GSTP1 and xenoestrogen detoxification: variants, electrophilic burden, and women's health

Stage: REPORT_EN
Project: L6-6
Research line: L6 — Hormonal nutrigenomics
Evidence cutoff: 2026-07-26
Status: complete scientific report, pending BILINGUAL_QA
Overall maturity: H0–H1; no hypothesis is ready for human translation

1. Executive scientific abstract

The proposition that a GSTP1 variant makes a woman a “poor detoxifier” of xenoestrogens and predisposes her to “estrogen dominance” does not survive mechanistic audit. GSTP1 rs1695 (Ile105Val) has no universal functional direction: Ile105 is more efficient toward the model substrate CDNB, whereas Val105 can conjugate several bulky polycyclic aromatic hydrocarbon diol epoxides more efficiently and reduce their DNA adducts. An allele acquires meaning only after the specific electrophilic species has been identified, active protein and GSH have been quantified, and the fraction of flux controlled by the enzyme has been demonstrated.

For bisphenol A (BPA), the xenoestrogen used as the main case, human evidence favors rapid glucuronidation and sulfation of the parent compound. In a study of 14 people—8 women and 6 men—less than 1% of total circulating BPA remained unconjugated after an acute oral dose. In human microsomes, UGT2B15 contributed more than 80% of hepatic glucuronidation below 5 µM. BPA can nevertheless undergo oxidative bioactivation: chemical studies identified BPA-3,4-quinone (BPAQ), its GSH and nucleotide adducts, and a 2026 study detected thiol-reactive metabolites of BPA and eight analogues in human liver microsomes. This demonstrates the possibility of bioactivation, not its magnitude in female tissue under usual exposure or any contribution from GSTP1.

The central difficulty is kinetic. BPAQ reacts rapidly, spontaneously, and reversibly with GSH; the reported bimolecular rate constant at 19.4 °C was 1547 M⁻¹s⁻¹. Observing BPAQ–GSH therefore does not identify GSTP1. Moreover, forming more conjugate is not equivalent to detoxification: if BPAQ–GSH reverts, only export or a terminal trap operating faster than that reversion can produce a net reduction in electrophile AUC and adducts. Even efflux can be misleading: it may protect the source cell while transferring reactive burden to another compartment.

The synthesis proposes three causal gates:

1. G1 — formation: a material amount of electrophile is produced from the parent compound;
2. G2 — incremental catalysis: GSTP1 accelerates capture above GSH alone and reduces competition with target nucleophiles;
3. G3 — fate: the conjugate is exported or terminally trapped before reversion and without transferring damage.

In parallel, a non-catalytic branch exists: GSTP1 binds JNK and TRAF2–ASK1 and can modify stress signaling without altering ligand clearance. GSTP1 could therefore change adducts, repair, apoptosis, or senescence without changing E1, E2, progesterone, or ER bioactivity. Such a result would refute “estrogen dominance” while preserving an electrophilic vulnerability.

Human evidence is proximal but not causal. A cohort of 300 women undergoing IVF found no significant associations between rs1695/rs1138272 and BPA in blood or follicular fluid; blood BPA also did not correlate with follicular BPA. The association between rs1695 and endometriosis is meta-analytically inconsistent: a 2013 synthesis was globally null, whereas a 2020 review estimated modest protection, OR 0.80 (95% CI 0.69–0.92). In breast cancer, population signals attenuate when analyses are restricted to higher-quality studies. None of these studies measured BPAQ, GSTP1 flux, efflux, specific adducts, and ER signaling in the same causal chain.

The favored hypothesis is that, for BPA, low BPAQ formation and UGT/SULT pathways dominate the balance; second, that spontaneous GSH chemistry dominates BPAQ capture. The lower-confidence primary hypothesis holds that GSTP1 can add useful catalysis only when coupled to a transporter that converts the reversible conjugate into a sink. The smallest discriminating experiment is D0a: stopped-flow with BPAQ, GSH, and GSTP1 *A/*B/*C, simultaneously measuring rate, conjugate, reversion, and a competing adduct. If material catalysis and protection are absent, the BPAQ–GSTP1 branch ends.

There is currently no direct human bridge from GSTP1 and xenoestrogens to women's healthspan or longevity. Distal relevance is plausible only as an unproven sequence—exposure, persistent tissue injury, organ-specific disease, and disease-free years—and does not authorize prescribing, clinical genotyping, or individual inference.

Concrete scientific delta

This report replaces the “slow detox allele” model with the falsifiable kinetic unit electrophile formation × GSTP1/GSH × conjugate fate, divides the previous null into limited formation and dominant spontaneous capture, and adds the intercompartmental shuttle alternative: reducing adducts in the source cell may not constitute whole-system detoxification if an exported reversible conjugate transfers BPAQ to a recipient.

2. Scientific question and relevance

Question

What mechanisms connect GSTP1 variants and activity to xenoestrogen disposition, women's health, and potentially longevity; which parts of this chain have been demonstrated, and what new hypotheses can be falsified?

Why it matters

GSTP1 sits at the intersection of three processes that are often conflated:

• GSH conjugation of electrophiles;
• non-catalytic regulation of stress signaling;
• genetic associations with hormone-sensitive diseases.

This conflation creates two errors. The first is converting activity toward CDNB into a general “detoxification” classification. The second is calling any outcome observed in the presence of an endocrine disruptor “estrogen dominance,” even when adducts or JNK change but estrogen signaling does not.

Relevance to women's health depends on separating four variables:

1. internal dose of the compound and reactive metabolite;
2. ERα/ERβ/GPER bioactivity;
3. electrophilic burden, redox state, and cell fate;
4. the context of E1/E2, progesterone, SHBG, and life stage.

