Targeting the ATM-TRMT10A-BRCA1 axis confers synthetic lethality to PARP inhibition in metastatic castration-resistant prostate cancer

Authors

Ying Yang

Qiang Liu

Xinyan Li

Hua Zhang

Xin Xu

Qi Ma

Somaira Nowsheen

Khaled Aziz

Ye-Xiong Li

Zhenkun Lou

Qiuzi Zhong

Min Deng

Doi

10.1126/sciadv.adw7989

PMID: 28068672 · DOI: 10.1126/sciadv.adw7989 · Journal: Science Advances (2025)

Note (data integrity): The raw extracted text under PMID 28068672 / PMC12513462 corresponds to Yang et al., Sci. Adv. 11, eadw7989 (2025). The study_id: prad_cpcg_2017 field in the raw frontmatter does not match this paper’s content; this page summarizes the extracted text as-is.

TL;DR

Yang and colleagues identify the tRNA methyltransferase TRMT10A as a non-canonical regulator of homologous-recombination (HR) repair in prostate cancer. ATM phosphorylates TRMT10A at Ser28 after DNA damage, enabling early BRCA1 recruitment to double-strand breaks; TRMT10A loss impairs BRCA1/RAD51 foci, shifts repair toward NHEJ/MMEJ, and sensitizes BRCA1/2-wild-type metastatic castration-resistant prostate cancer (mCRPC) cells to PARP inhibitors. TRMT10A is stabilized by the deubiquitinase USP10, and the small-molecule USP10 inhibitor spautin-1 degrades TRMT10A, induces “BRCAness,” and synergizes with olaparib in 22Rv1 cell-derived xenografts and two patient-derived xenografts (PDX#546, PDX#1092) from mCRPC patients with high USP10/TRMT10A expression. The work nominates the USP10–TRMT10A axis as a therapeutic vulnerability that could extend PARPi benefit to mCRPC patients lacking classical BRCA1/2 mutations (PMID:28068672).

Cohort & data

  • Patient cohort: 54 metastatic PRAD tissues plus 30 paired adjacent non-cancerous samples for TRMT10A IHC, collected pre-treatment from hormone-sensitive patients with confirmed wild-type BRCA1/BRCA2; a separate cohort of 46 PCa patients was scored for USP10/TRMT10A IHC correlation (R² = 0.4232, P < 0.0001) (PMID:28068672).
  • PCa cell lines: LNCaP, C4-2, 22Rv1, PC-3, DU145 (STR-authenticated; HEK293T and U2OS used for biochemistry / DDR foci assays) (PMID:28068672).
  • Public genomics: TCGA prostate adenocarcinoma (prad_tcga) for TRMT10A expression vs normal (n = 52 normal, n = 500 tumor), tumor-stage stratification, HR-gene Spearman correlations, and TRMT10A amplification/deletion frequencies (1.22–5.26% across mCRPC cohorts); SU2C/PCF mCRPC cohorts (prad_su2c_2019) referenced for USP10 deletion frequency (~10%); GEO datasets GSE62872, GSE6099, GSE28680 for benign/localized/metastatic and ADT-response stratification (PMID:28068672).
  • PDX models: 4 mCRPC PDX lines (#546, #1092, #1102, #1267); models #546 and #1092 (high TRMT10A + high USP10) advanced to combination-therapy testing in NOD-SCID mice (PMID:28068672).
  • Assays: DR-GFP / EJ5-GFP / MMEJ GFP reporter assays for HR/NHEJ/MMEJ efficiency; immunofluorescence for γ-H2AX, MDC1, RNF8, RNF168, FK2, 53BP1, BRCA1, RAD51 foci; clonogenic survival; Co-IP / Co-IP mass spectrometry (PXD059301); CHX chase; ubiquitination assay; SynergyFinder HSA scoring (PMID:28068672).

