Review Reports - BRCA1-Associated Protein 1 in Sarcoma

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This review provides a timely overview of the emerging role of BAP1 in soft tissue sarcoma. The manuscript covers the biology of BAP1, its alterations across sarcoma subtypes, and potential therapeutic implications. The topic is interesting and relevant, but several points should be addressed to improve the clarity and strength of the manuscript

The review is interesting, but the evidence for BAP1 varies greatly across sarcoma subtypes. A table summarizing the evidence for each subtype would improve the manuscript.
In Figure 3, please show the percentage of BAP1 loss for each sarcoma subtype, not only the number of cases. This would make comparisons easier.
It is not clear whether BAP1 copy-number loss always results in loss of BAP1 function. Please discuss this point in more detail.
The therapeutic section includes several potential treatments, but much of the evidence is still preclinical. Please make this limitation clearer.
The opposite relationship between BAP1 and PLK1 in sarcoma and lymphoid cells is interesting. A short explanation of this difference would strengthen the review.
Only a few sarcoma cases have been reported in patients with BAP1-TPDS. The discussion about changing surveillance recommendations should therefore be more cautious.
The immunosuppressive microenvironment in BAP1-deficient sarcoma is an important finding. Please provide a little more detail about the immune changes observed.
The manuscript has several grammatical errors and awkward sentences. Careful English editing is recommended to improve readability.

Author Response

The review is interesting, but the evidence for BAP1 varies greatly across sarcoma subtypes. A table summarizing the evidence for each subtype would improve the manuscript.

JH: Thank you for this suggestion. Table 1 and revised Figure 3 summarize the available evidence for BAP1 alterations in the indicated sarcoma subtypes.

In Figure 3, please show the percentage of BAP1 loss for each sarcoma subtype, not only the number of cases. This would make comparisons easier.

JH: Thank you for this suggestion. We have added the percentage of BAP1 loss for each sarcoma subtype to revised Figure 3.

It is not clear whether BAP1 copy-number loss always results in loss of BAP1 function. Please discuss this point in more detail.

JH: Thank you for this suggestion. We have clarified in Section 5.1 that copy-number loss can reduce BAP1 expression and contribute to functional loss, rather than invariably establishing complete inactivation. The added text reads: "Copy-number loss refers to genomic loss involving the BAP1 locus, which can reduce BAP1 expression and consequently contribute to loss of BAP1 function."

The therapeutic section includes several potential treatments, but much of the evidence is still preclinical. Please make this limitation clearer.

JH: Thank you for this suggestion. We have added the following limitation at the end of the therapeutic section: "However, these therapeutic strategies are currently supported primarily by preclinical studies, and their clinical relevance remains uncertain, highlighting the need for further validation in BAP1-deficient sarcoma."

The opposite relationship between BAP1 and PLK1 in sarcoma and lymphoid cells is interesting. A short explanation of this difference would strengthen the review.

JH: Thank you for this suggestion. We have added the following sentence to discuss this context-dependent observation while identifying the proposed mechanism as a hypothesis: "The mechanisms underlying this differential regulation of PLK1 by BAP1 across cell types remain unclear; one possibility is that additional TP53 mutations in sarcoma induce broader epigenetic reprogramming, although this hypothesis requires further investigation."

Only a few sarcoma cases have been reported in patients with BAP1-TPDS. The discussion about changing surveillance recommendations should therefore be more cautious.

JH: Thank you for this suggestion. The revised manuscript includes two additional reported sarcoma cases in patients with BAP1-TPDS. Nevertheless, the number of reported cases remains small and does not establish a basis for changing surveillance recommendations. Our aim is to draw attention to this emerging association and the need for further investigation.

The immunosuppressive microenvironment in BAP1-deficient sarcoma is an important finding. Please provide a little more detail about the immune changes observed.

JH: Thank you for this suggestion. We have added the following paragraph discussing the association between Bap1 loss and an immunosuppressive microenvironment in sarcoma: "Additionally, emerging evidence suggests that BAP1 loss can influence the tumor immune microenvironment in sarcoma. In our recent study, RNA-sequencing analysis showed that Bap1-deficient sarcomas exhibited downregulation of multiple immune-related pathways, including B-cell signaling, immune activation, and MHC class II-associated pathways. Multiplex immunohistochemistry and flow cytometry further demonstrated reduced T-cell infiltration and/or activity in Bap1-deficient tumors, including reduced CD3+ T-cell infiltration, decreased proliferation of CD8+ T cells, and an increased Treg-to-CD4+ T-cell ratio. Bap1 loss was also associated with altered myeloid composition, including increased PMN-MDSCs in the Trp53/Rb1/Bap1-deficient model compared with its Bap1-intact counterpart. Collectively, these findings support an association between BAP1 deficiency and an immunosuppressive tumor microenvironment in sarcoma, although the molecular mechanisms linking BAP1-dependent chromatin regulation to altered immune-cell recruitment and function remain to be determined [6]. Overall, the mechanistic study of BAP1 in regulating sarcoma initiation, progression, and immune suppression warrants future investigation."

The manuscript has several grammatical errors and awkward sentences. Careful English editing is recommended to improve readability.

JH: Thank you for this suggestion. We have edited the manuscript throughout to correct grammatical errors, improve sentence structure, and enhance readability.