Only then can a trajectory toward endometrial, breast, ovarian, hepatic, or systemic function be tested. Longevity is the final link, not a synonym for fewer adducts, absence of cancer, or age at menopause.

3. Scope, population, and life stage

Chemical scope

The core is frozen as BPA → BPAQ → BPAQ–GSH. Other compounds serve as comparators:

• estradiol quinones: endogenous positive control for GSTP1 conjugation;
• (+)-anti-BPDE: environmental control for allele-specificity and efflux;
• parabens and MEHP: controls for alternative pathways, not assumed GSTP1 substrates;
• BPA analogues: future compound-specific candidates, not post hoc rescues if BPAQ fails;
• parent PCB153/180: epidemiologic exposures, not GSTP1 direct substrates by definition.

Systems and population

• Biochemistry: recombinant human GSTP1*A (Ile105/Ala114), *B (Val105/Ala114), and *C (Val105/Val114).
• Metabolic source: primary hepatocytes or spheroids from female donors aged 35–65 years.
• Target tissue: non-malignant endometrial organoids; breast and ovary remain later extensions.
• Life stage: late premenopause, menopausal transition, and early postmenopause, ideally classified using STRAW+10.

Strata that must not be pooled

Pregnancy, lactation, puberty, prenatal exposure, IVF with ovarian stimulation, surgical menopause, premature ovarian insufficiency, and exogenous hormone use are biologically distinct contexts. IVF evidence can inform compartmentalization, but it does not represent women in the general population or the menopausal transition.

Exclusions

• active cancer or tumor cell lines as the sole causal foundation;
• spot urine or blood as substitutes for tissue;
• symptoms as proxies for GSTP1 activity;
• individual recommendations, prescribing, supplements, or clinical genotyping;
• current Lua data, private data, or private genomes.

4. Background knowledge and mechanism map

4.1 What GSTP1 does

GSTP1 facilitates nucleophilic attack by GSH on electrophiles. rs1695 modifies the hydrophobic H-site pocket, and rs1138272 contributes to haplotype geometry. Observed function can be represented as:

GSTP1 flux =
active protein × substrate-specific activity × available GSH
× network competition × conjugate fate

In normal human lung, GSTP1 abundance varied approximately sevenfold independently of genotype. An allelic difference in specific activity may therefore be smaller than variation in active protein, GSH, oxidation, or tissue regulation.

4.2 Three-gate, two-branch map

BPA EXPOSURE
  │
  ├─ UGT/SULT ─────────────> BPA-glucuronide/sulfate ──> elimination
  │                           predominant human route
  │
  └─ CYP/peroxidase ───────> BPAQ
                              G1: is a material amount formed?
                              │
                              ├─ DNA/protein adducts
                              ├─ reduction/other pathways
                              └─ + GSH
                                   ├─ spontaneous reaction
                                   └─ GSTP1 *A/*B/*C
                                        G2: added catalysis and protection?
                                        │
                                        ▼
                                    reversible BPAQ–SG
                                        │
                            ABCC/terminal trap
                            G3: elimination or transfer?
                                        │
                        ┌───────────────┴────────────────┐
                        ▼                                ▼
                 PROTECTIVE SINK                    SHUTTLE
                 total AUC/adduct ↓                 donor ↓,
                                                    recipient ↑

NON-CATALYTIC BRANCH
monomeric GSTP1 ─| JNK and TRAF2–ASK1
stress/modification/oligomerization
  └─> pJNK/pASK1 ─> repair, adaptation, apoptosis, or senescence

CONDITIONAL CONVERGENCE
electrophile + stress signaling + ER/hormonal context
  └─> tissue function ─> organ-specific disease
       └─> healthspan/longevity [not demonstrated]

4.3 Correcting “estrogen dominance”

The term would be operational only if a GSTP1 × exposure interaction attributable to the mechanism changes ligand concentration or ER signaling within a defined hormonal context. It is refuted if GSTP1 changes adducts or JNK while E1/E2, progesterone, and ER remain equivalent.

5. Evidence method

This report integrated scoping.md, evidence-map.md, evidence-verification.md, mechanism.md, compute-decision.md, hypotheses.md, adversarial-review.md, and experiment-design.md, together with the cumulative lifetime-memory packet. Decisive sources were verified by PMID, DOI, or primary text.

Evidence hierarchy

• Human: pharmacokinetic, cohort, case–control, and participant-sample evidence.
• Human ex vivo: human microsomes, cytosols, or tissues outside the body.
• Animal: non-human in vivo or ex vivo evidence.
• Cellular: transformed, non-transformed, or transfected models.
• In vitro: purified protein or chemical systems.
• Computational: atlas, prediction, or modeling; not used as causal support.
• Inferred: a bridge constructed across separate studies.

Rules

1. A GSH conjugate does not identify GSTP1.
2. Catalysis does not demonstrate a reduction in adducts.
3. Reducing adducts in one cell does not demonstrate system-level detoxification.
4. Genotype does not equal active protein or flux.
5. Expression does not equal transport or metabolic control.
6. Association does not equal mediation.
7. An adduct does not equal mutation, clone, disease, or longevity.

Search and update

Primary sources were sought on GSTP1 variants, BPA/BPAQ, reversible conjugation, GSH-conjugate transport, human pharmacokinetics, endometriosis, breast/endometrial cancer, IVF, and evidence from 2021–2026. Recent verification retained:

• Syrkasheva et al. 2021: 300 women undergoing IVF, BPA, and phase I/II SNPs;
• He et al. 2026: bioactivation of BPA and eight analogues in human liver microsomes at 50 µM;
• Fan et al. 2026: S9 metabolism of BPA/o,p'-BPA and urinary metabolites in 103 human samples.