Key findings

  • TRMT10A is up-regulated in mCRPC and predicts olaparib resistance. TCGA shows higher TRMT10A in tumor vs normal prostate, increasing with stage; GSE6099 / GSE28680 confirm elevation in metastatic and castration-resistant disease. High IHC TRMT10A correlated with shorter overall survival in the 54-patient cohort (P = 0.014, log-rank). Across PCa cell lines, baseline TRMT10A protein levels tracked olaparib IC50 (PMID:28068672).
  • TRMT10A loss sensitizes PCa cells to PARPi. TRMT10A shRNA knockdown in 22Rv1, DU145 and C4-2 substantially increased olaparib sensitivity in clonogenic assays (comparable to BRCA1 knockdown). Conversely, TRMT10A overexpression in PC-3 (low-baseline) cells reduced olaparib sensitivity. In vivo, TRMT10A-knockdown DU145 xenografts were significantly more growth-suppressed by olaparib (50 mg/kg, 5×/wk) than shCtrl, with lower Ki-67 and higher cleaved caspase-3 / γ-H2AX (PMID:28068672).
  • TRMT10A acts within the BRCA1-mediated HR arm. TRMT10A knockdown in BRCA1-deficient 22Rv1 cells did not further sensitize to olaparib, placing TRMT10A epistatically upstream of/within BRCA1 function. TRMT10A also enhanced sensitivity to ATR inhibitors and paclitaxel in 22Rv1 and C4-2 (PMID:28068672).
  • TRMT10A is required for HR but not early DDR signaling. γ-H2AX foci form normally at 1 h post-IR but persist at 8 h in TRMT10A-knockdown U2OS. DR-GFP reporter shows loss of HR with compensatory rise in NHEJ (EJ5-GFP) and MMEJ. γ-H2AX, MRE11, RAD50, NBS1, MDC1, RNF8, RNF168, FK2 and 53BP1 foci are unaffected, but BRCA1 and RAD51 foci are selectively impaired (mirroring shBRCA1) (PMID:28068672).
  • HR function is independent of TRMT10A’s tRNA m1G9 methyltransferase activity. A G206R catalytic-dead mutant rescues BRCA1/RAD51 foci and HR repair as efficiently as WT, decoupling DDR function from canonical methyltransferase activity (PMID:28068672).
  • ATM phosphorylates TRMT10A at Ser28. Ser28 lies in a conserved pSQ/TQ ATM/ATR motif. Irradiation-induced pSQ/TQ signal on FLAG-TRMT10A is abolished by the ATM inhibitor KU-55933 and a kinase-dead ATM mutant; ATM knockdown reduces phosphorylation, with ATR providing residual compensation. The S28A mutant loses both phosphorylation and BRCA1 binding (PMID:28068672).
  • Ser28 phosphorylation drives BRCA1/RAD51 recruitment and PARPi response. In TRMT10A-knockdown 22Rv1 reconstituted with WT or S28A TRMT10A, only WT restored HR efficiency, BRCA1/RAD51 foci, and olaparib resistance; S28A behaved like vector. Notably, ATM knockdown alone did not sensitize 22Rv1, C4-2 or DU145 to PARPi, consistent with prior reports of partial HR retention after ATM loss (PMID:28068672).
  • USP10 deubiquitinates and stabilizes TRMT10A. Co-IP MS identified USP10 as a TRMT10A interactor (table S1; PXD059301). USP10 knockdown reduced TRMT10A protein in HEK293T and 22Rv1, was rescued by MG-132, accelerated TRMT10A degradation in CHX chase, and increased TRMT10A polyubiquitination. USP10 and TRMT10A protein levels positively correlated in 5 paired PCa tumor/normal pairs and in a 46-patient IHC cohort (R² = 0.4232) (PMID:28068672).
  • USP10 inhibition phenocopies TRMT10A loss. USP10 shRNA reduced HR, increased NHEJ, and reduced BRCA1/RAD51 (but not γ-H2AX) foci in C4-2; clonogenic assays showed PARPi sensitization (PMID:28068672).
  • Spautin-1 degrades TRMT10A and synergizes with olaparib. spautin-1 (USP10 inhibitor) reduced TRMT10A protein dose-dependently in 22Rv1 (rescued by MG-132) without changing TRMT10A mRNA, decreased HR efficiency, and reduced BRCA1/RAD51 foci. Combined olaparib + spautin-1 produced HSA synergy scores >10 in 22Rv1 and C4-2; the strongest dose-response zone was olaparib 0.4–0.8 µM with spautin-1 2–4 µM. Synergy persisted in BRCA2-WT-overexpressing 22Rv1 (which corrects the line’s monoallelic BRCA2 T3033Nfs*11), arguing the effect is not driven by 22Rv1’s partial BRCA2 deficiency (PMID:28068672).
  • In vivo synergy in CDX and mCRPC PDX. In 22Rv1 subcutaneous CDX, the olaparib (50 mg/kg) + spautin-1 (20 mg/kg) combination outperformed either monotherapy (CI = 0.76). In PDX#546 and PDX#1092 (selected for high TRMT10A/USP10), the combination further suppressed tumor growth vs olaparib alone (PDX#1092: 42.9% additional volume reduction), with elevated cleaved caspase-3, cleaved PARP, γ-H2AX and reduced TRMT10A in tumor lysates (PMID:28068672).
  • Tolerability in C57BL/6 mice. No histopathological abnormalities (heart, brain, testis, spleen, liver, lung, kidney). Olaparib alone caused mild HGB/PLT/WBC/NEU reduction; combination did not exacerbate it. Bone-marrow lineage counts and splenic CD4/CD8/CD69 T-cell proportions were unchanged. Combination showed minimal cytotoxicity in non-tumor BEAS-2B lung epithelial cells (PMID:28068672).