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript reviews the roles of BRCA-associated protein BAP1 in soft tissue sarcomas. Presented manuscript’s central argument is plausible, but it contains several definite factual and citation errors, alongside interpretations that need qualification.

1) Mentioned in the text “>90% amino-acid identity” claim includes zebrafish, although reference 36 reports 66% human–zebrafish identity, versus 85% similarity

2) The described 2:2 PR-DUB structural study is Foglizzo et al., 2018. Reference 25 is an Ismail DNA-repair study, and the correct Foglizzo paper is absent from the bibliography (page 4, lines 135–139).

3) The manuscript narrates that PR-DUB removes excess H2Aub1 (page 4, lines 154-157), “thereby antagonizing” compaction. Bonnet instead found that excess H2Aub1 antagonizes compaction, and its removal preserves Polycomb repression.

4) Sturgill reports that 99% of copy-number losses were estimated single-copy losses, and many involved broad chromosome 3p regions; this does not establish complete BAP1 inactivation or BAP1-specific causality.

5) Activity in Bap1-deficient models is supported, but the primary study explicitly acknowledges that selective dependence relative to Bap1-intact tumors was not established across multiple systems.

6) The review should identify the cooperating Trp53 loss rather than imply that Bap1 loss alone was sufficient. It should also distinguish niraparib resistance in vivo from PARP-inhibitor sensitivity observed in vitro.

7) The subtype denominators total 290: 75+100+59+5675+100+59+56, whereas the stated cohort contains 244 samples. The deletion counts do total 30. The authors must reconcile overlapping categories, eligibility differences, or transcription errors; simply replacing 244 with 290 would not be justified.

Author Response

1) Mentioned in the text “>90% amino-acid identity” claim includes zebrafish, although reference 36 reports 66% human–zebrafish identity, versus 85% similarity

JH: Thank you for identifying this error. We misinterpreted the statement in the original publication and have corrected the text as follows: "BAP1 is highly conserved among vertebrates, with Xenopus Bap1 sharing 71% amino acid identity with human BAP1 and zebrafish Bap1 sharing 66% identity; notably, amino acid identity exceeds 90% within conserved functional regions, including the catalytic domain and binding motif."

JH: Thank you for identifying this citation error. We have corrected the citation in the revised manuscript.

JH: Thank you for pointing out this ambiguity. We agree that our original wording could be misinterpreted. Bonnet et al. showed that excessive H2Aub1 itself antagonizes PRC1-mediated chromatin compaction by interfering with nucleosome stacking and chromatin fiber folding. PR-DUB acts as a rheostat by removing excessive H2Aub1, thereby preventing this antagonistic effect and preserving Polycomb-mediated chromatin repression. We have revised the text accordingly: "Bonnet et al. demonstrated that excessive H2Aub1 following PR-DUB catalytic loss increases accessibility at Polycomb targets and impairs nucleosome stacking and chromatin-fiber folding in vitro, resulting in target-gene derepression despite retained H3K27me3 and Polycomb binding."

JH: Thank you for this important clarification. We agree that BAP1 copy-number loss should not be interpreted as evidence of complete BAP1 inactivation or BAP1-specific causality. As reported by Sturgill et al., 99% of BAP1 copy-number losses were estimated to be single-copy losses, and many involved broad regions of chromosome 3p encompassing additional tumor suppressor genes. Nevertheless, BAP1 copy-number loss was associated with reduced BAP1 RNA expression, and a subset of these events was also associated with reduced inferred BAP1 activity. We have therefore revised the text to distinguish copy-number loss from complete functional inactivation and acknowledge the limitations of broad 3p deletions: "Thus, BAP1-deficient biology cannot be defined solely by mutation status, an important consideration for biomarker development and for sarcoma, where complex copy-number alterations are common. Nevertheless, BAP1 copy-number loss should be interpreted cautiously: such events may represent single-copy loss rather than complete BAP1 inactivation and can occur as part of broader chromosome 3p deletions involving additional tumor suppressor genes. Therefore, although BAP1 copy-number loss is associated with reduced BAP1 expression, copy-number loss alone does not establish BAP1-specific functional inactivation or causality."

5) Activity in Bap1-deficient models is supported, but the primary study explicitly acknowledges that selective dependence relative to Bap1-intact tumors was not established across multiple systems.

JH: Thank you for this important clarification. We agree that, although activity has been demonstrated in Bap1-deficient models, the available primary study did not establish a selective dependence relative to Bap1-intact tumors across multiple experimental systems. Research on BAP1 function in sarcoma remains at an early stage, with relatively few mechanistic studies and limited cross-model validation. We have therefore revised the manuscript to acknowledge this limitation and to avoid implying that the observed therapeutic activity represents a broadly established BAP1-selective vulnerability. We also added discussion of our recent findings comparing Bap1-deficient sarcomas with their Bap1-intact counterparts. These data show reduced immune-related pathway activity, impaired T-cell infiltration and/or proliferation, an increased Treg-to-CD4+ T-cell ratio, and altered myeloid composition, including increased PMN-MDSCs in the Trp53/Rb1/Bap1-deficient model. The newly added text concludes that these findings support an association between BAP1 deficiency and an immunosuppressive tumor microenvironment, while emphasizing that the underlying mechanisms and the generalizability of these effects require further investigation.

