Chapter 4: Peptide Pharmacology: Why Peptide Drugs Are Hard to Make, Hard to Deliver, and Worth the Effort

"The dose makes the poison." — attributed to Paracelsus, 16th century

"Getting the dose there at all makes the drug." — the peptide chemist's amendment

Overview

There is a version of this book's subject that stops at Chapter 3. In that version, you understand what peptides are, how they signal, and how the endocrine system works, and you conclude that peptide therapeutics should be straightforward: identify the signal that is missing or misfiring, supply the right molecule, done.

That version is roughly a century behind, and the reason is contained in a single fact.

Native GLP-1 has a half-life of about one to two minutes.

Think about what that means for a drug. You inject it. Within two minutes, half of it is gone. Within ten, essentially all of it. There is no dose you could give, no schedule you could design, that would produce useful sustained therapy with that molecule. GLP-1's physiology was understood in the 1980s. Its therapeutic potential was obvious. And for roughly two decades, nobody could do anything with it, because the molecule that carries the signal so beautifully is destroyed almost the instant it is released.

This chapter is about that gap and how it was closed. It is, in a sense, the chapter where peptide science becomes peptide medicine — and it is also the chapter that gives you the single most efficient filter for evaluating peptide product claims. A very large fraction of what is sold in the consumer peptide market fails not on efficacy but on delivery: it proposes a route that the molecule cannot survive, at a dose that could not reach the target, in a form that would not last long enough to matter.

You do not need a clinical trial to notice that. You need Chapter 4.

In this chapter, you will learn to:

• Explain precisely why you cannot swallow insulin, barrier by barrier
• Define bioavailability, half-life, and clearance, and say what each one governs
• Describe the two elimination routes for peptides and how each is defeated
• Trace all three modifications that turned GLP-1 into semaglutide
• Compare delivery routes and name the trade-off each makes
• Explain immunogenicity and why it ends some peptide programs
• Break down what a peptide drug actually costs to produce
• Use delivery as a first-pass filter on any peptide claim you encounter

Learning Paths

💊 GLP-1 — §4.5 is the chapter. It is the full engineering story behind the drug you came here for, and §4.6 explains the oral tablet's strange administration instructions. 🏋️ Performance — §4.2, §4.4, and §4.9. Chapter 4 is where you acquire the question that will resolve most Part III claims before you reach any evidence: how is this supposed to survive contact with a body? 🔬 Science — full read; §4.3 and §4.7 are the pharmacokinetic and immunogenicity foundations for Chapters 32–36. 💄 Cosmetic — §4.1 and §4.4. A topical peptide faces a delivery problem that is a close cousin of the oral one, and Chapter 30 is the payoff. 🏥 Clinical — §4.6 (bioavailability and why the oral tablet's instructions are what they are) and §4.7 (immunogenicity) are the two most clinically actionable sections.

4.1 Why you cannot swallow insulin

Insulin was isolated in 1921. Oral insulin has been pursued, seriously and continuously, ever since. More than a century later, insulin is still injected.

This is not for lack of effort or money. It is because a swallowed peptide has to survive a sequence of barriers, each of which is independently sufficient to destroy it.

THE ORAL GAUNTLET — what a swallowed peptide meets, in order

  ① STOMACH           pH around 1.5–3.5, plus PEPSIN
                      Acid alone unfolds many peptides. Pepsin cleaves peptide bonds.
                      · A folded peptide loses its shape; a shape-dependent peptide loses its function
                      · Disulfide-stabilized peptides survive better — which is why some peptides
                        are more oral-tolerant than others, and it is still not enough

  ② SMALL INTESTINE   TRYPSIN, CHYMOTRYPSIN, ELASTASE, CARBOXYPEPTIDASES from the pancreas
                      A coordinated protein-dismantling system. This is what it is FOR.
                      · These enzymes have specificities — trypsin cuts after lysine and arginine,
                        chymotrypsin after aromatic residues — so a peptide's own sequence
                        determines how fast it is destroyed

  ③ BRUSH BORDER      PEPTIDASES on the intestinal lining itself
                      A final layer that chops surviving short peptides into
                      di- and tripeptides and free amino acids

  ④ THE WALL          Even an intact peptide must now CROSS the intestinal epithelium
                      · Too large to pass between cells (tight junctions)
                      · Too polar and too large to diffuse through cell membranes
                      · Dedicated transporters exist — but for di- and tripeptides,
                        not for anything drug-sized

  ⑤ THE LIVER         Anything absorbed from the gut goes FIRST to the liver via the portal vein
                      FIRST-PASS METABOLISM: the liver clears much of what arrives
                      before it ever reaches the general circulation

  Net result for an unmodified therapeutic peptide: effectively zero.
  Not "low." Not "variable." Functionally zero.

