CYP1A1/CYP1B1, estrogen metabolism, and differential risk: why 2-OH versus 16-OH is not a “detoxification” axis

Project: L6-3
Stage: REPORT_EN
Evidence cutoff: 2026-07-20
Classification: Hormonal nutrigenomics · estrogen metabolism · genotoxicity · women's health
Status: complete English scientific twin; pending bilingual QA
Overall maturity: H1 for the proximal catechol hypothesis; H0/H1 for its causal competitor; H3 only for the replicated observational association between parent estrogens and postmenopausal breast/endometrial cancer
Ethical scope: mechanistic and translational research. This report does not prescribe, diagnose, recommend changing enzymes or metabolites, or use current user data.

Scientific delta of the report

The central conclusion is not that one estrogen branch is “good” and another “bad.” The project replaces the 2-OH:16α-OH ratio with a causal chain containing two separable processes: possible local initiation by CYP1B1→4-OH/3,4-quinone flux that matters only if it escapes neutralization and repair, and tissue promotion/selection dominated by parent E1/E2 load and ER signaling. The new evidentiary requirement is stricter than the historical literature: a change in metabolites or adducts does not validate initiation unless orthogonal perturbation and rescue produce concordant changes in fixed mutations and clonal fate.

This criterion integrates and corrects the accumulated evidence. It explains why human CYP1B1 can biochemically favor 4-hydroxylation without a urinary ratio identifying its activity; why 4-OH forms acute adducts but was insufficient for mammary tumors in ACI rats under the tested regimen; why parent E1/E2 show more consistent prospective human associations than 2-OH, 16-OH, or 2:16; and why excretion can reverse the sign of an association. The connection to longevity remains limited to an untested multi-organ hypothesis: there is no direct evidence that modifying these pathways increases female survival or healthy life-years.

1. Executive scientific abstract

CYP1A1 and CYP1B1 catalyze phase-I reactions on estrone and estradiol, but “phase I” does not mean detoxification. CYP1A1 usually favors 2-hydroxylation; isolated human CYP1B1 favors 4-hydroxylation over 2-hydroxylation by approximately fivefold under recombinant conditions. Other CYPs, especially CYP1A2 and CYP3A4/3A7, participate in the 2, 4, and 16 branches. Therefore, neither expression of one gene nor the 2-OHE1/16α-OHE1 ratio describes the metabolic flux of a woman or tissue.

The 4-OH branch has the clearest differential chemical potential. 4-OH-E1/E2 can oxidize to semiquinones and 3,4-quinones, form depurinating adducts, apurinic sites, and oxidative damage. COMT, NQO1, GST/GSH, SULT/UGT, export, and repair compete with that fate. The causal variable would not be total 4-OH, but the fraction of intracellular flux that remains reactive long enough to damage DNA, escape repair, and become fixed during replication. Evidence is strong for enzyme regioselectivity, moderate for chemistry and acute lesions, and weak for persistent mutation or human disease.

The 2-OH branch is not intrinsically safe. Its catechols are often methylated and 2,3-quinone is less depurinating than 3,4-quinone in isolated DNA, but it can also participate in redox cycling and ROS generation. The 16α-OH branch retains estrogenic activity and can sustain proliferation in some models, but it was insufficient for mammary tumors in ACI rats and its human direction changes by organ and method. A small cross-sectional bone study even found a direction incompatible with “universally harmful 16α-OH.” None of these observations demonstrates bone benefit or systemic harm.

Human epidemiology supports a cautious interpretation. In PLCO, unconjugated E2 was associated with postmenopausal breast-cancer risk; metabolic ratios were exploratory, expressed as P90-versus-P10 contrasts, and not multiplicity-corrected. In B~FIT, HR 0.60 referred to the sum of the entire 2 pathway divided by the sum of the entire 16 pathway; the classical 2-OHE1/16α-OHE1 ratio was null. In NHS 2025, the 2 pathway was positive even after E2 adjustment, while 87–99% of 4-pathway analytes were below the limit of detection. The prediagnostic urinary adduct proxy was null in WHI for large effects. In endometrium, parent E2 produced much larger associations and the sign of 2:16 did not reproduce the breast pattern.

Recent data do not rescue the binary narrative. Among 449 ER−/PR− cases and 449 controls in a pregnancy cohort, total estrogens were modestly associated with later risk (OR 1.16 per doubling; CI95% 1.02–1.32), while the 2:16 pathway ratio was null (OR 0.99; 0.92–1.06). Pregnancy physiology includes fetal–placental production and does not identify maternal CYP activity. In normal human breast explants from 23 donors, 10 nM E2 for four days increased R-loops but did not significantly increase γH2AX, and proliferation only showed a trend. A preprint in four immortalized epithelial lines with inducible ESR1 found more γH2AX in three; this contradiction separates ER-mediated promotion from direct ER-mediated initiation.

The public computational analysis was also negative for the most ambitious extension. MCF-7 4OHEE2-damage peaks were not enriched in normal luminal ATAC chromatin: 2,739 observed versus 3,810.7 expected, fold 0.719, CI95% 0.697–0.740. The apparent older:younger difference lay within its conditioned null (two-sided p=0.5765). This prevents prioritizing loci or age using unpaired atlases, but does not show that open chromatin is protective: damage and accessibility came from different systems.

