supplementation Endogenous; engineering thread Direct AP regulation Clinical (autosomal disease compensation literature) Kopp A et al. PMC4030870; Sándor N et al. PMC10894998

Natural-product Factor H upregulator class: the surveyed corpus did NOT surface a small-molecule or natural-product compound with documented Factor H expression-upregulation activity in matched assay format. Factor H promoter modulation is dominated by endogenous regulators (NF-κB, IFNγ-responsive elements) rather than dietary or natural-product compounds in the literature scanned. Flagged as a coverage gap.

C.3 Factor I / Properdin / Clusterin

Corpus did NOT surface natural-product compounds with documented direct modulation of Factor I, properdin, or clusterin at IC50/Ki tier evidence.

Properdin caution: Properdin downregulation is NOT universally desirable. Low plasma properdin associates with increased CV mortality in carotid atherosclerosis (Louwe MC et al. PMC12074774, 2025). Compounds that broadly suppress AP convertase function via properdin destabilization may have adverse cardiovascular signal in long-term use. Flagged.


D. Membrane regulators — DAF / CD55 / CD59 / CR1

The surveyed corpus did NOT surface natural-product or dietary compounds with characterized expression-upregulation activity on CD55, CD59, or CR1 with comparable IC50/EC50 metric. CD55-related natural-compound literature focuses on (1) cancer biology contexts where high CD55 is detrimental and downregulation is studied, (2) endogenous transcription factor regulation (KLF4, An FQ et al. PMC10884306), (3) viral exploitation of CD55 by enteroviruses. Coverage gap; this is the engineering-thread (comp-012, H05) territory, not the natural-product thread.


E. C5 / C5aR1 axis (residual coverage)

Per brief: "Direct natural-product C5aR1 antagonist class confirmed empty by comp-014 + validation-experiments §1.21." This re-run did NOT re-execute the C5aR1 scan; confirmed by ruling out via comp-014 cross-reference per brief instructions. Avacopan is the synthetic FDA-approved C5aR1 antagonist; not fermentable.

The C5 axis ENGINEERING surface (anti-C5 antibodies eculizumab, ravulizumab; anti-C5a antibodies vilobelimab) is well-mapped in wiki/complement-c5a-gout.md; not re-done here.


F. Marine compounds — sulfated polysaccharides

Compound Source IC50 (CP, μg/mL) Assay Target Evidence tier Citation
Sulfated galactofucan SJW-3 Saccharina japonica (brown algae) 3.11 CP hemolysis CP (broad) In vitro Jin W et al. PMC4728500 (2015)
Crude fucoidan ANW Ascophyllum nodosum 0.98 CP hemolysis CP In vitro Jin 2015
SJS (acid-extracted S. japonica) Brown algae 4.51 CP hemolysis CP In vitro Jin 2015
HFW Hizikia fusiforme 5.51 CP hemolysis CP In vitro Jin 2015
SJW Saccharina japonica 7.26 CP hemolysis CP In vitro Jin 2015
HFS Hizikia fusiforme (acid-extract) 24.65 CP hemolysis CP In vitro Jin 2015
Sea cucumber polysaccharide (SC) Acaudina molpadioides (marine invertebrate) similar to ANW (~1 μg/mL range) CP hemolysis CP In vitro Jin 2015

Top-tier on CP-only metric: ANW (0.98 μg/mL) and SJW-3 (3.11 μg/mL) are within ~3× of each other in the same assay; both stand alone at top-tier for marine sulfated polysaccharides. SC (~1 μg/mL) joins this top tier per same paper. Per the no-headline rule, all three (ANW, SJW-3, SC) are surfaced as the top marine-polysaccharide tier.

Activator (negative finding): Floridoside from Mastocarpus stellatus (red alga) is a potent CP activator — recruits IgM to drive CP. Inverse hit, included for completeness; not a CP0 candidate. (Courtois A et al. PMC2579733)


G. Fungal metabolites

Per brief: "comp-014 catalogued ~6,800 fungal compounds with target activity. You can build on it for fungal sub-pass; don't redo the breadth scan."