Each barrier is separately fatal. A modification that protects against pepsin does nothing about trypsin. Surviving all the enzymes still leaves the physical problem of crossing the wall. And crossing the wall still leaves the liver.

This is why the oral peptide problem is genuinely hard rather than merely unsolved, and it is why the existence of an approved oral peptide drug is a real achievement rather than an obvious step (§4.6).

⚠️ Hype Check — "oral" and "sublingual" peptide products

A large category of consumer products offers peptides in oral, sublingual, buccal, or "liposomal" form. The claim is usually some version of:

"Our sublingual delivery bypasses digestion for direct absorption into the bloodstream."

What's true: sublingual and buccal absorption are real routes. The tissue under the tongue and inside the cheek is thin, well supplied with blood, and drains directly into the general circulation — genuinely bypassing both the stomach and the liver's first pass. Several small-molecule drugs use this route effectively, and it is the reason nitroglycerin works under the tongue.

Where it fails for peptides: the route bypasses barriers ①, ②, and ⑤. It does nothing about ④ — the physical problem of a large, polar molecule crossing a cell layer. Sublingual tissue is thin but it is still an epithelium with tight junctions, and it is a small surface area with a limited residence time before you swallow. The molecules that work sublingually are almost all small and lipophilic. A four-thousand-dalton peptide is neither.

The question to ask: what is the measured bioavailability, in humans, by this route, compared with injection? Not "is it absorbed" — everything is absorbed at some non-zero rate — but what fraction, measured how, in how many people? An approved oral peptide can answer this: oral semaglutide's answer is roughly 1%, established in a formal program, with a specific fasting protocol required to achieve even that.

A product that cannot answer has not measured it. And note the asymmetry: measuring bioavailability is not difficult or expensive by pharmaceutical standards. A company that has not done it has chosen not to.

4.2 The two-minute problem: proteolysis and renal clearance

Suppose you inject. You have now bypassed the entire oral gauntlet — and the peptide still disappears in minutes. Two mechanisms are responsible.

Proteolysis in circulation. Blood and tissue contain peptidases. Some are broad-spectrum; some are remarkably specific.

The specific ones matter enormously. DPP-4 — dipeptidyl peptidase-4 — cleaves two amino acids from the N-terminal end of peptides with a particular residue at position 2. GLP-1 has exactly that residue, and DPP-4 is abundant on cell surfaces and circulating in blood. The result is that native GLP-1 is being clipped essentially as fast as it is released, converting it into a fragment that no longer activates the receptor.

This specificity is the whole reason semaglutide's position-8 modification works (§4.5). Defeat one enzyme with one atom's worth of change, and a two-minute molecule becomes a stable one.

Renal clearance. Your kidneys filter blood continuously. The glomerular filter passes molecules below roughly 5,000 daltons freely, with the cutoff softening up to about 30,000–60,000 daltons depending on shape and charge.

Look back at Chapter 1's size spectrum and the implication is immediate: most peptides are small enough to be filtered out of blood on the first pass through the kidney. This is not degradation — the molecule is intact — it is simply removed.

TWO EXITS — and the two engineering strategies that close them

   PEPTIDE IN BLOOD
        │
        ├──▶  EXIT 1: PROTEOLYSIS            ├──▶  EXIT 2: RENAL FILTRATION
        │     enzymes cleave the chain       │     kidney passes molecules below
        │     · specific (DPP-4) or broad    │       roughly 5,000 Da
        │                                    │     · the molecule leaves INTACT
        │                                    │
        │  DEFEATED BY:                      │  DEFEATED BY:
        │  · substituting the residue the    │  · making the molecule EFFECTIVELY
        │    enzyme recognizes               │    larger — bind it to something big
        │  · D-amino acids (mirror images    │    (albumin, PEG, an antibody fragment)
        │    enzymes cannot process)         │  · the drug is not bigger; its
        │  · cyclization (no free ends)      │    COMPLEX is
        │  · N- or C-terminal capping        │
        ▼                                    ▼
   Both must be closed. Closing one leaves the other wide open — which is exactly
   why semaglutide needed both a substitution AND a fatty acid.