The proposed decisive test uses non-malignant human breast tissue, stable-isotope E2, orthogonal CYP1B1 and COMT perturbations, rescue, temporally restricted ER degradation, and duplex sequencing. The isotope can attribute metabolites and adducts to substrate, but not a final nucleotide substitution. Causality will be triangulated through labeled adduct, perturbation/rescue, mutation change, and clonal fate. If only products or adducts change, the catechol hypothesis fails as a material explanation of initiation.

2. Scientific question and relevance

Primary question

In non-malignant estrogen-sensitive female epithelium, does effective CYP1B1 versus CYP1A1 activity modify lesion, mutation, and clonal expansion at equal intracellular E1/E2 load, and does this effect depend more on complete 2/4/16 flux and its neutralization than on the 2-OH:16α-OH ratio?

Health and longevity question

If a reproducible proximal signal exists, does it cross the links of persistent mutation, tissue selection, incident disease, and multi-organ balance strongly enough to alter healthy life-years? Current evidence cannot answer yes. “Lower risk of one cancer” is not equivalent to longevity because bone, vasculature, brain, endometrium, and breast may respond differently to the same hormonal load.

Relevance

The problem matters because “estrogen detoxification” language compresses a tissue network into a urinary ratio. That simplification can confuse:

1. metabolite formation with elimination;
2. systemic concentration with tissue dose;
3. enzyme activity with genotype or expression;
4. a chemical adduct with a fixed mutation;
5. mutation with clonal expansion;
6. an organ-specific outcome with global health.

The scientific value of this project is to replace those equivalences with experiments capable of killing the mechanism.

3. Scope, population, and life stage

Primary biological population

The primary system is histologically normal human breast epithelium obtained during benign surgery, with a documented ERα+ luminal lineage. Non-malignant human endometrium is the mandatory organ replication if the mechanism passes its first gate. Tumor tissue and immortalized lines are system controls, not substitutes for normal physiology.

The relevant future human population spans approximately 35–65 years and must classify reproductive stage explicitly, not by age alone. Late reproductive life, menopausal transition, and early postmenopause change substrate source, amplitude, and timing. Pregnancy remains separate because of the fetal–placental unit; hormonal contraception, menopausal hormone therapy, ER modulators, chemotherapy, and aromatase inhibition require their own strata.

Exclusions

• No individual risk inference or personal intervention is proposed.
• Pregnancy is not used to represent non-pregnant metabolism.
• Rat or mouse findings are not extrapolated to human regioselectivity without measured products.
• An analyte is not called a biomarker before reliability, compartment validity, and incremental utility are shown.
• Menopause is not attributed to age-only groups.
• Breast findings are not generalized to endometrium, bone, vasculature, or brain.

4. Background knowledge and mechanism map

4.1 Substrate source

E1/E2 arise from ovary, peripheral and local aromatization, HSD17B interconversion, and enterohepatic recirculation. The ovarian pulse falls after menopause, but adipose tissue, breast stroma, skin, and other tissues can maintain local exposure. Adiposity and inflammation can simultaneously increase substrate, ER signaling, and redox stress; they are confounders and modifiers, not evidence of a specific CYP branch.

4.2 Phase-I partitioning

The minimum network is:

E1/E2 → CYP1A1/1A2/1B1/3A4/others → 2-OH | 4-OH | 16α-OH

Isolated human CYP1B1 favors 4-OH; CYP1A1 usually favors 2-OH; CYP1A2/3A4 contribute in liver and CYP3A4/3A7 participate in the 16 branch. The cell integrates coexpression, AhR/ER induction, substrate, and compartment. In induced MCF-7, 2- and 4-hydroxylation were approximately equal despite isolated CYP1B1 preference. A whole-body ratio cannot be derived from one enzyme.

4.3 Catechol fate and neutralization

Catechols compete among:

• O-methylation by COMT;
• quinone reduction by NQO1;
• GSH conjugation through GST;
• sulfation/glucuronidation and export;
• oxidation to semiquinone/quinone and redox cycling.

A phase-I reaction can be called “detoxification” only if the net sequence reduces reactive species or increases elimination. Formation of 2-OH or 4-OH alone does not meet that criterion.

4.4 Lesion, repair, and selection

4-OH-E2/3,4-quinone can produce depurinating N3Ade/N7Gua adducts, stable adducts, AP sites, and ROS. Those lesions can be repaired, trigger apoptosis/senescence, or become fixed during replication. ER may modify the process in two phases:

• during lesion formation, by changing metabolism, transcription, R-loops, or S-phase entry;
• during expansion, by favoring survival and growth of the injured lineage.

The integrated mechanism is not a whole-body sum but a cellular and temporal coincidence:

clonal opportunity ≈ unneutralized 4-OH flux
                     × lesion escaping repair
                     × divisions/survival
                     × permission for ER- and tissue-mediated expansion

This formulation is inferred and not yet quantitative. Its purpose is to make explicit what must be measured.

4.5 Roles of the 2 and 16 branches

The 2 branch may divert parent substrate and be efficiently methylated, but it is not chemically risk-free. The 16 branch may sustain contextual ER signaling, but its effect depends on AUC, receptor, organ, and parent exposure. Neither supports a single systemic valence.

