SEARCHCASE-CONTROLophthalmology, retina

case-control·ophthalmology, retina, neuro-ophthalmology, clinical trial, observational study·PMC10848380

The mechanistic functional landscape of retinitis pigmentosa: a machine learning-driven approach to therapeutic target discovery

Journal of Translational Medicine · 8 authors, 8 centres

AI SUMMARY

FIDELITY 100%

POPULATIONrd10 mice (murine model of autosomal recessive retinitis pigmentosa) and C57Bl/6J control mice

INTERVENTIONNot applicable (observational validation of gene/protein expression changes in rd10 mice compared to controls)

COMPARISONAge-matched control (C57Bl/6J) mice

This summary was generated by AI from a single paper. It has not been reviewed by a clinician and is not clinical advice. Verify against the source before acting on it.

AGENT 01 · STRUCTURED SUMMARY

TRACE TO SOURCE

“

TL;DR

This study constructed a mechanistic disease map of retinitis pigmentosa (RP) comprising 226 functional circuits from 40 KEGG pathways and used an explainable machine learning model to predict 109 drug targets with potential to modulate RP mechanisms. Five targets (Gabre, Gabra1, Slc12a5, Grin1, Glr2a) were experimentally validated in rd10 mice, showing dysregulation of GABAergic, glutamatergic, and glycinergic signaling during RP progression. The approach provides a resource for prioritizing therapeutic targets and drug repurposing candidates for RP.

Full summary

4,117 CHARS

**Background:** Retinitis pigmentosa (RP) is the most common inherited retinal disease, affecting over 1.5 million people worldwide (prevalence 1/2500 to 1/7000). It is a genetically heterogeneous disorder with mutations in over 100 known genes, yet the genetic cause remains unknown in 50% of patients. Current therapeutic options are limited. The authors aimed to construct a mechanistic disease map of RP and use machine learning to identify drug targets that could modulate RP-related signaling circuits, facilitating drug repurposing.

**Methods:** RP-associated genes were retrieved from Orphanet (93 genes), OMIM, and the Human Phenotype Ontology (HPO). Genes sharing at least 10 of 22 RP-specific HPO terms were added to expand the core set. These genes were mapped onto KEGG signaling pathways using the Hipathia R package, defining 226 signaling circuits (sub-pathways) across 40 KEGG pathways as the RP mechanistic map. Circuits were functionally annotated into nine RP hallmarks: Apoptosis, Necrosis, Stress Response, Inflammatory response, DNA integrity, Fatty acids and lipid metabolism, Sensory and stimuli response, Development processes, and Neuronal processes. A mechanistic model (Hipathia) estimated circuit activity from GTEx normal tissue transcriptomic data (17,382 samples from 948 individuals). An explainable multi-output random forest regressor was trained to predict circuit activity based on expression of 711 known drug targets (KDTs) from DrugBank (approved drugs with known pharmacological action). SHAP values identified the most predictive KDTs for each circuit. Stability was assessed using the Nogueira bootstrap procedure; 207 circuits with stability >0.4 were retained. KDTs were clustered into three groups (4, 15, and 90 KDTs) using hierarchical clustering and Gap statistics. Five KDTs (Gabre, Gabra1, Slc12a5, Grin1, Glr2a) were selected for experimental validation in rd10 mice (Pde6b mutant, autosomal recessive RP model) based on top relevance scores and involvement in neuronal/developmental hallmarks. Gene expression (qRT-PCR), protein levels (western blot), retinal histology (confocal microscopy), and GABA concentration (HPLC-MS) were assessed at postnatal days P15-P60.

**Key Results:** The RP mechanistic map comprised 226 circuits from 40 KEGG pathways. The machine learning model identified 109 KDTs (targeted by 284 drugs) with predictive relevance for circuit activity. Cluster 1 (4 KDTs: GABRA1, GRIN1, HRH3, SLC12A5) showed strong influence across most circuits. In rd10 mice, significant dysregulation was observed: Slc12a5 was downregulated at P15 but upregulated from P23 to P60; Gabra1 and Gabre were downregulated at P15, upregulated at P23-P60; Gabre was upregulated at P18 with decreased GABA content; Grin1 and Glr2a were upregulated from P23 to P60. Western blot confirmed increased KCC2 (Slc12a5), GABARα1 (Gabra1), and NR1 (Grin1) protein at P23. Immunostaining showed higher KCC2 and GABARα1 in the inner retina of rd10 mice at P23. Photoreceptor degeneration was evident at P23 with significantly fewer nuclei in the outer nuclear layer. Drug repurposing analysis identified candidates such as flumazenil (GABRA1 antagonist), taurine (GLRA2 agonist), and omega-3 carboxylic acids (ELOVL4 potentiator) with potential to reverse observed dysregulations.

**Clinical Implications:** This study provides a comprehensive, mechanistically grounded resource for understanding RP pathophysiology and identifying therapeutic targets. The validated dysregulation of GABAergic, glutamatergic, and glycinergic signaling components suggests that inner retinal remodeling involves neurotransmitter system alterations, which could be targeted to preserve visual function beyond photoreceptor loss. The approach demonstrates how integrating mechanistic models with machine learning can prioritize drug repurposing candidates for rare diseases with limited treatment options. The identified drugs (e.g., flumazenil, taurine, bumetanide) warrant further preclinical and clinical investigation for RP. The methodology is broadly applicable to other rare diseases.

