Quantitative Systems Pharmacology: A Case for Disease Models.

Quantitative systems pharmacology (QSP) has emerged as an innovative approach in model-informed drug discovery and development, supporting program decisions from exploratory research through late-stage clinical trials. In this commentary, we discuss the unique value of disease-scale "platform" QSP models that are amenable to reuse and repurposing to support diverse clinical decisions in ways distinct from other pharmacometrics strategies.

A quantitative systems pharmacology model of blood coagulation network describes in vivo biomarker changes in non-bleeding subjects.

Essentials Baseline coagulation activity can be detected in non-bleeding state by in vivo biomarker levels. A detailed mathematical model of coagulation was developed to describe the non-bleeding state. Optimized model described in vivo biomarkers with recombinant activated factor VII treatment. Sensitivity analysis predicted prothrombin fragment 1 + 2 and D-dimer are regulated differently. SUMMARY: Background Prothrombin fragment 1 + 2 (F1 + 2 ), thrombin-antithrombin III complex (TAT) and D-dimer can be detected in plasma from non-bleeding hemostatically normal subjects or hemophilic patients. They are often used as safety or pharmacodynamic biomarkers for hemostatis-modulating therapies in the clinic, and provide insights into in vivo coagulation activity. Objectives To develop a quantitative systems pharmacology (QSP) model of the blood coagulation network to describe in vivo biomarkers, including F1 + 2 , TAT, and D-dimer, under non-bleeding conditions. Methods The QSP model included intrinsic and extrinsic coagulation pathways, platelet activation state-dependent kinetics, and a two-compartment pharmacokinetics model for recombinant activated factor VII (rFVIIa). Literature data on F1 + 2 and D-dimer at baseline and changes with rFVIIa treatment were used for parameter optimization. Multiparametric sensitivity analysis (MPSA) was used to understand key proteins that regulate F1 + 2 , TAT and D-dimer levels. Results The model was able to describe tissue factor (TF)-dependent baseline levels of F1 + 2 , TAT and D-dimer in a non-bleeding state, and their increases in hemostatically normal subjects and hemophilic patients treated with different doses of rFVIIa. The amount of TF required is predicted to be very low in a non-bleeding state. The model also predicts that these biomarker levels will be similar in hemostatically normal subjects and hemophilic patients. MPSA revealed that F1 + 2 and TAT levels are highly correlated, and that D-dimer is more sensitive to the perturbation of coagulation protein concentrations. Conclusions A QSP model for non-bleeding baseline coagulation activity was established with data from clinically relevant in vivo biomarkers at baseline and changes in response to rFVIIa treatment. This model will provide future mechanistic insights into this system.

Making Every Subject Count: A Case Study of Drug Development Path for Medication in a Pediatric Rare Disease.

Approximately 50% of rare diseases are evident in children. Fatal disease prognosis and lack of treatments causes 30% of affected children to not live past their fifth birthday. This clear sense of urgency demands innovation and acceleration in drug development. A case study is discussed highlighting the need for data-rich phase I study design, extensive use of modeling and simulation, use of diverse data sources, and input from collaborators to respond to this urgent call.

Using a Systems Pharmacology Model of the Blood Coagulation Network to Predict the Effects of Various Therapies on Biomarkers.

A number of therapeutics have been developed or are under development aiming to modulate the coagulation network to treat various diseases. We used a systems model to better understand the effect of modulating various components on blood coagulation. A computational model of the coagulation network was built to match in-house in vitro thrombin generation and activated Partial Thromboplastin Time (aPTT) data with various concentrations of recombinant factor VIIa (FVIIa) or factor Xa added to normal human plasma or factor VIII-deficient plasma. Sensitivity analysis applied to the model revealed that lag time, peak thrombin concentration, area under the curve (AUC) of the thrombin generation profile, and aPTT show different sensitivity to changes in coagulation factors' concentrations and type of plasma used (normal or factor VIII-deficient). We also used the model to explore how variability in concentrations of the proteins in coagulation network can impact the response to FVIIa treatment.

A multiscale model of interleukin-6-mediated immune regulation in Crohn's disease and its application in drug discovery and development.

In this study, we have developed a multiscale systems model of interleukin (IL)-6-mediated immune regulation in Crohn's disease, by integrating intracellular signaling with organ-level dynamics of pharmacological markers underlying the disease. This model was linked to a general pharmacokinetic model for therapeutic monoclonal antibodies and used to comparatively study various biotherapeutic strategies targeting IL-6-mediated signaling in Crohn's disease. Our work illustrates techniques to develop mechanistic models of disease biology to study drug-system interaction. Despite a sparse training data set, predictions of the model were qualitatively validated by clinical biomarker data from a pilot trial with tocilizumab. Model-based analysis suggests that strategies targeting IL-6, IL-6Rα, or the IL-6/sIL-6Rα complex are less effective at suppressing pharmacological markers of Crohn's than dual targeting the IL-6/sIL-6Rα complex in addition to IL-6 or IL-6Rα. The potential value of multiscale system pharmacology modeling in drug discovery and development is also discussed.CPT: Pharmacometrics & Systems Pharmacology (2014) 3, e89; doi:10.1038/psp.2013.64; advance online publication 8 January 2014.