4.6 Excretion paradox

A high urinary concentration of 4-OH or conjugates may mean:

1. high formation and efficient neutralization;
2. high formation with reactive leakage;
3. low formation and high clearance;
4. hepatic/intestinal production without a relevant breast dose.

A low concentration may reflect low formation or tissue retention. Consequently, the sign of a spot-sample association can reverse after conditioning on formation and the fraction eliminated.

5. Evidence method

All verified project stages were integrated: scoping, evidence mapping, evidence verification, mechanistic synthesis, computational decision and execution, hypothesis generation, adversarial review, and experimental design. Primary sources with DOI, PMID, or public-deposit identifiers were prioritized.

For decisive studies, design, population or system, sex/life stage, sample size, exposure/comparator, outcome, effect, uncertainty, and generalization limit were extracted. Evidence was distinguished as:

• human prospective/observational: clinical temporality without causal pathway perturbation;
• human ex vivo: normal human tissue outside the organism;
• animal: experimental causality limited by species and pharmacokinetics;
• in vitro/biochemical: molecular resolution with low clinical sufficiency;
• computational: intersection of public datasets without activity or causal inference;
• inferred: connection across different systems.

Adversarial evidence was actively sought: null ratios, null common genetics, metabolites without tumors, null prospective adducts, human explants without significant γH2AX, and negative computation. Claims from 2025–2026 were rechecked in primary sources.

6. Evidence map

Resulting hierarchy

1. Strong: biochemical regioselectivity of human CYP1B1; non-specificity of 2:16.
2. Moderate: 4-OH/quinone chemistry, neutralization, and acute lesions.
3. Moderate-high observational: parent estrogens and postmenopausal breast/endometrial cancer risk.
4. Low: tissue 4-OH flux as a cause of human disease.
5. Very low/not established: mortality, healthspan, or longevity.

7. Contradictory evidence and null findings

7.1 Enzyme versus cell

Isolated CYP1B1 favors 4-OH, but cells coexpress enzymes, regulate substrate, and export products. Biochemical preference does not predict a systemic ratio.

7.2 Adduct versus tumor

4-OH/quinone forms acute mammary adducts in rats, but isolated 4-OH-E2 produced no detectable ACI mammary tumors over 36 weeks. Pellet recovery confirmed release, not equivalent tissue AUC. Rivals are pharmacokinetics, lack of promotion, repair/death, or true lack of sufficiency.

7.3 Retrospective versus prospective

A urinary adduct ratio separated groups in a small cross-sectional study, while prediagnostic WHI was null. Existing disease, selection, and matrix can explain the discrepancy.

7.4 “Protective” 2-OH versus positive or null

PLCO and some relative-pathway analyses suggested inverse associations; Shanghai attenuated after parent adjustment; classical B~FIT was null; ELISA studies were globally null; NHS 2025 found the 2 pathway positive even after E2. Constructs, LOD, matrices, and denominators are not equivalent.

7.5 Breast versus endometrium and bone

The sign of 2:16 may be inverse in breast, positive in a small endometrial analysis, or null. A cross-sectional bone study of 59 women produced directions incompatible with “bad 16.” This does not demonstrate a causal trade-off, but is enough to reject a universal valence.

7.6 ER in a line versus an explant

The HBEC-ESR1 preprint supports ER-dependent γH2AX in immortalized lines with inducible expression. Normal human explants showed R-loops without significant γH2AX or proliferation. Model, ER expression, S phase, repair, and cellular composition are the leading explanations; neither source should erase the other.

7.7 MCF-7 chromatin versus a normal atlas

Do 2025 found a DNase relationship in the experimental MCF-7 system. The normal-ATAC intersection showed conditioned depletion rather than enrichment. Because these were not the same systems, the result does not prove protection; it only kills prioritization through unpaired atlases.

8. Multiscale mechanistic synthesis

Molecular level

CYP1B1 converts part of E1/E2 to 4-OH; the catechol may be methylated, conjugated, or oxidized. In isolated DNA, 3,4-quinone has greater depurinating potential than 2,3-quinone. The relationship is not linear: extreme damage may kill cells and reduce clonal expansion.

Cellular level

The relevant cell must combine four conditions: reactive flux, insufficient neutralization/repair, capacity to survive, and opportunity to divide. An adduct released into medium does not show persistence in DNA. γH2AX, R-loops, AP sites, and 8-oxo-dG are supporting endpoints, not route-specific identifiers.

Tissue level

Breast and endometrium contain epithelium, stroma, adipose tissue, immune cells, and hormonal gradients. Tissue averages mix lineages with different CYP, COMT, ER, and repair. Cycling endometrium may couple substrate and proliferation differently from breast; bone and vasculature depend on other estrogen functions.

Systemic level

Liver, gut, adipose tissue, kidney, SHBG, and treatment determine what reaches plasma or urine. Postmenopause changes the relative source rather than eliminating local exposure. Pregnancy creates a fetal–placental source that invalidates simple comparisons with other life stages.

Disease and longevity

The defensible chain is:

1. human CYP1B1 favors 4-OH;
2. 4-OH/quinone can cause an acute lesion;
3. neutralization modifies that lesion in vitro;
4. persistent mutation in normal tissue remains a gap;
5. E1/E2–ER promotion/selection is plausible and parent estrogens associate with disease;
6. tissue flux, incidence, mortality, and healthspan have never been measured together.

