A QSP Modeling Approach for Neurodegenerative Disease Drug Development

Model-informed Drug Development approaches such as physiologically based pharmacokinetics (PBPK) and quantitative systems pharmacology (QSP) have proven to be effective for supporting and streamlining drug development. PBPK has been used  to inform dosing and clinical trial design through the understanding of the mechanisms of drug disposition and the prediction of: drug-drug interactions, variation in drug clearance and drug exposure.  QSP combines computational modeling and experimental data to examine the relationships between a drug, the biological system, and the disease process. This approach has enormous potential to improve pharma R&D productivity and inform decision-making across the drug development process from early discovery to Phase 3. QSP has been used to design First-in-Human clinical trials, optimize immuno-oncology drug discovery, and manage immunogenicity in biologic drugs.

Alzheimer’s (AD) is a devastating incurable neurodegenerative disease characterized by memory loss, confusion, and dementia. AD is sixth-leading cause of death in the US. Current treatments may help alleviate some dementia symptoms but limited understanding of the disease has thwarted developing treatments that can halt or reverse disease progression. QSP models can be used to validate targets and elucidate the molecular and cellular processes in disease pathways.1

However, mechanistic approaches to support developing neurodegenerative disease therapeutics poses challenges. Determining changes in brain volume is needed to estimate the concentration of specific biomarkers, e.g. amyloid beta, cholesterol, and others, and it is an important parameter evaluated for drug discovery of agents to treat these complex diseases. In general, quantitative models assume a constant brain volume. Thus, the development of a dynamic model is needed to consider the effect of aging on brain volume to accurately estimate biomarker concentrations.

The Relationship between Aging and Brain Size

Aging and gender are important factors to determine the size of major organs, e.g. the liver, and their impact on predicting safe and efficacious individualized dosing using PBPK models. In the case of neurodevelopment and disease, analyzing geriatric clinical data is a means to explore the effects of aging on brain size given that brain size varies with the lifespan. Full brain development occurs in our mid 20’s, after which brain size gradually decreases. The decline happens primarily in the frontal cortex due to neuronal death and/or degeneration. On average, the cerebral cortex loses 50,000 to 60,000 neurons per day. In AD patients, the brain is approximately 10% smaller compared to healthy elderly people. In advanced AD, the cerebral cortex atrophies.

Biomarkers for Alzheimer’s Disease Intervention

Biomarkers are used to pre-diagnose cases of AD and determine disease progression. These biomarkers include beta amyloid (Aβ), soluble Aβ precursor protein (sAPP), beta-site amyloid precursor protein cleaving enzyme 1(BACE-1), and Aβ autoantibodies.

Aβ belongs to a family of proteins released by proteolytic cleavage of the glycoprotein amyloid precursor protein (APP). Aβ peptides are sensitive and specific diagnostic biomarkers known to play a central role in Alzheimer’s disease pathology and may also be targets for treating AD.  Aβ is elevated in AD patients and aggregates in plaques. The plaques develop between cells and disrupt neuronal function.

Using QSP to Study Neurodegenerative Diseases

We initiated a study to design a dynamic QSP model that incorporates aging and brain size which would address the complexity of developing drugs for neurodegenerative diseases. Since Aβ is an important marker for AD we chose this for our modeling studies. Studying the accumulation of specific biomarkers, e.g. Aβ, using a dynamic model, which considers the patient’s brain volume change due to aging, allows us to see how the decrease in brain volume relates to the increasing biomarker concentration levels observed clinically.

Our goals for the model were:

  1. Develop a QSP model that considers brain volume changes when estimating biomarker concentrations
  2. Compare the results of the variable brain volume model with a constant volume model
  3. Evaluate the effect of the variability of brain volume changes in a given population on estimating biomarker concentrations

To include the specific biological mechanisms into the model, we needed to better understand the mechanisms that correlated to Aβ (both peptides, Ab40 and Ab42) production and build a multi-step biology plan. To do this, we built an initial biological map to demonstrate how the process starts with the cleavage of APP to the cell membrane by either the alpha-secretase or BACE-1 pathway, which produce soluble APP. Another component was to build an additional step which generates insoluble Aβ peptides which produce the plaques.

Once we built the biological map to describe APP processing using our findings and previously published models, we created the equations related to this map under two different case scenarios: (1) considering brain size constant without including any dynamic aging effect, and (2) a complete dynamic model for the brain development/aging, which covers the development stages through the lifespan and gradual decrease in size. 2,3,4

A two-step approach was used for our QSP modeling methodology as shown below.

Step 1. Describe the main mechanism for Aβ production and correlation in the brain

Step 2. Combine the QSP model of APP processing and Aβ metabolism with a model that considers the effects of aging on brain volume

 

In a future blog, I will review the results of our simulations using different parameter values for different types of populations, and how we implemented this integrated approach with the original model and the brain volume changing in time.

To learn more on the QSP model development for this project, review our presentation given during the 2018 PAGE conference in Montreux, Switzerland.

References

  1. Lloret-Villas A, Varusai TM, Juty N, et al. (2017). The Impact of Mathematical Modeling in Understanding the Mechanisms Underlying Neurodegeneration: Evolving Dimensions and Future Directions. CPT Pharmacometrics Syst. Pharmacol., 6(2), 73-86.
  2. Clausznitzer D, Pichardo-Almarza C, Relo, AL, et al. (2018). Quantitative Systems Pharmacology Model for Alzheimer Disease Indicates Targeting Sphingolipid Dysregulation as Potential Treatment Option. CPT Pharmacometrics Syst. Pharmacol., 7, 759-770.
  3. Ortega F, Stott J, Visser S, and Bendtsen C. (2012). Interplay between α-, β-, and γ-Secretases Determines Biphasic Amyloid-β Protein Level in the Presence of a γ-Secretase Inhibitor. Biol. Chem., 288, 785-792.
  4. Potter R., Patterson BW, Elbert DL, et al. (2013). Increased in Vivo Amyloid-β42 Production, Exchange, and Loss in Presenilin Mutation Carriers. Science Translational Medicine, 5 (189), pp. 189ra77.
  5. Borzage M, Blüml S, and Seri I. (2014). Equations to describe brain size across the continuum of human lifespan. Brain Struct. Funct. 219,141–150.