Abiraterone

證據等級: L5

預測適應症: 10 個

AuTxGNN Drug Repurposing Evidence Report

Abiraterone — Multi-Indication Prediction Analysis

Field	Detail
Candidate ID	TW-DB05812-multi
Report Version	v4
Date Generated	3 April 2026
Data Cutoff	3 April 2026
Data Sources Queried	DrugBank, ClinicalTrials.gov, PubMed, ICTRP, TFDA

1. Executive Summary

Abiraterone (DrugBank ID: DB05812) is a selective CYP17A1 inhibitor currently approved internationally for the treatment of metastatic castration-resistant prostate cancer (mCRPC). The TxGNN knowledge graph model generated 10 predicted new indications for abiraterone, spanning migraine subtypes, pulmonary hypertension, leprosy, rheumatoid arthritis, and several rare conditions.

Overall Assessment: No actionable repurposing candidates identified.

All 10 predictions received a recommendation of “Hold” (9 of 10) or “Research Question” (1 of 10 — pulmonary hypertension). The mechanistic rationale linking CYP17A1 inhibition to any of the predicted conditions ranges from non-existent to contradictory. Critically, for several predictions (rheumatoid arthritis, nephrogenic syndrome of inappropriate antidiuresis), the expected pharmacological effect of abiraterone would theoretically worsen rather than improve the condition.

TGA Registration Status: Abiraterone acetate is registered on the Australian Register of Therapeutic Goods (ARTG) under brand names including Zytiga® (Janssen) for use in combination with prednisone/prednisolone for metastatic castration-resistant prostate cancer and metastatic hormone-sensitive prostate cancer. Multiple generic formulations are also listed. The evidence pack data originates from the Taiwanese regulatory context (TFDA), where abiraterone is listed as “not marketed” with zero licences.

Key Data Gaps: TGA/TFDA label warnings, contraindications, and detailed mechanism of action data were not available in this evidence pack, representing a blocking gap for formal safety assessment.

2. Drug Information

Parameter	Detail
INN (Generic Name)	Abiraterone
DrugBank ID	DB05812
Australian Brand Names	Zytiga® (Janssen-Cilag); generic formulations available
Chemical Class	Steroidal CYP17A1 (17α-hydroxylase/C17,20-lyase) inhibitor
Mechanism of Action	[Data Gap — not provided in evidence pack]
Known MOA (from literature)	Irreversible inhibition of CYP17A1, blocking androgen biosynthesis at the adrenal and intratumoral level. Reduces circulating testosterone, DHEA, and androstenedione. Compensatory mineralocorticoid excess occurs due to upstream accumulation of steroid precursors.
TGA-Approved Indications	Metastatic castration-resistant prostate cancer (mCRPC); metastatic high-risk hormone-sensitive prostate cancer (mHSPC) — in combination with prednisone or prednisolone
ARTG Status	Registered (multiple entries)
Formulation	Oral tablets (250 mg, 500 mg)
TFDA (Taiwan) Status	Not marketed (0 licences)

3. Predicted New Indications

3.1 Summary Table

Rank	Predicted Indication	TxGNN Score	Evidence Level	Decision Stage	Recommendation	Mechanistic Plausibility
1	Migraine disorder	0.988	L5	S0	Hold	Very weak
2	Migraine with or without aura (susceptibility)	0.988	L5	S0	Hold	Very weak
3	Migraine with brainstem aura	0.986	L5	S0	Hold	Very weak
4	Leprosy	0.986	L5	S0	Hold	None
5	Pulmonary hypertension	0.984	L4	S1	Research Question	Weak–contradictory
6	Nephrogenic syndrome of inappropriate antidiuresis	0.982	L5	S0	Hold	Very weak (harmful direction)
7	Rheumatoid arthritis	0.981	L5	S0	Hold	Contradictory
8	Kyphoscoliotic heart disease	0.981	L5	S0	Hold	None
9	Atrophoderma vermiculata	0.977	L5	S0	Hold	Very weak, speculative
10	Ulerythema ophryogenesis	0.975	L5	S0	Hold	Very weak, speculative

