Pharmacogenomics Applied to Acute Leukemias: Identifying Clinically Relevant Genetic Variants

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This review article compiles pharmacogenomic variants from the ClinPGx/PharmGKB database associated with chemotherapy response and toxicity in Acute Myeloid Leukemia (AML) and Acute Lymphoblastic Leukemia (ALL). The authors identified 24 variants in AML and 57 in ALL, with the ABCB1 gene being the most frequently implicated in both leukemia types. The study highlights the prevalence of synonymous and nonsynonymous variants and their associations with drug efficacy and adverse effects such as myelotoxicity, hepatotoxicity, and cardiotoxicity. The authors advocate for pre-treatment genotyping to enable personalized therapy and reduce treatment-related toxicity. Collectively, the manuscript provides a useful compilation of pharmacogenomic variants in acute leukemias and underscores the potential of genetic profiling to improve treatment outcomes. However, the study is largely descriptive and would benefit from deeper mechanistic insights, broader discussion of therapeutic implications, and greater emphasis on the functional consequences of synonymous and non-coding variants. The following specific comments are offered to strengthen the paper:

(1) The manuscript would benefit from a discussion of how identified SNVs—particularly synonymous or non-coding variants—might influence RNA modifications (e.g., m6A methylation) and subsequent gene expression or splicing, which could mechanistically explain their impact on drug response.

(2) The clinical significance of the findings should be elaborated beyond prediction of drug response. For example, how might these variants inform the development of combination therapies or adjuvant strategies to overcome resistance?

(3) The authors should discuss potential pharmacological interventions that could counteract the effects of high-risk variants (e.g., P-gp inhibitors for ABCB1-mediated resistance, or allopurinol for NUDT15-related thiopurine toxicity).

(4) The role of synonymous and intronic variants in regulating gene expression or mRNA stability is underemphasized. The authors should expand on how these “silent” or non-coding variants may contribute to chemoresistance through non-canonical mechanisms.

(5) The authors should highlight the importance of incorporating SNV profiling—not just gene expression—into routine genomic analyses (e.g., RNA-seq, whole-genome sequencing) to better predict treatment outcomes and toxicity risks.

(6) The discussion on ABCB1 is thorough, but other clinically relevant genes (e.g., TPMT, NUDT15) deserve more detailed mechanistic and clinical context, especially given their high level of evidence (LOE 1A).

(7) The manuscript should address the limitations of using a database like PharmGKB, such as potential population biases, underrepresentation of rare variants, and the predominance of Level 3 evidence for many associations.

(8) A forward-looking perspective on how these findings could be translated into clinical trials or integrated into existing treatment protocols (e.g., via point-of-care genotyping) would enhance the impact of the conclusion.