That last line is the key insight of the chapter. Half-life extension requires defeating both exits simultaneously. A protease-resistant peptide that is still small enough to filter will last somewhat longer and still not last a week. A peptide bound to albumin but still cleavable by DPP-4 will be clipped while riding around on its carrier.

🔍 Check Your Understanding — the two exits

1. Name the two routes by which an injected peptide leaves the circulation. Which one destroys the molecule, and which one removes it intact?
2. A modified peptide has been made completely resistant to every protease in blood. Its molecular weight is around 3,000 daltons. How long should you expect it to last, and which exit are you reasoning from?
3. Binding to albumin changes nothing about the peptide's own chemistry. Why does it extend the half-life anyway?

4.3 Half-life, dosing interval, and what "once weekly" requires

Three terms, used constantly, worth pinning down.

Half-life is the time for the concentration of a drug in the body to fall by half. It is the single most useful number for predicting how a drug will behave.

Clearance is the volume of blood cleared of drug per unit time — the underlying process; half-life is what you observe.

Bioavailability is the fraction of an administered dose that reaches the general circulation intact. Intravenous administration is 100% by definition. Everything else is less, sometimes by a lot.

HALF-LIFE AND DOSING — why the numbers work out the way they do

  Roughly, after each half-life, half of what remains is gone:
     1 half-life  → 50% remains
     2            → 25%
     3            → 12.5%
     4            → 6.25%
     5            → ~3%      ← "essentially gone" by convention

  DOSING INTERVAL should be roughly comparable to half-life for stable levels:

  peptide            half-life        practical dosing
  ─────────────────────────────────────────────────────────────
  native GLP-1       ~1–2 minutes     impossible as a drug
  exenatide          ~2–3 hours       twice daily
  liraglutide        ~13 hours        once daily
  semaglutide        ~1 week          once weekly
  a typical antibody weeks            every few weeks to months

  STEADY STATE is reached after roughly 4–5 half-lives of regular dosing.
  For semaglutide that is roughly a month — which is why the full effect of a
  dose change is not visible for several weeks, and why trials run for a year.

Two practical consequences that recur throughout the book.

Titration exists because of the side effects, not the efficacy. GLP-1 receptor agonists are started low and increased stepwise. This is not because a low dose is therapeutically adequate; it is because the gastrointestinal effects are much worse if you begin at the target dose. The body accommodates over weeks. This is one of the clearest cases where the schedule is a clinical intervention in its own right.

A long half-life is not free. It is excellent for adherence — once a week beats twice a day, and adherence is one of the largest determinants of real-world effectiveness. But it means that if something goes wrong, you cannot simply stop. A drug with a one-week half-life takes roughly a month to clear substantially. There is no undo button.

That trade-off is worth holding onto when Part III examines long-acting compounds used without medical supervision. A short-acting compound that produces an unexpected effect is a bad afternoon. A long-acting one is a bad month.

🧬 The Molecule — albumin, the carrier a peptide borrows

Half-life extension by albumin binding is the most successful single trick in the field, and it works by borrowing a molecule that had already solved the problem.

Human serum albumin is the most abundant protein in blood plasma, at roughly 66,000 daltons — far above the kidney's filtration cutoff, so it is not filtered out. It is also actively rescued from degradation and returned to the circulation by a dedicated salvage receptor, which is why albumin persists in blood for weeks rather than minutes.

And its day job is carrying fatty acids. Albumin has binding sites evolved to grip a fatty acid chain — which is exactly what a lipidated peptide presents to it.

So the peptide itself does not get bigger. Its complex does. Riding albumin, it is too large to filter and largely shielded from proteases; because binding is reversible, a small free fraction is always available to reach receptors while the bound pool works as a slow-release reservoir. The drug hitches a ride on something that was going to still be there next week regardless.

4.4 The delivery toolkit

Every route makes a trade. Here is the complete set available to peptides.

Subcutaneous injection is the default for a reason. It is self-administrable with a very small needle, it produces a smooth absorption profile over hours rather than an immediate spike, and it bypasses every one of the oral gauntlet's barriers. Nearly every peptide drug in this book is given this way.

The routes with low bioavailability are not thereby useless. A route delivering 1% is perfectly viable if you can simply put 100 times more drug in the dose and the drug is cheap and safe enough to waste. That is precisely the oral semaglutide strategy. Low bioavailability is a cost problem and a variability problem, not automatically a viability problem.

