Research-priority portfolio

What deserves the next decisive experiment?

11 evidence-graded dossiers organize plausible paths toward prevention or durable treatment-free remission—while making the causal gaps, safety limits, and failure conditions impossible to miss.

Explore the portfolio How to read the grades TLR7 therapy review

Different roles, different claims

A portfolio—not a leaderboard of treatments

Initiators, reservoirs, reconstitution states, amplifiers, and selectors answer different questions. They may be complementary rather than competing.

01Before or at disease transitionInitiation & prevention4 dossiers 02What may preserve—or rebuild—diseaseMaintenance & reconstitution4 dossiers 03Clinically actionable, not stand-alone curesControl pathways & selectors3 dossiers

Research boundaryClinical response while receiving therapy is not the same as durable drug-free remission. A biomarker or selector is not itself a biological cure target.

Before or at disease transition

Initiation & prevention

Candidates that may precede classified SLE or define a state in which transition could be prevented.

Initiator candidate

Band BTier 2

Rare functionally confirmed TLR7 or UNC93B1 gain-of-function disease

Correct an upstream genetically causal signaling defect in the rare subgroup whose pathogenic TLR7-pathway gain of function is independently confirmed by variant interpretation and functional testing.

Population in scope: People with a pathogenic or likely pathogenic TLR7 or UNC93B1 variant and independently replicated functional gain of TLR7-pathway signaling.
Strongest causal gap: Human causal perturbation: 0/4 No human carrier study has selectively normalized the variant-linked pathway and demonstrated causal clinical rescue.

Decisive experiment

Does selective normalization of a confirmed TLR7-pathway gain-of-function defect produce durable correction in human carriers?

Minimum follow-up: Two years after confirmed engagement, including off-intervention observation only where ethically permissible.

Inspect evidence, falsifiers & sources

Evidence shape

Temporality / precedence4/4

Human causal perturbation0/4

Durable remission / prevention0/4

Relevant-tissue evidence2/4

Independent replication2/4

Population definition4/4

Experiment design

An international prospective multiple-baseline rare-disease study using the same selective pathway intervention, standardized background care, variant-specific functional assays, staggered starts where ethical, and matched sampling before treatment, at engagement, after clinical stabilization, and during a prespecified observation after withdrawal where permissible.

Comparator: Staggered-start within-carrier control and matched natural-history carriers when randomization is infeasible.
Primary outcome: Reproducible normalization of variant-linked signaling followed by sustained correction of autoimmune activity or prevention-state progression beyond the participant's baseline trajectory.

What would raise confidence

Variant-linked signaling normalizes reproducibly while unrelated protective sensing remains within prespecified margins.
Clinical or preclinical-state improvement follows engagement across independent carriers and persists beyond expected pharmacologic washout where ethically testable.
The magnitude of functional correction tracks the outcome within carriers and across alleles.
Alternative maintenance reservoirs are measured so persistent disease despite correction is interpretable.

What would lower or falsify it

Complete variant-linked signaling correction does not alter autoimmune state or clinical course.
Improvement precedes engagement or is attributable to background therapy rather than pathway correction.
Only broad antiviral or innate suppression, rather than selective normalization, accompanies benefit.
Disease persists through autonomous reservoirs despite durable correction, weakening the target's established-disease cure role.

Evidence and contradiction

Supports
A de novo TLR7 Y264H gain-of-function variant was found in severe childhood lupus; it increased ligand sensing, caused lupus when introduced into mice, and MyD88 deficiency rescued the mouse phenotypes.
TLR7 gain-of-function genetic variation causes human lupus
Supports
Rare UNC93B1 variants in childhood-onset SLE increased TLR7 or TLR8-linked cytokine responses, and an orthologous variant produced spontaneous lupus-like disease in mice.
Genetic variants in UNC93B1 predispose to childhood-onset systemic lupus erythematosus
Qualifies
The strongest causal observation comes from rare variants and cannot establish that ordinary polygenic SLE shares the same initiating or maintenance dependency.
TLR7 gain-of-function genetic variation causes human lupus
Contradicts
B-cell-intrinsic TLR9 was protective in murine lupus, showing that broad endosomal nucleic-acid-sensor blockade cannot be treated as equivalent to selective correction of causal TLR7 gain of function.
B cell-intrinsic TLR9 expression is protective in murine lupus

Source ledger

TLR7 gain-of-function genetic variation causes human lupusprimary study · verified Jul 11, 2026
Genetic variants in UNC93B1 predispose to childhood-onset systemic lupus erythematosusprimary study · verified Jul 11, 2026
Enpatoran in systemic and cutaneous lupus: randomized placebo-controlled phase 1b studyprimary study · verified Jul 11, 2026
Enpatoran in moderate-to-severe systemic lupus erythematosus: WILLOW Cohort Bprimary study · verified Jul 11, 2026
B cell-intrinsic TLR9 expression is protective in murine lupusprimary study · verified Jul 11, 2026
DS-7011a randomized phase 1b/2 SLE registry resultstrial registry · verified Jul 12, 2026
TLR7 in systemic lupus erythematosus: genetics and emerging therapiesreview · verified Jul 12, 2026

Review state: draft-agent-assisted. This dossier has not been assigned domain or expert review.

Initiator candidate

Band CTier 2

Nonmonogenic high-TLR7 functional state

Identify a nonmonogenic subgroup whose disease or pre-disease state remains preferentially dependent on excessive TLR7 activity and test selective pathway normalization without borrowing causal certainty from rare variants.