The project does not claim that the pathway determines longevity. It specifies how that claim could eventually be demonstrated or refuted.

9. Computational layer

Question

Is the MCF-7 4OHEE2-damage map enriched in accessible ATAC chromatin from normal mammary luminal epithelium, and does it differ between four younger and four older donors?

Data and method

GSE279116 was crossed with GSE158879. Of 13,863 valid damage peaks, 12,183 received ten exact controls by chromosome, length, genomic context, and GC/mappability bins. ATAC replicates were consolidated by donor; 3/4 consensuses and 10,000 null sets were used.

Result

• ATAC union: 2,739 observed versus 3,810.7 expected; fold 0.719; CI95% 0.697–0.740.
• Younger: fold 0.590.
• Older: fold 0.612.
• All eight donors showed depletion in the same direction.
• Raw older-specific:younger-specific ratio 1.378, but conditioned empirical two-sided p=0.5765.
• HMEC DNase control: fold 0.111.

Permitted interpretation

Extension of H8 to normal luminal chromatin is weakened; no age H8A is created, and atlas-driven prioritization is withdrawn.

Prohibited interpretation

The analysis does not show protection by open chromatin, a menopause effect, CYP activity, repair, tumor, or risk. Narrow intervals reflect genomic units, not independent biological lesion replicates.

10. Primary hypothesis

L6-3-AR-H1 v2 — Attributable catechol initiation and separable promotion

Falsifiable statement: at comparable intracellular E1/E2 AUC, ER signaling, and number of divisions, reducing CYP1B1 or increasing neutralization will reduce substrate-attributable estrogen adducts, a material fraction of fixed mutations, and surviving clones; ER will mainly control expansion, and interaction will count only on a prespecified scale.

Lineage: L6-3-H1/H3/H4/H9 → HG-H1 → AR-H1 v2. H8 is withdrawn as an assumed gate outside MCF-7.

Mechanism: CYP1B1 generates 4-OH/3,4-Q; COMT/NQO1/GST-GSH reduce the reactive fraction; repair or death competes with fixation; ER permits expansion.

Predictions:

1. Labeled E2 will generate labeled 4-OH and adducts that fall with CYP1B1-CRISPRi and rise with COMT-CRISPRi.
2. Guide-resistant COMT rescue will reverse adducts, mutations, and clones.
3. Labeled 4-OH bypass will restore the low-CYP1B1 phenotype, but not correct low neutralization.
4. Removing ER only during expansion will reduce clones more than initial adducts/mutation.
5. The adduct–clone relationship will be non-monotonic at high toxicity.

Evidence for: human regioselectivity; 3,4-Q chemistry; COMT/GST gate; acute animal adducts; plausible parent-estrogen promotion.

Evidence against: null chronic ACI result; null urinary WHI result; null common genetics; high-dose/model limitations; no persistent mutation in normal tissue; non-transportable ATAC computation.

Kill criteria:

• flux and adducts change, but mutation and clones remain equivalent;
• rescue or a second guide does not reproduce the effect;
• the effect disappears after equalizing AUC, S phase, divisions, or viability;
• the signal appears only in MCF-7, with a derivatized probe, or at an unanchorable exposure;
• the spectrum is compatible with nonspecific ROS/cytotoxicity and lacks coherence with the labeled adduct.

Status: weakened but prioritized.
Maturity: H1.
Confidence: 0.40 for proximal lesion; less than 0.25 for human disease.

11. Competing hypothesis

L6-3-AR-H2N v1 — Catechol chemistry without material mutational consequence

Falsifiable statement: at comparable E1/E2, ER, cell phase, and viability, CYP1B1/COMT will change metabolites and perhaps adducts, but not fixed mutations or clones beyond the material margin; relevant lesions will arise from ER transcription/replication or background sources.

Mechanism: E1/E2 activates ER, transcription, R-loops, and S-phase entry; defective repair may convert replication stress into lesions. Metabolites accompany the precursor without driving mutation.

Predictions:

1. CYP1B1/COMT will remain equivalent for mutation and clones despite chemical target engagement.
2. ER/S phase will explain γH2AX, R-loops, and mutations better than flux/adducts.
3. Adding the catechol axis will not improve blinded prediction across donors.
4. Repair defects will amplify ER-mediated damage without requiring more 4-OH.

Evidence for: no integrated human chain, metabolites without ACI tumors, null WHI adduct proxy, null common genetics, and the ER-damage preprint.

Evidence against: specific 3,4-Q chemistry; COMT blockade increases adducts; normal human explants showed R-loops without significant γH2AX.

Kill criteria:

• CYP1B1/COMT perturbation and rescue change mutation and clones at equal AUC/ER/divisions;
• 4-OH bypass specifically restores the lost phenotype;
• a tissue measure of unneutralized flux adds a reproducible effect beyond ER/replication.

Status: proposed.
Maturity: H0/H1.
Confidence: 0.42.