But variability is the real killer. A route that delivers 1% consistently is workable. A route that delivers somewhere between 0.2% and 4% depending on what you last ate is not, because the dose the patient actually receives is unpredictable. Intranasal delivery struggles largely on this axis, and it is the underappreciated reason so many nasal peptide products disappoint.

🩺 Safety and Risk — a safety record belongs to a route, not only to a molecule

It is natural to treat safety as a property of a molecule. It is not. It is a property of a molecule at an exposure, and the route sets the exposure in at least four ways.

How much arrives — bioavailability. How fast — the same total dose delivered as a sharp spike rather than absorbed over hours produces a different peak, and peaks drive a great many side effects. Which tissue sees the highest concentration — an inhaled peptide exposes lung tissue, a topical one exposes skin, an oral one sends whatever is absorbed to the liver first. And what else is in the product — a permeation enhancer is not inert packaging. It is an excipient with pharmacology of its own, and it has to be evaluated on its own terms.

The rule that follows is blunt: a compound with a long and reassuring safety record by injection has, by any other route, an unstudied safety record. New questions arise that the original data never addressed — delivering a peptide into the lung raises pulmonary questions that a subcutaneous injection of the same molecule never raises.

And variability compounds all of it. A route delivering an unpredictable fraction produces an unpredictable exposure, which in an amplifying system (§2.4) is not a minor inconvenience.

4.5 The semaglutide anchor: three modifications, two minutes to seven days

This is the book's central worked example, and it is worth understanding completely because it demonstrates everything in this chapter operating at once.

The starting material. Human GLP-1 is 30 amino acids. It binds its receptor beautifully. It has a circulating half-life of about one to two minutes, because DPP-4 clips it almost immediately and because at roughly 3,300 daltons it is well below the renal filtration cutoff.

As a drug, it is useless. Not marginal — useless.

The engineering. Three changes, each addressing a specific problem.

FROM GLP-1 TO SEMAGLUTIDE — three modifications, three problems solved

  NATIVE GLP-1 (7-36 amide)        30 amino acids, half-life ~1–2 minutes
   1                                            30
   H-A-E-G-T-F-T-S-D-V-S-S-Y-L-E-G-Q-A-A-K-E-F-I-A-W-L-V-K-G-R
     ▲                                  ▲
     │                                  │
     │ position 8                       │ position 26 (a lysine)
     │                                  │
     └── MODIFICATION 1                 └── MODIFICATION 3

  ─────────────────────────────────────────────────────────────────────────

  ① POSITION 8:  Ala  →  Aib  (2-aminoisobutyric acid, a non-natural amino acid)
     PROBLEM SOLVED: DPP-4 recognizes and cleaves after position 2 in molecules
     with this pattern. Aib is not a standard residue and the enzyme cannot
     process it. The cut site is gone.
     → defeats EXIT 1 (proteolysis)

  ② POSITION 34: Lys  →  Arg
     PROBLEM SOLVED: not stability — SPECIFICITY of manufacture. GLP-1 has two
     lysines, and the fatty acid in modification ③ attaches to a lysine. With two
     available, you would get a mixture attached at either site. Removing one
     leaves a single attachment point, so the reaction produces one defined product.
     → an ENGINEERING necessity, not a pharmacological one. Worth noticing:
       one of the three modifications exists purely to make manufacturing controllable.

  ③ POSITION 26: a C18 diacid FATTY ACID attached via a short spacer to the
     remaining lysine
     PROBLEM SOLVED: the fatty acid binds ALBUMIN, the most abundant protein in
     blood (~66,000 Da). The complex is far too large to be filtered by the kidney,
     and albumin-bound drug is also protected from enzymes. The peptide is
     released slowly from albumin to act, so the complex acts as a reservoir.
     → defeats EXIT 2 (renal filtration), and adds further protease protection

  ─────────────────────────────────────────────────────────────────────────

  RESULT: half-life ~1–2 minutes  →  ~1 week.       Once-weekly injection.
  Molecular weight ~4,100 Da. Receptor activity: preserved.

Three things about this deserve emphasis.

The activity was not improved. Semaglutide is not a better GLP-1 receptor agonist than GLP-1 in any pharmacological sense. Its whole advantage is persistence. This is the general pattern in peptide engineering: the receptor pharmacology is usually already optimal, because evolution did that part. What is being engineered is survival.

