Translational quantitative systems pharmacology; crossing borders between experimental and computational drug development using zebrafish as model organism

Rob C. van Wijk 1,2, Elke H.J. Krekels 1, Wanbin Hu 3, Astrid van der Sar 4, Sharka M. Dijkema 1, Dirk-Jan van den Berg 1, Rida Bahi 1, Jeremy Liu 1, Theo Verboom 4, Fons J. Verbeek 5, Ulrika S.H. Simonsson 2, Herman P. Spaink 3, Piet H. van der Graaf 1,6

1 Systems Biomedicine and Pharmacology, Leiden Academic Centre for Drug Research, Leiden University, Leiden, The Netherlands; 2 Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden; 3 Division of Animal Sciences and Health, Institute of Biology Leiden, Leiden University, Leiden, The Netherlands; 4 Department of Medical Microbiology and Infection Control, VU University Medical Center, Amsterdam, The Netherlands; 5 Section Imaging and Bioinformatics, Leiden Institute for Advanced Computer Science, Leiden University, Leiden, The Netherlands; 6 QSP, Certara, Canterbury, UK

Objectives: Quantitative systems pharmacology (QSP) is considered a computational discipline, but innovation in experimental methods is equally essential for its development [1]. QSP models require large datasets that go beyond traditional measurements in pharmacology. For this, high-throughput experimentation is desirable, which is generally limited to in vitro systems. Recently, the zebrafish larva has been proposed as whole organism to cross the borders between in vitro experiments and in vivo studies, as their high-throughput potential allows for the generation of the required data by for example non-invasive (fluorescence) imaging of infectious diseases [2,3]. However, at least two challenges exist for the zebrafish larva to reach its full potential as a model organism in QSP:

First, external drug concentrations upon waterborne treatment are generally taken as static exposure measure. Recently, we developed a methodology to measure internal drug exposure over in larval homogenates [4,5] as well as in blood samples [6], the latter of which is essential to quantify distribution volume and absolute clearance values, which is the basis for between-species pharmacokinetic (PK) translation.

Second, to translate drug effects between species, differences in disease and disease progression in these species need to be understood. The quantified internal exposure over time needs to be linked to observations of the disease and changes therein to determine the concentration-effect relationship. In addition, translational factors between species that correct for the between-species differences are required [7].

Here, we assess the principle of translatability of pharmacological findings in the zebrafish larva to higher vertebrates, by quantifying the PKPD of the antibiotic isoniazid in the zebrafish larva as disease model for tuberculosis (TB), and translate isoniazid effects to humans. To that aim, we developed and optimized novel experimental methods to measure drug concentration and effect in this new model organism with the potential of high-throughput application.

Methods: Zebrafish larvae infected with Mycobacterium marinum are commonly used as experimental system for TB research. To determine which laboratory strain is most similar to Mycobacterium tuberculosis, in vitro natural growth curves of different strains (E11, MUSA) over 221 days were quantitatively analysed with a multistate tuberculosis pharmacometric (MTP) model that has been previously developed for M. tuberculosis and applied to in vitro, murine and clinical data [8–10]. These fits were compared to reported in vitro M. tuberculosis natural growth curves [8].

To quantify the pharmacokinetics-pharmacodynamics (PK-PD) of isoniazid, zebrafish larvae were infected by microinjection with 200 CFU M. marinum E11 strain, which was most similar to M. tuberculosis, at 1 day post fertilization (dpf) and, after establishment of the infection, treated with isoniazid at different doses (7.5-150 mg/L) from 3 to 5 dpf. Internal exposure was quantified by sensitive LC-MS at different timepoints in whole homogenates and in nanoscale blood samples. In a separate group, at 3 dpf the bacterial burden was quantified by fluorescence microscopy imaging which was repeated per individual at 4 and 5 dpf. Non-linear mixed effects modelling using NONMEM 7.3 [11] through interface Pirana 2.9.6 [12] and PsN 4.7.0 [13] was used to analyse the PK and PD profiles. Linear and non-linear elimination were tested, as well as age as covariate on absorption and elimination. To quantify bacterial burden exponential, Gompertz, or logistic growth functions were tested. The drug effect was tested as a linear, Emax, or sigmoidal function on the bacterial growth.

Two translational factors were identified to correct for 1) the difference in sensitivity of M. tuberculosis and M. marinum to isoniazid based on their respective MICs (0.2 and 15 mg/L, respectively), and 2) the difference in stage of infection between an acute experimental infection in zebrafish larvae and a chronic clinical infection in patients. The latter was based on scaling the isoniazid drug response as quantified on the logarithmic phase and on the stationary phase (150 days of infection [7]) by taking the ratio of the maximal drug effect thereon [14].