Author Response

Dear reviewer, my co-authors and I would like to thank you for the suggestions made during this high-quality review and then we present the answers to the questions. We inform that with the reviews and suggestions we were able to improve the idea presented by our work and we appreciate the opportunity. We hope this review has left the article suitable for publication in this high-impact journal and respect in the area. Kind Regards. Response to Reviewer 1 (1) The manuscript would benefit from a discussion of how identified SNVs—particularly synonymous or non-coding variants—might influence RNA modifications (e.g., m6A methylation) and subsequent gene expression or splicing, which could mechanistically explain their impact on drug response. R: We appreciate this important insight. We have now incorporated a discussion on how synonymous and non-coding variants could influence drug response by potentially altering RNA modification sites (such as m6A methylation), thereby affecting gene expression and splicing. This addition, supported by recent literature, provides a plausible mechanistic layer to our findings. (2) The clinical significance of the findings should be elaborated beyond prediction of drug response. For example, how might these variants inform the development of combination therapies or adjuvant strategies to overcome resistance? R: We acknowledge the importance of exploring clinical implications beyond monotherapy responses. Our revised conclusion specifically explores how these pharmacogenomic findings could inform rational combination therapies and adjuvant strategies, potentially enabling the systematic elaboration of personalized treatment protocols that proactively optimize both efficacy and safety. (3) The authors should discuss potential pharmacological interventions that could counteract the effects of high-risk variants (e.g., P-gp inhibitors for ABCB1-mediated resistance, or allopurinol for NUDT15-related thiopurine toxicity). R: We are grateful to the reviewer for raising this valuable point. As suggested, we have incorporated a discussion on potential pharmacological strategies to counteract the effects of the high-risk variants identified in our study. The revised manuscript now includes a discussion on P-gp inhibitors and allopurinol supplementation, citing key recent literature to substantiate the proposed approaches. (4) The role of synonymous and intronic variants in regulating gene expression or mRNA stability is underemphasized. The authors should expand on how these “silent” or non-coding variants may contribute to chemoresistance through non-canonical mechanisms. R: We thank the reviewer for this observation. In response, we have expanded our Discussion to address the potential role of synonymous and intronic variants in chemoresistance. A new paragraph now discusses how such variants may influence gene expression, mRNA stability, and splicing, thereby contributing to resistance through non-canonical mechanisms, supported by recent literature on this topic. (5) The authors should highlight the importance of incorporating SNV profiling—not just gene expression—into routine genomic analyses (e.g., RNA-seq, whole-genome sequencing) to better predict treatment outcomes and toxicity risks. R: We thank the reviewer for this insightful comment. In response, we have added a new paragraph to the Discussion section highlighting the critical importance of incorporating SNV profiling alongside gene expression data in routine genomic analyses to better predict treatment outcomes and toxicity risks, citing relevant literature to support this integrated approach. (6) The discussion on ABCB1 is thorough, but other clinically relevant genes (e.g., TPMT, NUDT15) deserve more detailed mechanistic and clinical context, especially given their high level of evidence (LOE 1A). R: We are grateful to the reviewer for this valuable suggestion. We have expanded the Discussion to include a detailed mechanistic and clinical context for the high-evidence pharmacogenes TPMT and NUDT15, explaining their roles in thiopurine metabolism and the clinical standard of genotype-guided dosing, supported by the relevant CPIC guidelines and foundational literature. (7) The manuscript should address the limitations of using a database like PharmGKB, such as potential population biases, underrepresentation of rare variants, and the predominance of Level 3 evidence for many associations. R: We thank the reviewer for this important observation. In response, we have added a dedicated paragraph to the Discussion section explicitly addressing the limitations of using the PharmGKB database, including population biases, the underrepresentation of rare variants, and the predominance of Level 3 evidence for many associations, as suggested. (8) A forward-looking perspective on how these findings could be translated into clinical trials or integrated into existing treatment protocols (e.g., via point-of-care genotyping) would enhance the impact of the conclusion. R: We sincerely appreciate the reviewer's excellent suggestion to enhance the impact of our conclusion. In direct response, we have now added a forward-looking perspective to the concluding section of the manuscript. This new text discusses the potential for integrating our findings into clinical practice via point-of-care genotyping and outlines a framework for future clinical trials to validate a genotype-guided approach for improving safety and efficacy in acute leukemia treatment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

In this paper the authors investigate the relationship between genetic variants in selected genes and response/side effects following leukemia treatment. The manuscript is interesting and compelling. Following some minor observations:

• The methods of search and selection of studies/cases may be better explained for non-expert in the field.
• Among the variants with LOE 1 no one was synonymous. I think this is a point worth discussing.
• Are these variants polymorphism? How common are they in general population? Maybe a column with this information (easily retrievable) could be of interest.
• Mind quoting the sources not only in the table, but also in the text (e.g. “Some variants of the ABCB1 gene are common between AML and ALL, such as rs1045642 and rs1128503. The rs1045642 in AML patients treated with cytarabine is associated with higher chances of complete remission and 3-year EFS, in addition to also increasing the probability of vomiting and increased liver enzymes. In ALL patients treated with vincristine, methotrexate and etoposide, the absence of this same variant was associated with a higher probability of developing leukopenia, neutropenia, mucositis, hepatotoxicity, anemia, and thrombocytopenia.” These statements need citations. But also: “The survival rate in AML (32.9%) is generally lower than in ALL (92% for children and 40–60% for adults)” and several others).
• At list for LOE 1 variants, a summary/table of the putative mechanisms of actions would enrich the paper.
• “In this research, we identified ABCB1 as the gene with the highest frequency of single-nucleotide variants (SNVs) associated with chemotherapy response in acute leukemias, found in 12.5% of AML and 10.5% of ALL cases.” I don’t know if it is possible to retrieve this data, but how was the CR rate in patients harbouring this variant vs the ones who did not? Was the difference statistically significant?

