Aniruddha Amrite1, David McDougall2, Ralf Rosskamp1
1 Dicerna Pharmaceuticals,Inc. Lexington, MA; 2 Model Answers R&D Pty Ltd. Brisbane QLD
Background: Primary hyperoxaluria (PH) is a family of rare diseases characterized by hepatic overproduction of oxalate due to three distinct genetic mutations. Nedosiran (DCR-PHXC) is an investigational siRNA therapeutic targeting the LDHA enzyme, which is involved in the ultimate step of hepatic oxalate production and has the potential to treat all three known subtypes of PH.
Objectives: 1) To develop a population (POP) PK model for DCR-PHXC and determine the sources of inter-subject variability using covariate analyses. 2) To develop a POP PK-PD model based on the 24‑hour urinary oxalate excretion (UOx) data in PH patients. 3) To perform PK-PD simulations to inform the dosing strategy for the pivotal trial of nedosiran in PH patients.
Methods: PK and PD data from a single ascending dose study of nedosiran in normal healthy volunteers (NHV) (0.3 mg/kg to 12 mg/kg, n=15, adults only) and PH patients (1.5 mg/kg to 6 mg/kg, n= 17, adults and adolescents) were modeled using non-linear mixed-effects (NLME) methods. Initially, a POP-PK model was developed across NHV and PH patients. Allometric scaling was included on clearance and volume terms with exponents fixed to 0.75 and 1, respectively. The individual PK predicted concentrations were used to drive the PD response in PH patients. Base model, inter-subject variability, and covariate effects were estimated using standard POP-PK-PD modeling techniques. The parameters for the final PK and PD model are presented in Table 1 and Table 2 respectively. The models were evaluated for collinearity and model stability and qualified using prediction corrected visual predictive checks (pcVPC). A dataset of virtual subjects was generated from the National Health and Nutrition Examination Survey to simulate PK and PD following multiple doses of nedosiran. Simulations were performed for different dosing regimens (including loading and maintenance doses, flat dosing, and body weight dosing) for adults and adolescents, and for different degrees of renal impairment (normal renal function, mild renal impairment, moderate renal impairment) to inform a dosing strategy to achieve pre-defined PD targets.
Results: The best PK model to fit the available data was a 2-compartment disposition model with a parallel linear and non-linear clearance. Absorption was described by a dual zero order and first order input following an absorption lag. The only significant covariate was eGFR on the linear clearance. The pcVPC confirmed that the predictive performance of the model was acceptable. An indirect response model was used to describe the PD response in PH patients following a single dose of nedosiran, with the drug effect included as an Imax function on the production of UOx.. No statistically significant covariates were identified during the PD model development. Based on simulations from the final PK-PD model, a dosing regimen of 170 mg once a month (for participants ≥ 50 kg) and 136 mg once a month (for participants < 50 kg) was identified for adult participants in the pivotal studies. Simulations also indicated that no dose adjustments are required for mild and moderate renal impairment and the same dosing regimen can be used in adolescents (Aged ≥ 12 years to < 18 years).
Conclusions: The final PK-PD model adequately described the nedosiran concentration and UOx response following a single dose of nedosiran in adults and adolescents. PK-PD simulations supported
the dosing regimen selected for nedosiran in pivotal studies. The model simulations were based on single dose data and will be updated once multiple dose data are available.
Table 1: Final Population PK Model Parameters
|
Parameter |
Estimate (%RSE) |
|
Apparent clearance (CL/F, L/hr) ‡ |
8.3 (8.9) |
|
Maximum metabolic rate (Vmax, ng/hr) |
2.25 (33.2) |
|
Michaelis-Menten constant (KM, ng/mL) |
144.5 (42.5) |
|
Apparent inter-compartmental clearance (Q/F, L/hr) |
1.354 (27.2) |
|
Exponent for WT on CL/F and Q/F |
0.75 FIX |
|
Exponent for eGFR on CL/F |
0.905 (28.6) |
|
Apparent volume of distribution for the central compartment (Vc/F, L) |
86.2 (10.0) |
|
Apparent volume of distribution for the peripheral compartment (Vp/F, L) |
1588 (53.4) |
|
Exponent for WT on Vc/F & Vp/F |
1 FIX |
|
Fraction of the dose absorbed via the first pathway (FR1, %) † |
22.0 (10.5) |
|
Duration of zero-order input into the first compartment (D1, hr) |
0.722 (6.6) |
|
First-order absorption rate constant (KA, hr−1) |
0.122 (8.7) |
|
Absorption lag time (ALAG, hr) |
1.65 (4.9) |
|
Between subject variability on Vmax (%CV) |
27.0 (38.2) |
|
Between subject variability on Vc/F (%CV) |
35.1 (23.6) |
|
Between subject variability on KA (%CV) |
34.9 (20.8) |
|
Residual unexplained variability (%) |
40.7 (7.2) |
RSE = relative standard error, %CV = % coefficient of variation, WT = total body weight. †Converted out of logit domain. Clearance and volume parameters scaled using allometry centered on 70 kg human. ‡CL/F = 8.3 * (WT/70)0.75 * (eGFR/94)0.905
Table 2: Final Population PD Model Parameters
|
Parameter |
Estimate (%RSE) |
|
Baseline 24 hour urinary oxalate (UO24, µmol) |
1378 (12.2) |
|
First-order elimination rate of UO24 (Kout, hr-1) |
0.000997 (0.57) |
|
Half maximal inhibitory concentration (IC50, ng/mL) |
0.0136 (0.19) |
|
Maximum inhibitory effect (IMAX (%)) |
80.8 (8.5) |
|
Hill coefficient (γ) |
0.317 (1.3) |
|
Between subject variability on Kout |
81.7 (16.9) |
|
Between subject variability on Baseline UO24 |
37.0 (12.1) |
|
Additive residual unexplained variability (SD, µmol) |
275.5 (12.3) |
Note: PK parameters were fixed to the PK final EBE estimates in Objective 1 (Table 10). BSV = between-subject variability, SE = standard error, CV% = percent coefficient of variation, and where Kin (zero-order rate constant for production of response) = Baseline * Kout
Reference: PAGE () Abstr 9237 [www.page-meeting.org/?abstract=9237]
Poster: Clinical Applications