My piece at The Upshot today is about why biosimilars are hard and costly to make. Responding to an early draft, my friend John Methot, who has a decade of experience working on early drug discovery, explained in greater detail some of the science behind biologics:
Proteins have what’s called primary, secondary and tertiary structure. A protein is a chain of amino acids, and the primary structure is merely the sequence of amino acids (humans have 20 amino acids each represented by a letter of the alphabet). Secondary structure describes the “domains” of the protein — there are three basic domain types: alpha helices, beta sheets and coils. Each of those domain types has common high level properties in terms of the binding behavior important for therapeutics.
Tertiary structure is the actual 3D arrangement of the amino acids and it is very complicated. It is controlled by thermodynamics, electrostatic forces, etc. As proteins are being formed, in a cell or synthetically, they dynamically fold on complex ways. The folding is controlled or influenced by other factors including “chaperone proteins” that must be present to produce a specific folding regime. And it’s the specific folding arrangement that determines the protein’s binding behavior. You could produce two versions of the protein with identical amino acid sequence but folded differently due to different conditions and one might be therapeutically active and the other not.
You accurately described why they are called “biosimilars”; as background this environment-specific folding behavior is the reason you can’t produce an identically active protein under non-identical conditions even if you could produce identical amino acid sequence.
Using mass spectrometry and x-ray crystallography one could reverse engineer a therapeutic protein to determine its sequence and much of its structure but that doesn’t tell you how to produce it.
Very little, if any, of this complexity applies to small-molecule drugs. That’s why they’re so much easier to make and replicate as generics. More in my post.