JH: Thank you for this important clarification. We agree that the original wording could imply that Bap1 loss alone was sufficient for sarcomagenesis. In the reported study, Trp53 loss was present in two of the three UPS models, and the same study demonstrated that p53 loss was required for sarcoma initiation, indicating that Bap1 loss cooperated with Trp53 deficiency rather than acting as an independent initiating event. We have also clarified the context-dependent PARP-inhibitor findings by the lack of response to niraparib in vivo.

7) The subtype denominators total 290: 75+100+59+5675+100+59+5675+100+59+56, whereas the stated cohort contains 244 samples. The deletion counts do total 30. The authors must reconcile overlapping categories, eligibility differences, or transcription errors; simply replacing 244 with 290 would not be justified.

JH: Thank you for identifying this discrepancy. We agree that simply replacing the total cohort size of 244 with 290 would not be justified. The correct overall cohort size is 244 samples. The previously reported subtype denominators contained transcription errors and did not consistently account for samples that were excluded from subtype-level analysis because of inadequate data quality or an inability to complete the relevant analysis. We have re-evaluated the dataset and corrected the number of evaluable samples for each sarcoma subtype. The revised subtype denominators are now consistent with the 244-sample cohort, while the total number of BAP1 deletion cases remains 30. These corrections have been incorporated into revised Figure 3 and the corresponding main text.

Reviewer 3 Report

Comments and Suggestions for Authors

The abstract defines BAP1 as a critical regulator of chromatin DNA repair and cell cycle but does not specify which sarcoma subtypes have the strongest evidence. Adding subtype-level effect sizes would sharpen the review and prevent overgeneralization.

The introduction claims BAP1 loss is a functional driver of sarcoma based largely on one recent CRISPR study. This is important but should be balanced against older genomic data and the possibility of subtype-restricted effects.
The epidemiology section cites 2026 projections and SEER survival yet does not separate soft tissue sarcoma from bone sarcoma or GIST. Clarifying inclusion criteria is essential because incidence and survival vary widely.
The statement that STS accounts for less than 1 percent of adult cancers and 15 to 21 percent of pediatric malignancies needs a clear source definition. Some registries include GIST while others exclude it, changing the denominator.
The review lists more than 75 histologic subtypes without citing the WHO classification version. This matters because molecularly defined entities and reclassified tumors affect BAP1 association claims.
BAP1 domain architecture is described with precise amino acid boundaries but the text does not state which isoform is used. Isoform choice can shift UCH ULD NLS and interaction domain coordinates.
The claim of over 90 percent amino acid identity across vertebrates is broad and should be supported by alignment data. Human and mouse conservation alone cannot establish the same degree for zebrafish or dog.
The BRCA1 interaction section contains a likely nomenclature error where BRAD1 appears instead of BARD1. Correcting this is important because BARD1 is the obligate BRCA1 partner in DNA repair.
The review cites BAP1 tumor suppression independent of BRCA1 but does not reconcile this with BAP1 BRCA1 BARD1 complex data. A clearer model of separable versus overlapping functions would help.
PR-DUB composition is presented as a fixed mammalian complex but ASXL paralogs FOXK1/2 OGT and HCF1 may have context-dependent stoichiometry. Discussing assembly heterogeneity would deepen the mechanistic section.
The Drosophila Calypso ASX 2 to 2 complex is described as bidentate but the mammalian relevance remains uncertain. The review should distinguish structural findings from functional conservation.
The discussion of PR-DUB and PRC2 minimal overlap conflicts with other cited work showing H3K27me3 accumulation after BAP1 loss. This apparent contradiction needs explicit reconciliation.
BAP1 loss is said to cause broad H2AK119ub1 and H3K27me3 accumulation but causality versus secondary compensation is not resolved. Distinguishing direct deubiquitination effects from downstream chromatin remodeling is critical.
Pan-cancer analysis is used to argue that copy-number loss dominates BAP1 alteration, but sarcoma TCGA cohorts are small. Detection thresholds purity and ploidy effects should be addressed before extending this conclusion to STS.
Germline versus somatic BAP1 loss is described as clinically distinct in uveal melanoma, yet ascertainment and treatment era may confound this. The review should discuss prospective validation and mechanism.
Kury-Isidor syndrome is introduced as a neurodevelopmental disorder but its relevance to sarcoma is not developed. A short genotype-phenotype framework would clarify why these variants matter to cancer readers.
The review correctly notes that sarcoma is often grouped as other cancers, but this critique would be stronger with a systematic search strategy. Without methods the reader cannot judge selection bias in cited sarcoma cases.
Table 1 includes only three germline sarcoma cases and omits somatic BAP1 alterations. Expanding the table to include somatic copy-number loss mutation type and subtype would better reflect the evidence base.
Both MPNST cases involve splice-site germline variants and may be confounded by NF1 or other predisposition. The review should discuss differential diagnosis and co-occurring drivers in MPNST.
Mullerian adenosarcoma is cited as 25 percent BAP1 loss, but high-grade and low-grade tumors may differ biologically. Subtype stratification and independent validation are needed before calling BAP1 a driver.
The TCGA STS distribution reports 30 of 244 BAP1 copy-number losses but does not test enrichment versus other tumor suppressors. Statistical enrichment and co-deletion patterns would clarify whether BAP1 loss is selected or passenger.
The osteosarcoma functional study is limited to in vitro assays and a proposed miR-125 PI3K Akt axis. The review should note the lack of in vivo validation and potential cell line-specific effects.
The recent CRISPR UPS study provides strong in vivo evidence but BAP1 loss may cooperate with other mutations. Discussing whether Bap1 loss is sufficient or requires permissive chromatin and immune contexts is important.
PLK1 dependency is presented as a therapeutic opportunity but the BAP1 PLK1 relationship is context dependent in lymphoid cells. The review should propose mechanisms for sarcoma-specific PLK1 upregulation.
PARP inhibitor insensitivity in BAP1-deficient sarcoma challenges the assumption of homologous recombination deficiency. This should be integrated with DNA damage response biomarker discussions rather than treated as a footnote.
Epigenetic therapies are proposed for BAP1 loss but sarcoma-specific preclinical data remain sparse. Linking EZH2 HDAC and BET inhibitor sensitivity to PR-DUB loss and Polycomb imbalance would strengthen the therapeutic rationale.
The biomarker section recommends integrating DNA copy number RNA and immunohistochemistry but does not define cutoffs. Reproducibility and validation across sarcoma subtypes require standardized scoring criteria.
Germline surveillance recommendations are appropriately cautious but could be more actionable. The review should propose registry endpoints and triggers for sarcoma-focused imaging without overstating evidence.
Future directions mention patient-derived models and prospective profiling but omit equity and age considerations. Sarcoma spans pediatric and adult populations and trial access differences may shape BAP1 research.
The conclusion states BAP1 is emerging as a biomarker in numerous sarcoma subtypes, which may be too broad given current data. Specifying UPS MFS leiomyosarcoma and MPNST as leading candidates would better match the evidence.

