Adverse effects of dasatinib on glucose-lipid metabolism in patients with chronic myeloid leukaemia in the chronic phase

To explore the differences in glucose-lipid metabolism profiles among the 3 TKIs, we designed a retrospective study to compare the onset of hyperglycaemia, hypertriglyceridemia, hypercholesterolemia and hyper-low density lipoprotein (LDL)-cholesterolemia in the patients with normal baseline glucose-lipid profiles and had no medical record of cardio- or cerebro-vascular diseases and/or metabolic syndrome diseases, and identify variables associated with them. 370 chronic myeloid leukaemia patients receiving dasatinib, nilotinib or imatinib therapy ≥3 months were retrospectively reviewed. During TKI-therapy, the mean fasting glucose, triglyceride, cholesterol, and LDL-cholesterol levels increased significantly in both dasatinib and nilotinib cohorts compared with the imatinib cohort. In multivariate analyses, dasatinib was the factor significantly associated with both poor hyperglycaemia- and hypertriglyceridemia-free survival. In addition, nilotinib was significantly associated with more occurrences of hyperglycaemia and hypercholesterolemia; increasing age was significantly associated with more occurrences of hyperglycaemia and hypertriglyceridemia. We concluded that dasatinib, similar to nilotinib, has the adverse impact on glucose-lipid metabolism compared with imatinib.

Definitions. According to the upper limits of normal (ULN) in the lab of Peking University People's Hospital, hyperglycaemia, hypertriglyceridemia, hypercholesterolemia and hyper-LDL-cholesterolemia were defined as the laboratory test values of fasting glucose (GLU), triglyceride (TG), cholesterol (CHO), and LDL-cholesterol (LDL-C) higher than the ULN presented on two or more occasions at least 3 months apart, respectively. The normal range of GLU, TG, CHO and LDL-C were 3.3-6.1 mmol/L, 0.45-1.7 mmol/L, 2.9-6.2 mmol/L, and 1.9-4.1 mmol/L, respectively. The severity of hyperglycaemia, hypertriglyceridemia and hypercholesterolemia was assessed according to Common Terminology Criteria for Adverse Events (CTCAE) 4.0. As these AEs may occur over time on TKI-therapy, hyperglycaemia-, hypertriglyceridemia-, hypercholesterolemia-, and hyper-LDL-cholesterolemia-free survivals were defined as the time between TKI-therapy initiation and the onset of the AEs for analysing the glucose-lipid metabolic abnormality during TKI-therapy.
Statistical analyses. Descriptive analysis results are presented as median (range) or number (percent) as appropriate. Pearson Chi-squared (for categorical variables) and Mann-Whitney U/Kruskal-Wallis tests (for continuous variables) were used to measure between-group differences. Statistically significant differences in the levels of GLU, TG, LDL-C, and CHO were calculated with a paired-samples Student's T test between the values at baseline and at each time point in each cohort. The log-rank test was used to assess statistical significance in the time-to-event analyses. Univariate analyses including TKI used, sex, age, interval from diagnosis to starting TKI within 6 months or not, interval from diagnosis to taking this TKI, TKI used as first-line versus second-or third-line therapy and TKI-dosage change were performed to determine variables associated with the onset of hyperglycaemia, hypertriglyceridemia, hypercholesterolemia and hyper-LDL-cholesterolemia. Variables associated at a level of p < 0.2 were included in the Cox regression model to identify variables significantly associated with the abnormalities of glucose-lipid metabolism profiles. Factors with an effect significant at the p < 0.05 were interpreted as independently predicting outcomes. All analyses were conducted with SPSS version 22.0 software (SPSS Inc., Chicago, IL, USA). Informed consent. Informed consent was obtained from all individual participants included in the study.
Ethical approval. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
During the follow-up period, there was no cardio-cerebrovascular event occurred, and no concomitant medication was prescribed for the mild or intermediate glucose-lipid metabolic abnormalities in all patients.

