Oligonucleotide-based treatments capable of silencing specific genes have made it possible to tackle previously untreatable diseases ranging from muscular dystrophies (eteplirsen (2016), golodirsen (2019), vitolarsen (2020), casimersen (2021)) to certain forms of hereditary amyloidosis (patisiran, inotersen (2018), eplontersen (2024)).

But getting these therapies from the lab to the patient poses unique challenges. One of the most critical is defining the initial dose in humans during first-in-human (FIH) studies, that first clinical administration that must be safe but also active enough to generate useful information. Oligonucleotides have pharmacological properties that make this prediction difficult: they are widely distributed in tissues, have a short half-life in plasma but prolonged effects on their molecular target, and their elimination does not follow standard metabolic pathways. In addition, they face a critical barrier in the process of cellular internalisation: although they manage to enter cells via endocytosis, up to 99% are trapped in endosomes due to the lipid nature of their membrane, which severely limits their release into the cytoplasm, where they should exert their therapeutic effect. This “endosomal trapping” phenomenon represents one of the main bottlenecks in the efficacy of RNA-based therapies, which is why other routes of administration and delivery systems are being sought.

A recent analysis published by the FDA reviewed 89 investigational programmes (INDs) and 16 approved oligonucleotide therapies to understand how these complexities are being resolved. This practice reflects both a practical and ethical trend: to avoid exposing healthy volunteers to compounds with as yet uncertain mechanisms and potentially long-lasting effects.

In terms of trial design, single or multiple ascending dose schedules (SAD/MAD) with stepwise escalation based on preclinical studies predominated. Seventy per cent used a randomised design with placebo control, i.e. comparing the drug with an inactive substance to assess its actual effect, and more than half used sentinel cohorts (small groups receiving the first doses before moving on to the rest of the participants) in order to minimise risks. There was also a strong emphasis on the use of pharmacodynamic biomarkers, such as RNA levels or target proteins, which allows for more informed and agile dose adjustment.

But perhaps most interesting was the analysis of the cross-species scaling methods used to calculate the initial dose in humans. Two main approaches emerge:

Body surface area scaling (BSA): recommended by regulatory guidelines as more conservative, especially useful when toxicological risk is uncertain.
Body weight (BW) scaling: used more frequently in practice, especially when the chemical class is well characterised (such as ASOs with 2′-MOE modifications or GalNAc-conjugated siRNAs). In these cases, models in non-human primates (mainly cynomolgus monkeys) gave fairly accurate plasma exposure predictions, with concentration ratios close to 1:1 versus humans.

Another important conclusion is that there is still no absolute consensus on the best method of scale-up. Evidence suggests that the choice should be tailored to the specific chemistry of the oligonucleotide, its route of administration and the expected pharmacokinetic profile. In general, it is recommended to use at least two animal species and to validate the preclinical model with exposure and efficacy data.

Clinical development of oligonucleotides requires an integrated approach: relevant preclinical models, rational dose escalation, functional biomarker design and a strategic vision from the outset. It is not just about reaching the patient, but about getting it right, with scientific precision and clinical accountability.

Sources:

https://pubmed.ncbi.nlm.nih.gov/39311689/

https://www.nature.com/articles/s42003-024-06121-9

https://pubmed.ncbi.nlm.nih.gov/36669888/