Author Response

The abstract defines BAP1 as a critical regulator of chromatin DNA repair and cell cycle but does not specify which sarcoma subtypes have the strongest evidence. Adding subtype-level effect sizes would sharpen the review and prevent overgeneralization.

JH: Thank you for this important suggestion. We agree that the evidence for BAP1 varies across sarcoma subtypes. Because investigation of BAP1 in sarcoma is still at an early stage, subtype-level functional effect sizes are not yet consistently available. Our intent is therefore not to generalize the role of BAP1 across all sarcomas, but to summarize the emerging evidence and highlight areas requiring further study. Revised Figure 3 presents the frequency of BAP1 copy-number alterations across individual sarcoma subtypes, and we have updated the figure to include the full spectrum of somatic BAP1 copy-number alterations in each subtype. This revised subtype-level presentation more clearly shows the distribution of the available evidence and helps avoid overgeneralization.

The introduction claims BAP1 loss is a functional driver of sarcoma based largely on one recent CRISPR study. This is important but should be balanced against older genomic data and the possibility of subtype-restricted effects.

JH: Thank you for this important comment. We agree that the recent CRISPR study provides important functional evidence but, as a single mechanistic study, is not sufficient to establish BAP1 loss as a universal driver across sarcoma. Research on BAP1 function in sarcoma remains at an early stage, and our intent is not to generalize this conclusion to all subtypes. We have revised the Introduction to balance these findings with earlier genomic studies showing variable frequencies and patterns of BAP1 alterations across sarcoma subtypes and to clarify that the biological effects of BAP1 loss may be subtype- and context-dependent. Consistent with this revision, updated Figure 3 presents the full spectrum of somatic BAP1 copy-number alterations across individual sarcoma subtypes, providing a more cautious subtype-level view of the current evidence.

The epidemiology section cites 2026 projections and SEER survival yet does not separate soft tissue sarcoma from bone sarcoma or GIST. Clarifying inclusion criteria is essential because incidence and survival vary widely.

JH: We thank the reviewer for identifying this ambiguity. We have clarified that the 2026 incidence and mortality projections cited from Siegel et al. refer specifically to soft tissue sarcoma and do not include primary bone sarcomas, which are reported separately. We have also clarified that the quoted five-year relative survival is an aggregate SEER estimate for the category “soft tissue including heart cancer” rather than a uniform survival estimate for all sarcoma subtypes.

The statement that STS accounts for less than 1 percent of adult cancers and 15 to 21 percent of pediatric malignancies needs a clear source definition. Some registries include GIST while others exclude it, changing the denominator.

JH: Thank you for highlighting the need for clearer source definitions. We have corrected the pediatric estimate from 15–21% to up to 7% of cancers in children and adolescents younger than 20 years, citing Weiss et al. (Hematology/Oncology Clinics of North America, 2025). The adult estimate of less than 1% is supported separately by the adult STS literature, including Paudel et al. (Cancers, 2025). These percentages refer to STS rather than bone and soft tissue sarcomas combined.

The review lists more than 75 histologic subtypes without citing the WHO classification version. This matters because molecularly defined entities and reclassified tumors affect BAP1 association claims.

JH: Thank you for this suggestion. We have added the WHO classification citation in the revised manuscript.