One of the three modifications is a manufacturing concession. Position 34 does nothing for the patient. It exists so that the fatty acid attaches at one defined site rather than producing a mixture. That detail is worth remembering when Chapter 32 explains why peptide manufacturing is difficult and Chapter 34 explains what "purity" means — a mixture of two attachment isomers would be a quality control nightmare, and the molecule was designed to prevent it.

And the strategy is general. Lipidation — attaching a fatty acid for albumin binding — is now used across the field. Liraglutide uses a shorter fatty acid and achieves roughly 13 hours. Insulin degludec uses a related approach for ultra-long insulin action. Tirzepatide uses the same principle. Chapter 33 covers the full toolkit; this is its most successful single application.

🧬 The Molecule — why Aib?

2-aminoisobutyric acid is worth a moment because it illustrates something general.

It is essentially alanine with an extra methyl group on the alpha carbon. That extra group makes it a non-proteinogenic residue — not one of the twenty, not encoded by any gene, not incorporable by a ribosome. It has to be put in chemically, which is one reason semaglutide is made by synthesis rather than grown in bacteria (Chapter 32).

What it buys: DPP-4's active site expects a particular geometry at that position. Aib does not fit. The enzyme cannot process the substrate, and the cut site is functionally abolished.

What it costs: nothing pharmacologically — the receptor tolerates the change. The molecule still binds and still activates.

This is the ideal outcome of a modification, and it is rare. Most substitutions that block an enzyme also degrade receptor binding, because the features an enzyme recognizes and the features a receptor recognizes frequently overlap. Finding a change that one protein notices and the other does not is a genuine piece of medicinal chemistry, and it usually takes many attempts. The published version is the one that worked; the ones that did not are not in the paper.

4.6 Bioavailability, and the strangest instructions on any drug label

Oral semaglutide is one of the most instructive objects in this book, because its label makes the pharmacokinetics visible.

The problem was the oral gauntlet. The solution has two parts.

A permeation enhancer. The tablet contains a substance (SNAC, a salcaprozate salt) that transiently and locally raises pH in the immediate area and promotes absorption of the peptide across the stomach lining. This is an unusual strategy: the drug is absorbed from the stomach, in a small localized region, rather than from the intestine.

A strict administration protocol. Because absorption depends on a local environment that food and fluid disrupt, the tablet must be taken on an empty stomach with a small specified volume of water, followed by a waiting period before eating, drinking, or taking other oral medications.

The resulting bioavailability is roughly 1%.

Which is why the tablet contains dramatically more drug than the injection delivering a comparable effect. Approximately ninety-nine percent of the drug in that tablet is expected not to reach the circulation, and the formulation is designed around that expectation.

💊 In the Clinic — reading a label as pharmacology

Pull up the oral semaglutide label on DailyMed and read the administration section. Every line is Chapter 4:

• "On an empty stomach" → food disrupts the local absorption environment and destroys what little bioavailability exists
• "With no more than a specified small volume of water" → more fluid dilutes the enhancer and flushes the tablet out of the absorption window
• "Wait before eating, drinking, or taking other oral medications" → the absorption window is brief and easily closed
• "Swallow whole; do not split, crush, or chew" → the formulation's local geometry matters

Compare with the injection's instructions, which are essentially "once weekly, any day, with or without food."

The injection has almost no instructions because it has almost no delivery problem. The tablet's instructions are the delivery problem, written down. When a patient asks why the pill has so many rules, that is the answer, and it is more satisfying than "because the label says so."

As always: how any individual should take a prescribed medication is a matter for their prescriber and pharmacist. The point here is what the instructions reveal about the science.

4.7 Immunogenicity: when the body makes antibodies against the drug

A risk that has no counterpart for small molecules.

Your immune system is built to recognize foreign proteins and peptides. A therapeutic peptide is, structurally, exactly the kind of thing it is looking for. So sometimes the immune system responds and produces anti-drug antibodies.

Consequences range from nothing to program-ending:

Nothing. Antibodies are detected but do not bind in a way that matters. Common, and usually the outcome.

Reduced efficacy. Antibodies bind the drug and clear it faster, so the same dose produces less effect — a tolerance-like pattern arising from an entirely different mechanism than Chapter 2's.

Neutralization. Antibodies bind the drug's active region and block receptor binding. The drug stops working.

