Individual assessment of CYP enzyme activities can be challenging. Recently, the potato alkaloid solanidine was suggested as a biomarker for CYP2D6 activity. Here, we aimed to characterize the sensitivity and specificity of solanidine as a CYP2D6 biomarker among Finnish volunteers with known CYP2D6 genotypes. Using non-targeted metabolomics analysis, we identified 9152 metabolite features in the fasting plasma samples of 356 healthy volunteers. Machine learning models suggested strong association between CYP2D6 genotype-based phenotype classes with a metabolite feature identified as solanidine. Plasma solanidine concentration was 1887% higher in genetically poor CYP2D6 metabolizers (gPM) (n = 9; 95% confidence interval 755%, 4515%; P = 1.88 × 10–11), 74% higher in intermediate CYP2D6 metabolizers (gIM) (n = 89; 27%, 138%; P = 6.40 × 10–4), and 35% lower in ultrarapid CYP2D6 metabolizers (gUM) (n = 20; 64%, − 17%; P = 0.151) than in genetically normal CYP2D6 metabolizers (gNM; n = 196). The solanidine metabolites m/z 444 and 430 to solanidine concentration ratios showed even stronger associations with CYP2D6 phenotypes. Furthermore, the areas under the receiver operating characteristic and precision–recall curves for these metabolic ratios showed equal or better performances for identifying the gPM, gIM, and gUM phenotype groups than the other metabolites, their ratios to solanidine, or solanidine alone. In vitro studies with human recombinant CYP enzymes showed that solanidine was metabolized mainly by CYP2D6, with a minor contribution from CYP3A4/5. In human liver microsomes, the CYP2D6 inhibitor paroxetine nearly completely (95%) inhibited the metabolism of solanidine. In a genome-wide association study, several variants near the CYP2D6 gene associated with plasma solanidine metabolite ratios. These results are in line with earlier studies and further indicate that solanidine and its metabolites are sensitive and specific biomarkers for measuring CYP2D6 activity. Since potato consumption is common worldwide, this biomarker could be useful for evaluating CYP2D6-mediated drug–drug interactions and to improve prediction of CYP2D6 activity in addition to genotyping.

中文翻译：

---

茄碱是 CYP2D6 活性的敏感且特异的饮食生物标志物

对 CYP 酶活性的个体评估可能具有挑战性。最近，马铃薯生物碱茄碱被建议作为 CYP2D6 活性的生物标志物。在这里，我们的目的是在具有已知 CYP2D6 基因型的芬兰志愿者中表征茄碱作为 CYP2D6 生物标志物的敏感性和特异性。通过非靶向代谢组学分析，我们在 356 名健康志愿者的空腹血浆样本中鉴定出了 9152 种代谢物特征。机器学习模型表明基于 CYP2D6 基因型的表型类别与被确定为茄碱的代谢物特征之间存在很强的关联。 CYP2D6 遗传不良代谢者 (gPM) 的血浆茄碱浓度高出 1887%（n = 9；95% 置信区间 755%、4515%；P = 1.88 × 10–11），中间 CYP2D6 代谢者 (gIM) 的血浆茄碱浓度高出 74%（ n = 89; 27%, 138%; P = 6.40 × 10–4)，超快 CYP2D6 代谢者 (gUM) (n = 20; 64%, − 17%; P = 0.151) 比遗传正常者低 35% CYP2D6 代谢者（gNM；n = 196）。茄碱代谢物 m/z 444 和 430 与茄碱浓度比显示与 CYP2D6 表型的关联性更强。此外，这些代谢比率的接受者操作特征和精确回忆曲线下的面积显示出与其他代谢物、它们与茄碱的比率或单独的茄碱的比率相同或更好的识别 gPM、gIM 和 gUM 表型组的性能。人重组 CYP 酶的体外研究表明，茄碱主要由 CYP2D6 代谢，CYP3A4/5 的贡献较小。在人肝微粒体中，CYP2D6 抑制剂帕罗西汀几乎完全（95%）抑制茄碱的代谢。在一项全基因组关联研究中，CYP2D6 基因附近的几个变异与血浆茄碱代谢物比率相关。这些结果与早期研究一致，并进一步表明茄碱及其代谢物是测量 CYP2D6 活性的敏感且特异的生物标志物。由于马铃薯消费在世界范围内很常见，因此除了基因分型之外，该生物标志物还可用于评估 CYP2D6 介导的药物间相互作用，并改善对 CYP2D6 活性的预测。