PICO

PPOPULATION

rd10 mice (murine model of autosomal recessive retinitis pigmentosa) and C57Bl/6J control mice

IINTERVENTION

Not applicable (observational validation of gene/protein expression changes in rd10 mice compared to controls)

OOUTCOME

Gene expression (qRT-PCR), protein levels (western blot), retinal histology (confocal microscopy), GABA concentration (HPLC-MS) of selected targets (Gabre, Gabra1, Slc12a5, Grin1, Glr2a) at postnatal days P15-P60

STUDY TYPE

case-control

SPECIALTY

ophthalmology

SUMMARISED BY

AI pipeline

FIDELITY CHECK

100% · A

ORIGINAL PAPER

PMC10848380

The mechanistic functional landscape of retinitis pigmentosa: a machine learning-driven approach to therapeutic target discovery

Journal of Translational Medicine

READ THE FULL PAPER

JUMP TO SECTION

ABSTRACT

Background Retinitis pigmentosa is the prevailing genetic cause of blindness in developed nations with no effective treatments. In the pursuit of unraveling the intricate dynamics underlying this complex disease, mechanistic models emerge as a tool of proven efficiency rooted in systems biology, to elucidate the interplay between RP genes and their mechanisms. The integration of mechanistic models and drug-target interactions under the umbrella of machine learning methodologies provides a multifaceted approach that can boost the discovery of novel therapeutic targets, facilitating further drug repurposing in RP. Methods By mapping Retinitis Pigmentosa-related genes (obtained from Orphanet, OMIM and HPO databases) onto KEGG signaling pathways, a collection of signaling functional circuits encompassing Retinitis Pigmentosa molecular mechanisms was defined. Next, a mechanistic model of the so-defined disease map, where the effects of interventions can be simulated, was built. Then, an explainable multi-output random forest regressor was trained using normal tissue transcriptomic data to learn causal connections between targets of approved drugs from DrugBank and the functional circuits of the mechanistic disease map. Selected target genes involvement were validated on rd10 mice, a murine model of Retinitis Pigmentosa. Results A mechanistic functional map of Retinitis Pigmentosa was constructed resulting in 226 functional circuits belonging to 40 KEGG signaling pathways. The method predicted 109 targets of approved drugs in use with a potential effect over circuits corresponding to nine hallmarks identified. Five of those targets were selected and experimentally validated in rd10 mice: Gabre , Gabra1 (GABARα1 protein), Slc12a5 (KCC2 protein), Grin1 (NR1 protein) and Glr2a . As a result, we provide a resource to evaluate the potential impact of drug target genes in Retinitis Pigmentosa. Conclusions The possibility of building actionable disease models in combination with machine learning algorithms to learn causal drug-disease interactions opens new avenues for boosting drug discovery. Such mechanistically-based hypotheses can guide and accelerate the experimental validations prioritizing drug target candidates. In this work, a mechanistic model describing the functional disease map of Retinitis Pigmentosa was developed, identifying five promising therapeutic candidates targeted by approved drug. Further experimental validation will demonstrate the efficiency of this approach for a systematic application to other rare diseases. Supplementary Information The online version contains supplementary material available at 10.1186/s12967-024-04911-7.

CITATION EXPORT

Marina Esteban-Medina, Carlos Loucera, Kinza Rian et al.. The mechanistic functional landscape of retinitis pigmentosa: a machine learning-driven approach to therapeutic target discovery. Journal of Translational Medicine. (2024). https://doi.org/10.1186/s12967-024-04911-7

AUTHORS

8 · 8 CENTRES

Marina Esteban-Medina, Carlos Loucera, Kinza Rian, Sheyla Velasco, Lorena Olivares-González, Regina Rodrigo, Joaquin Dopazo, Maria Peña-Chilet

1Andalusian Platform for Computational Medicine, Andalusian Public Foundation Progress and Health-FPS, Seville, Spain
2grid.411109.c0000 0000 9542 1158Systems and Computational Medicine Group, Institute of Biomedicine of Seville, IBiS, University Hospital Virgen del Rocío/CSIC/University of Seville, 41013 Seville, Spain
3https://ror.org/05xr2yq54grid.418274.c0000 0004 0399 600XGroup of Pathophysiology and Therapies for Vision Disorders, Príncipe Felipe Research Center (CIPF), 46012 Valencia, Spain
4https://ror.org/00ca2c886grid.413448.e0000 0000 9314 1427Biomedical Research Networking Center in Rare Diseases (CIBERER), Health Institute Carlos III, 28029 Madrid, Spain
5https://ror.org/043nxc105grid.5338.d0000 0001 2173 938XDepartment of Physiology, University of Valencia (UV), 46100 Burjassot, Spain
6grid.440831.a0000 0004 1804 6963Department of Anatomy and Physiology, Catholic University of Valencia San Vicente Mártir, 46001 Valencia, Spain
7grid.476458.c0000 0004 0427 8560Joint Research Unit on Endocrinology, Nutrition and Clinical Dietetics UV-IIS La Fe, 46026 Valencia, Spain
8grid.84393.350000 0001 0360 9602BigData, AI, Biostatistics & Bioinformatics Platform, Health Research Institute La Fe (IISLaFe), 46026 Valencia, Spain