BAP1 domain architecture is described with precise amino acid boundaries but the text does not state which isoform is used. Isoform choice can shift UCH ULD NLS and interaction domain coordinates.

JH: Thank you for this suggestion. We have specified the BAP1 isoform in the main text and updated Figure 1 accordingly.

The claim of over 90 percent amino acid identity across vertebrates is broad and should be supported by alignment data. Human and mouse conservation alone cannot establish the same degree for zebrafish or dog.

JH: Thank you for highlighting this overgeneralization. We have replaced the broad claim of >90% identity across vertebrates with species-specific values from published alignments: Xenopus and zebrafish Bap1 share 71% and 66% identity with human BAP1, respectively, whereas >90% identity applies to specific conserved functional regions [19]. We also added a qualified discussion of canine BAP1: entries A0A8I3SCQ2 (752 amino acids) and A0A8I3QAS2 (729 amino acids) both contain an annotated UCH domain at residues 4-235. The revised text explains the inclusion of dogs based on their established use in comparative sarcoma research [28–30] and acknowledges that comprehensive annotation of canine BAP1 requires further investigation.

The BRCA1 interaction section contains a likely nomenclature error where BRAD1 appears instead of BARD1. Correcting this is important because BARD1 is the obligate BRCA1 partner in DNA repair.

JH: Thanks for pointing it out. We have corrected this typo.

The review cites BAP1 tumor suppression independent of BRCA1 but does not reconcile this with BAP1 BRCA1 BARD1 complex data. A clearer model of separable versus overlapping functions would help.

JH: Thank you for highlighting this distinction. Ventii et al. [34] provided evidence that BAP1-mediated growth suppression can occur independently of wild-type BRCA1, but did not establish whether this effect is also independent of BARD1. Thus, these findings do not exclude overlapping functions involving BRCA1/BARD1. BRCA1-independent growth suppression and BRCA1/BARD1-associated functions are not mutually exclusive, and their mechanistic overlap remains incompletely resolved. We have added the following qualification: "However, whether this effect also occurs independently of BARD1 was not established in that study."

PR-DUB composition is presented as a fixed mammalian complex but ASXL paralogs FOXK1/2 OGT and HCF1 may have context-dependent stoichiometry. Discussing assembly heterogeneity would deepen the mechanistic section.

JH: Thank you for this suggestion. We have clarified that mammalian PR-DUB represents alternative BAP1-ASXL assemblies rather than a fixed complex containing all listed paralogs. We now discuss mutually exclusive ASXL1/ASXL2 complexes, ASXL3-dependent BRD4 association, and FOXK-dependent chromatin recruitment and its regulation by O-GlcNAcylation. We also acknowledge that the occupancy and stoichiometry of associated factors, including OGT and HCF1, remain incompletely defined. The revised text reads: "Rather than constituting a single invariant complex, mammalian PR-DUB comprises alternative BAP1-ASXL assemblies associated with regulatory factors such as FOXK1/2, OGT, and HCF1. BAP1 forms mutually exclusive complexes with ASXL1 or ASXL2 [24], whereas ASXL3 can bridge BAP1 to BRD4 in ASCL1-high small-cell lung cancer [27]. Kolovos et al. demonstrated ASXL- and FOXK1/2-dependent recruitment of BAP1 to chromatin and maintenance of target-gene expression [38]."

The Drosophila Calypso ASX 2 to 2 complex is described as bidentate but the mammalian relevance remains uncertain. The review should distinguish structural findings from functional conservation.

JH: Thank you for this suggestion. We have added the following qualification to distinguish structural findings in Drosophila from mammalian functional conservation: "However, these structural and biochemical findings concern the Drosophila Calypso-ASX complex and do not establish that mammalian BAP1-ASXL complexes adopt the same 2:2 bidentate architecture or require this assembly for chromatin-associated activity. Conservation of H2A deubiquitinase function should therefore be distinguished from conservation of complex architecture."

The discussion of PR-DUB and PRC2 minimal overlap conflicts with other cited work showing H3K27me3 accumulation after BAP1 loss. This apparent contradiction needs explicit reconciliation.

JH: Thank you for this comment. We have revised the text to clarify the distinction between limited overlap and indirect functional coupling: "However, this does not exclude indirect coupling through H2AK119ub1 regulation. Conway et al. showed that BAP1 loss in mouse embryonic stem cells causes widespread H2AK119ub1 accumulation, PRC2 redistribution, and diffuse H3K27me3 accumulation accompanied by increased global chromatin condensation. Nevertheless, redistribution of PRC2 away from canonical Polycomb targets weakens their repression, permitting inappropriate target-gene activation [41]."

BAP1 loss is said to cause broad H2AK119ub1 and H3K27me3 accumulation but causality versus secondary compensation is not resolved. Distinguishing direct deubiquitination effects from downstream chromatin remodeling is critical.