To translate the isoniazid drug effect from zebrafish larvae to humans, the concentration-effect relationship and the two translational factors described above were combined with simulated concentration-time profiles for a week of daily dosing with 150, 300, and 450 mg daily based on a previously published PK model for isoniazid in humans [15]. Translated isoniazid effect in humans was compared with published bacterial burden after isoniazid monotherapy [16–18].

Results: Model fits with the MTP model showed different natural growths for the M. marinum strains. Strain E11 showed with a Gompertz functions for growth with a similar growth rate to M. tuberculosis a profile more similar to M. tuberculosis, shifting to slow- and non-multiplying sub-states after 30-60 days. This suggests that studies using E11 is most representative for M. tuberculosis behaviour.

Isoniazid blood concentrations in zebrafish larvae were with median 30.3 mg/L 20% of the external concentration and within the clinical range of isoniazid concentrations in patients [19,20]. A dose-linear increase in isoniazid internal exposure was observed, with steady state amounts ranging from 30 – 100 pmole/larva based on larval homogenates. A one compartment model with first order absorption, including age as covariate, and linear elimination fitted the PK data best. Absolute clearance and distribution volume were estimated at 0.189 uL/h and 1.08 L/kg, respectively. Bacterial growth was best described by an exponential growth model, and the drug effect by a linear model with inter-individual variability on the inoculum and slope.

Isoniazid treatment effects were translated to humans. Subtherapeutic and supratherapeutic predictions were found for the doses 150 and 450 mg, as expected. For the therapeutic dose of 300 mg, the observed bacterial burden was in good agreement with the median prediction, especially for the first 48 hours which was the duration of the zebrafish experiment.

Conclusion: The principle of translating drug effects from zebrafish larvae to humans has been shown here,  based on experimental innovation, and modelling and simulation. Internal drug exposure and antibacterial response in zebrafish larva as tuberculosis disease model were translated to predict the isoniazid exposure-response relationship in humans. Crossing borders between innovative experimentation and advanced quantitative computational analysis, the QSP potential of zebrafish and zebrafish larvae in translation to higher vertebrates is improved. This integration inspires continuous collaboration between experimental and computational scientists to answer the key questions in early drug development.