Genes & alterations

  • TRMT10A — overexpressed in PCa/mCRPC; somatic amplification/deletion in 1.22–5.26% of mCRPC cohorts (no recurrent mutations). Functions as an ATM-substrate scaffold (Ser28) for BRCA1 recruitment, independent of its tRNA m1G9 methyltransferase activity. High expression predicts shorter OS and PARPi resistance (PMID:28068672).
  • USP10 — deubiquitinase that stabilizes TRMT10A; co-overexpressed with TRMT10A in mCRPC tissue; ~10% deep deletions in SU2C/PCF mCRPC (prad_su2c_2019) cohorts. Druggable with spautin-1 (PMID:28068672).
  • BRCA1 — recruited to DSBs via TRMT10A pSer28; recruitment (but not protein abundance) is impaired by TRMT10A or USP10 loss. Knockdown phenocopies TRMT10A loss for PARPi sensitivity (PMID:28068672).
  • BRCA2 — 22Rv1 carries a monoallelic loss-of-function T3033Nfs*11; BRCA2-WT rescue does not abolish olaparib + spautin-1 synergy, indicating the effect is independent of this background (PMID:28068672).
  • ATM — kinase responsible for TRMT10A Ser28 phosphorylation after IR; ATM knockdown reduces phospho-TRMT10A and HR but does not sensitize 22Rv1/C4-2/DU145 to PARPi alone (consistent with clinical heterogeneity of ATM-loss PARPi response) (PMID:28068672).
  • ATR — provides compensatory TRMT10A phosphorylation when ATM is impaired; TRMT10A loss also confers ATR-inhibitor sensitivity (PMID:28068672).
  • RAD51 — recruitment to DSBs requires TRMT10A pSer28-dependent BRCA1 loading; foci reduced by TRMT10A or USP10 loss and by spautin-1 (PMID:28068672).
  • PALB2 — TRMT10A expression correlated positively with PALB2 (along with BRCA1, BRCA2, RAD51) in TCGA prostate samples (PMID:28068672).
  • PARP1 / PARP2 — direct targets of olaparib/rucaparib/niraparib/talazoparib; PARP1 knockdown used as positive control for PARPi sensitivity (PMID:28068672).

