Modeling the Influence of CYP2C9 and ABCB1 Gene Polymorphisms on the Pharmacokinetics and Pharmacodynamics of Losartan
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis manuscript describes a well-executed and comprehensive computational modeling study that evaluates the influence of polymorphisms in the CYP2C9 and ABCB1 genes on the PK of losartan. The authors expand upon a previously established model by incorporating the ABCB1 transporter and evaluating virtual patent response using a cardiorenal model. This manuscript is well-structured and contributes significantly to the field personalized therapy. However, several points need to clarification/revision before publication.
- Authors’ assumption that all individuals in the CYP2C9 studies were GT/CT for ABCB1 should be more explicitly justified and its limitations acknowledged. The assumption introduces a source of uncertainty that could significantly impact the model outputs.
- While the integration of data from different studies is important, the limitations of using such heterogeneous datasets (especially with single-patient data for the CYPs) must be emphasized more strongly in results and discussion.
- From the results, we can see that the model underpredict several pk parameters (Cmax et al). authors need to discuss more potential reasons. Please also include the fold change figures related to the ratio of Predic/observed.
- Please ensure all the supplemental tables and figures are clearly referenced and described in the manuscript. (noting related to Figure S3 and S4 in the manuscript, and Table S5 and Figure S1 should be explained more fully)
- There are plenty of typo and formatting issues (e.g. line breaks in formulas, inconsistent spacing, et al) should be addressed throughout the manuscript.
- The references for SNPs listed in Table 1, please cite the primary sources for these variants.
- For the sensitivity analysis, it would be better to use the simcyp to predict the sensitivity analysis.
Author Response
Point: This manuscript describes a well-executed and comprehensive computational modeling study that evaluates the influence of polymorphisms in the CYP2C9 and ABCB1 genes on the PK of losartan. The authors expand upon a previously established model by incorporating the ABCB1 transporter and evaluating virtual patent response using a cardiorenal model. This manuscript is well-structured and contributes significantly to the field personalized therapy. However, several points need to clarification/revision before publication.
Response: Dear reviewer, we would like to thank you for your positive assessment of our work and your constructive feedback. Please see our responses to your questions below, and refer to the revised version of the manuscript.
Point #1: Authors’ assumption that all individuals in the CYP2C9 studies were GT/CT for ABCB1 should be more explicitly justified and its limitations acknowledged. The assumption introduces a source of uncertainty that could significantly impact the model outputs.
Response #1: To support our assumption that the GT/CT genotype is the most prevalent among Koreans, we analysed an additional experimental trial in which Korean participants were studied to determine the frequency of different ABCB1 genotypes (Park et al., 2007). This study found that the GT/CT genotype was the most common, accounting for 40.5% of cases. For more details, please refer to lines 313-323 of the revised manuscript.
In the Discussion section, we addressed the limitations of this assumption, and recognised that considering all patients with an unknown ABCB1 genotype as GT/CT carriers could significantly impact the model's predictions (please see lines 547-552).
Point #2: While the integration of data from different studies is important, the limitations of using such heterogeneous datasets (especially with single-patient data for the CYPs) must be emphasized more strongly in results and discussion.
Response #2: We have added a paragraph to the Discussion section (lines 561-570 of the revised manuscript) regarding the consequences of using multiple datasets from different studies for model training.
Point #3: From the results, we can see that the model underpredict several pk parameters (Cmax et al). authors need to discuss more potential reasons. Please also include the fold change figures related to the ratio of Predic/observed.
Response #3: In the Discussion section of the revised manuscript, we have provided a few additional potential explanations for the discrepancies between the predicted and experimental values of the pharmacokinetic parameters. Please see lines 542-552.
We have also added a figure demonstrating the predicted/observed fold change ratio for all pharmacokinetic parameters across all CYP2C9 and ABCB1 genotypes (please see Figure 6 in the revised manuscript)
Point #4: Please ensure all the supplemental tables and figures are clearly referenced and described in the manuscript. (noting related to Figure S3 and S4 in the manuscript, and Table S5 and Figure S1 should be explained more fully)
Response #4: Supplementary Table S5 shows the AUC values for E-3174 and k_block for each ABCB1 genotype (GG/CC, GT/CT and TT/TT) at two losartan doses (50 mg and 100 mg). We have provided a more detailed description of how the k_block values were obtained based on the AUC values for each of the six cases (see lines 194-198 of the revised manuscript).
We have also provided a more precise description of how the curve shown in Supplementary Figure S1 was derived (lines 188-191).
To improve understanding of Supplementary Figures S3 and S4, we have provided a more detailed description of how we modeled the variability between subjects in losartan and E-3174 concentration-time curves in Section 3.4 (lines 408-413). We have also added Supplementary Table S11 to the supplementary material. This table provides the median and standard deviation values for each model parameter. These values were then utilized to derive a normal distribution, from which 100 values were randomly selected for each ABCB1 genotype.
Point #5: There are plenty of typo and formatting issues (e.g. line breaks in formulas, inconsistent spacing, et al) should be addressed throughout the manuscript.
Response #5: We apologize for the formatting errors. We have removed the extra spaces and corrected the line breaks after the formulas (see lines 132, 136, 139, 148, 153, 171, 224, and 309 of the revised manuscript for details).
