Therapeutics & platforms
Delivery is everything and the last step of the journey decides; the process defines the drug; specificity cuts both ways; and the trial, not the pharmacology, is where candidates go to fail.
Twenty-seven subjects, and the same three constraints keep deciding outcomes. Delivery is everything, and the last step of the journey decides. Nucleic-acid drugs pair trivial target selection with a delivery problem so hard that almost all of them are liver drugs; mRNA waited for an immune-silent transcript and a particle that escapes the endosome — barely; an antibody–drug conjugate is decided by its linker; a Gram-negative pathogen has two independent lines of defence; and an in vivo gene-therapy dose almost never gets a second attempt.
The process defines the drug. A protein made by a living cell is defined by how it was made — which is why quality cannot be tested into a batch, why validation argues from the process, and why fill–finish builds sterility by excluding a contaminant that is usually the operator.
Specificity cuts both ways. A phage that replicates at the infection site also hands bacteria an escape route; CAR-T inherits the limits of the surface it sees; TCR therapy opens the inside of the cell and removes the safety filter thymic selection built.
Start with biopharmaceuticals: it establishes process-defined medicine, the frame the platform, cell-therapy and validation pages all work inside.
Articles 28
| AI drug design and AlphaFold applications Why coevolution in a protein family predicts a fold — and why a correct fold still does not give you a binding affinity. |
| Bacteriophage therapy A drug that replicates at the site of infection — and why the same molecule that gives phage its precision also gives bacteria an escape route. |
| CAR-T cell therapy One synthetic receptor collapses antigen recognition and costimulation into a single molecule — and inherits every limitation of the surface it has to see. |
| Fill–finish and aseptic biomanufacturing A biologic cannot be sterilised after it is sealed, so sterility has to be built by exclusion — and the excluded contaminant is usually the operator. |
| mRNA platforms and LNP delivery Two separate inventions had to arrive together: an mRNA the immune system ignores, and a particle that can break out of an endosome — barely. |
| Process validation and equipment qualification Quality cannot be tested into a batch, because the defects that matter are too rare to sample — so assurance has to be argued from the process instead. |
| Radiopharmaceuticals and theranostics A targeting molecule decides where the dose goes; the physics of the isotope decides what the dose does — and half-life decides whether it arrives at all. |
| TCR-engineered T-cell therapy Reading peptides presented on MHC opens the inside of the cell as a target — and removes the safety filter that thymic selection built into natural receptors. |
| Next-generation antibiotics against AMR Why a Gram-negative bacterium has two independent lines of defence — and why defeating one of them buys nothing. |
| Biologics Why a protein drug has to be made inside a cell — and why it almost always stays outside one. |
| Linker chemistry: the conjugation science inside every ADC An antibody-drug conjugate is a delivery contract written in chemistry: the linker specifies where the payload rides, when it detaches, and what the manufacturing plant must contain. |
| Adenoviral vectors: the delivery workhorse of genetic vaccines and gene therapy An adenoviral vector is a virus stripped of its replication machinery and re-tasked as a delivery vehicle; the field's science is the capsid's tolerability, its industry is suspension-scale GMP, and its quiet precondition is an industrial plasmid supply. |
| Biosynthetic steroids Why cortisone cannot be assembled by chemistry alone — and why the answer turned out to be a mould. |
| Allogeneic editing: making donor cells invisible Allogeneic cell therapy is three knockouts and a bank: remove the markers that make donor cells attack or be attacked, and one manufacture serves a thousand patients. |
| Chemoenzymatic API synthesis Why an enzyme picks an enantiomer more reliably than a metal complex — and why every new molecule means paying for that again. |
| In vivo gene therapy: AAV and lentiviral vectors Why a single dose can work for years — and why it almost never gets a second attempt. |
| NK-cell and CAR-NK therapy Why a donor NK cell does not cause graft-versus-host disease — and why it survives in the patient for weeks rather than years. |
| PROTACs and targeted protein degradation Why destroying a protein is easier than blocking it — and why raising the dose here makes things worse. |
| ADAR RNA editing therapeutics Why correcting a letter in RNA needs neither a nuclease nor protein delivery — and why only one letter in four can be corrected. |
| TIL therapy Why the right T cells are already inside the tumour — and why growing enough of them makes them work less well. |
| Antibody-drug conjugates Why bolting a poison onto an antibody widens the therapeutic window — and why the linker, not the antibody, usually decides the outcome. |
| Biopharmaceuticals Why a drug made by a living cell is defined by its process rather than by its formula, and what follows from that for dosing, stability and copies. |
| Bispecific antibodies Binding two things at once turns an antibody from a blocker into a machine that forces two cells, or two proteins, into contact. |
| Drug and vaccine development Why roughly nine in ten candidates that enter human trials fail, and why the failure is concentrated in the step where efficacy is first tested. |
| Microbiome therapeutics and live biotherapeutics Why a defined community of gut bacteria stops recurrent C. difficile — the bile-acid chemistry behind it, and why the same logic has not yet transferred to other diseases. |
| Nucleic acid therapeutics Base pairing makes target selection trivial and delivery everything: why siRNA and antisense drugs are almost all liver drugs. |
| Psychedelic medicine and neuroplastic agents A receptor that has been known for decades, a plasticity effect that is measurable in cells — and a trial design problem that no amount of pharmacology fixes. |
| Stem cell therapy in autoimmune disease Infused mesenchymal cells mostly die in the lung within hours — and the current best account of how they work says that is the mechanism, not a failure. |