The translational gap: bridging basic research, clinical practices and society

The past few decades have seen a dramatic shift in how human disorders can be treated. Growing knowledge of the molecular dynamics underlying physiological processes and technological advances are driving the development of transformative therapies.

Spinal muscular atrophy, a genetic condition that causes muscle weakness and degradation, for example, was untreatable until 2016. The approval of nusinersen, an RNA-based drug that restores muscle function when injected into the spine, marked a significant milestone in the field of gene therapy. Since then, thousands of children have benefitted from the drug.

Efficient collaboration among researchers, clinicians, and industry partners is crucial to rapidly translate discoveries into clinical applications. “We are very good at doing research and publishing papers, but we need to motivate and raise awareness in young researchers of the importance of transferring scientific results from laboratories to market and society,” says Fabio Terragni, member of the management committee delegate for technology transfer at Human Technopole. Set up by the Italian government in 2018, Human Technopole is a research foundation located in Milan, Italy. Its mission is to conduct basic research in the life sciences, promoting people’s health and well-being.

In October 2023, with the help of its new director, Marino Zerial, who took office a month earlier, Human Technopole organized a conference on ‘Future Trends in Translational Medicine’ in collaboration with Nature Italy, which gathered leading scientists, industry and tech-transfer experts in four research areas that are bridging the gap between basic science and clinical practice.

Gene therapy and RNA: investment in basic science pays off

While gene therapies deliver functional copies of missing or faulty genes, RNA-based therapies can help restore normal protein production from dysfunctional genes. Several gene and RNA-based drugs, such as nusinersen, are now in the clinic.

“These types of therapies are the result of years of research into the mechanisms of gene expression,” says Irene Bozzoni, molecular biologist at Istituto Italiano di Tecnologia in Genoa, Italy, and a pioneer in the field of therapeutic RNA¹.

Bozzoni is very optimistic about the future of DNA and RNA-based therapeutics for improving the lives of patients with rare genetic diseases, despite the time progressing them to the clinic, which can be up to 20 years, and the challenge of delivering them to target tissues or cells. “For almost every disease we can find a molecular solution; what is proving harder, is finding effective delivery methods,” she said.

Adeno-associated viruses (AAVs) are the preferred method to deliver small payloads. Early-stage clinical trials of liver-directed, AAV-mediated gene therapies for two inherited metabolic disorders (Mucopolysaccharidosis type VI and Crigler Najjar) led by Nicola Brunetti-Pierri, physician-scientist at Telethon Institute of Genetics and Medicine, near Naples, are showing promising results. Some patients receiving higher doses experienced clinical improvements due to sustained production of the therapeutic gene. Brunetti-Pierri will be carrying out further assessments of the efficacy and long-term safety of the approach as part of the clinical development process.

Lipid nanoparticles (LNPs) offer an alternative delivery option to AAVs for DNA and RNA-based therapeutics. Michael Mitchell, professor of bioengineering at The University of Pennsylvania, Philadelphia, is using chemistry and microfluidics to formulate LNPs with ionizable lipids and polymers that can deliver (and effectively release) nucleic acids into specific cell types, including foetus cells in utero. With these LNPs it is possible to treat congenital diseases before birth, preventing irreversible tissue damage and dramatically extending lifespan. Mitchell’s team is exploring the use of LNPs to deliver base-editing therapies to foetal and neonatal mouse brains with the aim of correcting lethal lysosomal storage diseases. The future of these therapies looks promising thanks to the synergy between basic and applied research, clinics, patients and private donors.

Data science and genomics: fuelling discovery with transformative resources

The analysis of large, detailed and diverse datasets is providing actionable insights into disease and risk prediction in different populations. Nicole Soranzo, head of the Genomics Research Centre - Population & Medical Genomics, at Human Technopole, wants to use genomic data to inform tailored prevention and treatment strategies for common diseases. “A data-driven approach to genomics allows discovery without any preconceptions. It is incredibly powerful when it comes to uncovering new associations or correlations,” she says. “We need a bold, concerted initiative in Italy to realise the opportunities of large-scale genomic studies for personalized medicine.”

