Chapter 18 — Key Takeaways

TB-500, Thymosin Alpha-1, and Other Recovery and Immune Peptides

The one-sentence version

The "recovery and immune peptide" category groups molecules that have almost nothing in common, its favorite phrase — "immune modulation" — is unfalsifiable as usually stated, and the mechanisms these compounds are promoted for are, at the level of cell biology, indistinguishable from the mechanisms a tumor requires.

The molecules

Thymosin β4 is a 43-amino-acid endogenous human peptide. It binds G-actin and buffers actin polymerization, which is the cellular machinery of shape change and crawling — and therefore of wound closure. Real molecule, real biochemistry, real animal literature.

"TB-500" is a laboratory code for a marketed short synthetic fragment of it. They are not the same molecule. This is the single most common error about this compound. Vendors use the names interchangeably, and the effect — whether intended or inherited — is to attach one molecule's peer-reviewed literature to another molecule's product.

Thymosin alpha-1 is a different molecule entirely — 28 amino acids, derived from prothymosin alpha, immunomodulatory. The shared word "thymosin" is a historical accident of isolation, not a family resemblance that predicts anything.

LL-37 is a human cathelicidin, antimicrobial and separately immunomodulatory. Chapter 25 owns it; it appears here as a bridge because it demonstrates, in one molecule, the difference between a specifiable activity and an unspecifiable one.

KPV is a tripeptide fragment of α-MSH with reported anti-inflammatory activity in preclinical models. Larazotide is an octapeptide studied for celiac disease as a tight junction regulator.

The ratings

The two ❌ ratings are not the same failure. TB-500's means the claim has not been tested — a completed trial resolves it. "Immune modulation" means the claim cannot be tested as stated — only rewriting it helps.

The ⚠️ is a claim, not a hedge. It says: there is real human evidence here and it does not settle the question. Thymosin alpha-1 is this book's clearest case of a compound in the genuinely contestable middle.

The animal data

The thymosin β4 animal literature spans dermal wound healing, cardiac repair after injury, and corneal healing. It is real work and should be taken seriously. Chapter 17's discipline applies: never downgrade with distaste.

And it does not settle the human question. All five of Chapter 5 §5.3's translation problems apply: species differences, exposure scaling, model artificiality, endpoint mismatch, and publication filtering.

Publication filtering is the one people forget. Animal research is largely unregistered, so the published literature is a filtered sample of an unknown total. There is no denominator anywhere. This is why this book will never tell you how many animal studies exist for a compound — any number would imply a completeness nobody can verify.

A flawlessly executed animal study is still an animal study. Improving preclinical rigor does not convert it into clinical evidence. The categories do not convert.

Split regulatory status

Thymosin alpha-1 is approved in a number of countries — notably for hepatitis B, and as an immune adjuvant in some jurisdictions — and is not FDA-approved in the United States.

Both popular readings of that fact ("the FDA is behind" / "the FDA knows best") share a hidden premise: that one regulator has erred. Drop the premise and the split becomes informative.

Convergent approval implies an effect large and consistent enough that different decision rules converge. No approval anywhere usually means nobody assembled a dossier. A split is the characteristic external signature of evidence in the genuinely contestable middle — strong enough for some reasonable thresholds, weak enough to fail others.

An approval elsewhere is real and meaningful. It is also for a defined indication, does not transfer across borders, and does not report an effect size.

The gut peptides

Larazotide's non-absorption is a design specification, not a defect. Its target is at the apical surface of the intestinal epithelium, facing the lumen. Systemic absorption would add exposure with no therapeutic purpose.

The generalizable question is "where does it need to get to?" — asked before "can it be absorbed?" If the target is systemic, Chapter 4's constraints apply in full. If the target is the gut lumen, they do not.

This defense is not available to systemic claims. An oral peptide marketed for tendon repair or muscle growth cannot invoke larazotide logic. Those targets are not in the gut lumen.

The chapter's core idea: "immune modulation" is unfalsifiable

The claim, as usually stated, specifies no direction, no immune arm, no measurable endpoint, and no population. It is compatible with every possible result — and a claim compatible with every result cannot be supported by any of them.

Five things a real immune claim must specify:

1. Which arm, cell population, or mediator — not "the immune system"
2. Measured how — a named assay with a known reference range
3. In which direction — stated in advance; not "modulate," "balance," or "optimize"
4. In what population — healthy adults, chronic infection, post-transplant, and autoimmune have opposite therapeutic goals
5. Connected to what clinical outcome — because a shifted laboratory value is a surrogate

Missing 1–4: the claim cannot be tested at all. Missing 5: it can be tested, but tests something that may not matter.

"It doesn't suppress or stimulate — it balances" is the tell. That sentence explicitly declines to name a direction and presents the refusal as a superior property.

Same failure as "anti-inflammatory" (Ch. 10 §10.9): an umbrella term collapsing a heterogeneous, multi-directional system into one word with a positive valence, after which it can be attached to any finding.

The right response is not disagreement. It is a request for specification.

The symmetric risk

Compounds promoted for repair are promoted on angiogenesis, proliferation, and cell survival. Those are the three processes a tumor requires.

No causal link to human cancer is established for any compound in this chapter, and none is asserted. Writing "peptides may cause cancer" would be the same overclaiming this book criticizes, run in reverse.

And the concern is not resolved by that absence. It has not been shown and it has not been studied. The absence of evidence of harm, from a literature that never looked for harm, is silence.

"Promotes healing" and "promotes growth" are not separable claims at the mechanistic level. A signal saying proliferate, survive, build vessels carries no address label.

A mechanism cited as a benefit keeps all of its consequences. Rule 3 forbids upgrading a rating with mechanism, because mechanism is a story about what should happen. Mechanism can still generate legitimate safety questions — an efficacy claim asserts a benefit occurs and requires demonstration; a safety concern asserts a question is open and requires only that the question be well-formed.

Chapter 16 (IGF-1) and Chapter 14 (acromegaly) are the precedents. Neither establishes that these compounds carry the same risk. Both establish that the question is not paranoid.

Dossier skill: Field 5 when the literature is thin

"No evidence" collapses three completely different situations:

• (A) Studied and disappointing — someone looked and the answer was unfavorable. The most informative state.
• (B) Not studied — tells you about funding, incentives, and regulatory status; nothing about efficacy.
• (C) Studied in a different molecule — a parent, a different route, a different species. The most dangerous state, because the literature looks full.

TB-500 is primarily (C) with (B) underneath. Thymosin alpha-1 is in none of them — it is in the ordinary state of "studied, and the answer is contested."

Record two confidences, not one. How confident are you that you have characterized the evidence correctly? How confident are you about the compound? Merging them is how "I'm sure there's no good evidence" becomes "I'm sure it doesn't work."

Write the trigger sentence. If you cannot name the finding that would change your mind, your position is a preference, not a conclusion.

What a clinician adds here

Not a warning you could have read. Knowledge of your specific history, your surveillance status, your other treatments, and the ability to notice something you would not have thought to mention. For anyone with a current or prior malignancy, §18.7's concern changes in kind rather than degree, and it belongs in a conversation with an oncologist before rather than after.

Next: Chapter 19 asks why, if these compounds are so promising, nobody has run the trial. The answer is less about science than about who pays for evidence.