No new systematic review with dual screening or meta-analysis was performed. Negative searches are interpreted as localized evidence gaps, not proof of nonexistence.

6. Evidence map

Reading by evidence level

Human evidence

Human studies demonstrate the predominant disposition of parent BPA and heterogeneous clinical associations. They do not demonstrate GSTP1 control of BPAQ. The 2026 human evidence for urinary BPA-O confirms that oxidation is not nonexistent, but its mean concentration was lower than glucuronide and sulfate, and it does not identify BPAQ or the tissue of origin.

Human ex vivo evidence

Human ex vivo studies demonstrate hepatic/breast glucuronidation and microsomal bioactivation of bisphenols. Microsomes lack the full architecture of GSH, cytosol, transport, and compartments; 50 µM bisphenol and 5 mM GSH do not directly translate to usual exposure.

Animal evidence

Animal evidence supports covalent modification of defenses by BPA and JNK/UGT compensation after GstP loss. It is insufficient for common human variants or the menopausal transition.

Cellular and in-vitro evidence

This is the strongest layer for allele specificity, the BPAQ–GSH reaction, and JNK signaling, but it is also the most vulnerable to dose, overexpression, species, and absence of whole-body metabolism.

Computational evidence

No computation was run. GTEx, HECA, HPA, or BioNeMo could rank tissues or transporters after a positive D0, but cannot infer incremental catalysis, reversibility, direct transport, or net detoxification.

Inferred evidence

Every bridge from adducts/JNK to endometrial function, disease, healthspan, or longevity remains inferred.

7. Contradictory evidence and null findings

7.1 Ile105 versus Val105

The most productive contradiction is biochemical: Ile105 performs better with CDNB, whereas Val105 can perform better with certain PAH epoxides. This refutes universal “fast/slow” labels and requires measurement of each electrophile.

7.2 Conjugate as protection or reservoir

BPAQ–GSH forms rapidly but is reversible. Two explanations compete:

• conjugation immediately reduces competition with DNA/protein;
• the conjugate temporarily stores BPAQ and later releases it.

Without AUC, adducts, and mass balance, a conjugate peak supports both.

7.3 Dominant UGT versus bioactivation

Human pharmacokinetics shows extensive conjugation; 2026 human microsomes show thiol-reactive metabolites. These findings are not incompatible. The favored explanation is that the oxidative branch exists but represents a smaller or tissue-/dose-dependent fraction. The needed experiment is a balance from labeled BPA at anchored concentrations, not narrative selection of one result.

7.4 Discordant human associations

• In endometriosis, Chen et al. 2013 reported a global allelic OR 1.02 (0.97–1.07), whereas Méar et al. 2020 estimated OR 0.80 (0.69–0.92).
• In breast cancer, Egan et al. observed Val/Val OR 1.92 (1.21–3.04) in Chinese women, and Martínez-Ramírez et al. reported OR 2.39 (1.24–4.24) in 150 Mexican cases and 150 controls. Another Mexican study was null, and Miao et al. weakened the associations when restricting analysis to higher-quality studies.
• In PCB-associated endometriosis, the extreme stratified effect appeared in Ile/Ile, not Val carriers; the subset was small and intervals were very wide.

These tensions are compatible with specific chemical interactions, ancestry/LD, differential exposure, measurement error, selection bias, multiplicity, or false positives. None decides mechanism because none measured the metabolite and its flux.

7.5 Preserved null findings

• rs1695/rs1138272 were not significantly associated with blood or follicular BPA in 300 women undergoing IVF.
• GSTP1 alone did not predict PAH–DNA adducts in LIBCSP controls.
• No convincing modification by GSTP1 occurred in 718 BRCA1/2 carriers.
• A 2022 endometrial study detected no rs1695 association, but it is too small and contains reporting inconsistencies to count as a strong null.
• No direct female-longevity result is mediated by xenoestrogen–GSTP1.

8. Multiscale mechanistic synthesis

8.1 Molecular scale

The variant changes specific activity, but real capacity depends on active protein and GSH. An electrophile may also modify or inactivate GSTP1: 4-hydroxyequilenin inactivated purified GSTP1 and reduced GST activity in non-transformed MCF-10A cells. This creates a possible nonlinearity:

electrophile ↑
  → GSH demand ↑
  → modified/inactive GSTP1 ↑
  → catalytic reserve ↓

The chemistry is supported; the bisphenol threshold relevant to human tissue is not.

8.2 Network scale

For BPA, UGT/SULT compete upstream; spontaneous GSH, other GSTs, NQO1, and transporters compete downstream. GSTP1's flux-control coefficient can approach zero even if the enzyme accepts BPAQ. The network may also adapt: sustained GstP loss in mice increased JNK/AP-1, HO-1, and UGT1A6, but it is unknown whether acute loss or a common human SNP produces a similar shunt.

8.3 Cellular scale

Four outcomes must be distinguished:

1. no GSTP1 effect: UGT/formation/GSH dominate;
2. capture without protection: conjugate rises, AUC/adducts do not change;
3. intracellular protection: AUC/adducts fall in the source cell;
4. net detoxification: total donor+medium+recipient reactivity falls and a non-reactive terminal product appears.

The JNK branch must be measured through GSTP1–JNK/TRAF2 complexes, pASK1/pJNK, and separation-of-function rescues. pJNK alone is a nonspecific proxy.

8.4 Tissue scale

The liver is a source of transformation and elimination; endometrium, breast, or ovary can be recipient tissues. The IVF cohort shows that blood and follicular fluid are not interchangeable. Efficient hepatic metabolism might protect one tissue while distinct local bioactivation creates exposure in another. A universal index is therefore not proposed.