3.2 Detailed Indication Analyses

Rank 1 — Migraine Disorder

TxGNN Score: 0.988 (rank 11,869)
Evidence Level: L5 (computational prediction only)
Clinical Trials: None identified (ClinicalTrials.gov: 0; ICTRP: 0)
Literature: None identified (PubMed: 0 results)
Mechanistic Assessment: Very weak. CYP17A1 inhibition reduces androgens. While sex hormone fluctuations (particularly oestrogen) are associated with migraine, abiraterone primarily affects the androgen pathway rather than the core migraine pathophysiology involving CGRP and the trigeminovascular system. Furthermore, known adverse effects of abiraterone — including mineralocorticoid excess, hypokalaemia, and fluid retention — may exacerbate headache.
Recommendation: Hold. No supporting evidence. Mechanistic rationale is inadequate.

Rank 2 — Migraine With or Without Aura (Susceptibility)

TxGNN Score: 0.988 (rank 12,386)
Evidence Level: L5 (computational prediction only)
Clinical Trials: None identified
Literature: 20 PubMed results returned, however all 20 articles are false positives — they relate to epilepsy genetics, susceptibility to seizures, and paediatric epilepsy models, with no direct relevance to abiraterone or migraine treatment. The only tangentially relevant paper (PMID 33856647, Gotra et al. 2021) discusses shared genetic and molecular mechanisms between epilepsy and migraine but does not reference the CYP17A1 pathway.
Mechanistic Assessment: Very weak. Same rationale as Rank 1. The literature set does not support any mechanistic connection.
Recommendation: Hold. Retrieved literature is not relevant; likely a PubMed keyword matching artefact on the term “susceptibility.”

Rank 3 — Migraine With Brainstem Aura

TxGNN Score: 0.986 (rank 13,674)
Evidence Level: L5
Clinical Trials: None identified
Literature: None identified
Mechanistic Assessment: Very weak. Brainstem aura migraine involves brainstem dysfunction; no known link to androgen biosynthesis inhibition.
Recommendation: Hold.

Rank 4 — Leprosy

TxGNN Score: 0.986 (rank 13,923)
Evidence Level: L5
Clinical Trials: None identified
Literature: None identified
Mechanistic Assessment: No plausible connection. Leprosy is caused by Mycobacterium leprae infection and requires antimicrobial therapy (dapsone, rifampicin, clofazimine). CYP17A1 inhibition has no known antimycobacterial activity nor relevant immunomodulatory mechanism for this infection.
Recommendation: Hold.

Rank 5 — Pulmonary Hypertension ⭐ (Highest evidence level in this set)

TxGNN Score: 0.984 (rank 15,720)
Evidence Level: L4
Decision Stage: S1
Clinical Trials: 1 result (NCT01961843) — false positive. This is a prostate cancer circulating tumour cell study, unrelated to pulmonary hypertension treatment. Relevance grade: C.
Literature: 5 PubMed results — all are prostate cancer studies involving abiraterone. None investigate pulmonary hypertension as a therapeutic target. Key papers include STAMPEDE platform RCTs (PMID 34953525), ACIS trial (PMID 34600602), and a cardio-oncology retrospective study on angiotensin inhibitors during abiraterone therapy (PMID 33869068).
Mechanistic Assessment: Weak to contradictory.
- Sex hormones have a recognised but complex relationship with pulmonary arterial hypertension (PAH): female sex confers higher incidence, while male sex confers worse prognosis.
- DHEA (a CYP17A1 downstream metabolite) has shown pulmonary vasculoprotective effects in animal models. Abiraterone reduces DHEA, which could theoretically be harmful rather than beneficial.
- CYP17A1 is expressed in pulmonary vascular smooth muscle, but the directionality of effect is uncertain.
Recommendation: Research Question — but with significant caveats. The mechanistic direction is potentially counterproductive. This warrants literature monitoring only; no preclinical or clinical investigation is recommended at this stage.