Population in scope: People without a functionally causal TLR7 or UNC93B1 variant who meet a prospectively locked high-TLR7 functional classifier before treatment.
Strongest causal gap: Temporality / precedence: 0/4 A TLR7-specific nonmonogenic functional state has not been validated prospectively before SLE transition or established-disease activity.

Decisive experiment

Does a locked nonmonogenic TLR7-specific functional state predict preferential and durable benefit from selective pathway inhibition?

Minimum follow-up: One year of randomized treatment assessment plus at least one year of post-treatment or extension follow-up.

Inspect evidence, falsifiers & sources

Evidence shape

Temporality / precedence0/4

Human causal perturbation3/4

Durable remission / prevention0/4

Relevant-tissue evidence1/4

Independent replication2/4

Population definition0/4

Experiment design

Prospectively freeze a TLR7-selective functional assay and threshold in an observational run-in, then test it in a multicenter randomized placebo-controlled trial that enrolls both classifier-high and classifier-low participants, stratifies randomization by classifier, uses the same background-care and taper rules, and prespecifies the treatment-by-classifier interaction as the primary causal test.

Comparator: Placebo within classifier-high and classifier-low strata, using the same selective inhibitor and background care.
Primary outcome: A prespecified treatment-by-classifier interaction for a validated clinical response jointly requiring TLR7-specific target engagement and followed by persistence assessment after washout.

What would raise confidence

The frozen assay is reproducible across sites and distinguishes TLR7 responses from TLR8, TLR9, and generic interferon activity.
Classifier-high participants show materially greater randomized benefit than classifier-low participants without outcome-driven threshold changes.
Clinical benefit follows selective target engagement and is not explained by skin phenotype, baseline severity, exposure, or broad immunosuppression.
A meaningful subset sustains improvement or delayed rebound beyond expected pharmacologic washout while protective immunity remains intact.

What would lower or falsify it

The assay cannot reproduce across sites or collapses after controlling for cell composition and infection.
Classifier-high and classifier-low participants have similar randomized treatment effects.
Benefit occurs without TLR7-specific engagement or is attributable to TLR8 or generic anti-inflammatory activity.
Any improvement disappears immediately with exposure loss, refuting a durable cure role even if on-treatment control remains valid.

Evidence and contradiction

Supports
A randomized phase 1b study in active systemic or cutaneous lupus showed that dual TLR7/8 inhibition was feasible and produced exploratory pharmacodynamic and clinical signals.
Enpatoran in systemic and cutaneous lupus: randomized placebo-controlled phase 1b study
Supports
The randomized systemic WILLOW cohort tested multiple doses in 354 participants and reported secondary clinical and biomarker signals, establishing a platform for prospective interaction testing.
Enpatoran in moderate-to-severe systemic lupus erythematosus: WILLOW Cohort B
Contradicts
The systemic WILLOW cohort did not meet its primary objective of a statistically significant dose-response relationship for BICLA response, weakening an unselected broad-dependency claim.
Enpatoran in moderate-to-severe systemic lupus erythematosus: WILLOW Cohort B
Qualifies
DS-7011a had only 19 active and six placebo participants in the safety set, smaller available-data efficacy denominators, no dependency enrichment, and no peer-reviewed patient efficacy report at the cutoff.
DS-7011a randomized phase 1b/2 SLE registry results

Source ledger

Enpatoran in systemic and cutaneous lupus: randomized placebo-controlled phase 1b studyprimary study · verified Jul 11, 2026
Enpatoran in moderate-to-severe systemic lupus erythematosus: WILLOW Cohort Bprimary study · verified Jul 11, 2026
B cell-intrinsic TLR9 expression is protective in murine lupusprimary study · verified Jul 11, 2026
Progression from at-risk state to clinical and severe systemic lupus erythematosus involves molecular dysregulationsprimary study · verified Jul 11, 2026
Enpatoran for active cutaneous manifestations: WILLOW Cohort Aprimary study · verified Jul 12, 2026
DS-7011a randomized phase 1b/2 SLE registry resultstrial registry · verified Jul 12, 2026
E6742 dual TLR7/8 antagonist randomized phase 1/2 SLE studyprimary study · verified Jul 12, 2026
Afimetoran randomized phase 1b study in cutaneous lupusprimary study · verified Jul 12, 2026
TLR7 in systemic lupus erythematosus: genetics and emerging therapiesreview · verified Jul 12, 2026

Review state: draft-agent-assisted. This dossier has not been assigned domain or expert review.

Prevention state

Band CTier 2

Prospectively validated pre-disease transition trajectory

Define a reproducible high-absolute-risk window in which mechanism-matched prevention can be tested before classified SLE and irreversible organ injury, while protecting likely nonprogressors from unnecessary intervention.

Population in scope: Prospectively enrolled ANA-positive asymptomatic or symptomatic people without SLE, unaffected relatives, and incomplete-lupus populations analyzed as distinct strata.
Strongest causal gap: Durable remission / prevention: 0/4 No trajectory-guided intervention has demonstrated durable prevention of SLE or organ disease.

Decisive experiment

Can a locked longitudinal trajectory safely identify SLE-specific transition risk and predict differential benefit from a mechanism-matched prevention strategy?

Minimum follow-up: Five years for primary prevention and organ outcomes, with ten-year passive or registry follow-up for late transitions and harms.

Inspect evidence, falsifiers & sources

Evidence shape

Temporality / precedence4/4

Human causal perturbation1/4

Durable remission / prevention0/4

Relevant-tissue evidence1/4

Independent replication3/4

Population definition2/4

Experiment design

A prospective multicenter external-validation cohort with serial measurements and blinded outcome adjudication, followed by an embedded biomarker-enriched randomized mechanism-matched prevention trial. Freeze the model, thresholds, assay pipeline, missing-data handling, comparator clinical model, and subgroup analyses before outcomes are revealed.