The UK Biobank and The Human Cell Atlas initiative are examples of some revolutionary large-scale bioresources freely available to advance research. Emanuele Di Angelantonio, head of the Health Data Science Centre, at Human Technopole, and colleagues are developing more accurate risk prediction models for non-communicable diseases by combining different types of information including genetics and electronic health records (EHR). They have shown using UK Biobank data that adding polygenic risk scores, which estimate an individual’s disease risk due to multiple genetic variants, to conventional cardiovascular disease (CVD) risk factors can improve prediction of first-onset CVD².

Pharmaceutical companies recognize the higher likelihood of success for drug targets with human genetic evidence of disease association³. They are incorporating genetic and health data from the UK Biobank into their discovery pipelines. Astrazeneca researchers, for instance, analysed UK Biobank data to identify gene-phenotype relationships, sharing the results in the public PheWAS Portal⁴. Their findings suggested that inhibiting the MAP3K15 gene could be a novel therapeutic strategy for reducing diabetes risk.

The Human Cell Atlas contains single cell genomic data and spatial transcriptomic data from nearly 120 million human cells collected by scientists globally. Rasa Elmentaite and colleagues in Sarah Teichmann’s group are developing a suite of computational tools to integrate, interpret and extract biological insights from the single cell data. Tools such as drug2cell can be used to predict drug side-effects on pacemaker cells, whereas genes2genes can reveal differences in the developmental trajectory of cells growing in vitro and in vivo, for example.

AI for life sciences: making sense of the data

The main purpose of using artificial intelligence in research is to extract useful information from complex multidimensional data. “With AI we can boil down complexity to what really matters,” says Andrea Sottoriva, head of Computational Biology Research Centre at Human Technopole.

His lab is exploring cancer evolution and why some tumours develop resistance to treatment. He does this by combining experimental observations from patient samples and model systems with computational methods, to predict the course of disease. “If we can control cancer in the long-term, in the same way we can control HIV, that would be transformative,” says Sottoriva.

AI is already extensively used in drug discovery, not just to identify targets but also to design new, more effective drugs with fewer side effects — based on predictions of how the compound will interact with other proteins. ML algorithms analyse the chemical properties of drugs and then compare them to information about diseases and biological pathways to identify new uses for existing drugs. Drug repurposing drastically reduces the cost and time of bringing new treatments to patients and is an area where the impact of AI on healthcare is likely to be felt first.

Other AI applications that are expected to reach the clinic soon are in medical image and EHR analysis. ML algorithms can be trained to detect early signs of disease from medical images or spot patterns in EHR data. Riccardo Bellazzi, professor of bioengineering and biomedical informatics, and colleagues at the University of Pavia, Italy, are working on approaches to assess the reliability of AI systems based on EHR data. This type of work is key to ensure AI systems meet the EU’s regulatory requirements and to address concerns of potential users.

Advances in organoids: “ready-to-go” models

The last decade has seen tremendous advances in the production and use of organoids, tiny, self-organising, three-dimensional versions of organs or tissue cultures derived from stem cells. Organoids offer a viable alternative to animal testing as they can be made to recapitulate dysfunctions observed in disease and used to study treatment responses.

Giulio Pompilio, scientific director at Centro Cardiologico Monzino, one of Europe’s largest cardiac hospitals, is interested in how cardiac organoids are being used for modelling inherited heart conditions and drug screening. “With organoids we can carry out ‘clinical trials’ in a dish to explore personalized medicines and drug toxicity,” he explains.

Christine Mummery, professor of developmental biology at Leiden University, Netherlands, and her team are using small molecules to differentiate human pluripotent stem cells into different cardiovascular cell types that can self-assemble to form cardiac microtissues. These cardiac microtissues can be used to explore the non-cardiomyocyte contribution to heart disease⁵ and drug-induced cardiotoxicity in high-throughput calcium screens.

Patient-derived tumour organoids are proving to be useful for people with rare cancers, where there is no standard of care and limited clinical trial options. Alice Soragni’s lab at UCLA has developed a method to grow organoids from sarcoma tissue samples and screen hundreds of drugs to identify the most suitable ones within 6 days of surgery. Soragni will be leading the organoid-based functional precision medicine trial in osteosarcoma (PREMOST).

Leveraging the strengths of industry and academia

The Human Technopole’s Centre for Innovation and Technology Transfer, set up in 2020, is providing entrepreneurial training to Italian scientists and hosting national and international networking events involving academic organizations, public bodies and industry, to capture the value of Italian research. Terragni concludes: “This conference gave us an opportunity to reflect on ways to unleash the game-changing potential of our life sciences research on healthcare.”