Cross-reaction with the endogenous peptide. The most serious outcome: antibodies raised against the drug also recognize the patient's own version of the molecule, and neutralize that too. This can produce a deficiency worse than the original condition. It is rare, it is the reason immunogenicity is taken so seriously, and it has ended real drug programs.

What raises the risk: differences from the human sequence, aggregation of the drug in the vial, impurities from manufacture, the route (subcutaneous is generally more immunogenic than intravenous), and dosing frequency.

Which is why this section belongs in a book with a Chapter 19. Aggregation and impurities are manufacturing and storage problems. A peptide that has been improperly stored, reconstituted with inappropriate diluent, or manufactured without aggregation control is more immunogenic than the same molecule handled properly — not because the sequence changed, but because the physical state did. Regulated manufacturing controls for this explicitly. Nothing controls for it in an unregulated supply chain.

4.8 Why peptide drugs are expensive — the honest breakdown

Peptide drugs cost a great deal, and the reasons are a mixture of genuine cost and market structure. Being precise about which is which matters, because both "it's pure greed" and "it's just expensive to make" are wrong.

WHAT YOU ARE PAYING FOR                         [illustrative structure, not a specific drug]

  GENUINE PRODUCTION COSTS
  ├─ Synthesis or fermentation      real, and higher than small molecules — Ch 32
  ├─ Purification                   substantial; separating a 30-mer from 29-mers is hard
  ├─ Formulation                    stabilizers, buffers, preservatives, aggregation control
  ├─ Fill-finish, sterile           aseptic filling is a genuine capacity bottleneck
  ├─ Delivery device                pens and autoinjectors are precision devices
  ├─ Cold chain                     refrigerated storage and shipping, end to end
  └─ Quality control                identity, purity, potency, sterility, endotoxin — Ch 34

  DEVELOPMENT COSTS RECOVERED
  ├─ Discovery and preclinical      years, mostly unsuccessful candidates
  ├─ Clinical trials                the dominant cost; SELECT alone enrolled ~17,000 people
  ├─ Failed programs                the ones that worked pay for the ones that didn't
  └─ Regulatory                     submission, review, post-marketing commitments

  MARKET STRUCTURE
  ├─ Patent exclusivity             a legal monopoly for a defined period
  ├─ What the market will bear      pricing is set by willingness to pay, not by cost
  ├─ Rebates and net vs. list       the advertised price is not what most payers pay
  └─ Jurisdiction                   national price negotiation produces large differences

  The first two blocks are real and substantial. The third is why the same molecule
  costs roughly ten times more in one country than another. Both are true; Ch 12
  works the whole picture.

Two observations worth carrying forward.

A research-grade vial and a prescription are not comparable products. A gray-market vial's price reflects the first line of the first block and essentially nothing else — no sterile fill-finish, no formulation science, no quality control system, no clinical evidence, no regulatory oversight, no device, and no cold chain guarantee. It is cheaper because it is a different thing, not because somebody found efficiencies. Chapter 19.

And the manufacturing constraint is real. When GLP-1 drugs went into shortage, the binding constraint was not the peptide. It was aseptic fill-finish capacity and injector pen assembly — the last steps, which require enormous, highly regulated, slow-to-build facilities. Understanding this is what makes Chapter 12's compounding story comprehensible rather than mysterious.

💊 In the Clinic — what the needle costs, in practice

Chapter 1 established why nearly every peptide medicine is injected. What that costs the person receiving it is a separate subject, and it is larger than it looks from the prescribing side.

Needle fear is common and consequential. It is a documented barrier to starting injectable therapy and to staying on it, and it is not dissolved by explaining that the needle is small.

The cold chain follows the patient home. Refrigeration until first use, a defined in-use window at room temperature, protection from light and from freezing. That is a real constraint for someone traveling, working shifts, living somewhere very hot, or without dependable power — and a product that has been frozen or overheated is a quality problem the patient cannot see.

The device is part of the drug. Pens and autoinjectors demand training, dexterity, and eyesight, and they are a manufacturing bottleneck in their own right, which means a supply disruption can land on the device rather than on the molecule.

Then add sharps disposal, injection-site reactions, and site rotation.

Put it together and it is obvious why a needle-free version of anything is commercially irresistible — and why the products in §4.1's Hype Check exist. The demand they are answering is entirely real. The delivery claim is the part that has to be checked.

