Disease pharmacokinetic-pharmacodynamic (PKPD) modelling to support the development of gene therapy treatments for rare diseases
Zinnia P Parra-Guillén (1,2), Diego Vera-Yunca (1,2), Lei Jiang (3), Marjie Hard (3), Lin T. Guey (3), Iñaki F. Troconiz (1,2)
(1) Department of Pharmaceutical Technology and Chemistry, School of Pharmacy and Nutrition, University of Navarra, (2) Navarra Institute for Health Research (IdisNA), (3) Moderna, Inc, Boston, MA
Introduction: Acute intermittent porphyria (AIP) is a metabolic rare disease caused by the hepatic deficiency of the enzyme porphobilinogen deaminase (PBGD), third enzyme in the heme biosynthesis pathway. In this context, therapies that restore enzyme levels in the liver are an appealing option [1]. Different mRNA sequences encoding for the PBGD enzyme and encapsulated in different lipid nanoparticle formulations have been developed by Moderna, Inc [2].
Objectives: The goal of this analysis was to build a mechanistic computational model describing longitudinal pharmacokinetic (i.e. liver PBGD activity, PK) and pharmacodynamic (i.e. 24-h urinary heme precursors, PD) data obtained in the porphyric pre-clinical arena across different species and using different PBGD mRNA compounds in order to project the results to humans.
Methods: To mimic porphyric acute attacks, porphyrogenic drugs (e.g. phenobarbital) were daily administered for 2-5 days over one or more challenges. Then, treated animals received mainly one or up to 3 doses of different PBGD mRNA compounds -i.e., PBGD mRNA sequence & lipid formulation- on day 2 or day 3. In total, 8 different sequences encoding for the PBGD enzyme and encapsulated in 3 different lipid nanoparticle systems were available for the analysis.
The disease PKPD model for AIP C57BL mice is comprised of the following main processes: (i) the PBGD PK model describing mRNA release from the formulation, degradation and translation to the encoded PBGD protein in the liver, (ii) the disease model characterising the urinary excretion of heme precursors (ALA, PBG and porphyrins [POR]) during phorphyric acute attacks in the absence of treatment, and (iii) the PBGD activity model accounting for the normalisation of heme precursor in urine in the presence of PBGD enzyme.
To account for the additional liver PBGD activity data from wild type animals (C57BL mice, Sprague Dawley rats, New Zealand rabbits and cyno), and the PD data collected in wild type rats and rabbits, species-specific parameters were estimated without modifying the structure of the AIP mouse model. Finally, baseline PBGD activity levels of AIP patients were used to extrapolate preclinical results to clinical scenarios.
Results: Assuming formulation-specific release parameters and mRNA sequence-specific degradation parameters, the proposed disease PKPD AIP mouse model successfully described all available experimental scenarios for the different mRNA compounds using a common model structure. More than two-fold differences were observed between formulation release parameters, whereas larger variations were obtained across sequences with degradation values ranging between 2.56x10-4 to 8.9x10-3 h-1. Differences in the response (i.e. reduction of urinary precursor accumulation) were thus explained at the PK level, since PBGD efficacy was preserved across mRNA compounds. PBGD activity levels achieved during the initial acute attack were sufficient to inhibit more than 90 % of the drug-induced accumulation of precursors (compared to baseline) with the majority of the mRNA compounds except for one mRNA sequence.
An adequate data characterization was obtained when using the AIP mice model, but adjusting the PBGD activity at baseline for the different species (dose was adjusted per animal weight) and estimating species-specific excretion rate constants of the heme precursor. The final model was used to predict in silico the inhibition that the different mRNA compounds would offer in case of a theoretical acute attack in humans assuming a PBGD activity model similar to that of AIP mice. Under these assumptions, predicted liver PBGD activity levels in AIP patients would remain above the normal PBGD activity levels -quantified in healthy untreated donors- for up to two months after the administration of the standard dose of 1mg/kg of some evaluated mRNA compounds to AIP patients.
Conclusions: In summary, an integrative quantitative framework capable to describe the effects of novel mRNA compounds on the accumulation of heme precursors in urine during-induced acute attack across different animal species has been proposed. This framework has been used to project the time course of the different mRNA compounds to humans. Moreover, it has the potential to be expanded with additional information characterizing the time course of urine heme precursors in humans during acute attacks to predict in silico the pharmacodynamic response during PBGD mRNA treatment.
References:
[1] Fontanellas et al. Exp Rev Mol Med. 18: e17 (2016)
[2] Jiang et al. Nat Med. 24:1899-1909 (2018)