Happy 2022, Blopiggers!
My first post of the year is about another major change to the way the World Health Organisation will be assigning “International Non-proprietary Name”s (INNs) to antibody-based therapeutics. I haven’t seen this publicised widely, so I thought I’d share it here as it is an important consideration for anyone mining or exploiting this data.
When I first started working in this field, I didn’t appreciate the full extent of the information wrapped up in a therapeutic’s INN. Classically there were five components…
Example using pre-2017 nomenclature: Adalimumab
(a) The stem (chosen by the developer), e.g. Ada–
(b) A letter, or pair of letters, reflecting the nature of the therapeutic target, e.g. –li– (for “immunomodulating”, adalimumab targets TNFa)
(c) [optional] a diversifying consonant, which also serves to make the name (somewhat) pronouncable! e.g. –m–
(d) A letter, or pair of letters, reflecting the species origin/developmental journey of the therapeutic, e.g. –u– (for “fully human”)
(e) The invariant suffix, –mab*
In 2017, it was decided that component (d) would have to expand so much to reflect growing number of species origins and developmental technologies that it was no longer tenable. In Working Doc 17.416, the committee decided to remove it altogether, presumably allowing companies to use whichever letter(s) they like. Components (a), (c), and (e) remained unchanged, while component (b) was expanded to include -vet-, reflecting the growing field of veterinary biologics.
This remained the standard until a meeting in October 2021, when the WHO decided that the explosion in the variety of antibody formats should somehow be reflected in the INN (see Working Doc 21.531). For example, there’s no way to tell that adalimumab is a monoclonal therapy, while dilpacimab is a bispecific. To do this, they resolved to replace the historically static suffix “-mab” with four new stems:
- Unmodified immunoglobulins: –tug. As the name suggests, your standard monoclonal antibody.
- Antibody articial: –bart. This covers a whole range of options, but is perhaps summarised as: antibodies with any mutations/allelic crossovers to engineer the function of or stabilise full Fc constant domains.
- Multispecific immunoglobulins: –mig. Pretty self-explanatory, bispecifics etc. This trumps all other suffixes, i.e. a bispecific with modified full Fc would be -mig, not -bart (or -migbart!)
- Fragments: –ment. Any immunoglobulin without a full Fc domain, such as Fabs or scFvs.
In addition to this change, the committee also decided to unpack the former “immunomodulatory” (-li-) category into -ki- (cytokine/cytokine receptor targetting), -ler- (allegen targetting), -sto- (immunostimulatory), and -pru- (immunosuppressive).
While my understanding is that antibodies named under the older conventions won’t get new names, I thought it would be fun to note that under the new system [assuming the same choice of (c) and (d)], adalimumab would be adakimutug. As if INNs weren’t hard enough to pronounce already! Also worth noting that there will be edge cases that’ll mean using these new conventions for mining is imperfect (e.g. I want all the therapeutic antibodies that are fragments, but I can’t just search for the -ment suffix as I’ll miss all the bispecific fragments that end -mig).
Anyway, the main takehome is that today’s INNs are less expressive about species origin, but more expressive about target category and format, and should provide the naming diversity needed to cope with the expected wave of new diverse antibody therapeutics entering the clinic. Oh, and don’t be too concerned when you start seeing -tugs, -barts, -migs, and -ments showing up in Thera-SAbDab!
*Another format of interest not covered by these rules are fusion proteins that contain an antibody variable domain. These have the suffix -afusp (antibody fusion protein, see this review page 13/96), and we’ve been tracking these in Thera-SAbDab since their inception.
