Selecting the right therapeutic target for kidney disease

Abstract

Kidney disease is a complex disease with several different etiologies and underlying associated pathophysiology. This is reflected by the lack of effective treatment therapies in chronic kidney disease (CKD) that stop disease progression. However, novel strategies, recent scientific breakthroughs, and technological advances have revealed new possibilities for finding novel disease drivers in CKD. This review describes some of the latest advances in the field and brings them together in a more holistic framework as applied to identification and validation of disease drivers in CKD. It uses high-resolution 'patient-centric' omics data sets, advanced in silico tools (systems biology, connectivity mapping, and machine learning) and 'state-of-the-art' experimental systems (complex 3D systems in vitro, CRISPR gene editing, and various model biological systems in vivo). Application of such a framework is expected to increase the likelihood of successful identification of novel drug candidates based on strong human target validation and a better scientific understanding of underlying mechanisms.

Keywords: artificial intelligence; chronic kidney disease; drug discovery; machine learning; omics; systems biology; validation.

FIGURE 1

Framework for target identification and validation in CKD. Potential CKD targets identified from human data via in silico and SME approaches are prioritized using thorough validation in vitro and in vivo systems, facilitating target selection. Target identification starts with the collection of biopsies, urine, and blood from patients with CKD and controls from clinical trials and collaborations. Omics data (genetic, transcriptomic, proteomic, and metabolomic) are then generated from these samples and processed with integrative omics analyses followed by bioinformatic and statistical analyses and machine learning. This data processing results in a list of CKD targets that is assessed in silico to build further evidence of human target–disease associations. The shortlisted targets are then prioritized by applying biologically relevant in vitro validation in cultured cells and advanced in vitro models. The targets with the strongest supportive data are then validated further in vivo to build proof-of-mechanism and proof-of-principle in CKD before being presented for target selection and investment decision to enter the portfolio. CKD, chronic kidney disease; SME, subject-matter expert.

FIGURE 2

Human target validation and prioritization. After generation of CKD target lists, additional disease-relevant evidence is added to generate testable hypotheses and to prioritize the candidates. Tissue-specific expression enrichment and expression modulation in disease versus healthy states help to ascertain the role of targets in CKD. Correlation of targets with renal functional biomarkers and parameters is ascertained and target expression across CKD stages or etiologies is explored to add confidence around disease relevance. Prediction of target kidney cell type is useful for guiding downstream in vitro validation and assay selection. Pathway and network analyses can provide additional biological context for dysregulated cellular mechanisms and help infer potential mechanisms of action. The accumulated evidence supporting human target validation and mechanism of action results in a set of prioritized candidates for further experimental validation. CKD, chronic kidney disease; DN, diabetic nephropathy; eGFR, estimated glomerular filtration rate; eQTL, expression quantitative trait loci; FSGS, focal segmental glomerulosclerosis; HT, hypertensive nephropathy; IgA, immunoglobulin A nephropathy; MCD, minimal change disease; MGN, membranous glomerulonephritis; RPGN, rapid progressive glomerulonephritis; SLE, systemic lupus erythematosus.

FIGURE 3

In vitro target validation process. In vitro target validation in cultured cells is used as a first approach to validate and screen several targets, aiming to triage targets with strong support for further target validation in advanced in vitro systems. Data from all systems feed into each other to select the most translatable model and define the correct stressor that regulates target and pathway. An in vitro target validation toolbox may comprise various assays, stressors, and readouts, which are chosen based on the disease biology of the target. ADMA, asymmetric dimethylarginine (arginine metabolite); CKD, chronic kidney disease; ECAR, extracellular acidification rate; HGEC, human glomerular endothelial cell; HMOX, hem oxygenase; IHC, immunohistochemistry; iPSC, induced pluripotent stem cell; MMPs, matrix metalloproteinases; MSC, mesenchymal stem cell; OCR, oxygen consumption rate; PAN, pyromycin aminonucleoside; PTEC, proximal tubular epithelial cell; RPTEC, renal proximal tubular epithelial cell; SNP, single nucleotide polymorphism; TGF, transforming growth factor; TLR, toll-like receptor; ROS, reactive oxygen species. 1. Faivre A et al. Front Med (Lausanne) 2021; 8:742072; 2. Imasawa T et al. The International Journal of Biochemistry and Cell Biology 2013; 45:2109–2118.3. Oates JC et al. American Journal of Physiology-Renal Physiology 2022; 322:F309-F321.4. Tang SCW et al. Nature Reviews Nephrology 2020; 16:206–222.5. Lee HW et al. Journal of the American Society of Nephrology 2015; 26:2741–2752.6. Perico L et al. Nature Reviews Nephrology 2016; 12:692–710.7. Prozialeck WC et al. Pharmacology and Therapeutics 2007; 114:74–93.8. Slyne J et al. Nephrology Dialysis Transplantation 2015; 30:iv60-iv67.9. Wieser M et al. American Journal of Physiology-Renal Physiology 2008; 295:F1365-F1375.10. Jourde-Chiche N et al. Nature Reviews Nephrology 2019; 15:87–108.11. Sol M et al. Front Pharmacol 2020; 11:573557.12. Liu Y. Kidney International 2006; 69:213–217.13. Yun CW et al. International Journal of Molecular Sciences 2019; 20:1619.

FIGURE 4

Summary of in vitro validation models. Listed in vitro models arranged from the assay containing the most tissue complexity down to single cells, listing the pros and cons for the different models. The physiological complexity of the different models impacts both the screening capacity of the assay and what biology that translates. MPS, micro-physiological systems. ^aconsiderations are listed in a step-wise manner and main considerations only are stated.

FIGURE 5

Workflow for in vivo target validation and compound testing. Ex vivo models are chosen to validate genes of interest and assess the efficacy of compounds targeting these gene products, based on their cellular expression and target engagement, respectively. Data from the ex vivo experiments guide the design of mechanistic and PKPD in vivo studies that generate data regarding target engagement and proof of mechanism, and on exposure and proof of principle, respectively. Various mechanistic models are used to validate targets and measure efficacy of test compounds.​ Selection of the most relevant in vivo PKPD model for a specific target or signaling pathway is based on ex vivo experimental data, human omics data, and RNA sequencing data from our ‘rodent portal’, generated from a panel of rodent renal disease models. The bottom panel illustrates differential upregulation of a target gene in different disease models, which in turn allows selection of the most appropriate model for the tested protocol. ALAT, alanine aminotransferase; ANTN, accelerated/non-accelerated nephrotoxic nephritis; BTBR, black and tan brachyuric; eNOS, endothelial nitric oxide synthase; IRI, ischemic reperfusion injury; KIM1, kidney injury molecule one; LPS, lipopolysaccharide; NGAL, neutrophil gelatinase; PAN, pyromycin aminonucleoside; PKPD, pharmacokinetic–pharmacodynamic; POM, proof of mechanism; TE, target engagement; UUO, unilateral ureter obstruction.

FIGURE 6

Identification of the right CKD target. The target identification and validation framework relies on multiple data sources and validation models, integrating many disease-relevant data sets, to create a holistic scientific understanding of the mechanisms that link the target to disease biology. The in-depth scientific understanding of pathophysiology and target link to disease is, in our view, essential for delivering successful medicines to patients with CKD in the future. AI, artificial intelligence; CKD, chronic kidney disease; PD, pharmacodynamics; SME, subject-matter expert.