Periodic Reporting for period 1 - IMPACT (Cardiogenomics meets Artificial Intelligence: a step forward in arrhythmogenic cardiomyopathy diagnosis and treatment)
Período documentado: 2023-10-01 hasta 2024-09-30
Arrhythmogenic cardiomyopathy (ACM) is a rare genetic cardiac disease, phenotypically characterized by structural and functional alterations, with an estimated prevalence of approximately 1 in 5,000 in the general population. Clinical symptoms typically present between the second and fourth decades of life and include syncope, ventricular tachycardia, and sudden cardiac death (SCD), with the latter often being the sole manifestation. The pathological hallmark of ACM is the fibrofatty replacement of myocardial tissue, which is associated with ventricular atrophy. The right ventricle is most commonly affected, though biventricular and predominantly left ventricular forms are also observed. No therapy has been proven to attenuate these progressive myocardial changes. Therefore, the management of ACM focuses on preventing SCD and slowing disease progression. For patients with symptomatic ventricular arrhythmias, current treatment options include palliative anti-arrhythmic drugs, such as beta-blockers, or more invasive interventions, such as implantable cardioverter-defibrillators (ICD) or catheter ablation.
Approximately half of ACM patients have mutations in one or more genes encoding proteins of the cardiac intercalated discs, with the most commonly affected genes being plakophilin-2 (PKP2), desmoglein-2 (DSG2), and desmoplakin (DSP). Cardiac intercalated discs are responsible for mechanical and electro-metabolic coupling between cardiomyocytes. However, many of the identified variants in these disease-related genes are classified as variants of uncertain significance (VUS), limiting their clinical utility.
The overall aim of this project is to combine large-scale data from genomics, proteomics, and instrumental analyses from patients with structural and functional data from in vitro (3D microtissue) and in vivo (murine) models to establish the genotype/cardiac phenotype relationship. This could lead to a better understanding of the role and impact of known genes and epigenetic factors (e.g. miRNAs) on susceptibility, clinical progression, and treatment of ACM.
This is an interdisciplinary research program involving translational collaboration among experts in various fields, all working toward the common goal of developing an effective treatment for ACM.
Approximately half of ACM patients have mutations in one or more genes encoding proteins of the cardiac intercalated discs, with the most commonly affected genes being plakophilin-2 (PKP2), desmoglein-2 (DSG2), and desmoplakin (DSP). Cardiac intercalated discs are responsible for mechanical and electro-metabolic coupling between cardiomyocytes. However, many of the identified variants in these disease-related genes are classified as variants of uncertain significance (VUS), limiting their clinical utility.
The overall aim of this project is to combine large-scale data from genomics, proteomics, and instrumental analyses from patients with structural and functional data from in vitro (3D microtissue) and in vivo (murine) models to establish the genotype/cardiac phenotype relationship. This could lead to a better understanding of the role and impact of known genes and epigenetic factors (e.g. miRNAs) on susceptibility, clinical progression, and treatment of ACM.
This is an interdisciplinary research program involving translational collaboration among experts in various fields, all working toward the common goal of developing an effective treatment for ACM.
The project aims to achieve the following goals:
1. Using Artificial Intelligence to combine clinical data with genetic and molecular information from ACM patients. This will help us identify different patient groups, discover targets for new treatments, and improve how we classify patients for better care.
2. Creating 3D Heart Models to Study ACM: We will develop 3D mini heart tissues in the lab. These models will help us understand uncertain genetic variants and identify disease pathways that could be targeted with new treatments.
3. Testing New and Existing Drugs for ACM: We will explore the effectiveness of FDA-approved drugs and new experimental molecules in both lab and animal models to find promising treatments for ACM.
During the first year of the project, the following key achievements have been obtained:
- Pipeline Definition for AI-Assisted Clustering of ACM Patients: initial trials with sample MRI and ECG data successfully demonstrated the platform’s functionality, confirming that the software and libraries on the virtual machines are working as expected. With this foundation in place, we are now ready to launch official data analysis workflows as soon as the complete dataset becomes available.
- Development of Cardiac Cell Types: we successfully created three heart cell types from stem cells for two known ACM-causing genetic variants, DSG2 and DSP, along with genetically identical controls for comparison. For another significant gene, PKP2, we successfully developed two out of the three targeted cell types (cardiomyocytes and endothelial cells). All cell lines exhibit high purity (over 80%) and express key markers, making them ideal for constructing reliable 3D heart tissue models.
1. Using Artificial Intelligence to combine clinical data with genetic and molecular information from ACM patients. This will help us identify different patient groups, discover targets for new treatments, and improve how we classify patients for better care.
2. Creating 3D Heart Models to Study ACM: We will develop 3D mini heart tissues in the lab. These models will help us understand uncertain genetic variants and identify disease pathways that could be targeted with new treatments.
3. Testing New and Existing Drugs for ACM: We will explore the effectiveness of FDA-approved drugs and new experimental molecules in both lab and animal models to find promising treatments for ACM.
During the first year of the project, the following key achievements have been obtained:
- Pipeline Definition for AI-Assisted Clustering of ACM Patients: initial trials with sample MRI and ECG data successfully demonstrated the platform’s functionality, confirming that the software and libraries on the virtual machines are working as expected. With this foundation in place, we are now ready to launch official data analysis workflows as soon as the complete dataset becomes available.
- Development of Cardiac Cell Types: we successfully created three heart cell types from stem cells for two known ACM-causing genetic variants, DSG2 and DSP, along with genetically identical controls for comparison. For another significant gene, PKP2, we successfully developed two out of the three targeted cell types (cardiomyocytes and endothelial cells). All cell lines exhibit high purity (over 80%) and express key markers, making them ideal for constructing reliable 3D heart tissue models.