Rank 6 — Nephrogenic Syndrome of Inappropriate Antidiuresis (NSIAD)

TxGNN Score: 0.982 (rank 16,832)
Evidence Level: L5
Clinical Trials: None identified
Literature: None identified
Mechanistic Assessment: Very weak, potentially harmful. NSIAD results from gain-of-function mutations in AVPR2. Abiraterone’s known adverse effect of mineralocorticoid excess can cause fluid retention and electrolyte disturbance, which would likely worsen this condition.
Recommendation: Hold. Predicted pharmacological effect directionally harmful.

Rank 7 — Rheumatoid Arthritis

TxGNN Score: 0.981 (rank 17,720)
Evidence Level: L5
Clinical Trials: None identified
Literature: None identified
Mechanistic Assessment: Contradictory. Androgens generally exert immunosuppressive and anti-inflammatory effects. Low androgen states are associated with increased rheumatoid arthritis risk. By reducing androgens, abiraterone would theoretically promote autoimmune inflammation rather than suppress it.
Recommendation: Hold. Mechanistic direction opposes therapeutic intent.

Rank 8 — Kyphoscoliotic Heart Disease

TxGNN Score: 0.981 (rank 17,838)
Evidence Level: L5
Clinical Trials: None identified
Literature: None identified
Mechanistic Assessment: None. This is a structural/mechanical condition resulting from spinal deformity causing cardiopulmonary restriction. No pharmacological pathway exists through which CYP17A1 inhibition could address thoracic structural abnormalities.
Recommendation: Hold.

Rank 9 — Atrophoderma Vermiculata

TxGNN Score: 0.977 (rank 21,340)
Evidence Level: L5
Clinical Trials: None identified
Literature: None identified
Mechanistic Assessment: Very weak, speculative. This rare follicular keratinisation disorder could theoretically be influenced by reduced sebaceous activity from androgen suppression, but the core pathology is keratinisation abnormality, not androgen excess. Anti-androgen therapy has not been reported in the literature for this condition.
Recommendation: Hold.

Rank 10 — Ulerythema Ophryogenesis

TxGNN Score: 0.975 (rank 22,196)
Evidence Level: L5
Clinical Trials: None identified
Literature: None identified
Mechanistic Assessment: Very weak, speculative. This keratosis pilaris variant involving the eyebrow region shares the same follicular keratinisation disorder spectrum as Rank 9. The androgen connection is equally indirect and unsupported.
Recommendation: Hold.

4. Evidence Summary

4.1 Clinical Trial Evidence

Source	Total Hits	Relevant Hits	Notes
ClinicalTrials.gov	1	0	1 false positive (prostate cancer CTC study)
ANZCTR / ICTRP	0	0	No Australian or international registry matches

No clinical trials were identified investigating abiraterone for any of the 10 predicted indications.

4.2 Published Literature

Source	Total Hits	Relevant Hits	Notes
PubMed	25	0	20 epilepsy/genetics papers (false positives from “susceptibility” keyword); 5 prostate cancer papers

Critical Observation: The PubMed literature retrieval contains significant noise. The 20 papers associated with “migraine with or without aura, susceptibility to” are overwhelmingly epilepsy-focused, likely resulting from MeSH term overlap on “susceptibility.” The 5 papers associated with pulmonary hypertension are all prostate cancer treatment studies that mention abiraterone but do not investigate pulmonary hypertension as a therapeutic target.

4.3 Australian-Specific Studies

No Australian-specific studies were identified for any predicted indication.

5. Safety Considerations

5.1 Known Adverse Effects

⚠️ Data Gap (DG001 — Blocking): TGA/TFDA label warnings and contraindications were not available in this evidence pack. The information below is derived from general pharmacological knowledge.

Common adverse effects of abiraterone (from international prescribing information):

Mineralocorticoid-related: hypertension, hypokalaemia, fluid retention/oedema
Hepatotoxicity (elevated ALT/AST)
Fatigue
Joint discomfort
Hot flushes
Urinary tract infections
Cardiac disorders (including cardiac failure, atrial fibrillation)

5.2 Contraindications

[Data Gap] — TGA Product Information not parsed
Known international contraindications include: severe hepatic impairment (Child-Pugh C), pregnancy/women of childbearing potential (Category X — teratogenic)

5.3 Drug–Drug Interactions

Parameter	Result
DDI Query Status	Not found
Total Interactions	0 (data gap)

Note: Abiraterone is a known inhibitor of CYP2D6 and CYP2C8. Clinically significant interactions are expected with substrates of these enzymes. Full DDI profiling is a required data gap remediation step.