Cooperating module, not a fourth principal hypothesis: L6-3-AR-H2P

Human evidence supports E1/E2–ER as promoter/selector, not as a demonstrated initiator. It is strengthened if removing ER only during expansion reduces clones without changing the initial lesion. It may cooperate with AR-H1 or with background damage.

12. Translational hypothesis

L6-3-AR-HT1 v2 — Incremental utility without endpoint leakage

Falsifiable statement: only if AR-H1 demonstrates causality and tissue–system concordance exists, a small set of exclusively upstream predictors of parent load and flux/neutralization will improve out-of-sample prediction of an independent mutation/lesion endpoint beyond parent estrogens alone; 2:16 will be insufficient.

Architecture:

• axis A: free E1/E2, local source, ER, and stage;
• axis B: 4-OH formation, neutralization, and export, excluding the endpoint itself;
• comparators: parent estrogens; classical 2:16; summed-pathway 2:16; systemic panel.

Predictions:

1. Analytes will pass blinded CV, detectability, stability, and ICC gates.
2. Tissue, plasma, and urine will reach minimum concordance before modeling.
3. The two-axis model will reduce error and preserve external calibration.
4. Gain will not depend on eGFR, batch, hydration, BMI, stage, or treatment.

Evidence for: parent load and neutralization represent distinct dimensions; 2:16 is unstable and non-specific.

Evidence against: no panel has external validation; the 4 pathway is often below LOD; tissue is difficult to obtain; overfitting and endpoint leakage are major risks.

Kill criteria:

• analytic failure or detectability below the gate;
• insufficient tissue–system concordance;
• no external improvement over parent estrogens;
• gain explained by kidney function, batch, hydration, BMI, stage, or treatment;
• unstable variable selection across centers.

Status: weakened and subordinate.
Maturity: H0.
Confidence: 0.25.
Governance: HUMAN_QA_REQUIRED.

13. Falsifiable predictions and kill criteria

The most important criterion is conjunctive: mutation and clonal fate must change. A positive secondary endpoint does not rescue failure at this gate.

14. Discriminating experiment

14.1 Model and unit

Low-passage organoids derived from benign reduction mammoplasty, with a documented ERα+ luminal lineage. Donor—not well or organoid—is the biological unit. MCF-7/MCF-10F are controls. A surviving proximal endpoint must replicate in matched fresh explant and then in non-malignant endometrium before generalization.

14.2 D0 — Analytic qualification

• Stable-isotope E2; confirm absence of a material isotope effect.
• LC-HRMS/MS standards for parents, 2/4/16, methoxy products, and conjugates.
• GSH and NAC traps for quinone.
• Stable DNA adducts and depurinating effluent adducts measured separately.
• DuplexSeq/NanoSeq with ENU positive control.

Gates: within-assay CV ≤15%, inter-day CV ≤20%, recovery 80–120%, detectability ≥80% for 4-OH/product/adduct, mass balance ≥80%, and functional mutagen control.

14.3 D1 — Model qualification

Four donors and a 0.1, 1, and 10 nM labeled-E2 ladder. Select the lowest measurable exposure with viability ≥90%, apoptosis not elevated by more than 10 percentage points, anchored AUC or explicit designation as experimental, and at least two verified divisions. ERα, PGR/GREB1, CYP1B1/COMT activity, composition, and passage must remain preserved.

14.4 D2 — Pilot

Six evaluable donors in a within-donor design:

• CYP1B1 CRISPRi with two guides;
• COMT CRISPRi and guide-resistant rescue;
• non-targeting guide and mock;
• labeled 4-OH bypass as a downstream test.

The pilot estimates variance, duplex coverage, and attrition. It does not confirm the hypothesis.

14.5 D3 — Confirmation

Provisional plan: 12 evaluable donors, acquiring up to 15. Twelve provides approximately 80% power for a large within-donor effect near dz 0.9; the final size is frozen from the pilot.

Primary endpoint 1: SNV/indel per informative duplex base, adjusted for baseline DNA and measured before/after expansion.
Primary endpoint 2: clone number and size normalized to viable cells and divisions.

Support: labeled flux, adducts, AP sites, 8-oxo-dG, γH2AX, comet assay, R-loops with RNase H, S phase, repair, apoptosis, and senescence.

AR-H1 success: mutation reduction ≥20% (RR≤0.80), CI95% excluding 1, concordant clonal direction, correct rescue/bypass, and the same sign in at least 9/12 donors.
AR-H2N equivalence: CI90% for mutation and clone entirely within RR 0.80–1.25 despite validated flux/adduct change.

14.6 D3b — Temporal separation of ER

ER always active; ER removed only during lesion formation; ER removed only during expansion; repeat with low CYP1B1. Fewer clones after ER removal in expansion without changing initial lesion strengthens promotion AR-H2P. Fewer mutations after ER removal during lesion formation with low CYP1B1 strengthens ER initiation within AR-H2N.

14.7 D4 — Mass balance

Perfused tissue/organoid from six donors at two effluent flow rates. Measure input, tissue content, 4-OH formation, neutralization efficiency, exported fraction, and adducts. Reversal of the effluent–adduct sign after conditioning on formation strengthens the elimination hypothesis.

14.8 D5 — Human bridge

Pilot of 30 benign-surgery participants with normal tissue, plasma, and three first-morning urines. CV≤15–20%, detectability ≥80%, CCC≥0.60 as the target and CCC<0.40 as clear failure. This pilot estimates neither risk nor subgroups.

