Development of a quantitative systems pharmacology model of chronic kidney disease: metabolic bone disorder

Leveraging quantitative systems pharmacology and artificial intelligence to advance treatment of chronic kidney disease mineral bone disorder

Chronic kidney disease mineral bone disorder (CKD-MBD) is a complex clinical syndrome responsible for the accelerated cardiovascular mortality seen in individuals afflicted with CKD. Current approaches to therapy have failed to improve clinical outcomes adequately, likely due to targeting surrogate biochemical parameters as articulated by the guideline developer, Kidney Disease: Improving Global Outcomes (KDIGO). We hypothesized that using a Systems Biology Approach combining machine learning with mathematical modeling, we could test a novel approach to therapy targeting the abnormal movement of mineral out of bone and into soft tissue that is characteristic of CKD-MBD. The mathematical model describes the movement of calcium and phosphate between body compartments in response to standard therapeutic agents. The machine-learning technique we applied is reinforcement learning (RL). We compared calcium, phosphate, parathyroid hormone (PTH), and mineral movement out of bone and into soft tissue under four scenarios: standard approach (KDIGO), achievement of KDIGO guidelines using RL (RLKDIGO), targeting abnormal mineral flux (RLFLUX), and combining achievement of KDIGO guidelines with minimization of abnormal mineral flux (RLKDIGOFLUX). We demonstrate through simulations that explicitly targeting abnormal mineral flux significantly decreases abnormal mineral movement compared with standard approach while achieving acceptable biochemical outcomes. These investigations highlight the limitations of current therapeutic targets, primarily secondary hyperparathyroidism, and emphasize the central role of deranged phosphate homeostasis in the genesis of the CKD-MBD syndrome.NEW & NOTEWORTHY Artificial intelligence is a powerful tool for exploration of complex processes but application to clinical syndromes is challenging. Using a mathematical model describing the movement of calcium and phosphate between body compartments combined with machine learning, we show the feasibility of testing alternative goals of therapy for Chronic Kidney Disease Mineral Bone Disorder while maintaining acceptable biochemical outcomes. These simulations demonstrate the potential for using this platform to generate and test hypotheses in silico rapidly, inexpensively, and safely.

Machine learning progressive CKD risk prediction model is associated with CKD-mineral bone disorder

Background

Recently, we developed the machine learning (ML)-based Progressive CKD Risk Classifier (PCRC), which accurately predicts CKD progression within 5 years. While its performance is robust, it is unknown whether PCRC categorization is associated with CKD-mineral bone disorder (CKD-MBD), a critical, yet under-recognized, downstream consequence. Therefore, we aimed to 1) survey real-world testing utilization data for CKD-MBD and 2) evaluate ML-based PCRC categorization with CKD-MBD.

Methods

The cohort study utilized deidentified data from a US laboratory outpatient network, composed of 330,238 outpatients, over 5 years. The main outcomes were: 1) Laboratory testing utilization of eGFR, urine albumin creatinine ratio (UACR), parathyroid hormone (PTH), calcium, phosphate; and 2) PCRC categorization and biochemical abnormalities associated with CKD-MBD over 5 years.

Results

We identified significant under-utilization of laboratory testing for UACR, phosphate and PTH, which ranged from -40 % to -100 % against the minimum standard-of-care. At five years, the CKD progression group, as predicted by the PCRC, was associated with 15.5 % increase in phosphate (P value <<0.01) and 94.9 % increase in PTH (P value <<0.01), consistent with CKD-MBD.

Conclusions

We identified significant under-utilization of laboratory testing for CKD-MBD. Moreover, we demonstrated that CKD progression, as predicted by the PCRC, is associated with CKD-MBD, several years in advance of disease. To our knowledge, this investigation is the first to examine the role of predictive analytics for CKD progression on mineral bone disorder. While further studies are required, these findings have the potential to advance AI/ML-based risk stratification and treatment of CKD and CKD-MBD.

Applications of model simulation in pharmacological fields and the problems of theoretical reliability

The use of mathematical models has become increasingly prevalent in pharmacological fields, particularly in drug development processes. These models are instrumental in tasks such as designing clinical trials and assessing factors like efficacy, toxicity, and clinical practice. Various types of models have been developed and documented. Nevertheless, emphasizing the reliability of parameter values is crucial, as they play a pivotal role in shaping the behavior of the system. In some instances, parameter values reported previously are treated as fixed values, which can lead to convergence towards values that deviate substantially from those found in actual biological systems. This is especially true when parameter values are determined through fitting to limited observations. To mitigate this risk, the reuse of parameter values from previous reports should be approached with a critical evaluation of their validity. Currently, there is a proposal for a simultaneous search for plausible values for all parameters using comprehensive search algorithms in both pharmacokinetic and pharmacodynamic or systems pharmacological models. Implementing these methodologies can help address issues related to parameter determination. Furthermore, integrating these approaches with methods developed in the field of machine-learning field has the potential to enhance the reliability of parameter values and the resulting model outputs.

Application of artificial intelligence to chronic kidney disease mineral bone disorder

The global derangement of mineral metabolism that accompanies chronic kidney disease (CKD-MBD) is a major driver of the accelerated mortality for individuals with kidney disease. Advances in the delivery of dialysis, in the composition of phosphate binders, and in the therapies directed towards secondary hyperparathyroidism have failed to improve the cardiovascular event profile in this population. Many obstacles have prevented progress in this field including the incomplete understanding of pathophysiology, the lack of clinical targets for early stages of chronic kidney disease, and the remarkably wide diversity in clinical manifestations. We describe in this review a novel approach to CKD-MBD combining mathematical modelling of biologic processes with machine learning artificial intelligence techniques as a tool for the generation of new hypotheses and for the development of innovative therapeutic approaches to this syndrome. Clinicians need alternative targets of therapy, tools for risk profile assessment, and new therapies to address complications early in the course of disease and to personalize therapy to each individual. The complexity of CKD-MBD suggests that incorporating artificial intelligence techniques into the diagnostic, therapeutic, and research armamentarium could accelerate the achievement of these goals.

