Personalised regenerative medicine

When the medicine is a living cell taken from a particular patient, ordinary pharmaceutical logic stops applying. The active substance is not synthesised to a specification; it is harvested from material whose quality was set by the patient’s disease and prior treatment. This page is about the constraint that follows. The biology of individual cell products is covered in the pages on those therapies.

Autologous versus allogeneic is a trade, not a hierarchy

An autologous product removes the immunological problem: the cells are the patient’s own, there is no rejection, and conditioning immunosuppression is given for other reasons. The price is manufacturing. Every dose is its own batch, with its own incoming material, its own record set and its own release; scale comes from running more parallel lines, not a larger reactor, so unit cost barely falls with volume. An allogeneic product restores batch economics — hundreds of doses from one characterised donor cell bank — and restores the immunology with it. Foreign HLA is recognised by T cells, and a cell stripped of HLA class I to hide from them becomes a target for NK cells, released by the missing inhibitory ligand. Hence universal-donor engineering: B2M and CIITA knockouts plus forced expression of HLA-E, or CD47 as a don’t-eat-me signal. None of these schemes yet delivers persistence comparable to an autologous product.

Variability at the input

The starting material for an autologous product is not a reagent. A patient’s lymphocytes after several lines of chemotherapy may be exhausted, skewed toward terminally differentiated phenotypes and poor at expanding; the fraction of naive and central-memory cells in the apheresis product predicts both final yield and duration of response. Some manufacturing failures are not process failures but depleted donors. That is why early collection and cryopreservation are now standard practice, and why allogeneic platforms are of most interest precisely for heavily pretreated patients.

Releasing a batch of one

A conventional biologic sits in quarantine until a compendial sterility culture reads out — fourteen days. A living cell product with functional potency cannot wait that long, and neither can a patient already lymphodepleted. Hence rapid microbiological methods and conditional release with later confirmation. The same conflict appears in analytics: the sample volume available from a single dose is limited, and every test consumes product that would otherwise be infused. The practical consequence is a shift from product control to process control, where culture parameters and in-process characterisation carry part of the evidential load that a final-substance specification carries in conventional pharma.

What is unsolved

Comparability. Any change — a new vector, a different bioreactor, a site transfer, automation — changes the product, and equivalence cannot be shown by running two batches as it can for a small molecule, because a batch here is inseparable from a patient. That, rather than the youth of the field, is why the cost of these therapies falls so slowly.