14.9 Stopping rule

Do not proceed to incidence, mortality, or intervention cohorts if D3 does not demonstrate mutation and clone, or if D5 does not show concordance. A molecular signal cannot fill the absence of longevity evidence.

15. Biomarkers and stratification

Measurement candidates, not validated biomarkers

2:16 is retained as a negative historical comparator. It does not identify CYP1A1:CYP1B1, loses absolute amounts, omits 4-OH, and amplifies numerator/denominator error.

Future scientific stratification

• explicit reproductive stage and time since final menstrual period;
• cycle phase and time of day;
• exogenous hormonal exposure;
• parity;
• BMI/adiposity and inflammation;
• kidney/liver function;
• smoking/combustion and AhR activation;
• ancestry and geographic context;
• tissue and lineage.

No isolated SNP should classify a “fast/slow metabolizer” without activity or flux. rs4680 may modify COMT, but it does not replace tissue neutralization and does not prove that catechol flux moves the systemic SAM/SAH pool.

16. Individual variability

Variability is not accessory noise. It may arise from:

1. E1/E2 source and pulsatility;
2. cellular composition and ER;
3. inducible CYP activity, not genotype alone;
4. COMT/NQO1/GST-GSH and redox availability;
5. BER/HR/NHEJ repair and number of divisions;
6. adipose tissue, stroma, and inflammation;
7. AhR exposure, smoking, and drugs;
8. liver, microbiome, transit, kidney, and hydration;
9. life stage, pregnancy, and the fetal–placental unit;
10. matrix, batch, LOD, and analytic hydrolysis.

The human explant study showed donor heterogeneity and possible parity differences. The first confirmatory experiment will not be powered for interactions by stage, ancestry, parity, or BMI. These variables will be documented and balanced, not converted into post hoc subgroups.

17. Pharma relevance and maturity

Research opportunities

The initial opportunity is not a drug or the 2:16 ratio. It is a causal pharmacology assay that rejects candidates that change metabolites without reducing mutation/clone. Cytotoxicity with fewer clones is not equivalent to protection.

Maturity

• CYP1B1/COMT target biology: H1;
• causality in normal tissue: unproven;
• 2:16 biomarker: rejected as specific;
• two-axis panel: H0;
• chemical matter and selectivity: not evaluated;
• clinical indication: undefined;
• partnering: HUMAN_QA_REQUIRED and ineligible before tissue replication, multi-organ selectivity, and concordance.

18. Limitations

1. Catechol/adduct chemistry is concentrated in a small number of groups and non-physiological models.
2. The 4 pathway is often below LOD; an analytical null is not a biological null.
3. No specific 4-OH-E2 mutational signature is validated in normal human breast.
4. Isotope tracing attributes metabolites/adducts, not the final substitution.
5. MCF-7, MCF-10F, and immortalized HBECs alter ER, metabolism, repair, and survival.
6. Organoids lose circulation, immunity, and part of the stroma; explant replication is required.
7. ACI rats do not reproduce human regioselectivity and did not verify equivalent tissue AUC for metabolites.
8. Adjusting for E2 may control confounding, block a precursor, or induce collinearity; every human analysis needs a DAG.
9. Epidemiology centers on postmenopause and cancer; transition, LATAM, cognition, frailty, and multimorbidity are underrepresented.
10. The pregnancy cohort studies an extreme physiology and an ER−/PR− subtype; it does not identify a maternal CYP phenotype.
11. The computational analysis crosses unpaired experiments and its intervals are not biological replicates.
12. The mortality study is postdiagnostic, single-sample, chemotherapy-dependent, and ELISA-based.
13. The small bone result is cross-sectional and does not demonstrate benefit.
14. There is no direct evidence of causal healthspan or longevity.
15. No current result demonstrates that intervening on CYP1B1, COMT, ER, or metabolites produces net benefit.

19. Conclusions

1. Human CYP1B1 favors 4-hydroxylation, but person- or tissue-level activity cannot be inferred from 2:16.
2. 2-OH is not synonymous with detoxification, and 16α-OH is not a universal systemic toxin.
3. The 4-OH/3,4-quinone branch is the most specific genotoxic candidate, conditioned on neutralization, repair, survival, and replication.
4. Parent E1/E2 and ER/tissue context dominate the human epidemiologic signal, probably as exposure and promotion; ER has not been shown to be the main molecular initiator.
5. The classical 2-OHE1/16α-OHE1 ratio is unstable, non-specific, and methodologically heterogeneous; the inverse B~FIT HR did not refer to this ratio.
6. Adversarial evidence matters: metabolites without ACI tumors, null WHI adducts, null common genetics, explants without significant γH2AX, and non-transportable ATAC computation prevent closure of the mechanism.
7. The primary hypothesis survives only under a high standard: it must connect target engagement, attributable adduct, fixed mutation, and clone through perturbation and rescue.
8. Excretion can reverse the sign, so a spot urine is not a tissue dose.
9. Translation dies if tissue and system do not agree, even if ex vivo chemistry is causal.
10. There is no direct evidence of healthy longevity. The only honest formulation is a future multi-organ test after validating mechanism, compartment, and temporality.