Predictability of intranasal pharmacokinetics in man using pre-clinical pharmacokinetic data with a dopamine 3 receptor agonist, PF-219061.

(R)-3-(4-propylmorpholin-2-yl) phenol (PF-219061) is a potent, selective agonist of the dopamine 3 receptor for the treatment of female sexual dysfunction. In vivo, PF-219061 exhibits liver blood flow clearance in both rat and dog. Oral bioavailability was 0.7% in dog and less than 5% in rat. Intranasal dosing was investigated to improve bioavailability. Pre-clinical assessments in rat and dog demonstrated intranasal bioavailabilities of 16-38% in rat and 54-61% in dog with very rapid absorption. It was predicted that an intranasal dose in man would give approximately 25-50% bioavailability. The clinical data verified the preclinical predictions demonstrating rapid absorption and approximately dose-proportional increases in exposure. The intranasal bioavailability in man was estimated to be 26-38%. These findings indicate the potential utility of intranasal dosing as a route that circumvents the first-pass effects for PF-219061 resulting in high exposures.

A randomized, multicenter study of subcutaneous and intravenous darbepoetin alfa for the treatment of chemotherapy-induced anemia.

BACKGROUND: This randomized, open-label study evaluated the efficacy, safety and pharmacokinetics of darbepoetin alfa administered intravenously (i.v.) or subcutaneously (s.c.) in chemotherapy-induced anemia. PATIENTS AND METHODS: Patients received darbepoetin alfa i.v. (n=59) or s.c. (n=59) at a dose of 4.5 mug/kg once weekly for 6 weeks (correction phase) followed by 4.5 mug/kg once every 3 weeks for the remainder of the 18-week treatment period (maintenance phase). RESULTS: During the correction phase, the mean [95% confidence interval (CI)] change in hemoglobin (intention-to-treat) was 1.1 (0.6-1.5) g/dl in the i.v. group and 1.3 (0.9-1.7) g/dl in the s.c. group; using available data, the mean change was 1.4 (1-1.9) g/dl and 1.6 (1.2-2) g/dl, respectively. The percentage (95% CI) of patients maintaining hemoglobin (i.e. average decrease < or =0.5 g/dl) during the maintenance phase was similar between the i.v. (82%; 95% CI 66% to 92%) and s.c. (80%; 95% CI 66% to 90%) groups. Thirty-five per cent (95% CI 20% to 50%) of patients in the i.v. group and 32% of patients in the s.c. group (95% CI 18% to 45%) received red blood cell transfusions during week 5 to the end of the treatment period. Darbepoetin alfa was well tolerated in both groups. No significant difference (P=0.36) in weekly darbepoetin alfa serum concentrations was observed between groups. CONCLUSIONS: Darbepoetin alfa can be administered i.v. or s.c. at equal doses for the treatment of anemia in this setting.

Pharmacokinetics of novel erythropoiesis stimulating protein (NESP) in cancer patients: preliminary report.

Anaemia is a common occurrence in patients with cancer, and currently can be treated in several ways. Novel erythropoiesis stimulating protein (NESP, darbepoetin alfa) was created using site-directed mutagenesis to have 8 more sialic acid side chains than recombinant human erythropoietin (rHuEPO). The additional sialic acid content has resulted in an approximately 3-fold greater half-life relative to rHuEPO in patients with chronic renal failure. This study evaluates the pharmacokinetic profile of NESP in patients receiving multiple cycles of chemotherapy. Anaemic patients (haemoglobin < or = 11.0 g dl(-1)) who had non-myeloid malignancies received NESP weekly (2.25 mcg kg(-1) wk(-1)) under the supervision of a physician, starting on day 1 of chemotherapy for 3 chemotherapy cycles given at 3-week intervals. Blood samples were collected during chemotherapy cycles 1 and 3 for pharmacokinetic analysis. All patients were followed for 4 weeks after treatment. NESP was well tolerated by all patients. After a single dose during chemotherapy cycle 1, pharmacokinetic parameters (mean (SD), n) for the first 15 patients were: T(max)86.1 (22.8) h (n = 14); C(max)9.0 (5.1) ng ml(-1)(n = 14); t(1/2,z)32.6 (11.8) h (n = 7); CL/F 3.7 (1.0) ml h(-1) kg(-1)(n = 7). The subjects for whom all parameters could be calculated may represent a sub-group of the entire population. Similar results were obtained in cycle 3. In addition, haemoglobin response data suggests that, in this patient population, dosing less frequently than the 3 times weekly doses used for rHuEPO may be possible while improving anaemia.