JH: Thank you for this comment. We have revised the discussion to distinguish H2AK119ub1 regulation from downstream changes in Polycomb localization and chromatin organization: "However, this does not exclude indirect coupling through H2AK119ub1 regulation. Conway et al. showed that BAP1 loss in mouse embryonic stem cells causes widespread H2AK119ub1 accumulation, PRC2 redistribution, and diffuse H3K27me3 accumulation accompanied by increased global chromatin condensation. Nevertheless, redistribution of PRC2 away from canonical Polycomb targets weakens their repression, permitting inappropriate target-gene activation [41]. In Drosophila, Bonnet et al. demonstrated that excessive H2Aub1 following PR-DUB catalytic loss increases accessibility at Polycomb targets and impairs nucleosome stacking and chromatin-fiber folding in vitro, resulting in target-gene derepression despite retained H3K27me3 and Polycomb binding [42]. Thus, both studies support a role for PR-DUB in maintaining Polycomb target-gene silencing, although they identify distinct mechanisms and consequences for chromatin organization."

Pan-cancer analysis is used to argue that copy-number loss dominates BAP1 alteration, but sarcoma TCGA cohorts are small. Detection thresholds purity and ploidy effects should be addressed before extending this conclusion to STS.

JH: Thank you for highlighting these considerations. The reported alteration frequencies derive from processed TCGA data analyzed using the original study’s variant- and copy-number-calling criteria [5], rather than unprocessed genomic signals. Nevertheless, detection thresholds and residual uncertainty related to tumor purity and ploidy can influence alteration estimates. We also acknowledge that the limited size and histologic heterogeneity of the sarcoma cohort constrain generalization. Accordingly, the predominance of copy-number loss is presented as a finding within the analyzed pan-cancer dataset, not as definitive evidence that it is the dominant mechanism of BAP1 alteration across all STS subtypes.

Germline versus somatic BAP1 loss is described as clinically distinct in uveal melanoma, yet ascertainment and treatment era may confound this. The review should discuss prospective validation and mechanism.

JH: Thank you for highlighting these limitations. We agree that the reported difference in metastatic risk should not be interpreted as definitive evidence of intrinsically different tumor behavior. We propose clarifying that the small germline subgroup and potential differences in ascertainment, surveillance, follow-up, and treatment era may influence this association. Prospective, multicenter validation with standardized molecular classification and adjustment for these factors is needed. We updated in the revision: “However, the small germline subgroup and potential ascertainment, surveillance, follow-up, and treatment-era differences limit causal interpretation. Prospective, multicenter studies accounting for tumor stage, molecular risk features, surveillance, and treatment are needed to establish whether germline status independently influences outcome. Although BAP1 regulates lineage commitment and differentiation-associated chromatin states, whether the timing of biallelic inactivation or the surrounding tissue context explains the reported clinical difference remains unknown.”.

Kury-Isidor syndrome is introduced as a neurodevelopmental disorder but its relevance to sarcoma is not developed. A short genotype-phenotype framework would clarify why these variants matter to cancer readers.

JH: Thank you for this suggestion. We propose clarifying the genotype-phenotype framework by distinguishing the de novo missense variants associated with Küry-Isidor syndrome from the broader pathogenic variant spectrum associated with BAP1 tumor predisposition syndrome. Functional studies demonstrate impaired H2A deubiquitination and altered chromatin regulation for several syndrome-associated variants. A specific association between Küry-Isidor syndrome and sarcoma remains unestablished. We also updated in the revision: “From a genotype-phenotype perspective, Küry-Isidor syndrome is associated with specific de novo heterozygous missense variants, whereas BAP1 tumor predisposition syndrome encompasses truncating, splice-site, and pathogenic missense variants. Several Küry-Isidor-associated variants impair H2A deubiquitination and alter chromatin regulation, linking developmental dysfunction to molecular processes also relevant to cancer. However, these findings do not establish equivalent tumor-predisposing effects. For sarcoma research, the distinction underscores the need for variant-specific functional assessment rather than assuming that all pathogenic germline BAP1 variants confer the same cancer risk.”.

The review correctly notes that sarcoma is often grouped as other cancers, but this critique would be stronger with a systematic search strategy. Without methods the reader cannot judge selection bias in cited sarcoma cases.

JH: Thank you for this suggestion. We used the Mastermind Genomic Intelligence Platform to identify reported germline and somatic BAP1 variants associated with sarcoma, complemented by keyword-based literature searches using ‘BRCA1-associated protein 1’, ‘BAP1’, and ‘sarcoma.’ We cross-compared these results with the Mastermind findings to improve coverage and capture recent publications. Although this complementary approach aimed to minimize omissions, this remains a narrative review, and we cannot guarantee that every relevant study was identified or exclude selection and publication bias.

Table 1 includes only three germline sarcoma cases and omits somatic BAP1 alterations. Expanding the table to include somatic copy-number loss mutation type and subtype would better reflect the evidence base.

JH: Thank you for this suggestion. Table 1 summarizes the reported germline BAP1 variants associated with sarcoma, including a newly identified case. Figure 3 provides the complementary summary of somatic BAP1 copy-number alterations across the analyzed sarcoma subtypes.

Both MPNST cases involve splice-site germline variants and may be confounded by NF1 or other predisposition. The review should discuss differential diagnosis and co-occurring drivers in MPNST.

JH: Thank you for raising this concern. NF1 mutation status was not reported in the publications we reviewed. Consequently, these reports do not allow us to exclude NF1-associated predisposition or assess its contribution to the observed MPNST cases.