8.5 Endocrine scale

Parent BPA may cause ER effects in parallel with an electrophilic branch. The question is not whether BPA changes ER, but whether GSTP1 differentially modifies that change. A main exposure effect on ER and a GSTP1×exposure interaction on adducts can coexist.

8.6 Health and longevity scale

The required distal chain would be:

validated GSTP1 flux
→ persistent change in adducts/JNK
→ repair, senescence, mutation, or tissue function
→ organ-specific disease
→ disease-free years
→ longevity

Every arrow after proximal injury remains open for BPAQ–GSTP1 in women. Endometriosis, cancer, fertility, and all-cause mortality are not reciprocal substitutes. Longevity may integrate competing organ effects, so even reducing one risk does not automatically predict healthspan.

9. Computational layer

Decision: skip.

Public expression data can answer where GSTP1, UGT, GSH machinery, and ABCC coexist; they cannot answer whether:

• BPAQ forms materially;
• GSTP1 adds catalysis above GSH;
• a specific ABCC transports BPAQ–SG;
• the conjugate reverts or reaches a terminal fate;
• total reactivity falls;
• the mechanism affects health.

BioNeMo ESM-2 cannot convert a GSTP1 sequence into a validated BPAQ kcat/Km, and Geneformer cannot infer reversible chemistry from transcriptomes lacking exposure and metabolites. Computation reopens only after a positive D0a, to select at most two cell–transporter pairs for functional validation. No computational signal alone would raise maturity.

10. Primary hypothesis

L6-6-AR-H1, v2 — GSTP1 capture of BPAQ coupled to efflux

• Role: primary mechanistic.
• Status: weakened_narrowed.
• Maturity: H0.
• Confidence: 0.25.

Statement

For prespecified BPAQ, GSTP1 adds incremental catalysis above GSH alone and, when a directly validated transporter removes BPAQ–SG toward a terminal fate, reduces intracellular BPAQ AUC and adducts; blocking transport eliminates net protection, although it may preserve the allelic order of conjugate formation.

Evidence for

• substrate-dependent allelic inversion in GSTP1 systems;
• greater reduction of BPDE adducts by Val105 variants;
• GSTP1 conjugation of estrogen quinones;
• MRP2 efflux-dependent protection with BPDE.

Evidence against

• BPAQ–GSH is spontaneous and reversible;
• no BPAQ comparison ±GSTP1 *A/*B/*C exists;
• the BPAQ–SG transporter is unknown;
• no test exists in human female tissue;
• a positive D0 with BPAQ does not demonstrate formation from BPA.

Predictions

1. stopped-flow detects incremental catalysis and less competing adduct;
2. the *A/*B/*C order replicates with equal active protein;
3. a human transporter directly and ATP-dependently moves BPAQ–SG;
4. GSTP1×transporter reduces AUC/adducts in the source cell;
5. protection disappears when transport is blocked;
6. total mass balance demonstrates terminal fate rather than shuttle.

Kill criteria

• the upper bound of the catalytic effect is below the materiality threshold;
• the competing adduct does not fall;
• haplotypes are equivalent within 0.80–1.25;
• no transporter in the frozen panel moves BPAQ–SG;
• conjugate changes but total AUC/adduct does not;
• the effect occurs only at an unanchored concentration, with GSH depletion or cytotoxicity.

11. Competing hypothesis set

L6-6-AR-H2F, v1 — limited formation from BPA

• Role: principal competitor.
• Status: strengthened_split.
• Maturity: H1 for acute oral BPA; H0 for female tissue and repeated exposure.
• Confidence: 0.65.

Statement

With UGT/SULT intact and BPA at an internally plausible exposure, the fraction converted to BPAQ is too small for GSTP1 loss or rescue to materially change mass balance, BPAQ AUC, or adducts.

Predictions

• glucuronide/sulfate contain much more mass than BPAQ/BPAQ–SG;
• reducing UGT2B15 changes the oxidative branch more than perturbing GSTP1;
• a positive D0 with added BPAQ does not extend to parent BPA;
• the effect depends on tissue and exposure route.

Kill criteria

From labeled BPA at an anchored concentration, GSTP1 loss/rescue changes a material and reproducible fraction of BPAQ/AUC/adducts with UGT/SULT intact in two human systems.

L6-6-AR-H2C, v1 — dominant spontaneous capture

• Role: chemical competitor.
• Status: proposed_strengthened.
• Maturity: H0/H1.
• Confidence: 0.55.

Statement

Once BPAQ forms, its spontaneous reaction with GSH explains almost all initial capture, and GSTP1 *A/*B/*C does not materially reduce competition with target nucleophiles.

Predictions

• stopped-flow curves are nearly superimposable ±GSTP1;
• the early BPAQ–SG increment remains within the metrologic margin;
• the competing adduct does not fall;
• no reproducible allelic order exists.

Kill criteria

One variant increases incremental catalysis and reduces the competing adduct, with equal protein/GSH, on two kinetic platforms.

L6-6-AR-H3, v2 — predominantly electrophilic, not ER, effect

• Role: contrary to the “estrogen dominance” construct.
• Status: retained_reframed.
• Maturity: contextual H0; H1 only for generic GSTP1–JNK biology.
• Confidence: 0.30.

Statement

If GSTP1 modifies the response to BPA/BPAQ, the GSTP1×exposure interaction will be material for electrophile AUC, adducts, and/or the JNK complex but equivalent for ER bioactivity after fixing parent-ligand dose. A main ER effect of BPA may exist in parallel.

Kill criteria

No GSTP1 interaction occurs in any proximal endpoint, or GSTP1 consistently changes ligand dose and ER signaling while electrophilic/JNK mediation adds no information.