References:
[1] Van Der Graaf, P.H. Pharmacometrics and/or Systems Pharmacology. CPT Pharmacometrics Syst. Pharmacol. 2019, 8, 331–332.
[2] Van Wijk, R.C.; Krekels, E.H.J.; Hankemeier, T.; Spaink, H.P.; Van der Graaf, P.H. Systems pharmacology of hepatic metabolism in zebrafish larvae. Drug Discov. Today Dis. Model. 2016, 22, 27–34.
[3] Schulthess, P.; Van Wijk, R.C.; Krekels, E.H.J.; Yates, J.W.T.; Spaink, H.P.; Van der Graaf, P.H. Outside-in systems pharmacology combines innovative computational methods with high-throughput whole vertebrate studies. CPT Pharmacometrics Syst. Pharmacol. 2018, 7, 285–287.
[4] Kantae, V.; Krekels, E.H.J.; Ordas, A.; González, O.; Van Wijk, R.C.; Harms, A.C.; Racz, P.I.; Graaf, P.H. Van Der; Spaink, H.P.; Hankemeier, T. Pharmacokinetic modeling of paracetamol uptake and clearance in zebrafish larvae: Expanding the allometric scale in vertebrates with five orders of magnitude. Zebrafish 2016, 13, 504–510.
[5] Van Wijk, R.C.; Krekels, E.H.J.; Kantae, V.; Harms, A.C.; Hankemeier, T.; Van der Graaf, P.H.; Spaink, H.P. Impact of post-hatching maturation on the pharmacokinetics of exogenous compounds in zebrafish larvae. Sci. Rep. 2018, 9, 2149.
[6] Van Wijk, R.C.; Krekels, E.H.J.; Kantae, V.; Ordas, A.; Kreling, T.; Harms, A.C.; Hankemeier, T.; Spaink, H.P.; van der Graaf, P.H. Mechanistic and quantitative understanding of pharmacokinetics in zebrafish larvae through nanoscale blood sampling and metabolite modelling of paracetamol. J. Pharmacol. Exp. Ther. 2019, 371, 15–24.
[7] Wicha, S.G.; Clewe, O.; Svensson, R.J.; Gillespie, S.H.; Hu, Y.; Coates, A.R.M.; Simonsson, U.S.H. Forecasting Clinical Dose-Response From Preclinical Studies in Tuberculosis Research: Translational Predictions With Rifampicin. Clin. Pharmacol. Ther. 2018, 104, 1208–1218.
[8] Clewe, O.; Aulin, L.; Hu, Y.; Coates, A.R.M.; Simonsson, U.S.H. A multistate tuberculosis pharmacometric model: A framework for studying anti-tubercular drug effects in vitro. J. Antimicrob. Chemother. 2016, 71, 964–974.
[9] Chen, C.; Ortega, F.; Rullas, J.; Alameda, L.; Angulo-Barturen, I.; Ferrer, S.; Simonsson, U.S.H. The multistate tuberculosis pharmacometric model: a semi-mechanistic pharmacokinetic-pharmacodynamic model for studying drug effects in an acute tuberculosis mouse model. J. Pharmacokinet. Pharmacodyn. 2017, 44, 133–141.
[10] Svensson, R.J.; Simonsson, U.S.H. Application of the Multistate Tuberculosis Pharmacometric Model in Patients With Rifampicin-Treated Pulmonary Tuberculosis. CPT Pharmacometrics Syst. Pharmacol. 2016, 5, 264–273.
[11] Beal, S.; Sheiner, L.; Boeckmann, A.; Bauer, R.J. (eds) NONMEM 7.3.0 Users Guides. (1989-2013). ICON Development Solutions, Hanover, MD, USA.
[12] Keizer, R.; Van Benten, M.; Beijnen, J.; Schellens, J.; Huitema, A. Pirana and PCluster: a modeling environment and cluster infrastructure for NONMEM. Comput Methods Programs Biomed 2011, 101, 72–79.
[13] Keizer, R.J.; Karlsson, M.O.; Hooker, A. Modeling and Simulation Workbench for NONMEM: Tutorial on Pirana, PsN, and Xpose. CPT Pharmacometrics Syst. Pharmacol. 2013, 2.
[14] Clewe, O.; Wicha, S.G.; de Vogel, C.P.; de Steenwinkel, J.E.M.; Simonsson, U.S.H. A model-informed preclinical approach for prediction of clinical pharmacodynamic interactions of anti-TB drug combinations. J. Antimicrob. Chemother. 2018, 73, 437–447.
[15] Wilkins, J.J.; Langdon, G.; McIlleron, H.; Pillai, G.; Smith, P.J.; Simonsson, U.S.H. Variability in the population pharmacokinetics of isoniazid in South African tuberculosis patients. Br. J. Clin. Pharmacol. 2011, 72, 51–62.
[16] Li, L.; Mahan, C.S.; Palaci, M.; Horter, L.; Loeffelholz, L.; Johnson, J.L.; Dietze, R.; Debanne, S.M.; Joloba, M.L.; Okwera, A.; et al. Sputum Mycobacterium tuberculosis mRNA as a marker of bacteriologic clearance in response to antituberculosis therapy. J. Clin. Microbiol. 2010, 48, 46–51.
[17] Johnson, J.L.; Hadad, D.J.; Boom, W.H.; Daley, C.L.; Peloquin, C.A.; Eisenach, K.D.; Jankus, D.D.; Debanne, S.M.; Charlebois, E.D.; Maciel, E.; et al. Early and extended early bactericidal activity of levofloxacin, gatifloxacin and moxifloxacin in pulmonary tuberculosis. Int. J. Tuberc. Lung Dis. 2006, 10, 605–612.
[18] Hafner, R.; Cohn, J.A.; Wright, D.J.; Dunlap, N.E.; Egorin, M.J.; Enama, M.E.; Muth, K.; Peloquin, C.A.; Mor, N.; Heifets, L.B.; et al. Early bactericidal activity of isoniazid in pulmonary tuberculosis: Optimization of methodology. Am. J. Respir. Crit. Care Med. 1997, 156, 918–923.
[19] Hemanth Kumar, A.K.; Kannan, T.; Chandrasekaran, V.; Sudha, V.; Vijayakumar, A.; Ramesh, K.; Lavanya, J.; Swaminathan, S.; Ramachandran, G. Pharmacokinetics of thrice-weekly rifampicin, isoniazid and pyrazinamide in adult tuberculosis patients in India. Int. J. Tuberc. Lung Dis. 2016, 20, 1236–1241.
[20] Ribera, E.; Azuaje, C.; Lopez, R.M.; Domingo, P.; Curran, A.; Feijoo, M.; Pou, L.; Sánchez, P.; Sambeat, M.A.; Colomer, J.; et al. Pharmacokinetic interaction between rifampicin and the once-daily combination of saquinavir and low-dose ritonavir in HIV-infected patients with tuberculosis. J. Antimicrob. Chemother. 2007, 59, 690–697.


Reference: PAGE 99 (2020) Abstr 9455 [www.page-meeting.org/?abstract=9455]
Oral: Lewis Sheiner Student Session

Link to DDMoRe model repository:http://repository.ddmore.foundation/model/DDMODEL00000311
Click to open PDF poster/presentation (click to open)
Top