Comparator: Clinical and routine-laboratory risk assessment alone, plus placebo within the randomized high-risk prevention stratum.
Primary outcome: Well-calibrated adjudicated SLE-specific absolute risk and a prespecified treatment-by-trajectory interaction for incident SLE or clinically meaningful organ disease.

What would raise confidence

External calibration and discrimination meet prespecified thresholds at all time horizons and remain acceptable in major population strata.
The trajectory adds decision-curve net benefit beyond ordinary clinical and laboratory assessment while reducing false-positive enrollment.
The high-risk state predicts greater randomized benefit from the matched intervention, with mechanism-specific engagement mediating outcome improvement.
Low-risk participants have a sufficiently low observed event rate to support de-escalated research surveillance without missed organ-threatening disease.
The complete model, code, assay specification, and deidentified validation data are independently reproducible.

What would lower or falsify it

The model is miscalibrated or loses discrimination outside discovery cohorts or within major demographic groups.
Serial molecular features add no meaningful net benefit beyond inexpensive clinical measures.
The trajectory does not predict differential benefit from a mechanism-matched intervention.
False-positive labeling, intervention harm, or missed organ disease outweighs prevention benefit.
Apparent prediction is explained by outcome-component leakage, treatment, infection, or referral selection.

Evidence and contradiction

Supports
Archived serial samples showed that at least one tested SLE autoantibody preceded diagnosis in 115 of 130 future cases, establishing a measurable pre-diagnosis phase.
Development of autoantibodies before the clinical onset of systemic lupus erythematosus
Supports
Among 409 initially unaffected relatives, 45 transitioned over mean 6.4 years; baseline clinical burden and selected soluble mediators were associated with transition.
Discerning risk of disease transition in relatives of systemic lupus erythematosus patients
Contradicts
In the randomized SMILE trial, hydroxychloroquine did not reduce SLICC-criteria accumulation or transition to SLE, showing that identifying incomplete lupus does not by itself yield effective prevention.
A randomized placebo-controlled trial of hydroxychloroquine in incomplete lupus
Qualifies
Eighty-nine percent of the ANA-positive or undifferentiated cohort did not develop a systemic autoimmune rheumatic disease over two years, underscoring false-positive risk and the weakness of ANA alone.
Evaluation of progression from preclinical to systemic autoimmune rheumatic disease

Source ledger

Development of autoantibodies before the clinical onset of systemic lupus erythematosusprimary study · verified Jul 11, 2026
Discerning risk of disease transition in relatives of systemic lupus erythematosus patientsprimary study · verified Jul 11, 2026
Evaluation of progression from preclinical to systemic autoimmune rheumatic diseaseprimary study · verified Jul 11, 2026
Progression from at-risk state to clinical and severe systemic lupus erythematosus involves molecular dysregulationsprimary study · verified Jul 11, 2026
A randomized placebo-controlled trial of hydroxychloroquine in incomplete lupusprimary study · verified Jul 11, 2026

Review state: draft-agent-assisted. This dossier has not been assigned domain or expert review.

Initiator candidate

Band CWatch

EBV-positive autoreactive antigen-presenting B-cell state

Prevent formation or selectively remove an EBV-reprogrammed autoreactive B-cell state that could couple persistent viral latency to self-antigen presentation and systemic autoimmune amplification.

Population in scope: Research participants with directly measured EBV-positive B cells whose expressed B-cell receptors bind validated SLE nuclear autoantigens and whose cells show a prespecified antigen-presenting program.
Strongest causal gap: Temporality / precedence: 0/4 The defining cellular state has not been demonstrated prospectively before autoimmunity or first SLE transition.

Decisive experiment

Does the EBV-positive autoreactive antigen-presenting B-cell state precede SLE and mediate preventable transition in a biomarker-defined high-risk population?

Minimum follow-up: Five years for prevention outcomes, with longer surveillance for late viral and lymphoproliferative safety.

Inspect evidence, falsifiers & sources

Evidence shape

Temporality / precedence0/4

Human causal perturbation1/4

Durable remission / prevention0/4

Relevant-tissue evidence1/4

Independent replication1/4

Population definition1/4

Experiment design

First prospectively validate a locked cellular-state assay in serial pre-disease samples. Only after temporality, calibration, and a safe selective intervention are established, conduct a multicenter biomarker-enriched randomized prevention trial with mechanistic mediation analysis and a nested state-negative observational comparator.

Comparator: Matched placebo or control intervention within state-positive participants, plus a state-negative observational comparison cohort.
Primary outcome: Adjudicated incident SLE or clinically meaningful organ disease, jointly interpreted with selective elimination or durable normalization of the prespecified EBV-positive autoreactive B-cell state.

What would raise confidence

The state reproducibly appears before clinical transition and adds calibrated SLE-specific risk beyond ordinary clinical and serologic predictors.
A selective intervention reduces the state without broad loss of protective B-cell immunity or uncontrolled EBV reactivation.
State reduction precedes and statistically mediates a lower rate or delayed timing of SLE transition in randomized comparison.
The association and intervention effect replicate across independent centers and major demographic strata.

What would lower or falsify it

The state is absent before onset or appears only after disease activity, treatment, or EBV reactivation.
The state is equally prevalent and functional in persistent nonprogressors after adjustment for confounding.
Selective state reduction does not change transition risk or relevant autoimmune trajectories.
Clinical benefit occurs without changing the state, or state reduction causes unacceptable loss of protective immunity.