5.4 Safety Concerns Specific to Predicted Indications

Indication	Safety Concern
Migraine (Ranks 1–3)	Mineralocorticoid excess (hypertension, hypokalaemia) may worsen headache; unacceptable risk–benefit for a non-life-threatening condition
Pulmonary hypertension (Rank 5)	Fluid retention and cardiac adverse effects could exacerbate haemodynamic compromise
NSIAD (Rank 6)	Mineralocorticoid excess would worsen fluid retention — directly counterproductive
Rheumatoid arthritis (Rank 7)	Androgen reduction may worsen autoimmune inflammation
All indications	Abiraterone’s adverse effect profile is acceptable in the context of advanced prostate cancer but disproportionate for the non-oncological conditions predicted

6. Data Gaps

ID	Category	Item	Severity	Impact	Remediation
DG001	Drug Level	TGA/TFDA Label Warnings & Contraindications	Blocking	Cannot enter S1 safety assessment	Download and parse TGA PI document
DG002	Drug Level	Mechanism of Action (MOA) detail	High	Affects mechanistic association analysis	Query DrugBank API for complete MOA data

7. Recommendations

7.1 Evidence Strength Assessment

Metric	Value
Predictions at Evidence Level L5 (computational only)	9 of 10
Predictions at Evidence Level L4	1 of 10 (pulmonary hypertension)
Predictions with supporting clinical evidence	0 of 10
Predictions with plausible mechanistic rationale	0 of 10
Predictions with contradictory mechanistic direction	3 of 10 (NSIAD, RA, potentially PAH)

7.2 Overall Recommendation

No repurposing candidates from this prediction set warrant further investigation at this time.

The TxGNN model assigned high prediction scores (>0.975) to all 10 indications, yet none are supported by:

Clinical trial evidence
Relevant published literature
Plausible mechanistic rationale

This result likely reflects a limitation of graph-based prediction models when applied to drugs with highly specific mechanisms (selective CYP17A1 inhibition) that do not generalise well to non-oncological therapeutic areas.

7.3 Suggested Next Steps

Remediate data gaps — Obtain TGA Product Information for complete safety profiling (DG001) and DrugBank MOA data (DG002) before any further assessment.
No preclinical or clinical investigation recommended for any of the 10 predicted indications.
Refine PubMed search strategy — The high false-positive rate in literature retrieval (particularly for “susceptibility”-containing disease terms) suggests the evidence collection pipeline may benefit from improved keyword specificity and relevance filtering.
Model feedback — These results should be fed back into the TxGNN model evaluation pipeline to assess whether CYP17A1 inhibitors systematically generate low-quality predictions for non-oncological conditions.

7.4 Relevant Australian Research Context

Abiraterone is well-studied in Australian prostate cancer research through the ANZUP (Australian and New Zealand Urogenital and Prostate Cancer Trials Group) network.
No Australian research groups have published on abiraterone repurposing for any of the predicted indications.
Pulmonary hypertension research is active at centres including the Victor Chang Cardiac Research Institute (Sydney) and the National Heart Foundation, but there is no indication of interest in CYP17A1-targeted approaches.

8. Query Log Summary

A total of 33 queries were executed across 5 data sources on 10 March 2026:

Source	Queries	Results Found	Relevant Results
TFDA	1	0	0
DDI	1	0	0
DrugBank	1	1	1
ClinicalTrials.gov	10	1	0 (false positive)
ICTRP	10	0	0
PubMed	10	25	0 (all false positives)
Total	33	27	1 (DrugBank only)

Research Use Only: This report is generated for research purposes only and does not constitute medical advice. Drug repurposing predictions require clinical validation before therapeutic application. This is not a TGA-approved assessment.

Generated by AuTxGNN — Australian Drug Repurposing Predictions Report ID: TW-DB05812-multi | Version: v4 | Date: 3 April 2026

Disclaimer

This content is for research purposes only and does not constitute medical advice. Clinical validation is required before any clinical application.