The independent scan of "fungal complement inhibitor" / "Monascus complement" / "fumagillin complement" / "K76" surfaced no novel fungal natural product with documented direct complement-cascade IC50 in the matched paradigm. The Husakova & Patakova 2025 Monascus pigments review (PMC11877510) shows antioxidant/antimicrobial activity but not complement-cascade activity. Aspergillus-derived fumagillin targets MetAP2, not complement. Confirms comp-014's structural-gap finding for fungi at the C5aR1 node and extends it: no characterized fungal natural product directly inhibits any of the named upstream complement nodes in the surveyed corpus.

Flagged as a coverage finding. comp-014 verdict (zero direct fungal C5aR1 antagonists) is here independently extended: fungal natural products are systematically absent from upstream complement modulator literature with characterized IC50.


H. Bacterial metabolites

Survey of bacterial-metabolite anticomplement activity surfaced: - E. coli K5 polysaccharide (chemically O-sulfated) — non-natural-form, requires synthetic sulfation to gain activity - Klebsiella pneumoniae over-cleaves C5 to evade MAC formation — pathogen-evasion biology, not therapeutic - Gut commensal SCFA (butyrate/propionate) — well-characterized HDAC inhibitors per comp-007, but NOT documented direct upstream complement modulators

Coverage gap on natural bacterial-metabolite direct upstream complement modulators.


I. Top-tier compound list per node — depth-first compilation

For each upstream complement node, all compounds within ~20% of the lead metric (NO single headline):

Lead-metric tier per node

C1q-cascade direct (CP via C1q), ranking by CH50 in the matched Zhang/Chen + Yin/Chen 2008/2016 paradigm (μM-scale):

Rank tier Compounds
Top tier (CH50 < 50 μM) (7S,8R)-Dihydrodehydrodiconiferyl alcohol (9 μM); Machicendonal (40 μM)
Mid tier (CH50 50–500 μM) Luteolin (190 μM); Quercetin (500 μM); Quercitrin (530 μM); Rutin (580 μM); Helicteres compound 6 (kaempferol-3-O-glucoside, 143–1000 μM range across compounds 6–12)
Heparin reference ~2600 μM (38.5 μg/mL ÷ ~15 kDa MW)

MASP-2 / Lectin pathway selective, ranking by LP IC50 in WieLISA (μg/mL):

Rank tier Compounds
Top tier (LP IC50 < 5 μg/mL, LP-selective) Heparin octasaccharide (3); Heparin hexasaccharide (4); Heparin tetrasaccharide LP only (21)
Pan-pathway top tier (also CP+AP active) Unfractionated heparin (LP 2, CP 39, AP 76); LMW heparin (enoxaparin, fragmin, fraxiparin)
Plant polysaccharide top tier Bupleurum chinense polysaccharide BCPs (LP 98 μg/mL; CH50 350 μg/mL; AP50 337 μg/mL)

C3 convertase / C5 convertase indirect via C3b:

Rank tier Compounds
Lead Rosmarinic acid (C3b covalent IC50 34 μM; CP hemolysis 180 μM; AP hemolysis 160 μM)
Indirect: C3-binding cyclic peptide (engineered) Compstatin / pegcetacoplan family (sub-μM; FDA-approved 2021)

Marine sulfated polysaccharides (CP):

Rank tier Compounds
Top tier (CP IC50 < 5 μg/mL) ANW (Ascophyllum nodosum fucoidan, 0.98); SC (sea cucumber polysaccharide, ~1); SJW-3 (Saccharina japonica sulfated galactofucan, 3.11); SJS (4.51)

Factor B / Factor D / Factor H (direct natural-product modulators): COVERAGE GAP — no characterized natural-product direct modulators with IC50 evidence in the surveyed corpus. Engineering thread (mini-FH constructs) and synthetic small-molecule thread (iptacopan, danicopan) own these nodes.

CD55 / CD59 / CR1 expression upregulators: COVERAGE GAP — engineering thread territory (comp-012, H05).