The answer to the project question is therefore conditional. CYP1A1/CYP1B1 connect estrogen metabolism to women's health through substrate partitioning, quinone chemistry, repair, and tissue promotion—not through a simple 2-OH/16-OH axis. The new hypotheses are falsifiable because they can yield a verdict of causal catechol initiation, secondary catechol chemistry, or separable ER promotion. The correct experiment does not seek to confirm “detoxification”; it tests whether the chemistry actually reaches mutation and clonal fate.

20. References

1. Hayes CL, Spink DC, Spink BC, Cao JQ, Walker NJ, Sutter TR. CYP1B1-mediated hydroxylation of estradiol. PNAS. 1996;93:9776–9781. PMID 8790407. DOI: 10.1073/pnas.93.18.9776.
2. Spink DC et al. Differential expression of CYP1A1/CYP1B1 and estrogen metabolism in mammary cell lines. Carcinogenesis. 1998;19:291–298. PMID 9498279. DOI: 10.1093/carcin/19.2.291.
3. Dawling S, Roodi N, Mernaugh RL, Wang X, Parl FF. Catechol estrogen metabolism by CYP1A1/1B1, COMT and GSTP1. Chemical Research in Toxicology. 2004. PMID 15377160. DOI: 10.1021/tx0498657.
4. Nishida CR et al. Species differences and residue 395 in CYP1B1 regioselectivity. Molecular Pharmacology. 2013. PMID 23821647. DOI: 10.1124/mol.113.087700.
5. Lee AJ et al. Characterization of the oxidative metabolites of estradiol and estrone formed by 15 human CYP enzymes. Cancer Research. 2003. PMID 12865317. PubMed.
6. Yamazaki H et al. Roles of CYP1A2 and CYP3A4 in estrogen hydroxylation by human liver microsomes. Chemical Research in Toxicology. 1998. PMID 9625734. DOI: 10.1021/tx970217f.
7. Zahid M et al. COMT inhibition increases depurinating estrogen-DNA adducts in MCF-10F cells. Free Radical Biology and Medicine. 2007. PMID 17964424. DOI: 10.1016/j.freeradbiomed.2007.08.005.
8. Zahid M et al. Greater reactivity of estradiol-3,4-quinone versus estradiol-2,3-quinone with DNA. Chemical Research in Toxicology. 2006. PMID 16411670. PubMed.
9. Wang M et al. Oxidative DNA damage induced by estrogen metabolites. Carcinogenesis. 2003. PMID 12807746. DOI: 10.1093/carcin/bgg049.
10. Li KM et al. Metabolism and DNA adduct formation of 4-hydroxyestradiol in the mammary gland of ACI rats. Carcinogenesis. 2004. PMID 14578156. DOI: 10.1093/carcin/bgg191.
11. Turan VK et al. Lack of mammary carcinogenicity of isolated estradiol metabolites in ACI rats under the tested regimen. Journal of Endocrinology. 2004. PMID 15525577. DOI: 10.1677/joe.1.05802.
12. Liehr JG et al. Carcinogenicity of estradiol and metabolites in Syrian hamsters. Journal of Steroid Biochemistry. 1986. PMID 3009986. DOI: 10.1016/0022-4731(86)90080-4.
13. Swaneck GE, Fishman J. Covalent binding of 16α-hydroxyestrone to estrogen receptor. PNAS. 1988. PMID 3186693. DOI: 10.1073/pnas.85.21.7831.
14. Telang NT et al. Proliferative and transforming effects of 16α-hydroxyestrone in mammary cells. Journal of the National Cancer Institute. 1992. PMID 1556774. DOI: 10.1093/jnci/84.8.634.
15. Fuhrman BJ et al. Estrogen metabolism and postmenopausal breast cancer risk in PLCO. JNCI. 2012. PMID 22232133. PMCID PMC3283536.
16. Dallal CM et al. Estrogen metabolism and breast cancer risk in B~FIT. Carcinogenesis. 2014. PMID 24213602. PMCID PMC3908751.
17. Sampson JN et al. Association of estrogen metabolism with breast cancer risk in four prospective cohorts. Cancer Research. 2017. PMID 28011624. DOI: 10.1158/0008-5472.CAN-16-1717.
18. Moore SC et al. Urinary estrogen metabolites and breast cancer in the Shanghai Women’s Health Study. JNCI. 2016. PMID 27193440. DOI: 10.1093/jnci/djw103.
19. Arslan AA et al. Circulating estrogen metabolites and postmenopausal breast cancer. 2014. PMID 24769889. PubMed.
20. Dallal CM et al. Urinary 2-hydroxyestrone, 16α-hydroxyestrone and breast cancer across prospective studies. 2012. PMID 22865302. DOI: 10.5301/JBM.2012.9353.
21. Brantley KD et al. Circulating estrogens and postmenopausal breast cancer in the Nurses’ Health Study. Cancer Epidemiology, Biomarkers & Prevention. 2025. PMID 39190182. DOI: 10.1158/1055-9965.EPI-24-0577.
22. Reding KW et al. Prediagnostic urinary estrogen-DNA adducts and breast cancer in WHI. Cancer Epidemiology, Biomarkers & Prevention. 2020. PMID 32699078. DOI: 10.1158/1055-9965.EPI-20-0133.
