The second millennium has seen the surge of new approaches to drug development. The wide and cheap availability of gene editing techniques has favoured the increasing role of Advanced Therapy Medicinal Products (ATMPs, including gene and cellular therapies) over traditional medicines based on small, chemical molecules. Biotechnological last generation products are no more just limited to CAR-T therapies, as acknowledged by the new family of mRNA vaccines developed against the SARS-CoV-2 virus. Many opportunities are also present in the development of highly potent active pharmaceutical ingredients (HPAPIs) and in the use of fermentation as a source of new drug substances. We summarise here some of the most recent trends in APIs development.
Many opportunities for biotechnology products
CAR-T are one of the newest classes of therapeutics, characterised by a still limited range of indications. Bi- and tri-specific antibodies able to recognise different cancer epitopes can also represent very interesting products within development pipelines, according to an interview to Inthera Biosciences CEO’s, Martin Bonde, published in Labiotech.eu.
Personalisation of treatments is also expected to grow, thanks to the progress in human genomics and the availability of specific biomarkers useful to better stratify patients that may respond to a certain treatment. According to the article, the third decade might see the first gene therapies based on CRISP-Cas9 and other gene editing techniques reaching the market, even if some technical issues are still to be solved. Single-cell technologies able to target specific cell types/organs based on receptor expression may also gain increased importance, said Alexandra Bause (Apollo Health Ventures).
The market for highly potent APis
Antibody-drug conjugates (ADC) are a class of highly potent APIs those manufacturing poses many challenges from the point of view of the management of the complex supply chain of the different starting materials and intermediates needed to obtain the conjugate. As discussed in an article by Brian Clark published in Pharmaceutical Online, this may include different manufacturing sites for the antibody, the small cytotoxic molecule and the conjugate, as well as a dedicated filling facility for the drug product, a maybe also a labeler/packager, and a logistics company for distribution. Current trends for the ADC sector indicates the preference for highly specialised and integrated CDMO contractors, able to run at a single manufacturing site all these different phases up to the finished product.
This market sector continues to offer many opportunities, explained three executives from Lonza in an interview published on Outsourcing-Pharma.com. Highly potent APIs pose challenges also from the formulation point of view, due to the need to obtain uniform dispersions of the active ingredient, and require the complete containment of the plants to protect operators from the possible health hazards. According to an article by Cynthia A. Challener published in Pharmaceutical Technology, the increasing molecular complexity of HPAPIs represents a further issue that adds to the ones previously mentioned. This complexity is reflected by the great fragility of many molecules, thus requiring formulation as liquid dosage forms for parenteral or i.v. administration instead as the more traditional oral dosage forms. Another possible conflict arises from the need to offer patients HPAPIs’ oral dosage forms for chronic or frequent administration, in indications different than oncology. Microdosing, liquid-filling of capsules, and wet granulation are examples of the possible solutions, according to Challener’s article. Many of these techniques can also be used for the preparation of low-dosage formulations of HPAPIs, while microdosing may prove useful to formulate poorly bioavailable highly potent ingredients.
The potential of artificial intelligence
Artificial intelligence (AI) is playing an increasing role in drug discovery and development. Many tasks typical of the discovery lab can now be completely automated, as well as the collection and analysis of the resulting data using machine learning algorithms. The first drug candidate completely developed by the AI to reach the clinical phase for the treatment of obsessive-compulsive disorder (OCD) has been developed by the British firm Exscientia, and it is now under a phase I experimentation run by Japanese Sumitomo Dainippon Pharma, after a development process that took just one year instead of the usual five.
Exscientia is also leading the small molecule drug design activities in the CARE consortium (Corona Accelerated R&D in Europe), an initiative funded under the IMI public-private partnership framework to speed up the development of treatments against the SARS-CoV-2 virus.
Screening of large libraries of molecules and rational-based drug design are just two of the main application of AI and big data analysis in drug development. A big multinational company such as Merck, for example, has already developed some 300 AI predictive models for compound properties, in order to reduce both time and costs of development.
A new life for fermentation-based processes
Fermentation has a long history in the manufacturing of pharmaceutical active ingredients, both small molecules and biotechnology products, thanks to the simple reaction conditions and the use of cheap starting materials. Many new, engineered micro-organisms are today available thanks to the advancements in molecular and synthetic biology, adding to the more traditional ones used to run fermentative processes. The current situation of fermentation for the production of APIs has been examined by Cynthia A. Challener from the pages of Pharmaceutical Technology.
E. coli and P. pastoris (Komagataella phaffii) are the micro-organisms more frequently used in the pharmaceutical industry, according to the paper. Engineered strains, for example at the level of the biosynthetic pathways, benefit of a higher stability and a more predictable profile supporting the easier development of the fermentation process and leading for example to overproduction of the desired API and reduction of the related impurities.
Artificial intelligence is increasingly used in this case too to design and optimise the characteristics of the bacterial strain in order to produce a specific target molecule. Automation is growing also in this field, and many steps are now run automatically upon checking of the right conditions by mean of in-process controls and reals-time sensors able to keep monitored all process conditions.