Mullerian adenosarcoma is cited as 25 percent BAP1 loss, but high-grade and low-grade tumors may differ biologically. Subtype stratification and independent validation are needed before calling BAP1 a driver.

JH: Thank you for this comment. Our intent is to report the observed frequency of BAP1 copy-number loss in this subtype, not to establish BAP1 as a driver. This descriptive observation does not resolve potential differences between high-grade and low-grade tumors; functional studies and independent validation would be needed to support a driver role.

The TCGA STS distribution reports 30 of 244 BAP1 copy-number losses but does not test enrichment versus other tumor suppressors. Statistical enrichment and co-deletion patterns would clarify whether BAP1 loss is selected or passenger.

JH: Thank you for this comment. We agree that BAP1 copy-number loss may co-occur with other driver alterations. The analysis presented in this review is descriptive and reports the frequency of BAP1 copy-number alterations in the analyzed STS cohort; it does not establish statistical enrichment, selection, or BAP1-specific causality. Although our recent JCI Insight study provides functional evidence in experimental models, further analyses of co-deletion patterns and functional studies are needed to distinguish driver effects from passenger alterations across sarcoma subtypes.

The osteosarcoma functional study is limited to in vitro assays and a proposed miR-125 PI3K Akt axis. The review should note the lack of in vivo validation and potential cell line-specific effects.

JH: Thank you for this suggestion. We have highlighted this limitation in the revised manuscript: "Although this study provided initial in vitro evidence that BAP1 may suppress osteosarcoma cell growth and migration through a proposed miR-125-PI3K/Akt axis, the lack of in vivo validation and potential cell line-specific effects limit the generalizability of these findings."

The recent CRISPR UPS study provides strong in vivo evidence but BAP1 loss may cooperate with other mutations. Discussing whether Bap1 loss is sufficient or requires permissive chromatin and immune contexts is important.

JH: Thank you for this important comment. We have clarified that sarcoma development in the reported models required concomitant Trp53 loss. Thus, the in vivo evidence supports Bap1 loss as a cooperating driver for UPS initiation in a Trp53-deficient setting. We also added a paragraph discussion of the role of Bap1 in sarcoma immune microenvironment: “Importantly, sarcoma development in these models required concomitant Trp53 loss, supporting cooperation between Bap1 and Trp53 loss rather than demonstrating that Bap1 loss alone is sufficient to initiate UPS.” and “Additionally, emerging evidence suggests that BAP1 loss can influence the tumor immune microenvironment in sarcoma. In our recent study, RNA-sequencing analysis showed that Bap1-deficient sarcomas exhibited downregulation of multiple immune-related pathways, including B-cell signaling, immune activation, and MHC class II-associated pathways. Multiplex immunohistochemistry and flow cytometry further demonstrated reduced T-cell infiltration and/or activity in Bap1-deficient tumors, including reduced CD3+ T-cell infiltration, decreased proliferation of CD8+ T cells, and an increased Treg-to-CD4+ T-cell ratio. Bap1 loss was also associated with altered myeloid composition, including increased PMN-MDSCs in the Trp53/Rb1/Bap1-deficient model compared with its Bap1-intact counterpart. Collectively, these findings support an association between BAP1 deficiency and an immunosuppressive tumor microenvironment in sarcoma, although the molecular mechanisms linking BAP1-dependent chromatin regulation to altered immune-cell recruitment and function remain to be determined [5,6,45,46,60].”.

PLK1 dependency is presented as a therapeutic opportunity but the BAP1 PLK1 relationship is context dependent in lymphoid cells. The review should propose mechanisms for sarcoma-specific PLK1 upregulation.

JH: The underlying mechanism is still unknown. We have touched on this topic with one additional sentence: “The mechanisms underlying this differential regulation of PLK1 by BAP1 across cell types remain unclear; one possibility is that additional TP53 mutations in sarcoma induce broader epigenetic reprogramming, although this hypothesis requires further investigation.”.

PARP inhibitor insensitivity in BAP1-deficient sarcoma challenges the assumption of homologous recombination deficiency. This should be integrated with DNA damage response biomarker discussions rather than treated as a footnote.

JH: Thank you for this important comment. We acknowledge that the limited response to PARP inhibition in BAP1-deficient sarcoma challenges the assumption that BAP1 loss necessarily confers homologous recombination deficiency or predicts PARP inhibitor sensitivity. We have retained this discussion in its current position in Section 6, ‘Clinical and Therapeutic Implications,’ because it directly contextualizes the therapeutic relevance of BAP1-associated DNA damage responses.

Epigenetic therapies are proposed for BAP1 loss but sarcoma-specific preclinical data remain sparse. Linking EZH2 HDAC and BET inhibitor sensitivity to PR-DUB loss and Polycomb imbalance would strengthen the therapeutic rationale.