Discussion
We explored the effects of dasatinib on fasting blood GLU, TG, CHO and LDL-C in relatively young CML-CP patients with normal baseline and no cardio-vascular disease or metabolic syndrome, and compared with nilotinib and imatinib. We found that dasatinib was significantly associated with an increased onset of hyperglycaemia and cholesterol (CHO) (d) levels in the dasatinib, nilotinib and imatinib cohorts. Statistically significant differences were calculated with a paired-samples Student's T test between the baseline and each time point, solid dot indicate P < 0.05. imatinib, nilotinib, dasatinib. * statistically significant differences among the 3 cohorts (P < 0.05), △ statistically significant differences between the dasatinib and imatinib cohorts (P < 0.05), statistically significant differences between the dasatinib and nilotinib cohorts (P < 0.05). (2019) 9:17601 | https://doi.org/10.1038/s41598-019-54033-0 www.nature.com/scientificreports www.nature.com/scientificreports/ and hypertriglyceridemia compared with imatinib. In addition, nilotinib was significantly associated with more occurrences of hyperglycaemia and hypercholesterolemia; increasing age was significantly associated with more occurrences of hyperglycaemia and hypertriglyceridemia.
There are several studies about the adverse impacts of nilotinib on glucose-lipid metabolism 7,20,23,24 . However, data regarding glucose-lipid metabolic abnormalities focused on dasatinib is limited. Iurlo et al. reported in a study which included 107 patients (40 treated with dasatinib; 36, nilotinib; and 92, imatinib) at a median age of 56 years, although fasting plasma glucose, insulin, C-peptide, and Homeostasis Model Assessment-Insulin Resistance were significantly higher in the nilotinib group than those in the dasatinib and imatinib groups, LDL-C did not differ significantly between the dasatinib and nilotinib groups, which was higher than that in the imatinib group 20 . We confirmed TKI-related hyper-LDL-cholesterolemia by using more sensitive and accurate assessment methods of higher than the ULN presented on two or more occasions at least 3 months apart and event-free survival rate (refers to hyper-LDL-cholesterolemia-free survival), because drug-related glucose lipid abnormalities may develop over time. In our current study, dasatinib was also associated with a higher onset of hyperglycaemia and hypertriglyceridemia compared with imatinib. Because dasatinib and nilotinib share the same mechanism of targeting ABL and BCR-ABL, we speculate that dasatinib might have the similar effect and mechanism on IR and the IR metabolic pathway. It has been reported that nilotinib was associated with hyperinsulinaemia and insulin resistance, which likely occurred on the post-receptor level. A study including 10 CML patients without a medical history of diabetes mellitus clarified that the mechanism of impaired glucose metabolism and dyslipidaemia occurs via rapidly developed tissue insulin resistance and compensatory hyperinsulinaemia. Some studies in vitro indicate c-ABL was involved in the insulin receptor (IR) signalling pathway and enhances the IR metabolic pathway. However, the effect on the IR pathway of each TKI has not yet been well understudied 9 . Various animal models have been used to evaluate what role tyrosine kinases play in the regulation of glucose-lipid levels. Krishnamurthy et al. illustrated with in vitro studies that c-kit tyrosine kinase was essential for β-cell survival in www.nature.com/scientificreports www.nature.com/scientificreports/ the pancreas. A mouse with a c-kit point mutation that diminishes the receptor's kinase activity exhibited glucose intolerance, impaired insulin secretion and a reduction in β-cell mass 25 . However, the real, accurate mechanism has not been found in human studies.
Some studies reported that dasatinib can decrease blood glucose levels 19,26,27 . Keiko et al. reported a rapid amelioration of hyperglycaemia facilitated by dasatinib in a CML patient with diabetes mellitus 27 . However, in our current study dasatinib was associated with a higher onset of hyperglycaemia and hypertriglyceridemia. This finding may be due to the different population regarding age, comorbidities, ethnic and genetics which are associated with different risk or susceptibility to glucose-lipid dysfunction or prediabetes, and the assessing method of event-free survival used in our study. All patients in our study had a normal glucose-lipid baseline and no medical record of diabetes. The mechanism might be different between the patients with or without the comorbidity of diabetes.
Even in these younger patients with no medical record of pre-existing metabolic syndrome and a short follow-up period, abnormal glucose-lipid metabolism occurred in the dasatinib and nilotinib cohorts. Therefore, for those receiving dasatinib or nilotinib therapy in the pursuit of treatment-free remission, the risk of metabolic syndrome or cardio-cerebrovascular events should be taken into consideration, and laboratory index of glucose-lipid metabolism should be closely monitored.
Our study has some limitations. First, there were some differences in baseline characteristics among the 3 cohorts in this retrospective study. Second, there were relatively small patient numbers in the dasatinib and nilotinib cohorts. Third, patients with no medical history record of cardio-or cerebro-vascular diseases and/or metabolic syndrome were included in this study; however, some patients in very early stage of these diseases could not be excluded. Fourth, several other factors which had not been collected may contribute to modifications of glucose and blood lipids equilibrium, such as persistent diet or physical activity change, personal predisposition, and a gain of weight.
We concluded that dasatinib, similar to nilotinib, has adverse impact on glucose-lipid metabolism when compared with imatinib. These data favour using imatinib over dasatinib and nilotinib in older patients, even in younger CML patients with normal baseline glucose-lipid levels and without cardio-or cerebro-vascular diseases and/or metabolic syndrome who want to pursue TFR, laboratory index of glucose-lipid metabolism should be monitored closely.