Health Tech World explores the latest research and developments in health technology.
Scientists develop a method for engineering ultra-selective aptamers
Inspired by the way viruses attach to cells, EPFL scientists have developed a method for engineering ultra-selective aptamers – short segments of DNA or RNA that are designed to bind, like antibodies, to specific targets, which make them useful for biomedical diagnostics and treatments.
When new aptamer binders are needed, for example to detect a new virus, they are developed from libraries of millions of nucleic acid sequences from which the best matches for a given target are selected and amplified.
Maartje Bastings, head of the Programmable Biomaterials Lab in EPFL’s School of Engineering, said: “You can’t control where a monovalent binder interacts with its target: for example, it may bind to the side of a protein, rather than the binding interface, reducing its functionality.”
Bastings and her team have recently reported the first technique for producing multimeric aptamers, which target protein complexes with unprecedented precision and functionality.
The binders developed with the lab’s approach, dubbed MEDUSA (Multivalent Evolved DNA-based Supramolecular Assemblies), yields binding affinities that are between 10 and 1,000 times stronger than those achieved with monovalent binders.
“We have retro-engineered the natural paradigm seen in viruses, in which multivalent molecular complexes co-evolve, and translated it into a new binder discovery method that allows us to select multivalent binders that can block such viruses,” said PhD student and first author Artem Kononenko.
Once a first batch of binders is identified, candidates with increasing affinity for their target are developed through an iterative process of selection and amplification called ‘evolution’.
Although designing new scaffolds can take a matter of hours, the evolution process can take weeks. Looking ahead, the research team aims to shorten this timeframe to better suit the needs of biomedical diagnostics and therapeutics.
Another goal is to develop multimeric binders targeting pathogens with even more complex configurations, like Dengue fever (six binding subunits) or anthrax (seven).
“Ultimately, we want to use this new multivalent sequence space to train generative artificial intelligence models to do this for us,” Bastings said.
First-of-its-kind eye implant that slows vision loss in rare eye disease
The FDA has approved ENCELTO (revakinagene taroretcel-lwey), a surgically implanted device designed to preserve sight in people with macular telangiectasia type 2 (MacTel) – a rare and slowly progressing retinal disease.
Developed by the biotech company Neurotech Pharmaceuticals, ENCELTO’s origins stem from a long-standing partnership between the Scripps Research lab of physician-scientist Martin Friedlander and the Lowy Medical Research Institute.
The device is expected to be available to U.S. patients in June 2025.
MacTel affects the central part of the retina, which is essential for sharp, detailed vision. People with the disease often have trouble reading, recognizing faces or seeing fine detail – symptoms that gradually worsen over time.
To slow this progression, ENCELTO delivers a steady dose of ciliary neurotrophic factor (CNTF), a naturally occurring protein that supports the survival and health of nerve cells – including photoreceptors in the retina.
CNTF acts as a neuroprotectant, meaning it helps shield these cells from damage to delay the degenerative process.
“This is the first time a cell-based therapy delivering a neuroprotectant has been approved to treat a neurodegenerative disease,” said Scripps Research lab physician-scientist, Martin Friedlander.
ENCELTO is also being evaluated for neurovascular degenerative conditions beyond MacTel. Friedlander’s lab is now exploring its potential to deliver other therapeutic molecules for diseases such as glaucoma and age-related macular degeneration.
“The retina is an extension of the brain, so this also implies that a neuroprotectant could be used to prevent neurodegeneration in other diseases,” said Friedlander.
“There’s no greater satisfaction for a clinician-scientist than to be able to find a treatment that will potentially impact hundreds of thousands of patients. It’s tremendously rewarding to take something from the lab bench and actually bring it to the bedside.”
Enhanced learning strategies can enhance AI model effectiveness in hospitals
A new study has found that proactive, continual and transfer learning strategies for AI models could be key in mitigating data shifts and subsequent harms to patients from the use of these models.
If data used to train AI models for medical applications, such as hospitals, differs from the real-world data, it could lead to patient harm.
To determine the effect of data shifts, the team from York University, Toronto, built and evaluated an early warning system to predict the risk of in-hospital patient mortality and enhance the triaging of patients at seven large hospitals in the Greater Toronto Area.
“As the use of AI in hospitals increases to predict anything from mortality and length of stay to sepsis and the occurrence of disease diagnoses, there is a greater need to ensure they work as predicted and don’t cause harm,” said senior author York University assistant professor Elham Dolatabadi of York’s School of Health Policy and Management, Faculty of Health.
“Building reliable and robust machine learning models, however, has proven difficult as data changes over time creating system unreliability.”
