PCSK9 — when antibodies beat the peptide programme
In 2003 a French genetics group discovered a protein regulating the number of LDL receptors on liver cells: PCSK9. It was one of the most attractive target structures of modern medicine — a knockout reduced LDL cholesterol by over 50%. Several substance classes — small molecules, peptides, antibodies, later RNA interference — competed for the market. The antibodies won. The peptides did not. A story about the sometimes unfair advantages of a different substance class.
An unusual genetic discovery
In 2003 Nabil Seidah and co-authors published in PNAS a new pro-protein convertase: PCSK9 (Proprotein Convertase Subtilisin/Kexin Type 9). The original function remained unclear. Only in 2005 did Marianne Abifadel and Catherine Boileau in Paris identify gain-of-function mutations in the PCSK9 gene in families with autosomal-dominant hypercholesterolaemia: these families had extremely high LDL cholesterol levels and very early myocardial infarctions.
A few years later Helen Hobbs delivered the mirror observation. While investigating the Dallas Heart Study, she identified loss-of-function variants in PCSK9 — people who, due to a genetic variant, produced hardly any PCSK9. These people had LDL cholesterol levels about 30-50% below average, their cardiovascular event rates were correspondingly reduced. And — crucially — they were otherwise healthy. PCSK9 was a protein one could pharmacologically inhibit without safety concerns. It was a gift to the pharma industry: a genetically validated target with built-in safety guarantee.
Four substance classes compete
Between 2008 and 2015 industry and academia developed in parallel four substance classes against PCSK9. First small molecules: extremely difficult because PCSK9 binds the LDL receptor through a protein-protein interaction surface — no classical enzyme pocket that small molecules address well. Second peptides and peptidomimetic inhibitors: conceptually elegant but the same protein-protein interaction hurdles plus the general peptide problems (oral bioavailability, short half-life). Third monoclonal antibodies: large enough to directly block the PCSK9-LDL-receptor interaction, with long plasma half-lives thanks to FcRn recycling. Fourth RNA interference: inhibition of PCSK9 synthesis directly in the liver cell.
The small molecules failed essentially immediately in the preclinical phase. The peptides reached phase 1/2 but efficacy results lagged behind antibodies and safety profiles were less clean. The antibodies — alirocumab (Praluent, Sanofi/Regeneron, FDA approval 2015) and evolocumab (Repatha, Amgen, FDA approval 2015) — won the market: bi-weekly subcutaneous injection, LDL reduction of 50-60% vs. statin background, excellent safety profile. RNA interference followed: inclisiran (Leqvio, Novartis, FDA approval 2021) with twice-yearly administration.
Why peptides lost
Four structural reasons are methodologically valuable. First: PCSK9 acts extracellularly — as a circulating protein in plasma. Antibodies reach extracellular targets with high affinity and long half-life. Peptides could in principle do this too, but their affinity typically lags behind monoclonal antibodies and their half-life is shorter.
Second: the PCSK9-LDL receptor interaction is an extended protein-protein interaction without a clear 'hotspot' amino acid that a short peptide could efficiently block. An antibody can cover a large interaction surface with its Fab fragment; a 15-amino-acid peptide cannot.
Third: the economic threshold for a new hypercholesterolaemia therapy is low (statins as generics), but the patient population is large. An antibody at €1,000-1,500/year therapy cost is economically tolerable for high-risk patients; a peptide at similar cost and worse efficacy is not. Fourth: the patent investment climate 2010-2015 was substantially more attractive for antibody programmes than for peptides — the large biological platforms (Regeneron, Amgen, Genentech) had antibody expertise, not peptide expertise.
„If the target is extracellular, the mechanism a protein-protein interaction and the patient population large enough for antibody economics — then the antibody almost always wins. Peptides remain competitive where the antibody path fails: intracellular, with body-identical mimicry, in smaller indications."
What the PCSK9 story methodologically shows
Three structural lessons from this line. First: substance-class selection is its own pharmacological design goal, not a consequence of target selection. The same biological target can be pharmacologically addressed through four different substance classes, with different clinical, economic and regulatory profiles. Second: peptides are not universally superior, even if their mechanisms appear elegant. In constellations where antibodies structurally fit better (extracellular, protein-protein interaction, large patient population), they are typically more successful.
Third: the PCSK9 story is also the story of an unusually smooth regulatory approval. The genetically validated loss-of-function carriers in the Dallas Heart Study gave the authorities a safe platform: people with lifelong strongly reduced PCSK9 are healthy. This genetic validation has accelerated the phase 3 programmes and simplified the safety argument. Many peptide programmes lack this genetic foundation.
Open questions
- Will the RNA interference line (inclisiran, twice-yearly administration) displace antibodies long-term — and what lessons follow for other extracellular targets?
- Which next-generation PCSK9 modulators (oral small molecules, bispecific antibodies, in-vivo CRISPR) are in the pipeline?
- Can lessons from the peptide failure at PCSK9 be transferred to other protein-protein interaction targets — e.g. PD-1/PD-L1 in oncology?
- Which peptide-specific advantages (e.g. the possibility of oral or topical administration) could become relevant in future PCSK9 iterations?