L6-6-AR-H5, v1 — efflux as shuttle

• Role: fate competitor.
• Status: proposed.
• Maturity: H0.
• Confidence: 0.25.

Statement

Efflux of reversible BPAQ–SG reduces adducts in the source cell without reducing total reactive burden because the exported conjugate reverts and transfers BPAQ or adduct formation to a recipient compartment.

Kill criteria

Total mass balance demonstrates irreversible loss of reactivity, formation of a terminal product, and no transfer to the recipient.

12. Translational hypothesis

L6-6-AR-HT1, v2 — organ-specific net clearance outperforms genotype

• Role: translational/metrologic.
• Status: parked.
• Maturity: H0.
• Confidence: 0.08.
• Requirement: HUMAN_QA_REQUIRED.

Statement

Only after D1/D2 are demonstrated will a reproducible net-clearance measure—defined by formation, flux, and efflux—predict an independent tissue endpoint better than rs1695/rs1138272, GSTP1 protein, or GSH alone; it must be validated by tissue rather than as a universal index.

Predictions

1. ICC ≥0.75, lower bound ≥0.60, technical CV ≤10%, and controlled batch drift;
2. active protein/GSH/flux explain more variance than genotype;
3. the metric improves out-of-sample calibration and error;
4. the repair or barrier endpoint is not part of the predictor;
5. the gain replicates externally and is not explained by hepatic/renal function, batch, or storage.

Kill criteria

• metrology or stability fails;
• predictor–endpoint leakage;
• no external gain;
• signal depends on an unanchored dose;
• no reproducible tissue consequence.

It is not termed a biomarker before these gates are passed.

13. Falsifiable predictions and kill criteria

Global kill criteria

1. No useful BPAQ catalysis: close the GSTP1–BPAQ branch.
2. Catalysis without lower AUC/adducts: do not call it detoxification.
3. Transport without a fall in total reactivity: classify it as redistribution.
4. Effect only with unanchored dose or toxicity: do not translate it.
5. Equivalent ER despite electrophilic change: remove “estrogen dominance.”
6. No tissue consequence: do not study disease or longevity.

14. Discriminating experiment

Sequence

D0a useful catalysis
  → D0b direct transport
    → D0c sink vs shuttle
      → D1 formation from BPA
        → D2 female tissue: adduct/JNK vs ER
          → T0 translational metrology

Each failed gate terminates the branch.

D0a — BPAQ stopped-flow

Objective: separate H2C from H1.

System: isotopic BPAQ; GSH 0.5, 1, 2, and 5 mM; BPAQ 0.5, 1, 2.5, and 5 µM; GSTP1 *A/*B/*C 0.05–0.5 µM; competing nucleophile; milliseconds to 10 s and chase to 60 min.

Units: six independent lots per haplotype, nested technical duplicates; blinded maximum expansion to ten if variance requires it and the method passes.

Primary endpoints:

1. rate ratio GSTP1/GSH alone;
2. 0–60 s AUC of the competing adduct.

Go: rate >1.20 with 95% CI lower bound >1.20 and adduct <0.80 with 95% CI upper bound <0.80. These are analytical prioritization margins, not clinical thresholds.

Controls: no enzyme, inactive enzyme, no GSH, no BPAQ, vehicle, negative parent BPA, and a positive validated GSTP1 quinone.

D0b — direct transport

Human inside-out vesicles or proteoliposomes with ABCC1/2/3/4, using a frozen panel. Compare ATP-Mg with AMP, empty vector, ATPase mutant, vanadate, osmolarity, and positive substrate. Three preparations for screening and six to confirm at most two transporters.

Go: ATP-dependent uptake with 95% CI lower bounds for transporter/vector and ATP/AMP >1.5, ≥70% loss with ATPase/inhibition control, and valid controls.

D0c — sink or shuttle

Donor–recipient system with GSTP1 and transporter present/absent, plus terminal trap present/absent. First, an isogenic polarized monolayer; then confirmation with hepatocytes/spheroids from six female donors connected to an endometrial recipient or chamber containing a traceable nucleophile.

Endpoints: BPAQ/BPAQ–SG AUC, specific adducts, terminal products, GSH/GSSG, and isotopic recovery in donor, medium, and recipient.

Sink: total AUC/adduct falls ≥20% and a non-reactive terminal product appears.
Shuttle: donor adduct falls, recipient adduct rises, and total burden remains equivalent.

D1 — formation from BPA

Primary female hepatocytes, eight donors with a maximum expansion to twelve, paired control/GSTP1 loss/rescue design. Exposure to ^13C-BPA at 1, 10, and 100 nM; 1 µM only as an analytical control. UGT/SULT remain intact in the primary contrast; reduced UGT2B15 is an orthogonal control.

Endpoints: mass recovered as BPA, glucuronide, sulfate, BPAQ, BPAQ–SG, and adducts; AUC and rescue; GSH/GSSG, active GSTP1, and efflux.

Kill: no BPAQ species above LLOQ in at least six of eight donors with valid controls; GSTP1 equivalence 0.80–1.25; effect only at 1 µM or with loss of hepatic function.

D2 — female tissue and the endocrine construct

Non-malignant endometrial organoids from 24 donors: eight late premenopause, eight transition, and eight early postmenopause. The primary contrast is paired GSTP1×exposure; the groups of eight estimate heterogeneity only.

Measure:

• BPAQ/adducts/GSH;
• GSTP1–JNK/TRAF2 complexes, pASK1/pJNK;
• ERα/ERβ/GPER and target genes;
• post-washout repair, barrier function, apoptosis, and senescence.

Separation mutants are interpreted only if abundance, stability, dimerization, and GSH binding are matched. If clean separation cannot be achieved, that argument is abandoned.

