Love hearts

Precision medicines have revolutionised the treatment of some cancers and rare genetic diseases, yet progress in chronic diseases such as heart failure has lagged.

The challenge is that chronic diseases, like heart failure, are biologically complex and driven by multiple mechanisms. They are heterogenous; from how they present over time with an array of different symptoms and comorbidities, to the way they are diagnosed over the course of disease.

Heart failure affects 64 million people worldwide and has several underlying causes. Current heart failure treatments follow a ‘one-size-fits-all’ approach. Due to the wide range of mechanisms by which heart failure can occur, however, this is not always the best approach.

Currently, diagnosis relies on clinical symptoms and standard biomarkers – patient characteristics such as a gene, molecule, or other (e.g. blood pressure) – however, tests are often limited, imprecise and late on disease progression.

Targeting the underlying molecular cause of an individual patient’s disease in heart failure including widespread inflammation, fibrosis and microvascular dysfunction is a fundamental change from current clinical management. This growing understanding of the genetic drivers of heart failure is laying the foundations for precision medicine.

AstraZeneca is looking to change this, harnessing the many benefits of precision medicine and using multi-omics, novel technologies, imaging, artificial intelligence and machine learning to dig deep into the biology of chronic diseases.

By collaborating with other experts in precision medicine to identify novel biomarkers to guide and treat life-threatening diseases of heart muscle, including ischaemic cardiomyopathy (ICM) and idiopathic dilated cardiomyopathy (IDCM) and the inherited muscle wasting condition, Duchenne muscular dystrophy (DMD).

Fingerprints of heart failure

By harnessing the power of artificial intelligence, machine learning and gene expression analysis, AstraZeneca’s aim is to unravel the complex disease biology of heart failure at the molecular level in individual patients.

To achieve this, researchers are using machine learning to analyse large quantities of gene expression data from cardiac biopsy samples and stratify patients with heart failure into novel molecular sub-classes, irrespective of their clinical signs and symptoms.

Inspiration can sometimes come from unexpected places. Anna Reznichenko, Senior Principal Scientist, Research and Early Development, Cardiovascular, Renal and Metabolism, BioPharmaceuticals R&D, has devised a novel method based on market segmentation analysis for handling massive data sets drawn from the financial industry.

This approach has revealed distinct new patient categories, and could be used as the base for precision medicine in the future.

She comments: “Through this process, we have shown that the ‘molecular fingerprints’ shared by people do not always reflect the current classifications generally used for diagnosis. What we’re starting to do is link sub-class-specific gene expression profiles to dysregulated molecular pathways and processes indicative of distinct disease biology across the different sub-classes.

“Using all this new information, we aim to identify novel therapeutic targets that could form the basis of a precision medicine approach to treat patients with heart failure,” she adds.

Distinct disease categories based on molecular data that are different from current clinical classifications for heart failure is emerging. The data also provides the team with new information to explore potential biomarkers that could be used to reveal patients’ molecular disease classes non-invasively.

This allows us greater precision when allocating the right patients to the right trials and in the future, potentially provide tailored treatments based on scientifically determined individual disease categories.

Identifying novel targets with AI

By combining rich data sets, including patient data, and applying AI and machine learning, we can identify novel associations between data and disease. AstraZeneca has partnered with BenevolentAI to analyse vast amounts of scientific data to see the potential interactions between gene targets, expression and disease.

The collaboration combines AstraZeneca’s disease expertise with BenevolentAI’s Knowledge Graph and machine learning tools to identify novel insights and relationships to reveal new hypotheses and propose drug targets that have never been considered for a disease before.

The collaboration between the two companies has resulted in successful identification of potential novel targets in chronic kidney disease and idiopathic pulmonary fibrosis. In heart failure the team has identified potential targets and is currently experimentally validating them using novel experimental procedures like CRISPR.

Dr Anne Phelan, Chief Scientific Officer at BenevolentAI, explains: “Our collaboration with AstraZeneca brings together sophisticated drug discovery with innovative AI-driven technologies to integrate and analyse vast amounts of scientific data from diverse sources. In doing so, we can uncover new ways to understand complex disease biology and identify novel drug targets.”

Heart of the matter

Cardiomyopathy is a leading cause of heart failure and uncovering genetic insights that hold a key to transforming disease remains a crucial part of the drug discovery process. Research has revealed a significant hereditary component to heart failure, and variants in the genome have been identified to play a role in the development of the disease.

One gene of interest that plays a role in causing stretched and weakened heart muscle, as seen in dilated cardiomyopathy (DCM), produces a protein called phospholamban (PLN). Excessive PLN activity is linked to faulty calcium cycling and impaired heart muscle contraction and relaxation, but this mechanism has proved hard to target with conventional drugs.

Kenny Hansson, Head of Bioscience Cardiovascular, Early CVRM, at AstraZeneca, reflects: “Encouraging laboratory data has demonstrated the potential of antisense oligonucleotides (ASOs) to target PLN activity in DCM.

“The research, carried out in collaboration with Ionis Pharmaceuticals and international heart failure scientists at University Medical Center Groningen and Karolinska Institute, shows that ASOs – strands of synthetic DNA – can be used to deplete the formation of PLN linked to DCM.”

Encouraging results with ASOs in other heart failure models has shown this could be a promising precision medicine approach in cardiomyopathy and possibly other forms of heart failure.

The final analysis

Advances in the care of children born with DMD have improved the outlook for those living with the disease, but progressive wasting of heart muscle can lead to life-limiting DCM and heart failure when individuals reach their 20s.

Progress with gene therapy targeted at heart muscle has been limited. Using their well-established CRISPR-Cas9 gene editing expertise, however, the teams at AstraZeneca are investigating the removal of faulty sequences from the dystrophin gene, and using adeno-associated viruses to efficiently deliver targeted treatment into heart muscle cells. If this works, there is also the potential to extend this approach to other inherited diseases.

By exploring subtle genetic mutations, variations in gene expression and gene-environment interactions in more common forms of heart failure, there is a potential to stratify patients for clinical trials of biomarker-guided targeted treatment.

Kenny says: “Drawing on innovations in clinical trial design and using an expanding toolkit of novel drug modalities, we’re aiming to target almost any type of underlying disease biology in heart failure, so that one day the right drug is always available for the right patient at the right time.”