Author Response

Dear reviewer, my co-authors and I would like to thank you for the suggestions made during this high-quality review and then we present the answers to the questions.

 

We inform that with the reviews and suggestions we were able to improve the idea presented by our work and we appreciate the opportunity. We hope this review has left the article suitable for publication in this high-impact journal and respect in the area.

 

Kind Regards.

 

Response to Reviewer 2

 

• The methods of search and selection of studies/cases may be better explained for non-expert in the field.

R: We thank the reviewer for this valuable suggestion. In response, we have expanded and clarified the description of the search and selection methods for studies and cases to make this section more accessible to non-expert readers. Specifically, we provided additional context about the purpose of the PharmGKB database, detailed how the searches were conducted, and described the sections (“Clinical Annotations” and “Variant Annotations”) used to identify and extract relevant data on genetic variants associated with drug response in acute leukemias.

• Among the variants with LOE 1 no one was synonymous. I think this is a point worth discussing.

R: We appreciate the reviewer’s insightful observation. We agree that the fact that none of the LOE 1A variants were synonymous is noteworthy and deserves further discussion. Accordingly, the Results section was revised to highlight that all variants with the highest level of evidence are non-synonymous or affect splicing regions, consistent with their functional relevance. In addition, a brief discussion was added to emphasize that such variants are more likely to alter enzyme structure or activity, thereby influencing drug metabolism and toxicity.

• Are these variants polymorphisms? How common are they in general population? Maybe a column with this information (easily retrievable) could be of interest.

R: We thank the editor for this valuable suggestion. Yes, these variants correspond to polymorphisms. In response, we have added a new column to the table indicating the altered allele frequency of each variant across five populations (African, Admixed American/Latino, East Asian, European, and South Asian), as retrieved from public genomic databases.

• Mind quoting the sources not only in the table, but also in the text (e.g. “Some variants of the ABCB1 gene are common between AML and ALL, such as rs1045642 and rs1128503. The rs1045642 in AML patients treated with cytarabine is associated with higher chances of complete remission and 3-year EFS, in addition to also increasing the probability of vomiting and increased liver enzymes. In ALL patients treated with vincristine, methotrexate and etoposide, the absence of this same variant was associated with a higher probability of developing leukopenia, neutropenia, mucositis, hepatotoxicity, anemia, and thrombocytopenia.” These statements need citations. But also: “The survival rate in AML (32.9%) is generally lower than in ALL (92% for children and 40–60% for adults)” and several others).

R: Thank you for your valuable comments. Specifically, we have incorporated the table citations into the main text as requested. Regarding the survival rates for AML and ALL, these were already supported by reference “118 - Health, N.I. of National Cancer Institute: Survaillance, Epidemiology, and End Results Program Available online: https://seer.cancer.gov/statistics-network/explorer/”, at the end of the paragraph.

• At list for LOE 1 variants, a summary/table of the putative mechanisms of actions would enrich the paper.

R: We thank the reviewer for the valuable suggestion to summarize the putative mechanisms of action for LOE 1 variants. We agree that this is an interesting aspect; however, the primary aim of our manuscript is to catalog and present the variants themselves. To address the reviewer’s comment while maintaining the focus of the paper, we have added several paragraphs in the Discussion section highlighting the putative mechanisms of action for LOE 1 variants. This allows us to provide additional context without shifting the central objective of the manuscript.

• “In this research, we identified ABCB1 as the gene with the highest frequency of single-nucleotide variants (SNVs) associated with chemotherapy response in acute leukemias, found in 12.5% of AML and 10.5% of ALL cases.” I don’t know if it is possible to retrieve this data, but how was the CR rate in patients harbouring this variant vs the ones who did not? Was the difference statistically significant?

R: We thank the reviewer for this important suggestion. The ClinPGx database primarily provides pharmacogenomic associations between variants and drug response, but it does not include individual patient-level clinical outcomes such as complete remission (CR) rates. For example, the study by Han et al. (2021) (doi:10.3389/fonc.2021.759805) established correlations between certain ABCB1 and MTHFR variants and methotrexate-related myelotoxicity in a cohort of patients with hematological malignancies, including patients with ALL. However, this study did not focus on analyses that would provide information on CR rates. Therefore, we are unable to directly report CR rates in patients harboring the ABCB1 variant versus those without it. Such analyses could be addressed in a separate future study.

 

Author Response File: Author Response.pdf