T0 — metrology

Thirty donors, blinded test–retest, one frozen metric, and an independent endpoint. External validation would be justified only afterward. This pilot is not powered for cancer, healthspan, or longevity.

15. Biomarkers and stratification

Candidate analytes, not validated biomarkers

• BPA, BPA-glucuronide, and BPA-sulfate;
• BPAQ and BPAQ–SG regioisomers;
• isotopically attributable adducts;
• GSH/GSSG;
• active GSTP1 protein, not RNA alone;
• terminal BPAQ–SG product, if one exists;
• GSTP1–JNK/TRAF2 complex as a mechanistic readout.

None should be called a clinical biomarker. A urinary metabolite may show that a pathway exists, but it does not by itself quantify endometrial exposure or GSTP1 control.

Required stratification

• *A/*B/*C haplotype;
• active protein and GSH separate from genotype;
• STRAW+10 stage, age, and exogenous hormones;
• hepatic and renal function;
• source and recipient tissue;
• exposure route, dose, and timing;
• genetic ancestry separate from geography;
• batch, storage, and analytical contamination.

Variables that must not be merged

• BPA, BPAQ, and BPAQ–SG;
• blood, urine, and follicular fluid;
• liver and endometrium;
• ER effect and electrophilic damage;
• acute, repeated, and prenatal exposure;
• organ-specific disease and longevity.

16. Individual variability

Expected variability arises from interactions across multiple layers:

1. Genetics: rs1695, rs1138272, haplotypes, LD, and ancestry.
2. Abundance: active GSTP1 protein may vary more than allele-specific activity.
3. Redox network: GSH, GSSG, synthesis, and regeneration.
4. Phase I/II: CYP/peroxidases, UGT, SULT, NQO1, and other GSTs.
5. Transport: ABCC expression, localization, and function.
6. Tissue: hepatic source versus local bioactivation.
7. Life stage: hormones, adiposity, inflammation, and exposure change across the menopausal transition.
8. Time: acute injury, repair, and possible chronic adaptation.

This yields a strong prediction: genotype alone will be a weak stratifier unless a substrate shows a large allelic difference and the enzyme controls a material fraction of flux. Allele frequency in a population does not inform risk without exposure and substrate.

Mexico/LATAM requires its own validation of ancestry, exposure, entry routes, and metabolic context. Effect sizes from China, Europe, or IVF populations should not be transported. This does not justify use of private data or current user genomes.

17. Pharma relevance and maturity

Assessment

There is a conditional discovery opportunity, not a development-ready opportunity. No hypothesis exceeds H1, and no strategy is ready for partnering.

Target risks

• GSTP1 can protect normal tissue and also support tumor-cell survival;
• increasing efflux can create a shuttle;
• ABCC transports multiple drugs and metabolites;
• blocking JNK may interfere with repair, immunity, or apoptosis;
• a BPAQ-specific effect cannot be generalized to other xenoestrogens.

De-risking sequence

1. target validity: D0a–D1;
2. net fate: D0b–D0c;
3. specificity: catalysis versus JNK;
4. tissue and safety: D2;
5. proximal human relevance and metrology: T0;
6. HUMAN_QA_REQUIRED before any external assessment.

No compound, dose, administration, or use in people is proposed.

18. Limitations

1. The decisive literature is mostly older, biochemical, and not specific to women.
2. No direct BPAQ ± GSTP1 *A/*B/*C experiment exists.
3. D0a with added BPAQ does not represent formation, compartmentalization, or human exposure.
4. Human oral-BPA studies are small and acute.
5. The 2026 microsomal study used 50 µM bisphenols and 5 mM GSH; it identifies pathways, not in-vivo magnitude.
6. Urinary BPA-O does not equal tissue BPAQ.
7. MRP2/BPDE and GSTP1/E2-quinones are analogies, not BPAQ validation.
8. The ABCC1/2/3/4 panel may omit a transporter, although freezing it prevents a moving target.
9. The donor–recipient system simplifies circulation, bile, kidney, and mercapturic metabolism.
10. Endometrial organoids do not represent breast, ovary, vasculature, or the whole organism.
11. rs1138272 is less functionally characterized than rs1695.
12. Disease associations are vulnerable to exposure error, selection, multiplicity, population structure, and reverse causality.
13. The 2020 endometriosis meta-analysis contains few studies and did not measure chemical mediation.
14. No functional evidence exists across the menopausal transition or specifically for Mexico/LATAM.
15. No systematic review with dual screening or full formal risk-of-bias assessment was performed.
16. The 0.80–1.25 and 20% margins are analytical prioritization gates, not clinical thresholds.
17. JNK–UGT compensation derives from a murine knockout and may reflect development or selection.
18. No proximal endpoint permits inference about healthspan or longevity.

19. Conclusions

1. GSTP1 rs1695 does not define universal detoxification capacity. Its direction changes with the substrate, and active-protein abundance may exceed the allele effect.
2. For BPA, the best-supported explanation is UGT/SULT dominance and limited BPAQ formation. Oxidation exists, but its relevant human tissue fraction is not quantified.
3. BPAQ–GSH demonstrates neither GSTP1 nor detoxification. The reaction is rapid, spontaneous, and reversible.
4. The potential causal unit is GSTP1–efflux–terminal trap. It must reduce total AUC and adducts, not merely increase conjugate or protect a source cell.
5. The shuttle is a real, falsifiable alternative. Exporting a reversible conjugate may transfer damage.
6. GSTP1 also regulates JNK/TRAF2–ASK1 without catalysis. A stress phenotype does not demonstrate xenoestrogen retention.
7. “Estrogen dominance” is rejected as the current construct. GSTP1 could modify electrophilic burden without changing E1/E2, progesterone, or ER.
8. Human disease evidence is heterogeneous and lacks mediation. IVF, endometriosis, cancer, and mortality do not close the chain.
9. The smallest decisive experiment is D0a stopped-flow. A negative result should close the GSTP1–BPAQ attribution, not trigger a post hoc search for another compound.
10. No causal connection with female longevity is currently demonstrated. Healthspan should be evaluated only after flux, tissue consequence, mediation, and an organ-specific trajectory are established.