4.9 The delivery filter: what this chapter buys you

Here is the practical payoff. Before you evaluate any peptide product's evidence, ask five questions that Chapter 4 has equipped you to answer:

THE DELIVERY FILTER — five questions, asked before any evidence

  1. WHAT ROUTE, and does the molecule survive it?
     Oral or topical claim for an unmodified peptide? The burden is on the seller
     to explain what they solved. (4.1, 4.4)

  2. WHAT IS THE MEASURED BIOAVAILABILITY, in humans, by this route?
     Not "is it absorbed" — what fraction, measured how, in how many people? (4.3, 4.6)

  3. WHAT IS THE HALF-LIFE, and does the dosing schedule match it?
     A short-acting peptide dosed weekly is not being dosed. (4.3)

  4. WHAT MODIFICATION MAKES IT SURVIVABLE — and if none, why is that acceptable?
     Every successful peptide drug has an answer. (4.5)

  5. HOW IS IT STORED AND HANDLED?
     Peptides aggregate, adsorb to surfaces, and degrade. Cold chain is not
     optional decoration. (4.7, 4.8)

  A product that cannot answer these has not done the work — and that is a finding
  you can reach without a single clinical trial.

This filter is not a substitute for evidence. It is a prefilter, and its value is that it is cheap, fast, and settles a surprising number of cases. Chapter 2 established that mechanism is strong evidence against and weak evidence for. Delivery is the sharpest instance of that asymmetry: a molecule that cannot reach its target in useful quantity cannot work, regardless of what it does at the receptor.

🔍 Check Your Understanding — bioavailability

1. Oral semaglutide's bioavailability is roughly 1%, and it is an approved medicine that works. Why does a very low figure not by itself make a route a failure — and what has to be true for it to be workable anyway?
2. Two nasal products deliver, on average, the same small fraction of the dose. One does so consistently; the other ranges widely from dose to dose depending on what the person last ate. Which is the more serious problem, and why?
3. A product page states that its peptide is "highly absorbed." What exactly does that sentence fail to answer, and what would an adequate answer have to contain?

📋 Your Evidence Dossier

This chapter fills Field 4: Pharmacology.

FIELD 4 — PHARMACOLOGY
  Route(s)              how it is actually administered
  Bioavailability       by that route, in humans, if measured. If not measured, say so
  Half-life             with the route attached
  Dosing interval       and whether it matches the half-life
  Modification          what makes this molecule survivable as a drug — and if nothing,
                        what that implies
  Storage               refrigerated? light-sensitive? stable in solution for how long?
  Immunogenicity        reported? monitored? unknown?

The modification line is the one that separates a serious entry from a product description.

Worked demonstration — the anchor, both halves

FIELD 4 — NATIVE GLP-1                                   [worked demonstration]
  Route              Endogenous — released into portal circulation from intestinal L cells
  Bioavailability    Not applicable (not administered)
  Half-life          ~1–2 minutes
  Dosing interval    Not applicable — this is precisely why it cannot be a drug
  Modification       None. It is the natural molecule
  Storage            Not applicable
  Immunogenicity     Not applicable — it is a self-protein

FIELD 4 — SEMAGLUTIDE                                    [worked demonstration]
  Route              Subcutaneous injection (Ozempic, Wegovy); oral tablet (Rybelsus)
  Bioavailability    Subcutaneous: high. ORAL: roughly 1%, with a strict fasting protocol
  Half-life          ~1 week, subcutaneous
  Dosing interval    Once weekly (injection); daily (oral). Steady state after ~4–5 weeks
  Modification       THREE: (1) position 8 Ala→Aib blocks DPP-4 cleavage; (2) position 34
                     Lys→Arg leaves a single attachment site for manufacturing control;
                     (3) C18 diacid fatty acid at position 26 binds albumin, defeating
                     renal filtration and adding protease protection
  Storage            Refrigerated before first use; specified in-use period at room
                     temperature. Cold chain matters
  Immunogenicity     Anti-drug antibodies reported at low rates in trials; not generally
                     associated with loss of efficacy in the reported experience

Put the two entries side by side and the entire discipline of peptide pharmacology is visible in one page. Same receptor. Same activity. One is biologically perfect and pharmaceutically useless. The other is the same molecule, three atoms' worth of change and one fatty acid later, taken by millions of people once a week.

Your task

Complete Field 4 for every peptide in your dossier.

The modification line is where most entries will get interesting. For approved drugs, the answer is documented and often elegant. For research-code compounds, the answer is frequently "none — it is the native sequence or a fragment of one," which raises an immediate question: if the molecule is unmodified, what is its half-life, and does the dosing pattern people actually use make any sense against it?