The data to train clinical AI models for hospitals and other healthcare settings need to accurately reflect the variability of patients, diseases and medical practices, she adds. Without that, the model could develop irrelevant or harmful predictions, and even inaccurate diagnoses.
Differences in patient subpopulations, staffing, resources, as well as unforeseen changes to policy or behaviour, differing health-care practices between hospitals or an unexpected pandemic, can also cause these potential data shifts.
To mitigate these potentially harmful data shifts, the researchers used transfer learning strategies which allowed the model to store knowledge gained from one domain and apply it to a different but related domain, and continual learning strategies where the AI model is updated using a continual stream of data in a sequential manner in response to drift-triggered alarms.
Although machine learning models usually remain locked once approved for use, the researchers found models specific to hospitals which leverage transfer learning, performed better than models that use all available hospitals.
Depending on the data it was trained on, the AI model could also have a propensity for certain biases leading to unfair or discriminatory outcomes for some patient groups.
“We demonstrate how to detect these data shifts, assess whether they negatively impact AI model performance, and propose strategies to mitigate their effects,” said Dolatabadi.
“We show there is a practical pathway from promise to practice, bridging the gap between the potential of AI in health and the realities of deploying and sustaining it in real-world clinical environments.”
The study is a crucial step towards the deployment of clinical AI models as it provides strategies and workflows to ensure the safety and efficacy of these models in real-world settings.
Diabetes drug shows benefits for patients with liver disease
The sodium glucose cotransporter 2 (SGLT-2) inhibitor drug – dapagliflozin – widely used to treat type 2 diabetes, shows improvements for patients with progressive liver disease, a new clinical trial has found.
The results show that treatment with dapagliflozin improved metabolic dysfunction-associated steatohepatitis (MASH) – a condition where excess fat accumulates in the liver, leading to inflammation – and liver fibrosis (a build up of scar tissue) compared with placebo.
Several studies have reported that SGLT-2 inhibitors can improve liver fat content, liver enzymes, and liver stiffness, but this is the first trial to be carried out among patients with MASH.
Almost half (45 per cent) had type 2 diabetes, and almost all had liver fibrosis (33 per ent stage 1, 45 per cent stage 2, 19 per cent stage 3). After an initial screening biopsy, participants were randomly assigned to receive 10 mg of dapagliflozin or matching placebo once daily for 48 weeks and attended health education sessions twice a year.
After an end of study biopsy at week 48, 53 per cent of participants in the dapagliflozin group showed improvement in MASH without worsening of fibrosis compared with 30 per cent in the placebo group.
Resolution of MASH without worsening of fibrosis occurred in 23 per cent participants in the dapagliflozin group compared with 8 per cent in the placebo group. Fibrosis improvement without worsening of MASH was also reported in 45 per cent participants in the dapagliflozin group compared with 20 per cent in the placebo group.
The percentage of participants who discontinued treatment because of adverse events was 1 per cent in the dapagliflozin group and 3 per cent in the placebo group.
The researchers said: “Our findings indicate that dapagliflozin may affect key aspects of MASH by improving both steatohepatitis and fibrosis.”
Large scale and long term trials are needed to further confirm these effects, they add.
Targeting mitochondria to fight leukemia
A research team is working to turn cancer cells’ own energy systems against them.
Acute myeloid leukemia (AML) remains one of the most aggressive and deadly forms of blood cancer, even as treatments have advanced in recent years.
Standard approaches like high-dose chemotherapy and bone marrow transplants can extend life but often come at the cost of severe side effects, and many patients still relapse due to drug-resistant cancer cells.
“Cancer cells grow so rapidly that they place a tremendous burden on their mitochondria – the cell’s energy producers – but they also disrupt the mechanisms that keep these mitochondria healthy,” said Kirienko, associate professor of biosciences and Cancer Prevention and Research Institute of Texas (CPRIT) Scholar.
“We’ve discovered that targeting this mitochondrial dysfunction can selectively kill AML cells while leaving healthy blood cells unharmed.”
The team’s plan is to optimise new mitochondria-targeting drugs, evaluate patient-specific responses and develop personalised drug combinations that minimise side effects and maximise survival.
Another key part of the project involves preclinical testing using mice injected with human leukemia cells. These models allow researchers to better predict how new therapies will work in real patients.
Ultimately, the team said they hope their work will lead to safer, more effective options for people with AML.
If successful, the team’s findings could yield a new class of cancer therapies with broader implications for other treatment-resistant cancers beyond leukemia. And for the thousands of Texans who will face a leukemia diagnosis this year, it represents a promising step forward.