Clinical implications

  • Predictive biomarker. TRMT10A IHC (and possibly USP10 co-staining) could stratify mCRPC patients without classical BRCA1/2 alterations for PARPi benefit; high TRMT10A predicts intrinsic PARPi resistance whereas low TRMT10A or USP10 deep-deletion tumors may be intrinsically PARPi-vulnerable (PMID:28068672).
  • Drug-combination strategy. The authors propose spautin-1 (USP10 inhibitor) + olaparib as a synthetic-lethal combination for BRCA1/2-WT mCRPC, supported by HSA synergy >10 in cell lines, CI = 0.76 in 22Rv1 CDX, and additive activity in two TRMT10A/USP10-high PDX models. The combination did not worsen olaparib’s hematologic toxicity in C57BL/6 mice (PMID:28068672).
  • Prognostic marker. High tumoral TRMT10A is associated with shorter overall survival (P = 0.014) and higher-grade (Gleason 8–10) disease in the 54-patient mCRPC cohort (PMID:28068672).
  • Mechanism distinct from existing PARPi-combination strategies. Unlike direct HR inhibitors (BRCA1/2, RAD51, ATR), the USP10–TRMT10A axis induces BRCAness indirectly by degrading a BRCA1-recruitment scaffold, potentially limiting compensatory HR pathway activation (PMID:28068672).

Limitations & open questions

  • Discrepancy with prior PARPi-sensitivity literature in PCa cell lines. The authors find DU145 and PC-3 more PARPi-sensitive than LNCaP and C4-2B, opposite to refs (36, 48); they attribute this to genetic drift and lab conditions but it should be independently reproduced (PMID:28068672).
  • No clinical-grade USP10 inhibitor. Spautin-1 is a tool compound; pharmacokinetics, on-target specificity, and translational dosing in humans remain to be established (PMID:28068672).
  • Cross-talk with established BRCA1 regulators (BARD1, RAP80–ABRAXAS) and DNA end-resection factors (CtIP, MRE11, EXO1) was not delineated; potential redundancy or context-specific compensation is acknowledged but unmeasured (PMID:28068672).
  • Patient cohort size and geography are limited (54 mCRPC for OS / IHC; 46 for USP10–TRMT10A correlation; 4 PDXs total, 2 advanced to therapy). Validation in external cohorts and prospective biomarker studies is required before clinical translation (PMID:28068672).
  • ATM-loss tumors are not addressed by this strategy. ATM knockdown alone did not sensitize PCa lines to PARPi, suggesting USP10–TRMT10A-axis targeting may not benefit ATM-deficient mCRPC; predictive biomarker development should account for this (PMID:28068672).
  • USP10 has multiple substrates (p53, Beclin-1, AMPK). While the authors argue HR effects are TRMT10A-dependent, off-target consequences of long-term USP10 inhibition in proliferating tissues need fuller characterization (PMID:28068672).
  • Raw-file metadata mismatch. The PMID/PMCID assigned to this extracted text (28068672 / PMC12513462) and the study_id: prad_cpcg_2017 field do not appear to correspond to the Yang et al. 2025 Sci. Adv. paper whose text is in the file. Downstream linking and provenance should be reconciled with the source PDF.

Citations from this paper used in the wiki

  • “TRMT10A is a tRNA methyltransferase responsible for catalyzing the formation of N1-methylguanosine (m1G) at position 9 (m1G9) in tRNAs … In our study, we identify TRMT10A as a crucial regulator in the HR pathway, independent of its known methyltransferase activity.” — Introduction.
  • “Genomic alteration analysis revealed that TRMT10A amplification and deletion occur in 1.22 to 5.26% of mCRPC cases across multiple cohorts, with no observed mutations.” — Results, Fig. 1.
  • “Combination of olaparib and spautin-1 resulted in highest single agent (HSA) synergy scores greater than 10 in 22Rv1 cells … The CI of 0.76 indicated an obvious synergy effect.” — Results, Fig. 7.
  • “ATM knockdown did not sensitize PCa cells (22Rv1, C4-2, and DU145) to PARPis (fig. S5, A to F), likely reflecting that ATM loss only partially impairs HR repair capacity.” — Results.
  • “Given the ~10% frequency of USP10 deletions observed in PCa cohorts (SU2C/PCF mCRPC), it becomes critical to consider genetic contexts that might influence therapeutic efficacy.” — Discussion.
  • “All patients in our study carried WT BRCA1/2, as confirmed through genomic profiling conducted before this analysis.” — Methods, Immunohistochemistry.

This page was processed by crosslinker on 2026-05-14.