Chapter 4 — Key Takeaways

Peptide Pharmacology

The core claims

Peptides are excellent signals and difficult drugs, and the difficulty is entirely delivery.

The oral gauntlet has five barriers, each independently fatal: stomach acid and pepsin; pancreatic proteases; brush-border peptidases; crossing the intestinal epithelium; and first-pass hepatic metabolism. Protecting a peptide from enzymes does nothing about barrier four — a fully intact peptide still cannot cross a cell layer it is too large and too polar to traverse.

Injected, a peptide has two exits. Proteolysis in circulation (specific enzymes like DPP-4, plus broad-spectrum peptidases) and renal filtration (anything below roughly 5,000 Da passes freely). Both must be closed to achieve a long half-life. Closing one leaves the other wide open.

Half-life governs everything. Dosing interval, steady state (4–5 half-lives), and whether a therapy is once-weekly or twice-daily. A long half-life buys adherence and costs reversibility — there is no undo button.

Every route trades something. Subcutaneous is the default because it bypasses the entire oral gauntlet and is self-administrable. Variability is more damaging than low bioavailability — 1% consistently is workable; 0.2%-to-4% depending on conditions is not.

Immunogenicity is a peptide-specific risk with no small-molecule counterpart. Anti-drug antibodies can do nothing, reduce efficacy, neutralize the drug, or — rarely and seriously — cross-react with the patient's own endogenous peptide. Aggregation and impurities raise the risk, which makes storage and manufacturing quality an immunological question, not just a potency one.

Cost is genuine production plus recovered development plus market structure. All three are real. Only the third explains why the same molecule costs ten times more in one country than another.

The anchor: three modifications, two minutes to seven days

NATIVE GLP-1 → SEMAGLUTIDE

  ① position 8   Ala → Aib (non-natural residue)
                 → abolishes the DPP-4 cut site          [defeats proteolysis]

  ② position 34  Lys → Arg
                 → leaves ONE fatty-acid attachment site  [manufacturing control only —
                                                           no patient benefit]

  ③ position 26  C18 diacid fatty acid via a spacer
                 → binds albumin (~66,000 Da), too large to filter
                                                          [defeats renal clearance]

  half-life  ~1–2 minutes  →  ~1 week        receptor activity: unchanged

The receptor pharmacology was never the problem. Evolution had optimized it. What was engineered was survival — and one of the three modifications exists purely to make manufacturing controllable.

The rule to carry forward

Delivery is the sharpest form of "mechanism is strong evidence against."

A molecule that cannot reach its target in useful quantity cannot work, whatever it does at the receptor. That conclusion requires no trial. But it licenses only a narrow claim — this route, this dose, this product — never "the molecule does nothing." Oral semaglutide exists precisely because someone solved a delivery problem everyone had written off.

The delivery filter — five questions, before any evidence

  1. WHAT ROUTE, and does the molecule survive it?
  2. WHAT IS THE MEASURED BIOAVAILABILITY, in humans, by this route?
  3. WHAT IS THE HALF-LIFE, and does the dosing schedule match it?
  4. WHAT MODIFICATION makes it survivable — and if none, why is that acceptable?
  5. HOW IS IT STORED AND HANDLED?

  A product that cannot answer these has not done the work.

Evidence ratings issued in this chapter

One, in Case Study 1, deliberately narrow:

And explicitly limited: it establishes nothing about cardiovascular benefit, weight loss magnitudes comparable to later agents, or once-weekly dosing. The first drug in a class is rarely the best one; its job is to show the target is real.

Numbers worth remembering

Key terms

pharmacokinetics · pharmacodynamics · bioavailability · half-life · clearance · first-pass metabolism · parenteral · subcutaneous · depot · steady state · titration · albumin binding · lipidation · permeation enhancer · immunogenicity · anti-drug antibody · cold chain · Aib

What you can now evaluate

• ✅ whether an oral, sublingual, or topical peptide claim faces a physical problem
• ✅ what question to ask a seller (measured bioavailability, in humans, by this route)
• ✅ whether a dosing schedule is consistent with a molecule's half-life
• ✅ why the Ozempic tablet has strange instructions and the injection does not
• ✅ what a gray-market vial's low price actually represents
• ✅ why storage and handling are immunological questions, not just potency ones
• ❌ not yet: whether the drug helps anyone. Chapter 5.

The one-sentence version

A peptide's biology is usually already perfect; what has to be engineered is its survival — and a molecule that cannot survive contact with a body cannot be a drug, no matter how good the mechanism looks.

Next: Chapter 5 — the most important chapter in this book. How to read the evidence.