Evidence and contradiction

Supports
EBV-positive B cells in SLE showed a CD27-positive CD21-low memory phenotype, EBNA2-linked antigen-presentation programs, antinuclear autoreactivity, and the capacity to activate helper and related autoreactive B-cell responses.
Epstein-Barr virus reprograms autoreactive B cells as antigen-presenting cells in systemic lupus erythematosus
Supports
Human SLE samples showed defective control of latent EBV infection, supporting a context in which infected B-cell states may be abnormally retained or expanded.
Defective control of latent Epstein-Barr virus infection in systemic lupus erythematosus
Contradicts
Prospective flare monitoring in established SLE found that disease flares preceded EBV reactivation, supporting reactivation as a consequence or amplifier rather than the initiating cause in that context.
Exhausted cytotoxic control of Epstein-Barr virus in human lupus
Qualifies
The defining study analyzed established SLE and experimental cell systems, so it cannot determine whether the state precedes disease or is required for transition.
Epstein-Barr virus reprograms autoreactive B cells as antigen-presenting cells in systemic lupus erythematosus

Source ledger

Epstein-Barr virus reprograms autoreactive B cells as antigen-presenting cells in systemic lupus erythematosusprimary study · verified Jul 11, 2026
Exhausted cytotoxic control of Epstein-Barr virus in human lupusprimary study · verified Jul 11, 2026
Defective control of latent Epstein-Barr virus infection in systemic lupus erythematosusprimary study · verified Jul 11, 2026

Review state: draft-agent-assisted. This dossier has not been assigned domain or expert review.

What may preserve—or rebuild—disease

Maintenance & reconstitution

Candidate reservoirs and post-reset immune states that could help explain persistence, relapse, or durable remission.

Maintenance reservoir

Band CTier 2

Plasma-lineage reservoir

Remove or durably silence pathogenic plasma-lineage cells that escape CD19-only depletion and may preserve autoantibody production or humoral memory capable of restarting SLE.

Population in scope: Treatment-refractory SLE represented in dual CD19/BCMA cohorts
Strongest causal gap: Population definition: 1/4 Cohorts are refractory and often renal, but no locked plasma-reservoir classifier defines who should benefit.

Decisive experiment

Does a baseline plasma-lineage reservoir identify an SLE subgroup that gains greater durable remission from dual rather than CD19-only targeting?

Minimum follow-up: 5 years

Inspect evidence, falsifiers & sources

Evidence shape

Temporality / precedence2/4

Human causal perturbation3/4

Durable remission / prevention3/4

Relevant-tissue evidence3/4

Independent replication2/4

Population definition1/4

Experiment design

Prospective biomarker-stratified direct comparison of CD19-only and CD19/BCMA strategies using harmonized conditioning, clinical endpoints, sampling, protective-immunity testing, and late follow-up.

Comparator: CD19-only versus CD19/BCMA targeting within reservoir-positive and reservoir-negative strata.
Primary outcome: Treatment-by-reservoir interaction for durable treatment-free DORIS remission and time to relapse.

What would raise confidence

Dual targeting improves durability specifically in the prespecified reservoir-positive subgroup
Residual pathogenic plasma-lineage clones predict relapse after CD19-only targeting
Benefit replicates without disproportionate loss of protective immunity

What would lower or falsify it

No treatment-by-reservoir interaction is observed
CD19-only and dual-target durability are equivalent in reservoir-positive patients
Plasma-lineage clearance does not reduce relapse or causes unacceptable sustained immune deficiency

Evidence and contradiction

Supports
A 15-person dual CD19/BCMA trial reported 12 DORIS remissions with naive reconstitution and reduced interferon and BAFF programs.
Co-infusion of CD19-targeting and BCMA-targeting CAR-T cells for treatment-refractory systemic lupus erythematosus
Supports
Extended dual-target follow-up reported ten stringent complete remissions, eleven drug-free serologic complete remissions, and repeat kidney-biopsy improvement.
BCMA-CD19 armored compound CAR T cells in systemic lupus erythematosus: extended follow-up
Contradicts
CD19-only treatment produced drug-free responses despite persistent anti-Ro and anti-La antibodies interpreted as partly plasma-cell derived.
CD19 CAR-T cells for treatment-refractory autoimmune diseases: the CASTLE basket trial
Qualifies
The apparent dual-target advantage used five new participants and nonconcurrent public comparator data, so clinical superiority is unproven.
Dual-Targeting BCMA-CD19 CAR-T Cell Therapy in SLE: Single-Cell and Immune Repertoire Analysis

Source ledger

Co-infusion of CD19-targeting and BCMA-targeting CAR-T cells for treatment-refractory systemic lupus erythematosusprimary study · verified Jul 11, 2026
BCMA-CD19 armored compound CAR T cells in systemic lupus erythematosus: extended follow-upprimary study · verified Jul 11, 2026
Dual-Targeting BCMA-CD19 CAR-T Cell Therapy in SLE: Single-Cell and Immune Repertoire Analysisprimary study · verified Jul 11, 2026
CD19 CAR-T cells for treatment-refractory autoimmune diseases: the CASTLE basket trialprimary study · verified Jul 11, 2026

Review state: draft-agent-assisted. This dossier has not been assigned domain or expert review.

Maintenance reservoir

Band CTier 2

Relapse-capable autoreactive clonotypes

Identify and eliminate or durably silence functionally pathogenic immune clonotypes that survive depletion or re-emerge to reconstruct the autoimmune network causing relapse.

