Population pharmacokinetic analysis of doxecitine and doxribtimine in patients with thymidine kinase 2 deficiency and in healthy volunteers
Salim Bouchene1, Aravind Mittur2, Himanshu Naik3, Pinky Dua4, Ida Neldemo5, Jakob Ribbing5, Martin Bergstrand5, Akash Khandelwal6
1UCB, 2UCB, 3UCB, 4UCB, 5Pharmetheus AB, 6UCB
Introduction: Doxecitine and doxribtimine, a pyrimidine nucleoside therapy consisting of the two nucleosides doxecitine (deoxycytidine [dC]) and doxribtimine (deoxythymidine [dT]), is in development for the treatment of thymidine kinase 2 deficiency (TK2d). TK2d is an ultra-rare and severe myopathic mitochondrial DNA (mtDNA) depletion and multiple deletions syndrome characterized by reduced mtDNA copy number [1,2]. TK2d is associated with proximal myopathy and respiratory weakness, with many patients losing the ability to walk, eat and breathe independently [3]. Data suggest that treatment with doxecitine and doxribtimine improves survival and can positively change TK2d disease trajectory, with improvement and/or stabilization across key clinical outcomes. Objectives: To describe the pharmacokinetics (PK) of dC and dT, quantifying the associated interindividual variability (IIV) and investigating the effect of relevant covariates on their systemic exposure in healthy volunteers and patients with TK2d. Methods: Population PK modeling analyses of dC and dT plasma data in adult and pediatric healthy and patient populations were performed using NONMEM [4]. Body weight (WT) and food intake were included in the models as mechanistic covariates. Other covariate-parameter relationships were evaluated using stepwise covariate model building. The evaluated covariates included demographics, biomarkers for eliminating organ impairment, and disease-related variables. Results: dC PK was adequately described by a one-compartment disposition model with two parallel 1st-order absorption processes and 1st-order elimination. IIV was included on relative bioavailability (Frel), mean absorption time associated with the 1st and 2nd absorption compartments (MAT1 and MAT2, respectively), fraction absorbed in the 1st absorption compartment (FRACTION), estimated baseline plasma concentration (BASE), volume of distribution (V) and residual unexplained variability (RUV). The following covariate-parameter relationships were included: WT-based allometric relationships with estimated exponent for clearance (CL) and fixed exponent for V. Alanine aminotransferase and moderate-to-severe renal impairment had a statistically significant effect on CL, food status and dose on Frel, food status on MAT1, time after dose on MAT2, and age on BASE. Compared to healthy volunteers, patients with TK2d had higher IIV for Frel and RUV. The largest covariate impact on dC exposure was due to food status, with a median predicted 24-hour area under the curve at steady state (AUC0-24,ss) more than 2-fold higher in fed than in fasted conditions. The dose effect on Frel was such that increase in dC exposure was less than proportional to dose. dT PK was best described by a one-compartment disposition model with two parallel 1st-order absorption processes and 1st-order elimination. A bolus input with lag time (Tlag1) was added to the 1st absorption compartment whereas a zero-order input with lag time (Tlag2) was applied to the 2nd absorption compartment. IIV was included on Frel, MAT1, MAT2, FRACTION (going into the fast-absorption compartment), BASE, CL and RUV. The model included the following covariate-parameter relationships: WT-based allometric relationships with estimated exponent on CL and fixed exponent on V. Food status and ethnicity were added on Frel, food status was added on Frel, MAT1, MAT2, Tlag1, Tlag2 and FRACTION, and age was added on BASE. In addition, patients with TK2d had a higher magnitude of IIV for Frel than healthy volunteers. dT median AUC0-24,ss in fed conditions was approximately 2-fold higher than in fasted conditions. Unlike dC, for dT the dose effect on Frel was such that increase in exposure was more than proportional to dose. Conclusions: To our knowledge, this is the first PK analysis performed in patients with TK2d. Models characterized the PK of dC and dT in pediatric and adult populations effectively. High IIV was estimated for the PK of both nucleosides but was larger for dT than dC. Food intake had a substantial effect on the PK of dC and dT, leading to a median predicted AUC0-24,ss at least 2-fold higher in fed conditions compared to fasted conditions for both nucleosides. This model provides an understanding of the PK of dC and dT following oral administration of doxecitine and doxribtimine to support potential use in patients with TK2d. Study funded by UCB.
[1] Garone C et al. J Med Genet. (2018) 55, 515-521. [2] Berardo A et al. J Neuromuscul Dis. (2022) 9, 225-235. [3] Dominguez-Gonzalez et al. Ann Neurol. (2019) 860, 293-303. [4] Beal SL et al. 1989-2011. NONMEM 7.5 Users Guides. Gaithersburg, MD: ICON plc; 2020.