Percutaneous absorption and disposition of [14C]chlordecone in young and adult female rats.

The objective of this study was to evaluate the effect of age and dosage on percutaneous absorption and disposition of [14C]chlordecone (Kepone) and to describe results using a physiological based pharmacokinetic (PBPK) model. Female Fischer 344 rats 33 and 82 days old were used as the young and adult animal models, respectively, and were studied over a 10-fold dose range. [14C]Chlordecone (0.286 micromol/cm2) was applied to dorsal skin (2. 3% BSA) and radioactivity was quantified in selected tissues and excreta up to 120 h. Absorption and disposition were also determined at three dose levels up to 2.68 micromol/cm2; fraction absorbed decreased as dose increased. In vitro percutaneous absorption was measured by static and flow-through methods; these yielded similar penetration rates, which were lower than those obtained in vivo. In vivo percutaneous absorption over 120 h was 14.4+/-0.99 and 14.2+/-1. 5% dose in young and adults, respectively. Organ and tissue content increased over time (carcass>liver>kidney), indicating prolonged absorption. Statistical differences between young and old were found for liver, skin, and urine, but not for absorption. Excretion occurred primarily in feces, but also in urine. A biophysically based percutaneous model was fitted to both young and adult in vivo absorption data. This was embedded in a whole body PBPK model which, upon optimization with SAAM II, estimated apparent tissue partition coefficients, urinary and fecal excretion rates, and parameters characterizing hepatic nonlinear uptake of bound chlordecone. The model reasonably predicted tissue chlordecone content at higher doses, when decreased absorption was accounted for.

A pharmacokinetic/pharmacodynamic comparison of SAAM II and PC/WinNonlin modeling software.

This paper presents a detailed comparison of the kinetic analysis software packages SAAM II and PCNonlin/WinNonlin, based on benchmark modeling problems reported in "Pharmacokinetic andPharmacodynamic Data Analysis: Concepts and Applications" (Gabrielsson and Weiner, 1994) and seven additional models. For each model, both software packages were presented with identical implementations. Models were initially executed in PCNonlin or WinNonlin and automated comparisons with SAAM II made using Microsoft Test. Models investigated included one- and multicompartment models with nonlinearities, multiple inputs and samples, multiple simultaneous experiments, and linear equations. Maximum number of compartments, data sets, and parameters were 9, 5, and 10, respectively. We compared 88 different models, many of them in different configurations, e.g., different weighting schemes or different parameter limits. The total number of attempted comparisons between SAAM II and PCNonlin was 161, of which 142 executed without problems. Parameter estimates, their precision (standard errors), and model predictions were compared; a difference of 1% or less was considered "agreement". Observed differences, mainly in parameter standard errors, can be accounted for in terms of different optimization algorithms, convergence criteria, and individual capabilities. In general, there was good agreement (<1% difference) between SAAM II and PCNonlin in terms of parameter estimates and model predictions. However, due to differences in the optimization procedure, parameter standard errors showed considerable differences. Additionally, there were differences when multiple data sets were fitted, indicating the importance of different fitting procedures for interpreting multiple kinetic data sets. The full results of the comparison and the model files in SAAM II and PCNonlin/WinNonlin formats are available from the authors.

Clearance of postprandial and lipolytically modified human HDL in rabbits and rats.

Triglyceride (TG) enrichment of high density lipoproteins (HDL) in hypertriglyceridemic states renders the particles vulnerable to lipolysis, which reduces their size. In the present study we modified the size and composition of HDL in vivo in hypertriglyceridemic humans by administering a bolus of intravenous heparin, and tested the subsequent clearance of the isolated HDL particles in rabbits and rats. HDL was isolated by ultracentrifugation from 21 moderately hypertriglyceridemic humans, 5 h after ingestion of a high fat meal and then 15 min after an intravenous heparin bolus (60 U/kg). Postprandial large TG-rich preheparin HDL and small, TG-poor postheparin HDL were labeled with either 125I or 131I. The clearance of apoA-I associated with each HDL tracer was determined by injecting the tracers 1) simultaneously (n = 13) and 2) sequentially (n = 8) into male New Zealand White rabbits, an hepatic lipase-deficient animal, and 3) by injecting the tracers simultaneously into male Sprague-Dawley rats (n = 8), an animal that has hepatic lipase. Die-away curves of each radiolabeled tracer were analyzed using a two-pool model that assumes the existence of an intravascular pool in dynamic equilibrium with an extravascular pool. In the rabbit studies, the fractional catabolic rate (FCR) of small, postheparin TG-poor HDL was greater than the FCR of the larger TG-rich HDL (11% greater in the simultaneous study, P < 0.001, and 45% greater in the sequential study, P < 0.001). Opposite results were observed in rats as large TG-rich preheparin particles showed a greater FCR (1.8-fold) than smaller TG-poor postheparin HDL (P < 0.05). These data suggest that although size and composition of HDL can influence its catabolism, the effect is not always in the same direction and depends on other factors present in vivo.