The most important conclusion is not that GSTP1 lacks relevance, but that its relevance must be earned molecule by molecule and gate by gate. The broad popular hypothesis has been refuted; a narrow, low-maturity kinetic question with high falsification value survives.

Stable claims added by REPORT_EN

• L6-6-REPORT-EN-C1: rs1695 cannot be classified as a universally slow or adverse allele; functional direction depends on the electrophile and becomes causally relevant only after active protein, GSH, and flux control are quantified.
• L6-6-REPORT-EN-C2: for BPA, two nulls must be distinguished: insufficient BPAQ formation from the parent and dominant spontaneous capture once formed; D0a discriminates only the latter.
• L6-6-REPORT-EN-C3: with a reversible conjugate, reducing adducts in the source cell does not demonstrate net detoxification; donor+medium+recipient balance, total loss of reactivity, and a terminal fate are required.
• L6-6-REPORT-EN-C4: human detection of BPA-O and microsomal bioactivation of bisphenols in 2026 support an oxidative branch but neither quantify tissue BPAQ nor attribute flux to GSTP1.
• L6-6-REPORT-EN-C5: no direct human chain exists from defined xenoestrogen → GSTP1 flux → ER signaling/damage → health → female longevity; “estrogen dominance” remains refuted as an operational construct.

20. References

1. Ali-Osman F, Akande O, Antoun G, Mao JX, Buolamwini J. Molecular cloning, characterization, and expression in Escherichia coli of full-length cDNAs of three human glutathione S-transferase Pi gene variants. J Biol Chem. 1997;272:10004–10012. DOI: 10.1074/jbc.272.15.10004. PMID: 9092542.
2. Watson MA, Stewart RK, Smith GBJ, Massey TE, Bell DA. Human glutathione S-transferase P1 polymorphisms: relationship to lung tissue enzyme activity and population frequency distribution. Carcinogenesis. 1998;19:275–280. DOI: 10.1093/carcin/19.2.275. PMID: 9498276.
3. Johansson AS, Stenberg G, Widersten M, Mannervik B. Structure–activity relationships and thermal stability of human glutathione transferase P1-1 governed by the H-site residue 105. J Mol Biol. 1998. DOI: 10.1006/jmbi.1998.1708. PMID: 9600848.
4. Sundberg K, Johansson AS, Stenberg G, Widersten M, Seidel A, Mannervik B, Jernström B. Differences in the catalytic efficiencies of allelic variants of glutathione transferase P1-1 towards carcinogenic diol epoxides of polycyclic aromatic hydrocarbons. Carcinogenesis. 1998;19:433–436. DOI: 10.1093/carcin/19.3.433. PMID: 9525277.
5. Hu X, Xia H, Srivastava SK, et al. Activity of four allelic forms of glutathione S-transferase hGSTP1-1 for diol epoxides of polycyclic aromatic hydrocarbons. Cancer Res. 1999;59:2358–2362. PMID: 10344744.
6. Coles BF, Morel F, Rauch C, et al. Effect of polymorphism in the human glutathione S-transferase P1 gene on enzyme activity and expression in lung. Cancer Lett. 2000. DOI: 10.1016/S0304-3835(00)00458-4. PMID: 10880766.
7. Chang M, Zhang F, Shen L, et al. Inhibition of glutathione S-transferase activity by the catechol estrogen 4-hydroxyequilenin. 2001. PMID: 11294649.
8. Yao J, Chang M, Li Y, Pisha E, Liu X, Yao D, El-Bayoumy K, Bolton JL. Inhibition of cellular enzymes by equine catechol estrogen-4-hydroxyequilenin. Chem Res Toxicol. 2002. DOI: 10.1021/tx020018i. PMID: 12119004.
9. Dawling S, Roodi N, Mernaugh RL, Wang X, Parl FF. Catechol-O-methyltransferase and glutathione S-transferase modulation of estrogen quinone formation. 2003. PMID: 14679015.
10. Qiu SX, Yang RZ, Gross ML, et al. Synthesis and reactivity of bisphenol A-3,4-quinone with DNA and glutathione. Chem Res Toxicol. 2004. DOI: 10.1021/tx049953r. PMID: 15310235.
11. Stack DE, Byun J, Gross ML, Rogan EG, Cavalieri EL. Structural identification and kinetic analysis of the in vitro products formed by reaction of bisphenol A-3,4-quinone with N-acetylcysteine and glutathione. Chem Res Toxicol. 2018;31:81–87. DOI: 10.1021/acs.chemrestox.7b00239. PMID: 29281792.
12. Jaeg JP, Perdu E, Dolo L, et al. Characterization of new bisphenol A metabolites produced by CD1 mice liver microsomes and S9 fractions. 2004. DOI: 10.1021/jf049762u. PMID: 15264938.
13. Völkel W, Colnot T, Csanády GA, Filser JG, Dekant W. Metabolism and kinetics of bisphenol A in humans at low doses following oral administration. Chem Res Toxicol. 2002. PMID: 12387626.
14. Thayer KA, Doerge DR, Hunt D, et al. Pharmacokinetics of bisphenol A in humans following a single oral administration. Environ Int. 2015. PMID: 26115537.