Population in scope: SLE relapse after apparently complete depletion
Strongest causal gap: Relevant-tissue evidence: 1/4 Most clonotype evidence is blood-based and cannot exclude rare persistent clones in marrow, lymph node, kidney, or skin.

Decisive experiment

Do functionally pathogenic pre-treatment clonotypes reproducibly reappear before SLE relapse but remain absent or silent in durable remission?

Minimum follow-up: 5 years

Inspect evidence, falsifiers & sources

Evidence shape

Temporality / precedence2/4

Human causal perturbation2/4

Durable remission / prevention2/4

Relevant-tissue evidence1/4

Independent replication2/4

Population definition1/4

Experiment design

Prospective standardized cellular-therapy cohort with dense serial blood, matched baseline tissue, mandatory pre-rescue relapse sampling, recombinant clone testing, and blinded old-clone versus de-novo relapse analysis.

Comparator: Same-product durable responders versus participants with adjudicated relapse.
Primary outcome: Pre-relapse return or expansion of functionally validated baseline autoreactive clonotypes relative to durable responders.

What would raise confidence

The same functionally pathogenic clones reproducibly precede relapse
Clone dynamics add prediction beyond activity, autoantibodies, and total memory-cell counts
Selective old-clone loss tracks remission while unrelated protective repertoires remain intact

What would lower or falsify it

Relapses are usually driven by new unrelated clones
Verified old clones persist equally in durable remission and relapse
Clone return follows rather than precedes relapse
Functional assays show tracked clones are not pathogenic

Evidence and contradiction

Supports
Antigen-experienced B cells, particularly in nonresponders, persisted through rituximab depletion and included early persistent clones.
Single-Cell Level Characterization of B Cell Depletion and Repopulation Following Rituximab in SLE
Supports
Clonally related and evolving BCR sequences occurred in active SLE, while regenerating repertoires contained new clonal expansions.
Clonal expansion of B-cells in human systemic lupus erythematosus before and after B-cell depletion
Qualifies
Early persistent-clone and diversity effects were not long-lasting, and no tracked clone was shown to cause relapse.
Single-Cell Level Characterization of B Cell Depletion and Repopulation Following Rituximab in SLE
Contradicts
One relapse tracked activated memory T-cell return without a memory-B-cell-number change, suggesting heterogeneous relapse mechanisms.
Phenotypic changes of lymphocytes in SLE in longterm remission after rituximab

Source ledger

Single-Cell Level Characterization of B Cell Depletion and Repopulation Following Rituximab in SLEprimary study · verified Jul 11, 2026
Clonal expansion of B-cells in human systemic lupus erythematosus before and after B-cell depletionprimary study · verified Jul 11, 2026
Dual-Targeting BCMA-CD19 CAR-T Cell Therapy in SLE: Single-Cell and Immune Repertoire Analysisprimary study · verified Jul 11, 2026
B cell biomarkers of rituximab responses in systemic lupus erythematosusprimary study · verified Jul 11, 2026
Phenotypic changes of lymphocytes in SLE in longterm remission after rituximabprimary study · verified Jul 11, 2026
CD19 CAR T-Cell Therapy in Autoimmune Disease - A Case Series with Follow-upprimary study · verified Jul 11, 2026

Review state: draft-agent-assisted. This dossier has not been assigned domain or expert review.

Maintenance reservoir

Band CTier 2

Tissue-resident antigen-experienced B-cell reservoir

Remove or durably silence antigen-experienced B cells in lymphoid and affected tissues that may survive peripheral depletion and preserve the capacity to restart autoimmune disease.

Population in scope: Severe treatment-refractory SLE represented in existing cellular-therapy cohorts
Strongest causal gap: Independent replication: 1/4 The decisive paired-tissue evidence comes from one small program and disease-mismatched rituximab comparators.

Decisive experiment

Does residual tissue antigen-experienced B-cell burden causally distinguish durable treatment-free remission from relapse?

Minimum follow-up: 5 years

Inspect evidence, falsifiers & sources

Evidence shape

Temporality / precedence2/4

Human causal perturbation3/4

Durable remission / prevention2/4

Relevant-tissue evidence4/4

Independent replication1/4

Population definition2/4

Experiment design

Prospective multicenter same-product cohort with fixed conditioning and paired blood and tissue sampling before treatment, at nadir, after reconstitution, and at relapse, with blinded pathology and prespecified mediation analysis.

Comparator: Durable treatment-free responders versus relapsers receiving the same product and conditioning.
Primary outcome: Time to adjudicated relapse as a function of prespecified residual tissue antigen-experienced B-cell burden.

What would raise confidence

Low tissue-reservoir burden reproducibly precedes durable remission
Tissue burden adds relapse prediction beyond blood counts, baseline activity, and damage
The result replicates in an independent cohort and relevant organ tissue

What would lower or falsify it

Durable remission commonly coexists with persistent tissue reservoirs
Complete tissue depletion commonly occurs without remission
The association disappears after controlling for conditioning, endotype, and organ damage

Evidence and contradiction

Supports
Paired lymph nodes had complete CD19/CD20 B-cell depletion after CAR-T but retained B cells after rituximab despite absent blood B cells.
CD19-CAR T-cell therapy induces deep tissue depletion of B cells
Supports
Incomplete high-sensitivity depletion included every nonresponder, and faster memory/plasmablast repopulation accompanied earlier relapse.
B cell biomarkers of rituximab responses in systemic lupus erythematosus
Qualifies
Rituximab tissue comparators had systemic sclerosis or GPA rather than SLE, and every CAR-T recipient also received conditioning.
CD19-CAR T-cell therapy induces deep tissue depletion of B cells
Contradicts
Drug-free remission continued after peripheral B-cell return, so permanent global B-cell absence is not required for remission.
CD19 CAR T-Cell Therapy in Autoimmune Disease - A Case Series with Follow-up

Source ledger

CD19-CAR T-cell therapy induces deep tissue depletion of B cellsprimary study · verified Jul 11, 2026
B cell biomarkers of rituximab responses in systemic lupus erythematosusprimary study · verified Jul 11, 2026
Efficacy and safety of rituximab in patients with active proliferative lupus nephritisprimary study · verified Jul 11, 2026
CD19 CAR T-Cell Therapy in Autoimmune Disease - A Case Series with Follow-upprimary study · verified Jul 11, 2026

Review state: draft-agent-assisted. This dossier has not been assigned domain or expert review.