A physiologically based pharmacokinetic model incorporating dispersion principles to describe solute distribution in the perfused rat hindlimb preparation.

A physiologically based pharmacokinetic model incorporating dispersion principles has been developed to describe outflow data from the isolated perfused rat hindlimb preparation, for the three reference markers 14C-sucrose, 14C-urea, and 3H-water and three 14C-labeled 5-n-alkyl-5-ethyl barbiturates; the methyl, butyl, and nonyl homologues. Also 51Cr-RBC and 125I-albumin were studied. The model consists of four parallel components representing each of the tissues comprising the hindlimb: skeletal muscle, skin, bone, and adipose. Attempts to simplify the model by using the principle of tissue lumping were made by examining the tissue equilibration rate constant k tau for each of respective tissues for each compound. It was found that simplification was only possible in the case of 3H-water data. The model took into account a possible shunting component in the skin tissue and incomplete mass but not volumetric recovery from the system. The dispersion model characterizes the relative spreading of solute on transit through a tissue bed by a dimension-less parameter DN. The estimated dispersion numbers (DN) obtained were in the region of 2.7-4.72, 8.39-15.54, 0.61-2.74, and 6.02-14.0 for skeletal muscle, skin, bone, and adipose, respectively, and were independent of the compound studied. These values are much larger than the range reported in the literature for hepatic outflow data, DN = 0.2-0.5, and suggest a greater heterogeneity of vascular flow in the different component tissues of the rat hindlimb.

Pharmacokinetic-pharmacodynamic modeling of stimulatory and sedative effects of alprazolam: timing performance deficits.

Alprazolam decreased the reinforcement rate and increased the shorter-response rate of contingency-controlled timing behavior under a differential reinforcement of low-rate schedule (DRL 45-s) in rats. An integrated pharmacokinetic-pharmacodynamic (PK-PD) model was developed to describe and characterize the effects of i.v. and s. c. administration of alprazolam. The onset, peak and disappearance of alprazolam effects were evaluated during a 3-hr session. After s. c. alprazolam administration, two peak increases in shorter-response rate occurred at moderate alprazolam serum levels, first in the ascending and then in the descending limb of the concentration-time profile. We used a stimulation-sedation PD model incorporating two opposing effect-link sigmoidal Emax functions to model the two peaks after s.c. alprazolam administration. The model suggested that alprazolam possesses both stimulatory and sedative effects in a continuous but sequential fashion, which corresponded to low- and high-concentration effects as indicated by the EC50 values of 0.09 and 0.18 microg/ml, respectively. Owing to the rapid onset of i.v. administration, the first peak (a transition phase before the onset of the sedative effect) was absent, with the presence of the second peak again coinciding with the offset of the sedative effect. The reinforcement rate (IC50 = 0.02 microg/ml) characterized by the indirect response model to account for the initial hysteresis is an index for evaluating the deficit in timing performance. Although the effects of alprazolam can be described in behavioral terms, simultaneous PK-PD optimization numerically defines the performance and hypothesizes the coexistence of stimulation and sedation components for alprazolam. The stimulation-sedation model may help in delineating the possible mechanisms for adverse rebound side effects and of tolerance in humans.

Discrepancies in pharmacokinetic parameter estimation between bolus and infusion studies in the perfused rat hindlimb.

Isolated, perfused rat hindlimb consists of skeletal muscle, skin, bone, and adipose. Hence, it is a heterogeneous preparation composed of slowly equilibrating tissues of different characteristics and fractional flow rates. This paper shows how caution should be exercised in interpreting the results following bolus administration and subsequent statistical moment analysis of intravascular markers (51Cr-erythrocytes and 125I-albumin) and lipophilic barbiturates. For the intravascular markers, the events in the hindlimb are overshadowed by events in the connecting tubing and cannulas, due to their comparable volumes. For the barbiturates, these estimates appear to apply to short-term effects as the volume estimates obtained following infusion to steady state are greater than after bolus administration. For the extravascular markers, 14C-sucrose, 14C-urea, and 3H-water, no such time dependency was shown. However, it is only from the outflow profiles following bolus administration that events in the tissue beds can be elucidated.