2.6. Statistical analysis
The Kolmogorov-Smirnov tests was used to assess the normality of continuous variables. Descriptive statistics of normal data were expressed as mean ± standard deviation (SD), and non-normal data were expressed as median and interquartile range (IQR). Independent sample T test or Mann-Whitney U test were used to determine the difference in continuous values between the two groups, and one-way ANOVA or Kruskal-Wallis H test is used to detect the difference of continuous variables between three groups or more. Categorical data were expressed as frequency and percentage, using χ2 tests. The Hardy-Weinberg equilibrium (HWE) was evaluated on each SNPs using χ2 tests. Propensity score matching was used by caliper matching method (caliper value 0.05) to exclude confounding factors other than methylation, and propensity matching score was used to evaluate the balance between the two groups after matching. SPSS 22.0 software was used for statistical analysis, STATA17.0 software was used for PSM, and GraphPad Prism 9 software was used for mapping.. The results with P value < 0.05 were considered statistically significant.
Results
Patients characteristics
In this study, 116 concentration points were included, which were divided into three groups according to the standard range of voriconazole Cminss (1.0-5.5mg/L) . There were 31 concentration points in the low concentration group (Cminss<1.0mg/L), 55 concentration points in the standard concentration group (Cminss =1.0-5.5mg/L), and 30 concentration points in the high concentration group (Cminss>5.5mg/L). The CYP2C19phenotype of the included patients was normal metabolizer (NM) or intermediate metabolizer (IM). The daily dose of voriconazole is 400mg. There were no significant differences in age, sex, BMI, route of administration, purpose of administration, and information related to drug combination among the three groups. In laboratory examination, CRP, Abl and T-BIL were different between the high concentration group and the standard concentration group, while the other indexes were not statistically different between the three groups.The patient characteristics are detailed in Table.1.
Distribution and intergroup differences of CYP2C19 activity in NM/IM population
VNO is the main product of voriconazole metabolism by CYP2C19. The ratio of VNO to VRC concentration indicates the activity of CYP2C19. In the NM/IM population, the distribution of CYP2C19 activity is relatively wide (as shown in Fig.1A), and there may be incompatibility betweenCYP2C19 genotype and phenotype.
As shown in Fig.1B, the activity of CYP2C19 in the low concentration group (Cminss<1.0mg/L) was significantly higher than that in the standard concentration group (Cminss=1.0-5.5mg/L), and that in the high concentration group (Cminss >5.5mg/L) was significantly lower than that in the standard concentration group (Cminss=1.0-5.5mg/L). This suggests that CYP2C19gene polymorphism may not be used as a factor to explain the VRC Cminss in all patients in the low concentration group or high concentration group, and there may be other factors affecting the VRC Cminss in patients.
Effect of CYP2C19 DNA methylation on VRC Cminss
Effect of CYP2C19 DNA methylation on VRC Cminss in low concentration group
There was no significant difference in patient characteristics between the low concentration group (Cminss<1.0mg/L) and the standard concentration group (Cminss =1.0-5.5mg/L). The methylation level of CpG site of CYP2C19between the two groups is shown in Table2, where the methylation level of CYP2C19 CpG25 is significantly different between the two groups. The methylation level of CYP2C19 CpG25 in the low concentration group was higher than that in the standard concentration group, as shown in Fig.2.
Effect of CYP2C19 DNA methylation on VRC Cminss in high concentration group
There were statistical differences in CRP, Alb and T-BIL between the high concentration group (Cminss>5.5mg/L) and the standard concentration group (Cminss =1.0-5.5mg/L) (as shown in Table.1). CRP, Alb and T-BIL may influence the VRC Cminss in high concentration group. In order to exclude the above confounding factors and highlight the effect ofCYP2C19 DNA methylation, Propensity score matching was performed by caliper matching method. There were no significant differences in CRP, Alb and T-BIL between the two groups after PSM (as shown in Table.3). As shown in Table2.7, there was no significant difference in methylation levels at CpG sites between the two groups after PSM(as shown in Supplementary Table4). This suggests that CYP2C19 DNA methylation levels may not be a factor in VRC Cminss in the high concentration group.
Effect of other factors on VRC Cminssin high concentration group
There were statistical differences in CRP, Alb and T-BIL between the high concentration group (Cminss>5.5mg/L) and the standard concentration (Cminss=1.0-5.5mg/L) before PSM (as shown in Table.1 and Table.3). These results suggest that CRP, Alb and T-BIL may be the factors influencing the VRC Cminss in the high concentration group. Compared with the standard concentration group, CRP in the high concentration group was significantly higher (as shown in Fig.3A), Alb was significantly lower in the high concentration group (as shown in Fig.3B), and T-BIL was significantly higher in the high concentration group (as shown in Fig.3C).
Discussion
In this study, patients with CYP2C19 normal metabolizer (NM) and intermediate metabolizer (IM) were included, and the effect of CYP2C19 gene polymorphism on voriconazole Cminss was excluded. In the NM/IM population, CYP2C19 activity is widely distributed, and the predicted phenotype of CYP2C19 genotype may not be completely consistent with the measured phenotype. Burns et al. reported that 1/3 of gastrointestinal tumor patients have ”dynamic changes” in liver CYP2C19 function, and the enzyme activity loss of the seemingly ”poor metabolizer” phenotype is inconsistent with its genotype (CYP2C19*2, CYP2C19*3 )13. This divergence between genotype and phenotype of CYP2C19 has also been reported in solid tumors such as breast cancer, lung cancer, kidney cancer, and ovarian cancer14 and hematological malignancies15. Our research is consistent with that. Compared with standard concentration group (Cminss=1.0-5.5mg/L), CYP2C19 activity was significantly higher in patients with low concentration group (Cminss<1.0mg/L) and significantly lower in patients with high concentration group (Cminss>5.5mg/L). In summary, CYP2C19 activity may be determined by more than its metabolic function.