Reconstitution state

Band CTier 2

Tolerant immune reconstitution

Guide post-depletion immune recovery toward a stable state that does not rebuild pathogenic memory while restoring broad protective humoral and cellular immunity.

Population in scope: Treatment-refractory SLE after CD19 CAR-T, dual-target cellular therapy, rituximab, or AHSCT
Strongest causal gap: Relevant-tissue evidence: 2/4 Some marrow and kidney data exist, but most state definitions rely on peripheral blood rather than matched tissues.

Decisive experiment

Does a prespecified tolerant-reconstitution state causally distinguish durable treatment-free remission from relapse?

Minimum follow-up: 5 years

Inspect evidence, falsifiers & sources

Evidence shape

Temporality / precedence3/4

Human causal perturbation3/4

Durable remission / prevention3/4

Relevant-tissue evidence2/4

Independent replication3/4

Population definition2/4

Experiment design

Prospective multicenter same-product cohort that locks a composite state score before outcome analysis, samples matched blood and tissue, and externally validates prediction and mediation of relapse.

Comparator: Same-product durable treatment-free responders versus participants with adjudicated relapse.
Primary outcome: External-cohort time to relapse predicted by the locked reconstitution score beyond baseline endotype, depletion depth, conditioning, and organ damage.

What would raise confidence

The prespecified state reproduces across platforms and populations
It predicts durable remission before outcome is known
It distinguishes protective recovery from global immune deficiency
A component intervention changes relapse risk in the predicted direction

What would lower or falsify it

The state is equally common in relapsers and durable responders
Prediction disappears after adjustment for activity or conditioning
Favorable blood reconstitution coexists with pathogenic tissue persistence and relapse
The state requires unacceptable persistent immune deficiency

Evidence and contradiction

Supports
B cells returned after about 110 days with a naive non-class-switched phenotype while drug-free remission continued.
Anti-CD19 CAR T cell therapy for refractory systemic lupus erythematosus
Supports
Reconstitution was over ninety percent naive in evaluable participants, with virtual loss of memory and plasmablast compartments and no reported relapse during follow-up.
CD19 CAR-T cells for treatment-refractory autoimmune diseases: the CASTLE basket trial
Qualifies
A broadly diverse regenerating repertoire still contained clonal expansions, so naive-leaning recovery does not establish tolerance.
Clonal expansion of B-cells in human systemic lupus erythematosus before and after B-cell depletion
Contradicts
Relapse-associated changes were heterogeneous, involving memory B cells in one participant and activated memory T cells in another.
Phenotypic changes of lymphocytes in SLE in longterm remission after rituximab

Source ledger

Anti-CD19 CAR T cell therapy for refractory systemic lupus erythematosusprimary study · verified Jul 11, 2026
CD19 CAR-T cells for treatment-refractory autoimmune diseases: the CASTLE basket trialprimary study · verified Jul 11, 2026
Long-term follow-up after lymphodepleting autologous haematopoietic cell transplantation for treatment-resistant SLEprimary study · verified Jul 11, 2026
Single-Cell Level Characterization of B Cell Depletion and Repopulation Following Rituximab in SLEprimary study · verified Jul 11, 2026
B cell biomarkers of rituximab responses in systemic lupus erythematosusprimary study · verified Jul 11, 2026
Clonal expansion of B-cells in human systemic lupus erythematosus before and after B-cell depletionprimary study · verified Jul 11, 2026
Phenotypic changes of lymphocytes in SLE in longterm remission after rituximabprimary study · verified Jul 11, 2026

Review state: draft-agent-assisted. This dossier has not been assigned domain or expert review.

Clinically actionable, not stand-alone cures

Control pathways & selectors

Validated disease amplifiers and selection tools that can strengthen experiments without being mistaken for root causes.

Validated amplifier

Band BTier 3

BAFF and BLyS signaling

A clinically validated B-cell survival control pathway that can reduce disease activity, but is not established as a universal initiating cause, complete maintenance reservoir, or stand-alone cure target.

Population in scope: Patients aged 5 years and older with active SLE receiving standard therapy, matching the June 2025 U.S. label.
Strongest causal gap: Durable remission / prevention: 0/4 The cited evidence evaluates continued add-on therapy and does not establish durable treatment-free remission or disease prevention.

Decisive experiment

Can BAFF blockade produce durable immune reprogramming in a prospectively identified BAFF-dependent subgroup rather than benefit only during continued inhibition?

Minimum follow-up: 5 years

Inspect evidence, falsifiers & sources

Evidence shape

Temporality / precedence1/4

Human causal perturbation4/4

Durable remission / prevention0/4

Relevant-tissue evidence2/4

Independent replication4/4

Population definition3/4

Experiment design

A prospective biomarker-stratified randomized mechanistic trial using a locked continuous BAFF and B-cell-state panel, a mechanism-distinct active comparator, paired blood and ethically available tissue measurements, B-cell receptor lineage tracking, plasma-cell and memory-state assays, standardized remission outcomes, and supervised withdrawal after sustained control.