23. Dallal CM et al. Estrogen metabolism and endometrial cancer in B~FIT. 2016. PMID 26728471. PMCID PMC8900527.
24. Brinton LA et al. Serum estrogens and endometrial cancer risk in WHI. 2016. PMID 27197275. PMCID PMC4930692.
25. Faupel-Badger JM et al. Comparison of immunoassay and LC-MS/MS for urinary estrogen metabolites. 2010. PMID 20056650. PMCID PMC2836837.
26. Rinaldi S et al. Long-term reproducibility of urinary estrogen metabolites in postmenopausal women. 2003. PMID 12889688. DOI: 10.1023/A:1024209122412.
27. Taioli E et al. Estrogen metabolites in breast tumor tissue and urine. 2010. PMID 20678202. DOI: 10.1186/1477-7827-8-93.
28. Wang T et al. Urinary estrogen metabolites and long-term mortality following breast cancer. JNCI Cancer Spectrum. 2020. PMID 32455334. DOI: 10.1093/jncics/pkaa014.
29. Kaur-Knudsen D et al. CYP1B1 genotype and cardiovascular, pulmonary and cancer risk. Pharmacogenetics and Genomics. 2009. DOI: 10.1097/FPC.0b013e32833042cb.
30. Dunphy KA et al. Inter-individual variation in response to estrogen in human breast explants. Journal of Mammary Gland Biology and Neoplasia. 2020; n=23. PMID 32152951. DOI: 10.1007/s10911-020-09446-3.
31. Troisi R et al. Estrogen metabolism pathways in pregnancy and subsequent breast cancer risk: a prospective follow-up study. Breast Cancer Research. 2026; 449 cases/449 controls. PMID 41546058. DOI: 10.1186/s13058-025-02204-5.
32. Majhi PD et al. Inducible estrogen receptor alpha in normal breast epithelial cells demonstrates estrogen receptor-dependent DNA damage. bioRxiv preprint, 2025. DOI: 10.1101/2025.03.13.643100.
33. Do Q-T et al. Genome-wide mapping and quantification of DNA damage induced by catechol estrogens using Click-Probe-Seq and LC-MS2. Communications Biology. 2025;8:357. PMID 40069327. DOI: 10.1038/s42003-025-07657-0. GEO GSE279116.
34. Senapati P et al. Loss of epigenetic suppression of retrotransposons with oncogenic potential in aging mammary luminal epithelial cells. Genome Research. 2023;33:1229–1241. PMID 37463750. DOI: 10.1101/gr.277511.122. GEO GSE158879.
35. Rosenbluth JM et al. Organoids from normal and cancer-prone human breast preserve epithelial lineages. Nature Communications. 2020. PMID 32249764. DOI: 10.1038/s41467-020-15548-7.
36. Meng P et al. Propagation of functional ER-positive normal human mammary epithelial cells in 3D culture. Breast Cancer Research and Treatment. 2019. DOI: 10.1007/s10549-019-05229-5.
37. Chabi K, Sleno L. Stable-isotope estrone/estradiol metabolism and GSH/NAC trapping by LC-HRMS/MS. Metabolites. 2022. PMCID PMC9611524. DOI: 10.3390/metabo12100931.
38. Zhang Q, Gross ML. Synthesis and tandem MS of estrogen-modified DNA bases with stable standards. Chemical Research in Toxicology. 2008. DOI: 10.1021/tx800067s.
39. Cho E et al. Duplex sequencing for direct mutagenicity assessment in TK6 cells across two laboratories. Mutation Research/Genetic Toxicology and Environmental Mutagenesis. 2023. PMID 37491114. DOI: 10.1016/j.mrgentox.2023.503649.
40. Kucab JE et al. Mutational signatures of environmental carcinogens in human tissue organoids revealed by duplex sequencing. Cell Reports. 2026. DOI: 10.1016/j.celrep.2026.117406.
41. Kostecka A et al. Duplex sequencing of subclonal variants in non-tumoral breast tissue. npj Breast Cancer. 2022. DOI: 10.1038/s41523-022-00443-9.
42. Lim SK et al. Altered hydroxylation of estrogen in patients with postmenopausal osteopenia. Journal of Clinical Endocrinology & Metabolism. 1997;82:1001–1006. PMID 9100564. DOI: 10.1210/jc.82.4.1001.
43. Wen W et al. Cytochrome P450 1B1 and catechol-O-methyltransferase genetic polymorphisms and breast cancer risk in Chinese women: results from the Shanghai Breast Cancer Study and a meta-analysis. Cancer Epidemiology, Biomarkers & Prevention. 2005;14:329–335. PMID 15734954. DOI: 10.1158/1055-9965.EPI-04-0392.
44. Rylander-Rudqvist T et al. Cytochrome P450 1B1 gene polymorphisms and postmenopausal endometrial cancer risk. Cancer Epidemiology, Biomarkers & Prevention. 2004;13:1515–1520. PMID 15342454. DOI: 10.1158/1055-9965.1515.13.9.

Integrity note

This report synthesizes verified evidence while preserving null results, contradictions, and transportability limits. It was not written to the promoted reports directory. The artifact must pass bilingual QA before promotion.