The expression of CYP2C19 is moderately associated with its activity16. The DNA methylation level of specific ADME gene affects its gene expression9.This study detected the CYP2C19 DNA methylation level to explore whether it is a factor for the VRC Cminss. DNA methylation atCYP2C19 CpG25 site was significantly down-regulated in low concentration group (Cminss<1.0mg/L) compared with standard concentration group (Cminss=1.0-5.5mg/L). In previous studies, the CYP2C19 mRNA expression in primary hepatocytes10, HepG2 cells, and HCT116 cells11 was up-regulated after treatment with DNA methyltransferase inhibitor 5-aza-DC. This suggests that there may be a negative correlation between CYP2C19 DNA methylation and mRNA expression. In this study, the decrease of VRC Cminss in the low concentration group may be due to the down-regulation of DNA methylation at CYP2C19 CpG25 site, which may lead to the up-regulation of CYP2C19 expression and increase of activity. This is consistent with the relationship between CYP2C19 DNA methylation and its expression in previous studies. A clinical study17 reported the relationship betweenCYP2C19 methylation and clopidogrel resistance. WhenCYP2C19 is hypomethylated, clopidogrel resistance may be due to the reduction of products metabolized by CYP2C19, which seems to be inconsistent with our study. This may be because the CYP2C19 methylation site detected in this study is located in the gene body, while the CYP2C19 methylation site detected in this study is located in the promoter. There was no significant difference in the CYP2C19 DNA methylation level in the high concentration group (Cminss>5.5mg/L) compared with the standard concentration group (Cminss=1.0-5.5mg/L) after PSM. This suggests that CYP2C19 DNA methylation may not be a factor affecting the VRC Cminss in the high-concentration group.
There were significant differences in CRP, Alb and T-BIL between the high concentration group (Cminss>5.5mg/L) and the standard concentration group (Cminss=1.0-5.5mg/L) before PSM. CRP is a biomarker of inflammatory state18. Compared with CRP [31.40 (10.40,71.80) mg/L] in the standard concentration group, CRP [100.75 (44.50,175.28) mg/L] in the high concentration group was increased. Most patients were in a state of moderate to severe inflammation (CRP≥40mg/L)19. A large number of inflammatory factors (such as IL-6, IL-1β, TNF-α, etc.) are usually produced in inflammatory states20. Some studies have reported that increasing the inflammatory factors concentration will reduce the expression and activity of drug metabolic enzymes21-23. Therefore, the VRC Cminss may be caused by the inflammatory state affecting the expression and activity of VRC metabolic enzymes such as CYP2C19 in the high concentration group. The plasma protein binding rate of VRC is 58%24, and the main binding protein is albumin25. Compared with the standard concentration group (34.27±4.539), the Alb in the high concentration group (31.61±5.088) was significantly decreased, and the free concentration of VRC in the high concentration group may be increased. In addition, it has been reported that the change of Alb may affect the clearance rate of VRC26. Therefore, the VRC Cminss may also be due to the effect of Alb level on the distribution and elimination of voriconazole in the body in the high concentration group. T-BIL is one of the indicators indicating liver function. Compared with the T-BIL of the standard concentration group [9.60 (6.20, 14.10) μmol/L], the T-BIL of the high concentration group [11.45 (8.80, 23.80) μmol/L] was significantly higher. The liver function of the high concentration group may be worse than that of the standard concentration group. It has been reported that decreased liver function may affect the expression of drug metabolic enzymes27. In addition, T-BIL was found to be related to VRC clearance28. Therefore, the VRC Cminss in the high concentration group may also be due to the effect of T-BIL level on the metabolism and elimination of VRC. In addition, the down-regulation of Alb and up-regulation of T-BIL may also be affected by inflammatory states29. In summary, the VRC Cminss may be related to the up-regulation of CRP, down-regulation of Alb and up-regulation of T-BIL in the high-concentration group.
In this study, factors influencing the VRC Cminss in the low concentration group (Cminss<1.0mg/L) and the high concentration group (Cminss>5.5mg/L) were investigated, and the relationship between CYP2C19 DNA methylation and the VRC Cminss was investigated for the first time, providing a new perspective for exploring the individual administration of voriconazole. The limitation of this study may be that CYP2C19DNA methylation levels in blood are used to replace CYP2C19 DNA methylation levels in liver, and there may be some differences in DNA methylation levels between blood and liver30. At the same time, some studies17,31,32used blood to measure the DNA methylation level of ADME gene to explore its role in pharmacokinetics. Due to the rarity of liver samples and the limitation of detection techniques, the use of blood samples instead of liver to explore the role of DNA methylation in pharmacokinetics is also an option to be considered.
Conclusion
In this study, the VRC Cminss in patients in the low concentration group (Cminss<1.0mg/L) may be related to methylation levels at the CYP2C19 CpG25 site . Based on the above possible qualitative relationships, it seems possible to obtain a better therapeutic effect of voriconazole by adjusting certain lifestyles associated with DNA methylation. More studies are expected to establish a quantitative relationship between DNA methylation and VRC Cminss, and it is possible to optimize the dosing regimen by adjusting the degree of DNA methylation in the future. The VRC Cminss in high concentration group (Cminss>5.5mg/L) may not be related toCYP2C19 DNA methylation level. The VRC Cminss in patients in high concentration group may be related to the levels of CRP, Alb and T-BIL of patients. CYP2C19 DNA methylation may be influenced by CRP, Alb, and T-BIL, and it is more meaningful to explore the interaction between DNA methylation and some non-genetic factors in the future.