Point #6: The references for SNPs listed in Table 1, please cite the primary sources for these variants.
Response #6: We have added references to the primary studies in which these polymorphisms were identified (see lines 49-51): (Hoffmeyer et al., 2000) for C1236T and C3435T, and (Mickley et al., 1998) for G2677T/A.
Point #7: For the sensitivity analysis, it would be better to use the simcyp to predict the sensitivity analysis.
Response #7: Thanks for your recommendation. However, there are a few things that make it difficult to perform sensitivity analyses of our model using Simcyp software. Simcyp is a specialized platform for PBPK modelling and is not suitable for general tasks. Furthermore, it does not support the SBML standard for model representation. This made it difficult to transfer our model from BioUML, where development and simulation had been performed, to Simcyp. Finally, since BioUML can perform sensitivity analyses, no additional software was required.
Reviewer 2 Report
Comments and Suggestions for Authors1. Line 56-57, it would be informative to expand the introduction on whether the 2 most common haplotypes of ABCB1 are associated with altered protein expression or function. Similarly, it's worth mentioning that CYP2C9*3 is known to have reduced enzyme activity and leading to lower clearance of its substrate/lower E-3174 in this case.
2. Please fix the language in Table 2 and line 79.
3. In figure 4, please specify the color code for losartan and E-3174
4. In line 306, is GT/CT the most common genotype in Korean to back up the assumption? The deviation from CYP2C9 data between GG/CC and GT/CT is relatively minor.
5. Please fix the language in line 316.
6. Could you clarify how the model was validated in 3.3? It looks like the model is built based on Shin 2020 dataset for ABCB1. Is there another dataset for validation or did you use half of the dataset to build the model and the other half for validation? It's unclear how the k_m and k_ent_int value validate the model.
Author Response
Response: Dear reviewer, thank you very much for your comments. We have tried to respond to all your concerns. Please refer to the responses provided below, as well as the revised version of the manuscript.
Point #1: Line 56-57, it would be informative to expand the introduction on whether the 2 most common haplotypes of ABCB1 are associated with altered protein expression or function. Similarly, it's worth mentioning that CYP2C9*3 is known to have reduced enzyme activity and leading to lower clearance of its substrate/lower E-3174 in this case.
Response #1: We expanded the Introduction section to include information regarding the impact of haplotype on the expression of the ABCB1 gene, along with the underlying mechanisms that facilitate this phenomenon (please see lines 57-61 in the revised manuscript). Furthermore, the information regarding the CYP2C9*2 and CYP2C9*3 variant alleles and their reduced activity was incorporated (lines 77-81)
Point #2: Please fix the language in Table 2 and line 79.
Response #2: We apologize for this mistake. The language has been corrected to English; please refer to Table 2 and line 83 of the revised manuscript.
Point #3: In figure 4, please specify the color code for losartan and E-3174
Response #3: Thank you for your attention. The legend has been included in Figure 4, and the colour code has been specified in the caption of this figure.
Point #4: In line 306, is GT/CT the most common genotype in Korean to back up the assumption? The deviation from CYP2C9 data between GG/CC and GT/CT is relatively minor.
Response #4: Indeed, the distances from CYP2C9*1/CYP2C9*1 dataset between GG/CC and GT/CT genotypes are very similar (see Table 5, 322.501 vs. 278.345). In order to corroborate the hypothesis that GT/CT may be the most common genotype in Koreans, we analysed an additional experimental trial in which the frequencies of different ABCB1 genotypes in patients who shared the same ethnicity were studied (Park et al., 2007). The analysis showed that the GT/CT genotype was more prevalent than the other genotypes (40.5 %). For further details, please refer to lines 313-323 of the revised manuscript.
Point #5: Please fix the language in line 316.
Response #5: If we understand your comment correctly, you meant the following sentence: "CYP2C9 genotypes with CYP2C9*2 were not examined due to the absence of this variant in the Korean population". We have amended it as follows: "The CYP2C9*2 allele was excluded from the analysis due to its rarity in the Korean population, as confirmed by previous studies" (please see lines 337-338 of the revised manuscript).
Point #6: Could you clarify how the model was validated in 3.3? It looks like the model is built based on Shin 2020 dataset for ABCB1. Is there another dataset for validation or did you use half of the dataset to build the model and the other half for validation? It's unclear how the k_m and k_ent_int value validate the model.
Response #6: The values of the model parameters were estimated based on two datasets. The first dataset (Bae et al., 2011) was used to estimate k_m, which indicates CYP2C9 activity (the first estimation step), while the second dataset (Shin et al., 2020) was used to estimate all other model parameters.
On the first step, using only one dataset (Bae et al., 2011) we determined two values of the k_m (for CYP2C9*1/CYP2C9*1 and CYP2C9*3/CYP2C9*3 genotypes). Secondly, we fixed the already identified values of the k_m and using both datasets, (Bae et al., 2011) and (Shin et al., 2020), we identified the values of all other parameters. Note, that on this estimation step we received three values for the k_ent_int parameter (for each ABCB1 genotype: GG/CC, GT/CT, and TT/TT), as for the k_m on the first estimation step.
We have expanded section 3.3 for better understanding of the whole process (please, see lines 334-336, 339, and 347-350 in the revised manuscript).
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsThanks for addressing my comments. I don't have any further comments.