Comparator: Mechanism-distinct active therapy plus the same protocolized background-care framework.
Primary outcome: Sustained protocol-defined clinical remission after supervised withdrawal with persistent correction of the prespecified pathogenic B-cell and tissue state.

What would raise confidence

A prespecified treatment-by-BAFF/B-cell-state interaction replicates independently.
A clinically meaningful fraction remains in remission after withdrawal with disappearance rather than redistribution of pathogenic lineages.
Tissue and blood states improve concordantly while protective humoral immunity is preserved.

What would lower or falsify it

Deep BAFF engagement and peripheral B-cell remodeling leave pathogenic lineages or tissue states intact.
Disease reliably returns after withdrawal without durable repertoire correction.
No reproducible treatment interaction exists across the prespecified BAFF/B-cell-state continuum.
Protective immune impairment outweighs any durable-remission gain.

Evidence and contradiction

Supports
The June 2025 U.S. label indicates belimumab for active SLE and active lupus nephritis from age five and documents coherent B-cell and serologic pharmacodynamics.
BENLYSTA belimumab U.S. Prescribing Information, June 2025
Supports
At week 104 in active lupus nephritis, primary efficacy renal response was 43% with belimumab plus standard therapy and 32% with placebo plus standard therapy; OR 1.6, 95% CI 1.0 to 2.3, p=0.031.
Two-Year, Randomized, Controlled Trial of Belimumab in Lupus Nephritis
Qualifies
At week 76 in one adult IV trial, SRI-4 response was 39% with belimumab 10 mg/kg and 32% with placebo and was not statistically significant.
BENLYSTA belimumab U.S. Prescribing Information, June 2025
Contradicts
A dedicated 448-participant trial in Black adults found a numerical primary response difference, 49% versus 42%, that was not statistically significant; earlier pooled Black subgroup findings were unfavorable.
BENLYSTA belimumab U.S. Prescribing Information, June 2025

Source ledger

Two-Year, Randomized, Controlled Trial of Belimumab in Lupus Nephritisprimary study · verified Jul 11, 2026
The role of baseline BLyS levels and type 1 interferon-inducible gene signature status in determining belimumab responseprimary study · verified Jul 11, 2026
BENLYSTA belimumab U.S. Prescribing Information, June 2025regulatory · verified Jul 11, 2026

Review state: draft-agent-assisted. This dossier has not been assigned domain or expert review.

Validated amplifier

Band BTier 3

Type I interferon signaling through IFNAR1

A clinically validated control pathway that may suppress an interferon-dominant disease state, but is not established as a universal root cause or stand-alone cure target.

Population in scope: Adults with moderate-to-severe active SLE receiving standard therapy, matching the April 2026 U.S. label and pivotal generalized-SLE trials.
Strongest causal gap: Durable remission / prevention: 0/4 The cited trials assessed response during continued therapy and do not demonstrate sustained drug-free remission or prevention of SLE.

Decisive experiment

Does IFNAR1 blockade create durable immune correction in a prospectively identified interferon-active subgroup rather than reversible suppression?

Minimum follow-up: 5 years

Inspect evidence, falsifiers & sources

Evidence shape

Temporality / precedence1/4

Human causal perturbation4/4

Durable remission / prevention0/4

Relevant-tissue evidence2/4

Independent replication4/4

Population definition3/4

Experiment design

A biomarker-stratified randomized mechanistic trial using a locked continuous IFN-state assay, a mechanism-distinct active comparator, paired blood and affected-tissue measurements where ethically available, repertoire tracking, standardized remission outcomes, and a supervised withdrawal phase after sustained disease control.

Comparator: Mechanism-distinct active therapy plus the same protocolized background-care framework.
Primary outcome: Sustained protocol-defined clinical remission after supervised withdrawal, coupled to persistent correction of the prespecified tissue and blood immune state.

What would raise confidence

A prespecified treatment-by-continuous-IFN-state interaction replicates in an independent cohort.
A clinically meaningful fraction remains in remission after withdrawal without recovery of the pathogenic tissue state.
Pathogenic repertoire or cellular states decline beyond what is seen with the active comparator while protective immunity is preserved.

What would lower or falsify it

Deep blood signature neutralization fails to correct the affected-tissue state.
Disease reliably returns as the signature recovers after withdrawal despite adequate on-treatment engagement.
No reproducible treatment interaction is observed across the continuous baseline IFN state.
Protective immune impairment outweighs durable-remission benefit.

Evidence and contradiction

Supports
Human peripheral-blood transcriptomics identified an interferon-inducible expression pattern in a substantial subset of SLE, establishing a replicated molecular state.
Interferon-inducible gene expression signature in peripheral blood cells of patients with severe lupus
Supports
TULIP-2 reported week-52 BICLA response in 47.8% with anifrolumab and 31.5% with placebo, a 16.3 percentage-point difference with 95% CI 6.3 to 26.3.
Trial of Anifrolumab in Active Systemic Lupus Erythematosus
Contradicts
TULIP-1 failed its prespecified week-52 SRI-4 primary endpoint: 36% responded with anifrolumab 300 mg and 40% with placebo; difference -4.2 points, 95% CI -14.2 to 5.8, p=0.41.
Type I interferon inhibitor anifrolumab in active systemic lupus erythematosus: TULIP-1
Qualifies
In pooled TULIP data, 82.6% were IFN-signature-high and subgroup efficacy was generally consistent; the small IFN-low group prevents a validated binary treatment-selection rule.
Anifrolumab efficacy and safety by type I interferon gene signature and clinical subgroups