And where bioavailability is unmeasured, write "not measured." Do not write "unknown" if the truth is that nobody has looked; those are different states and the distinction matters. A compound whose oral bioavailability has been measured and found to be 1% is in a completely different epistemic position from one where nobody has run the study.

Conclusion

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

Swallowed, a peptide meets five sequential barriers — stomach acid and pepsin, pancreatic proteases, brush-border peptidases, an epithelium it cannot cross, and a liver that clears what does get through — each independently sufficient. Injected, it faces two exits: proteolysis in circulation, and renal filtration for anything below roughly 5,000 daltons. Both exits must be closed to achieve a long half-life, which is why successful modifications come in combinations.

Semaglutide is the demonstration. A non-natural residue at position 8 abolishes the DPP-4 cut site; a substitution at position 34 leaves a single attachment point so manufacturing produces one defined product; a C18 fatty acid at position 26 binds albumin and defeats renal filtration. Half-life goes from about two minutes to about a week. The receptor pharmacology was never the problem — evolution had already optimized it. What was engineered was survival.

Every route trades something. Subcutaneous injection is the default because it bypasses the entire oral gauntlet and can be self-administered. Oral peptide delivery is possible — oral semaglutide achieves roughly 1% bioavailability with a permeation enhancer and a strict protocol — and that approximately-1% figure, plus the label's unusual instructions, is the clearest available illustration of how hard this is.

And the payoff is a filter. Route, bioavailability, half-life, modification, storage. Five questions you can ask about any peptide product, answerable without a clinical trial, that settle a surprising number of claims before evidence is consulted at all.

Which brings us to the evidence. Chapter 5 is the most important chapter in this book.

Key Terms

Pharmacokinetics — what the body does to a drug: absorption, distribution, metabolism, excretion.

Pharmacodynamics — what the drug does to the body: receptor binding and effect.

Bioavailability — the fraction of an administered dose reaching the general circulation intact. Intravenous is 100% by definition.

Half-life — the time for drug concentration to fall by half; the most useful single predictor of how a drug will behave.

Clearance — the volume of blood cleared of a drug per unit time.

First-pass metabolism — clearance by the liver of drug absorbed from the gut, before it reaches the general circulation.

Parenteral — any route bypassing the gastrointestinal tract; in practice, injection.

Subcutaneous — injected into the fat layer beneath the skin; the default route for peptide drugs.

Depot — a formulation releasing drug slowly over weeks to months after a single administration.

Steady state — the stable concentration reached after roughly four to five half-lives of regular dosing.

Titration — stepwise dose increase, usually to allow tolerance to side effects to develop.

Albumin binding — attachment of a drug to the most abundant blood protein, dramatically extending half-life by defeating renal filtration and shielding from proteases.

Lipidation — attaching a fatty acid to a peptide, typically to enable albumin binding.

Permeation enhancer — an excipient that transiently promotes absorption of a poorly absorbed drug across a membrane.

Immunogenicity — the tendency of a drug to provoke an immune response.

Anti-drug antibody — an antibody raised against a therapeutic protein or peptide, which may reduce efficacy, neutralize the drug, or cross-react with the endogenous molecule.

Cold chain — the refrigerated storage and transport required to keep a temperature-sensitive product stable from manufacture to use.

Aib (2-aminoisobutyric acid) — a non-natural amino acid used to block enzymatic cleavage; the position-8 modification in semaglutide.

Spaced Review

1. (Ch 1) A product claims to deliver an unmodified 30-residue peptide orally. Using §4.1, list the barriers it must overcome, and state which one cannot be addressed by protecting the molecule from enzymes.

2. (Ch 2) Semaglutide's long half-life defeats both elimination routes. Using Chapter 2's four termination mechanisms, explain what it does not defeat, and what consequence that has.

3. Native GLP-1 and semaglutide have essentially the same receptor pharmacology. Explain, in plain language, why one is useless as a drug and the other is taken by millions of people.

4. (Ch 3) Oral semaglutide is absorbed from the stomach rather than the intestine, in a small local region, at roughly 1% bioavailability. Using Chapter 3, explain why the route matters for a drug that acts on the incretin system — and whether bypassing the normal site of GLP-1 release should concern us.

5. Explain to a friend why a peptide product sold as a nasal spray might be legitimate and might be worthless, and what single piece of information would distinguish them.