J. Assay-format heterogeneity log

For compounds with reported IC50 spanning >5× across the literature:

Rosmarinic acid

Assay IC50 Source
Cell-based C3b covalent attachment 34 μM Sahu 1999
Classical pathway hemolysis (human serum) 180 μM Sahu 1999
Alternative pathway hemolysis (human serum) 160 μM Sahu 1999
Direct C5 convertase enzymatic 1500 μM Sahu 1999

Range: 34 → 1500 μM = 44× variation. Format-driven explanation: rosmarinic acid acts via covalent C3b modification — most efficient when C3b is being deposited (cell surface assay, IC50 34 μM), less efficient at later cascade steps where C3b is already bound (hemolysis IC50 180 μM), and very inefficient when forced to act on assembled C5 convertase directly (1500 μM). Mechanism is upstream of C5 convertase, not at it. Future use-case selection should match assay paradigm to intended mechanism:

  • Soft-tissue gout MSU surface C3b deposition → Sahu's 34 μM number is the operative one
  • Serum-pathway broad inhibition → 160-180 μM range
  • Late-cascade C5/C5b-9 quench → 1500 μM (rosmarinic acid is NOT the right tool)

Heparin

Assay IC50 Source
LP via WieLISA (1:100 serum dilution) 2 μg/mL Talsma 2020
CP via WieLISA 39 μg/mL Talsma 2020
AP via WieLISA (1:50 serum, 200 μg/mL ceiling) 76 μg/mL Talsma 2020
C4 cleavage MASP-mediated 102 μg/mL Talsma 2020
Sheep-erythrocyte CH50 (Zhang & Chen paradigm) 38.5 μg/mL Zhang & Chen 2008

Range: 2 → 102 μg/mL = 50× variation. Format-driven explanation: heparin is more potent vs LP because the LP assay uses 1:100 serum (less competing serum protein) AND because MASP-2 has a heparin-binding active-site geometry (Talsma 2020 Kd ~2 μM for direct binding). CP/AP IC50 values are higher because they involve more steps of the cascade and use 1:100 / 1:50 serum dilutions. Sheep-erythrocyte CH50 ≈ WieLISA CP IC50 within order of magnitude (38 vs 39 μg/mL).

Luteolin

Surveyed Zhang & Chen 2008 paper reports CH50 0.19 mM, AP50 0.17 mM. CNKI / WanFang searches (via WebSearch redirect) did not surface additional luteolin anti-complement IC50 in the matched paradigm, though luteolin is widely studied for other anti-inflammatory targets. Limited heterogeneity data in surveyed corpus — flagged as a Phase 2 follow-up: a focused CNKI / WanFang search in Chinese for 木犀草素 + 补体 (luteolin + complement) is the indicated next step.


K. Multilingual coverage analysis

Source language Hits surfaced Novel hits not in English Notes
English (China-based authors, English-language journals: Mol Pharmacol, J Ethnopharmacol, Molecules, Int Immunopharmacol) All structured CH50/AP50/IC50 data above Most TCM-derived complement work IS published in English-language journals by Chinese groups (Daofeng Chen group at Fudan; Quanbin Zhang group at OUC).
Chinese (CNKI / WanFang direct query) Targeted searches via WebSearch did not return primary-paper IC50 numbers in the time budget UNKNOWN — could be additional luteolin/quercetin/Bupleurum work in Chinese-only journals Phase 2 follow-up: dedicated CNKI/WanFang query for 补体抑制 + flavonoids/lignans/polysaccharides
Japanese (J-STAGE) Sho-saiko-to (TJ-9) Kampo paper PMC2276037 surfaced with IL-12 modulation; not direct complement Unclear — Kampo formulary likely contains anti-complement data Phase 2 follow-up: dedicated J-STAGE query for 補体 + 漢方 + Kampo formulas
Other (Korean KISS, Russian eLIBRARY) Not surveyed in time budget Phase 2 follow-up

Honest finding: the surveyed Western/English-language literature ALREADY contains substantial TCM-derived natural-product complement-modulator data — Daofeng Chen's group at Fudan publishes the canonical TCM anti-complement work in English (Mol Pharmacol, J Ethnopharmacol). The "Western-research bias" risk in this specific subfield is partially mitigated by the fact that the leading group publishes in English. However, the assumption is tested only at a coarse level — Kampo formulary work and CNKI-only Chinese-language work were NOT systematically scanned in this re-run's time budget. Phase 2 follow-up flagged.