Source ledger

Interferon-inducible gene expression signature in peripheral blood cells of patients with severe lupusprimary study · verified Jul 11, 2026
Type I interferon inhibitor anifrolumab in active systemic lupus erythematosus: TULIP-1primary study · verified Jul 11, 2026
Trial of Anifrolumab in Active Systemic Lupus Erythematosusprimary study · verified Jul 11, 2026
Anifrolumab efficacy and safety by type I interferon gene signature and clinical subgroupsprimary study · verified Jul 11, 2026
SAPHNELO anifrolumab-fnia U.S. Prescribing Information, April 2026regulatory · verified Jul 11, 2026

Review state: draft-agent-assisted. This dossier has not been assigned domain or expert review.

Enabling selector

Band CTier 3

Eight-endotype whole-blood transcriptomic selector

An enabling target for matching biologically heterogeneous SLE states to mechanism-specific experiments; it is an assay and decision tool, not a biological cure target.

Population in scope: Research participants with classified SLE and standardized baseline whole-blood RNA suitable for the published gene-module workflow.
Strongest causal gap: Durable remission / prevention: 0/4 No evidence shows classifier-guided care produces durable drug-free remission or prevents SLE.

Decisive experiment

Does a frozen eight-endotype classifier reproducibly predict comparative treatment benefit beyond ordinary clinical and serologic variables?

Minimum follow-up: 104 weeks

Inspect evidence, falsifiers & sources

Evidence shape

Temporality / precedence2/4

Human causal perturbation2/4

Durable remission / prevention0/4

Relevant-tissue evidence1/4

Independent replication4/4

Population definition3/4

Experiment design

Apply the published classifier without refitting to blinded, nonoverlapping banked samples from randomized phase III trials of mechanism-distinct therapies; preregister one interaction hierarchy, multiplicity control, calibration analysis, and an independent replication set before outcomes are unlocked.

Comparator: Concurrent randomized placebo or mechanism-distinct active comparator within each parent trial.
Primary outcome: A prespecified, replicated treatment-by-endotype interaction on the parent trial clinical endpoint with incremental decision value beyond clinical and serologic predictors.

What would raise confidence

Frozen assignments meet prespecified cross-laboratory reproducibility and failure-rate thresholds.
The same endotype-treatment interaction is clinically meaningful and statistically replicated in an independent trial.
The classifier improves calibration or net benefit beyond disease activity, organ involvement, anti-dsDNA, complement, ancestry, and background therapy.
Performance is acceptable across prespecified ancestry, sex, age, and organ subgroups.

What would lower or falsify it

Assignments are unstable across laboratory, platform, timepoint, or routine medication exposure.
Treatment interactions disappear under a frozen rule or fail independent replication.
Differences are explained by placebo prognosis rather than active-treatment effect modification.
Clinical and serologic variables equal or outperform the transcriptomic classifier.
Material subgroup miscalibration or access inequity cannot be mitigated.

Evidence and contradiction

Supports
Analysis across 3,166 samples in 17 datasets developed eight whole-blood molecular endotypes with distinct enrichment patterns and exploratory associations with flare and tabalumab response.
Analysis of transcriptomic features reveals molecular endotypes of SLE with clinical implications
Supports
A previously trained classifier was prospectively applied to 101 independent baseline samples and recovered all eight expected molecular patterns without outcome-based refitting.
Validation of eight endotypes of lupus based on whole-blood RNA profiles
Qualifies
The independent cohort validated staging-pattern reproduction, not treatment-effect prediction; some endotypes contained only three participants and no prospective treatment interaction was tested.
Validation of eight endotypes of lupus based on whole-blood RNA profiles
Qualifies
The main iberdomide analysis re-clustered the trial into five groups, used multiple within-group comparisons, included only women, and found higher placebo response in groups without an active-drug differential.
Responsiveness of systemic lupus erythematosus subjects to iberdomide based on molecular endotypes

Source ledger

Analysis of transcriptomic features reveals molecular endotypes of SLE with clinical implicationsprimary study · verified Jul 11, 2026
Validation of eight endotypes of lupus based on whole-blood RNA profilesprimary study · verified Jul 11, 2026
Responsiveness of systemic lupus erythematosus subjects to iberdomide based on molecular endotypesprimary study · verified Jul 11, 2026
Subsets of Patients With Lupus Identified by Gene Expression Profiles Exhibit Differential Clinical Responsiveness to Baricitinibprimary study · verified Jul 11, 2026

Review state: draft-agent-assisted. This dossier has not been assigned domain or expert review.

How to read the portfolio

High actionability cannot erase a causal gap

Grades are dimensional and deliberately not averaged. One weak dimension can define the next experiment.

A–D

Evidence band

A narrative summary of the evidence shape. Band A requires direct, replicated human evidence tied to prevention or durable remission; none of these 11 dossiers currently reaches it.

Tier 1–3

Research tier

A priority for the next causal experiment—not a treatment ranking. Tier 1 means a decisive study is ready enough to pursue under proper oversight.

0–4

Dimension grade

Temporality, perturbation, durable outcome, tissue relevance, replication, and population definition each retain their own rationale and uncertainty.

Fail well

Falsification first

Every dossier names results that should lower confidence. A useful target program must be able to discover that its favored hypothesis is wrong.

Research use only. This portfolio does not diagnose lupus, recommend a treatment, rank present clinical options, or establish that any candidate will lead to a cure.

Read the editorial methods