JH: Thank you for this valuable suggestion. We agree that linking candidate epigenetic therapies to BAP1-dependent chromatin regulation strengthens the therapeutic rationale. We have expanded Section 6 to summarize evidence from other cancers connecting BAP1 loss with EZH2-dependent Polycomb repression, HDAC inhibitor-mediated reversal of transcriptional and differentiation abnormalities, and increased BET inhibitor sensitivity linked to H2AK119ub1 regulation and BRD4-MYC-dependent transcription. We also explicitly acknowledge that sarcoma-specific preclinical evidence remains sparse and that these findings provide a rationale for investigation, not established predictive biomarkers or validated treatments for BAP1-deficient sarcoma: “Evidence from other cancers provides a mechanistic rationale for investigating epigenetic therapies in BAP1-deficient sarcoma. Loss of BAP1-mediated PR-DUB activity disrupts H2AK119ub1 homeostasis and can alter Polycomb-associated transcriptional regulation, although the resulting therapeutic dependencies are context dependent. In mesothelioma models, BAP1 loss was associated with increased EZH2 expression and H3K27me3-mediated repression, and genetic or pharmacologic EZH2 inhibition suppressed BAP1-deficient tumor growth, supporting a vulnerability linked to altered Polycomb regulation [76]. HDAC inhibition offers a complementary approach to reversing BAP1-loss-associated transcriptional repression: in uveal melanoma, a functional rescue screen identified quisinostat, which reversed BAP1-deficiency-associated phenotypes and inhibited BAP1-mutant tumor growth in mice [77]. These findings support targeting the altered differentiation and transcriptional states associated with BAP1 loss. A more direct link to H2AK119ub1 regulation has been demonstrated for bromodomain and extraterminal (BET) inhibitors in melanoma and clear-cell renal cell carcinoma models [75]. BAP1 loss enhanced BET inhibitor sensitivity, whereas BAP1 deubiquitinase activity reduced sensitivity; conversely, expression of the H2AK119 ubiquitin ligases RING1A or RING1B increased sensitivity. Mechanistically, BET inhibition preferentially suppressed BRD4-dependent proliferative transcription, particularly MYC, in BAP1-deficient cells. Collectively, these studies connect BAP1-associated chromatin dysregulation to potentially targetable dependencies involving Polycomb repression, histone deacetylation, and transcriptional maintenance. Nevertheless, sarcoma-specific preclinical evidence remains sparse, and neither BAP1 loss nor PR-DUB disruption alone should be assumed to predict sensitivity to these agents. To address this gap, we are systematically evaluating epigenetic inhibitors in BAP1-deficient sarcoma models to identify potential therapeutic vulnerabilities and determine whether BAP1 status influences treatment response.”.

The biomarker section recommends integrating DNA copy number RNA and immunohistochemistry but does not define cutoffs. Reproducibility and validation across sarcoma subtypes require standardized scoring criteria.

JH: Thank you for this important comment. We agree that standardized scoring criteria and validated cutoffs are essential for reproducible assessment of BAP1 status across sarcoma subtypes. We have expanded the biomarker discussion to acknowledge this requirement and highlight how clinical experience with BAP1 immunohistochemistry in other cancers, particularly mesothelioma, can inform assay development in sarcoma. Specifically, complete loss of tumor-cell nuclear staining should be distinguished from reduced or heterogeneous expression and interpreted only in the presence of positive internal controls. We also emphasize the need for standardized staining protocols, interobserver and interlaboratory reproducibility testing, and correlation with genomic findings. Rather than proposing arbitrary thresholds, we identify the establishment and independent validation of subtype-specific DNA copy-number, RNA-expression, and immunohistochemical criteria as prerequisites for clinical implementation.

Germline surveillance recommendations are appropriately cautious but could be more actionable. The review should propose registry endpoints and triggers for sarcoma-focused imaging without overstating evidence.

JH: Thank you for this constructive suggestion. We agree that the discussion would benefit from more actionable research priorities without implying that sarcoma-specific surveillance is already evidence based. We have expanded the Future Directions section to propose prospective registry endpoints, including sarcoma incidence per carrier-year, histologic subtype, age and stage at diagnosis, mode of detection, diagnostic delay, and imaging-related false positives and downstream procedures. We also distinguish symptom-triggered diagnostic evaluation from routine screening: an unexplained enlarging soft-tissue mass, persistent focal pain, or new focal neurologic symptoms should prompt clinical assessment and consideration of site-directed imaging. These represent general clinical warning signs rather than validated BAP1-specific imaging criteria. Routine sarcoma-focused imaging in asymptomatic carriers remains a question for prospective investigation.

Future directions mention patient-derived models and prospective profiling but omit equity and age considerations. Sarcoma spans pediatric and adult populations and trial access differences may shape BAP1 research.

JH: Thank you for highlighting this important omission. We agree that age representation and equitable access should be explicit considerations in future BAP1-focused sarcoma research. We have expanded the Future Directions section to recommend coordinated recruitment across pediatric and adult sarcoma programs, representation of children, adolescents and young adults, and older adults in prospective profiling and patient-derived model collections, and age-stratified analyses that account for histologic subtype and treatment history.

The conclusion states BAP1 is emerging as a biomarker in numerous sarcoma subtypes, which may be too broad given current data. Specifying UPS MFS leiomyosarcoma and MPNST as leading candidates would better match the evidence.

JH: Thank you for this suggestion. To avoid overgeneralization, we have revised the conclusion to identify eRMS, UPS, MFS, leiomyosarcoma, and MPNST as leading candidates for further investigation rather than implying that BAP1 is an established biomarker across numerous sarcoma subtypes.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I have no further concerns, and I recommend the paper for acceptance in its current form.