L. ChEMBL coverage gap analysis

Survey of compounds for ChEMBL anti-complement records:

Compound ChEMBL ID search via PubChem cross-ref ChEMBL anticomplement assay coverage Comments
Luteolin CHEMBL156 Multiple records; HMG-CoA, MAPK, sirtuin, but no direct C1q/C3/CH50 assay records Gap
Quercetin CHEMBL50 Hundreds of records; multiple kinases and oxidoreductases; no direct complement-pathway IC50 records as primary endpoint Gap
Rutin CHEMBL222302 Records present but no direct complement-cascade IC50 assays Gap
Rosmarinic acid CHEMBL165102 Records present; no direct C3 convertase IC50 record despite published 1988/1991/1999 papers (PMID 3198307, 1761351, 10353266) Significant gap — published in primary literature for 30+ years but absent from ChEMBL anticomplement curation
Helicteres lignans (machicendonal, dihydrodehydrodiconiferyl alcohol) Not in ChEMBL No records Total absence
Bupleurum polysaccharides BPs/BCPs Not in ChEMBL (polysaccharides are systematically absent from ChEMBL by design) None Structural gap
Marine fucoidans (SJW, SJS, HFW, ANW) Not in ChEMBL None Structural gap
Heparin Multiple ChEMBL records (clinical drug) Anticoagulation primary; complement-inhibition secondary Anticomplement records present but secondary annotation
Iptacopan CHEMBL5314125 Factor B records present Covered
Danicopan CHEMBL4794868 Factor D records present Covered
Compstatin CHEMBL2105656 C3 records present Covered

ChEMBL coverage gap finding (matches comp-013 / comp-014 pattern): ChEMBL anticomplement assay curation is structurally biased toward synthetic small-molecule clinical-stage compounds (iptacopan, danicopan, compstatin derivatives) and absent for:

  1. Plant flavonoids/lignans with documented anti-complement IC50 in primary literature (rosmarinic acid is the canonical example — 30+ year primary-literature record, ChEMBL does not curate the C3-convertase IC50)
  2. Polysaccharides as a structural class (Bupleurum, marine fucoidans, heparin oligosaccharides — systematically absent from ChEMBL by design, since polysaccharide structure ≠ small-molecule SMILES)
  3. Natural-product hits from non-Western lit
  4. Most pre-2000 publications

Operational implication: any computational triage on upstream-complement compounds that uses ChEMBL as the primary IC50 source will systematically miss the natural-product class. The ChEMBL coverage gap is even more severe here than for the gout-relevant transporter chokepoints (URAT1, ABCG2) per comp-013's TCM coverage observation.

n compounds surveyed in this re-run with documented anticomplement IC50: ~25 n with corresponding ChEMBL anticomplement-curated record: ~5 (synthetic clinical drugs only) n absent from ChEMBL entirely as a chemical entity: ~10 (polysaccharides + lignans + several flavonoid glycosides)

Coverage rate: ~20%. Significantly below ChEMBL's coverage rate for kinases or GPCRs (typical >70% for clinical-stage compounds).


M. Translation-disagreement summary

Only one source in this re-run was non-English (PMC7126446 — Zhang & Chen 2008, J Ethnopharmacol, English journal but China-based authors using TCM materia medica references). The IC50/CH50/AP50 numerical data are reported directly in the English text; no translation step required.

CNKI / WanFang queries did not return primary-paper full-text in the time budget for translation cross-check. Translation-disagreement summary: not applicable to this re-run's data; deferred to Phase 2.

If the Phase 2 CNKI/WanFang follow-up surfaces primary-language IC50 papers (e.g., 补体经典途径抑制 luteolin numerical data), the two-model cross-check protocol (Claude Anthropic + DeepSeek per CLAUDE.md §Translation protocol) applies.