Cross-Domain Text Mining of Pathophysiological Processes Associated with Diabetic Kidney Disease

Diabetic kidney disease (DKD) is the leading cause of end-stage renal disease worldwide. This study's goal was to identify the signaling drivers and pathways that modulate glomerular endothelial dysfunction in DKD via artificial intelligence-enabled literature-based discovery. Cross-domain text mining of 33+ million PubMed articles was performed with SemNet 2.0 to identify and rank multi-scalar and multi-factorial pathophysiological concepts related to DKD. A set of identified relevant genes and proteins that regulate different pathological events associated with DKD were analyzed and ranked using normalized mean HeteSim scores. High-ranking genes and proteins intersected three domains-DKD, the immune response, and glomerular endothelial cells. The top 10% of ranked concepts were mapped to the following biological functions: angiogenesis, apoptotic processes, cell adhesion, chemotaxis, growth factor signaling, vascular permeability, the nitric oxide response, oxidative stress, the cytokine response, macrophage signaling, NFκB factor activity, the TLR pathway, glucose metabolism, the inflammatory response, the ERK/MAPK signaling response, the JAK/STAT pathway, the T-cell-mediated response, the WNT/β-catenin pathway, the renin-angiotensin system, and NADPH oxidase activity. High-ranking genes and proteins were used to generate a protein-protein interaction network. The study results prioritized interactions or molecules involved in dysregulated signaling in DKD, which can be further assessed through biochemical network models or experiments.

Artificial intelligence-guided precision treatment of chronic kidney disease-mineral bone disorder

Chronic kidney disease (CKD)-mineral bone disorder (MBD) is a complex clinical syndrome that begins early during CKD and evolves into one of the deadliest complications of CKD through its effects on the cardiovascular and skeletal systems. Achievement of treatment goals to decrease the risk of accelerated cardiovascular events and fractures has been challenging. We hypothesized that application of quantitative systems pharmacology (QSP) modeling combined with artificial intelligence techniques could improve the management of CKD-MBD with the goal of improving outcomes for patients with CKD. We present the implementation of a reinforcement learning (RL) approach to achieve the prescribed goals for serum calcium, phosphorus, and parathyroid hormone through concurrent dosing of phosphate binders, vitamin D analogs, and calcimimetics by simulation in 80 subjects in Matlab. In silico simulation results demonstrate that the application of a QSP model coupled with RL more effectively and quickly achieves treatment goals even in the setting of inferior simulated subject compliance with medical therapy and identifies key decision variables for therapeutic recommendations.

Mathematical Models of Parathyroid Gland Biology: Complexity and Clinical Use

Altered parathyroid gland biology is a major driver of chronic kidney disease-mineral bone disorder (CKD-MBD) in patients with chronic kidney disease. CKD-MBD is associated with a high risk of vascular calcification and cardiovascular events. A hallmark of CKD-MBD is secondary hyperparathyroidism with increased parathyroid hormone (PTH) synthesis and release and reduced expression of calcium-sensing receptors on the surface of parathyroid cells and eventually hyperplasia of parathyroid gland cells. The KDIGO guidelines strongly recommend the control of PTH in hemodialysis patients. Due to the complexity of parathyroid gland biology, mathematical models have been employed to study the interaction of PTH regulators and PTH plasma concentrations. Here, we present an overview of various model approaches and discuss the impact of different model structures and complexities on the clinical use of these models.

Use of Artificial Intelligence to Identify New Mechanisms and Approaches to Therapy of Bone Disorders Associated With Chronic Kidney Disease

Chronic kidney disease (CKD) leads to clinically severe bone loss, resulting from the deranged mineral metabolism that accompanies CKD. Each individual patient presents a unique combination of risk factors, pathologies, and complications of bone disease. The complexity of the disorder coupled with our incomplete understanding of the pathophysiology has significantly hampered the ability of nephrologists to prevent fractures, a leading comorbidity of CKD. Much has been learned from animal models; however, we propose in this review that application of multiple techniques of mathematical modeling and artificial intelligence can accelerate our ability to develop relevant and impactful clinical trials and can lead to better understanding of the osteoporosis of CKD. We highlight the foundational work that informed our current model development and discuss the potential applications of our approach combining principles of quantitative systems pharmacology, model predictive control, and reinforcement learning to deliver individualized precision medical therapy of this highly complex disorder.

Recent applications of quantitative systems pharmacology and machine learning models across diseases

Quantitative systems pharmacology (QSP) is a quantitative and mechanistic platform describing the phenotypic interaction between drugs, biological networks, and disease conditions to predict optimal therapeutic response. In this meta-analysis study, we review the utility of the QSP platform in drug development and therapeutic strategies based on recent publications (2019-2021). We gathered recent original QSP models and described the diversity of their applications based on therapeutic areas, methodologies, software platforms, and functionalities. The collection and investigation of these publications can assist in providing a repository of recent QSP studies to facilitate the discovery and further reusability of QSP models. Our review shows that the largest number of QSP efforts in recent years is in Immuno-Oncology. We also addressed the benefits of integrative approaches in this field by presenting the applications of Machine Learning methods for drug discovery and QSP models. Based on this meta-analysis, we discuss the advantages and limitations of QSP models and propose fields where the QSP approach constitutes a valuable interface for more investigations to tackle complex diseases and improve drug development.