15. Street CM, Zhu Z, Finel M, Court MH. Bisphenol-A glucuronidation in human liver and breast: identification of UDP-glucuronosyltransferases and influence of genetic polymorphisms. 2017. PMID: 26999266.
16. Srivastava SK, Hu X, Xia H, et al. Metabolism and efflux of benzo[a]pyrene-7,8-diol-9,10-epoxide glutathione conjugate and protection from DNA adduct formation by MRP2/cMOAT. 2002. PMID: 11870881.
17. Leier I, Jedlitschky G, Buchholz U, Center M, Cole SPC, Deeley RG, Keppler D. ATP-dependent glutathione disulphide transport mediated by the MRP gene-encoded conjugate export pump. 1998. PMID: 9679551.
18. Loe DW, Stewart RK, Massey TE, Deeley RG, Cole SPC. ATP-dependent transport of aflatoxin B1 and its glutathione conjugates by the product of the multidrug resistance protein gene. 1997. PMID: 9187270.
19. Adler V, Yin Z, Fuchs SY, et al. Regulation of JNK signaling by GSTp. EMBO J. 1999. PMID: 10064598.
20. Wang T, Arifoglu P, Ronai Z, Tew KD. Glutathione S-transferase P1-1 inhibits c-Jun N-terminal kinase activity. J Biol Chem. 2001. DOI: 10.1074/jbc.M101355200.
21. Wu Y, Fan Y, Xue B, et al. Human glutathione S-transferase P1-1 interacts with TRAF2 and regulates TRAF2–ASK1 signals. 2006. PMID: 16636664.
22. Elsby R, Kitteringham NR, Goldring CE, et al. Increased constitutive c-Jun N-terminal kinase signaling in mice lacking glutathione S-transferase Pi. J Biol Chem. 2003. DOI: 10.1074/jbc.M301211200.
23. Abbas S, Greige-Gerges H, Karam N, Piet MH, Netter P, Magdalou J. Metabolism of parabens by hepatic microsomes and plasma of humans and rats. 2010. PMID: 20930423.
24. Hanioka N, et al. Glucuronidation of mono(2-ethylhexyl) phthalate in human liver and intestinal microsomes. 2017. PMID: 27071666.
25. Song Y, Wagner BA, Witmer JR, et al. Nonenzymatic displacement of chlorine and formation of free radicals upon reaction of glutathione with PCB quinones. 2009. PMID: 19497881.
26. Vichi S, Medda E, Ingelido AM, et al. Glutathione transferase polymorphisms and risk of endometriosis associated with polychlorinated biphenyl exposure. Fertil Steril. 2012. DOI: 10.1016/j.fertnstert.2012.02.027.
27. Chen L, et al. Association between GSTP1 polymorphism and endometriosis risk: a meta-analysis. Eur J Obstet Gynecol Reprod Biol. 2013. DOI: 10.1016/j.ejogrb.2013.10.005. PMID: 24289900.
28. Méar L, et al. Glutathione S-transferase P1 polymorphism and endometriosis: systematic review and meta-analysis. 2020. PMID: 31821471.
29. Egan KM, Cai Q, Shu XO, et al. Genetic polymorphisms in GSTM1, GSTP1, and GSTT1 and the risk for breast cancer: results from the Shanghai Breast Cancer Study and meta-analysis. 2004. DOI: 10.1158/1055-9965.EPI-03-0294. PMID: 14973092.
30. Martínez-Ramírez OC, Pérez-Morales R, Castro C, et al. Polymorphisms of catechol estrogens metabolism pathway genes and breast cancer risk in Mexican women. Breast. 2013. DOI: 10.1016/j.breast.2012.08.004. PMID: 23000097.
31. Jaramillo-Rangel G, et al. GST polymorphisms and breast cancer in northeastern Mexico. 2015. PMID: 26125851.
32. McCarty KM, et al. PAH-DNA adducts, GST polymorphisms, and breast cancer risk. Environ Health Perspect. 2009. DOI: 10.1289/ehp.0800119. PMID: 19440493.
33. Spurdle AB, et al. GST polymorphisms and breast cancer risk in BRCA1/2 mutation carriers. 2010. DOI: 10.1007/s10549-009-0601-0. PMID: 19921428.
34. Miao J, et al. Glutathione S-transferase genetic variation and breast cancer susceptibility: an assessment of cumulative evidence. PLoS One. 2020. DOI: 10.1371/journal.pone.0216147. PMID: 32155154.
35. Syrkasheva A, et al. Association of bisphenol A exposure and detoxification gene polymorphisms with follicular fluid concentrations and IVF parameters in women. 2021. PMID: 34834443.
36. Michalczyk K, et al. Metalloestrogens, GSTP1 polymorphism and endometrial pathology. 2022. PMID: 35893933.
37. He Q, Hu X, Li X, Li N, Wu JL. Covalently active metabolites of bisphenol A analogs by mass spectrometry diagnostic ions: possible mechanisms of their toxicity. Chem Res Toxicol. 2026. DOI: 10.1021/acs.chemrestox.5c00417. PMID: 41493763.
38. Fan W, Weng Y, Fang Z, et al. Isomer-specific metabolism and novel metabolite identification of bisphenol A and its isomer o,p'-BPA in humans. Ecotoxicol Environ Saf. 2026; DOI: 10.1016/j.ecoenv.2026.119879. PMID: 41690093.
39. Nørskov MS, et al. GSTP1 genotype and risk of tobacco-related cancer and mortality in a large prospective cohort. 2017. DOI: 10.1016/j.jtho.2017.07.008. PMID: 28739440.

Maturity register and next decision

Next project stage: BILINGUAL_QA.