1. Using Artificial Intelligence to combine clinical data with genetic and molecular information from ACM patients.
Using data from the Netherlands ACM Registry, UMCU researchers identified a cohort of ACM patients with ECG and CMR imaging data, that is currently being analyzed by Lutech. This is all performed on a shared cloud-based computing platform hosted by UMCU. An in-depth study of the literature on ACM and discussions with medical professionals in Utrecht led to the development of an analysis pipeline by Lutech where CMR parameters and ECGs are processed using pre-trained algorithms to extract key morphological and electrocardiographic features. Patients were clustered based on these features,and further analysis will incorporate clinical, demographic, and omics data to reveal specific patterns in each group. Additionally, tissue samples from ACM patients are being studied for cellular details, RNA, and protein expression to gain deeper insights. A genome-wide association study (GWAS) is also underway, focusing on genetic markers associated with ACM, with promising early results showing several risk markers. These findings together with proteomics datasets produced on cardiac tissue from selected patients will help refine AI-based clustering methods.
2. Creating 3D heart models to study ACM.
Our project is making significant progress in studying ACM by developing advanced 3D cardiac microtissues derived from patient-specific stem cells. These models replicate the complex cellular environment of the human heart more accurately than traditional 2D cultures, enhancing our understanding of ACM’s molecular mechanisms and identifying potential therapeutic targets.
UNIPD has successfully created cardiac cells with pathogenic mutations in key ACM-related genes (DSG2, DSP, PKP2) and genetically identical controls. These cell lines are undergoing phenotypic and molecular characterization to ensure they are suitable for 3D tissue formation.
Moreover, microtissues harboring variants of unknown significance (VUS) will be used to evaluate the relative pathogenic potential, aiding in improved patient management and risk stratification.
Additionally, Lutech is employing artificial intelligence for comparative analysis of the microtissues alongside in vivo ACM models. This integrated approach, combining 3D models, animal studies, and AI analysis, aims to provide new insights into ACM mechanisms and develop innovative treatment strategies.
3. Testing New and Existing Drugs for ACM.
Objective 3 of our project is focused on exploring potential treatments for ACM through two approaches: testing small-molecule drugs by KSILINK and Italfarmaco, and developing antagomirs—molecules that block specific RNAs linked to the disease. Although the main experimental work will start in mid-2025, the groundwork is already underway.
KSILINK is currently optimizing the procedures needed to create 3D cardiac microtissues from various cell types, using automated methods and high-throughput 384-well plates.
In parallel, Maastricht University is studying microRNAs and long non-coding RNAs that may play a role in ACM. Early results have identified several promising targets, which could be blocked using antagomirs or gapmers to reduce disease impact.
These efforts lay the foundation for innovative treatment approaches that could potentially halt or slow the progression of ACM in affected individuals.
Using data from the Netherlands ACM Registry, UMCU researchers identified a cohort of ACM patients with ECG and CMR imaging data, that is currently being analyzed by Lutech. This is all performed on a shared cloud-based computing platform hosted by UMCU. An in-depth study of the literature on ACM and discussions with medical professionals in Utrecht led to the development of an analysis pipeline by Lutech where CMR parameters and ECGs are processed using pre-trained algorithms to extract key morphological and electrocardiographic features. Patients were clustered based on these features,and further analysis will incorporate clinical, demographic, and omics data to reveal specific patterns in each group. Additionally, tissue samples from ACM patients are being studied for cellular details, RNA, and protein expression to gain deeper insights. A genome-wide association study (GWAS) is also underway, focusing on genetic markers associated with ACM, with promising early results showing several risk markers. These findings together with proteomics datasets produced on cardiac tissue from selected patients will help refine AI-based clustering methods.
2. Creating 3D heart models to study ACM.
Our project is making significant progress in studying ACM by developing advanced 3D cardiac microtissues derived from patient-specific stem cells. These models replicate the complex cellular environment of the human heart more accurately than traditional 2D cultures, enhancing our understanding of ACM’s molecular mechanisms and identifying potential therapeutic targets.
UNIPD has successfully created cardiac cells with pathogenic mutations in key ACM-related genes (DSG2, DSP, PKP2) and genetically identical controls. These cell lines are undergoing phenotypic and molecular characterization to ensure they are suitable for 3D tissue formation.
Moreover, microtissues harboring variants of unknown significance (VUS) will be used to evaluate the relative pathogenic potential, aiding in improved patient management and risk stratification.
Additionally, Lutech is employing artificial intelligence for comparative analysis of the microtissues alongside in vivo ACM models. This integrated approach, combining 3D models, animal studies, and AI analysis, aims to provide new insights into ACM mechanisms and develop innovative treatment strategies.
3. Testing New and Existing Drugs for ACM.
Objective 3 of our project is focused on exploring potential treatments for ACM through two approaches: testing small-molecule drugs by KSILINK and Italfarmaco, and developing antagomirs—molecules that block specific RNAs linked to the disease. Although the main experimental work will start in mid-2025, the groundwork is already underway.
KSILINK is currently optimizing the procedures needed to create 3D cardiac microtissues from various cell types, using automated methods and high-throughput 384-well plates.
In parallel, Maastricht University is studying microRNAs and long non-coding RNAs that may play a role in ACM. Early results have identified several promising targets, which could be blocked using antagomirs or gapmers to reduce disease impact.
These efforts lay the foundation for innovative treatment approaches that could potentially halt or slow the progression of ACM in affected individuals.