N. Recommendations

N.1 Prioritize for in vitro confirmation

Tier 1 — strongest combined evidence (matched assay format + multiple groups + in vivo data where available):

  1. Rosmarinic acid — C3 convertase / C3b covalent inhibitor, 34 μM (PMID 10353266); 30+ year record across 3 independent papers (1988, 1991, 1999); Lamiaceae family (rosemary, perilla, lemon balm) is dietary-tier; in vivo activity in 3 models documented. Highest-leverage candidate for upstream-complement natural-product intervention. Specific gout-relevant follow-up: does rosmarinic acid reduce C3b deposition on MSU crystal surface in vitro? (Answer not in surveyed literature; specific Phase 2 test.)
  2. Luteolin — broad CP+AP inhibitor, 190 μM CH50 (PMC7126446); flavonoid common to Apium graveolens (celery), Petroselinum crispum (parsley), Camomilla, dietary-tier; multiple TCM clinical contexts. Already partially in OE corpus per comp-013 (luteolin XO IC50 550 nM, URAT1 expression downregulation). The convergence of XO + URAT1 + complement-inhibitor activity makes luteolin a multi-mechanism candidate worth elevating.
  3. Helicteres benzofuran lignans (compounds 4 and 5, machicendonal + (7S,8R)-dihydrodehydrodiconiferyl alcohol) — most potent CH50 in the surveyed corpus (9–40 μM), single-paper documentation (Yin 2016), C1q + C2 + C3 + C9 mapping done. Tier-1-with-replication-needed: rerun the assay before in vivo work; single-paper finding is fragile.

Tier 2 — strong in vitro but mechanism gap or limited replication:

  1. Bupleurum chinense polysaccharide (BCPs) — pan-pathway active including LP, oral GI use long history in TCM; concrete IC50 numbers (CH50 0.35 mg/mL); polysaccharide MW heterogeneity is a downside for reproducibility.
  2. Marine fucoidans — sulfated GAG class; structurally similar to heparin (which is FDA-approved); broad-pathway activity. Cardiovascular safety profile is a known risk (fucoidan alginate microbeads cause coagulation per PMC11783016 — caution).

N.2 Drop or de-prioritize

  • Direct natural-product Factor B / Factor D / Factor H modulators — corpus is empty at this resolution. Engineering thread (mini-FH per Kamala 2021) and synthetic small-molecule thread (iptacopan, danicopan) own these nodes for now. Natural-product CP0 is C1q/C3-targeted, not Factor B/D-targeted.
  • CD55 / CD59 / CR1 expression upregulator natural-product class — also empty at this resolution. Engineering thread (comp-012, H05) is the right tool here.
  • Floridoside (red alga) — confirmed CP activator, not inhibitor; do not pursue as a CP0 candidate.

N.3 Phase 2 follow-ups (explicit)

  1. CNKI / WanFang Chinese-language deep dive for 补体抑制剂 + 黄酮 / 多糖 / 木脂素 (complement inhibitor + flavonoid / polysaccharide / lignan) — verify there is no large untapped Chinese-only corpus. Spot-check via 2-model translation per CLAUDE.md.
  2. J-STAGE Kampo formulary search for 補体 + 漢方医学 (complement + Kampo). Sho-saiko-to and Juzen-taiho-to (already known TCM/Kampo formulas) may have direct anti-complement work in Japanese-only literature.
  3. Helicteres benzofuran lignan replication — Yin 2016 is single-paper; rerun the CH50/AP50 with independent compound stocks before further work.
  4. Rosmarinic acid + MSU crystal surface assay — operative gap: does the 34 μM C3b-covalent IC50 translate to MSU-crystal-induced C5a generation suppression at gut-luminal-tractable doses? Concrete Phase 2 in vitro test.
  5. Multi-level comp-021 candidate: compound × upstream-complement chokepoint × matched-assay-format mapping. The 44× rosmarinic-acid IC50 spread (34 → 1500 μM) is the canonical example of why future scans need format-stratified IC50 reporting.

N.4 What this re-run does NOT do (scope discipline)

  • Does NOT triage compounds by oral bioavailability / gut-luminal tractability (that's the comp-004/comp-013 framework, future Phase 2 follow-up via comp-NNN extension)
  • Does NOT propose engineering decisions (e.g., "express compound X biosynthetic pathway in koji") — that's downstream of mechanism confirmation
  • Does NOT replace the comp-012/H05 engineering thread — protein-engineering CD55 SCR1-4 is a separate exploration vector, complementary to natural-product work catalogued here