Chapter 18: TB-500, Thymosin Alpha-1, and Other Recovery and Immune Peptides

"The intensity of a conviction that a hypothesis is true has no bearing on whether it is true or not." — Peter Medawar, Advice to a Young Scientist (1979)

Overview

Chapter 17 gave you the template for an honest ❌. Take the animal literature seriously, describe the human situation accurately, explain how both can be true at once, and never sneer at the people using the compound. This chapter applies that template four more times, and then does something Chapter 17 did not have to do: it rates a claim form rather than a molecule.

The compounds here are sold together under a single umbrella word — recovery, repair, regeneration, immune support — and the umbrella is doing enormous work. Underneath it sit molecules with almost nothing in common. One is a genuine 43-residue human peptide with a substantial animal literature, whose fragment is what the market actually sells. One is a 28-residue peptide that is an approved medicine in a number of countries and not approved in the United States. One is a human antimicrobial peptide that belongs to a different chapter. Two are early-stage compounds proceeding through development the way compounds are supposed to, which makes them the least exciting and most encouraging things here.

What binds them is not chemistry. It is a marketing category, and the category's favorite phrase — "immune modulation" — is the vaguest claim in the entire field. Section 18.6 is the intellectual core of this chapter, and it is not really about peptides at all. It is about what makes a claim capable of being wrong. A claim specifying no direction, no immune arm, no measurable endpoint, and no population is compatible with every possible result, and a claim compatible with every result cannot be supported by any of them. You met this failure once already, in Chapter 10's treatment of "anti-inflammatory." Here it recurs in a purer form.

Then §18.7 asks the question the "recovery peptide" category avoids. If a compound works by promoting angiogenesis, proliferation, and cell survival, it is promoting the exact processes a tumor requires. That is not an accusation. No causal link to human cancer has been established for any compound here, and this book will not assert one. But "promotes healing" and "promotes growth" are not separable claims at the level of mechanism, and a reader who wants to think clearly about repair peptides has to hold both halves of that sentence at once.

In this chapter, you will learn to:

• Tell thymosin β4 apart from the fragment sold as TB-500, and explain why vendors' use of the names interchangeably is the single most consequential error about this compound
• Read the animal repair literature for what it is: real work, on the parent molecule, in models with known translation problems
• Interpret a split regulatory status — approved there, not approved here — as evidence about where a question sits rather than as a verdict from either direction
• Place LL-37 and the gut peptides correctly, and recognize when a chapter is handing a topic off
• Diagnose an unfalsifiable claim in one reading, and say precisely what would have to be added to make it testable
• State the symmetric risk of pro-healing signals accurately, without inflating a theoretical concern into a finding
• Document evidence when there is very little of it, which is a distinct skill from documenting evidence when there is a lot

Learning Paths

This chapter matters most to the Performance and Clinical paths, but §18.6 is required reading for everyone. It is the most portable idea in Part III.

💊 GLP-1 — §18.6 and §18.7 only. You will meet "immune modulation" language in longevity-clinic marketing that also sells GLP-1 agonists, and the ability to spot an unfalsifiable claim transfers completely. 🏋️ Performance — read in full. §18.1 is the section that will save you the most money and confusion, and §18.4 of Chapter 31 explains why "used in racehorses" is a question rather than a credential. 🔬 Science — §18.1, §18.2, and §18.7 are the mechanistically dense sections; §18.3's analysis of regulatory divergence is the closest this book comes to philosophy of evidence. 💄 Cosmetic — §18.2's corneal and dermal wound-healing material is the relevant background for Chapter 30's peptide-in-a-serum claims, and §18.7 is the reason "cell-proliferating" is a phrase to read carefully on a label. 🏥 Clinical — §18.3 and §18.6 are where your patients' questions will land. §18.7 is what you will be asked about after a cancer history comes up in the history-taking, and it deserves a prepared answer rather than an improvised one.

18.1 Thymosin β4 and the fragment sold as TB-500 are not the same molecule

Start with the sentence that this whole section exists to install, because it is the single most common error about this compound and it is made by sellers, by forum posts, by clinic websites, and by a startling number of otherwise careful articles:

Thymosin β4 is a 43-amino-acid human peptide. "TB-500" is marketed as a short synthetic fragment of it. They are not the same molecule.

Read that twice. Here is why it is not a pedantic distinction.

Thymosin β4 (Tβ4) is a real, well-characterized, endogenous peptide found in essentially every human cell type and in extracellular fluids. Its best-established biochemical job is binding G-actin — the free, unpolymerized form of actin, the protein that builds the cell's internal scaffolding. By sequestering G-actin, Tβ4 buffers the cell's control of actin polymerization, and polymerization is how cells change shape, crawl, extend processes, and close gaps. That is not a speculative mechanism. It is textbook cell biology, and it is why Tβ4 appears in any discussion of cell motility — and by extension of wound closure, where cells have to crawl.

At 43 residues, Tβ4 also sits right on the peptide/protein boundary Chapter 1 §1.5 described. You will see it called both. Nothing hinges on which.

"TB-500" is a laboratory code, not a generic drug name. What is sold under that code is described by sellers as a short synthetic fragment corresponding to the actin-binding region of Tβ4 — a handful of residues, not 43. The rationale offered is that the actin-binding region is "the active part," so a short peptide reproducing it should reproduce the activity while being far cheaper and easier to synthesize.

That rationale is not absurd. Fragment-based drug design is a real discipline, and there are genuine cases where a short stretch of a larger molecule carries most of its activity. But notice what the rationale is: it is an argument from mechanism, and Rule 3 of this book's rating system says you may never upgrade a rating with mechanism. The claim that a fragment reproduces the parent's activity is a hypothesis. It is testable. Whether it has been tested — for this fragment, for the endpoints people actually care about, in humans — is the question, and §18.2 gives the answer.

WHAT THE NAMES ACTUALLY REFER TO

  THYMOSIN β4 (Tβ4)                        "TB-500"
  ─────────────────────────────            ─────────────────────────────
  43 amino acids                           a short synthetic fragment
  endogenous — your cells make it          synthesized; not a natural product
                                             as sold
  binds G-actin; buffers actin             marketed as reproducing the
    polymerization                           actin-binding region
  substantial ANIMAL literature on         essentially no literature of its
    wound, cardiac, corneal repair           own on repair endpoints
  studied under its own name in            never received a generic drug
    the research literature                  name; known only by a code

  ┌──────────────────────────────────────────────────────────────────┐
  │  THE MOVE TO WATCH:  animal evidence generated with the LEFT      │
  │  column is routinely cited as evidence for the RIGHT column.      │
  │  Those are different molecules. The transfer is an assumption,    │
  │  not a finding, and it is almost never stated as an assumption.   │
  └──────────────────────────────────────────────────────────────────┘

That diagram is the whole section. The consumer market for this compound rests on an evidentiary transfer that nobody performing it acknowledges as a transfer. The animal work was done on the parent molecule. The product is a fragment. Between those two facts sits an unstated assumption — that the fragment does what the parent does, at whatever exposure a person achieves, for the endpoints a person cares about — and that assumption is exactly the kind of thing drug development exists to check.

A second transfer sits on top of the first. Even if the fragment reproduced the parent's actin-binding activity perfectly, actin-binding is a biochemical activity, not a clinical outcome. "Binds G-actin" and "heals your rotator cuff" are separated by several layers of biology, each of which can fail. Chapter 6 gave you the vocabulary: binding is a biochemical endpoint, tissue repair on imaging is a surrogate endpoint, and returning to sport without pain is a clinical endpoint. Evidence at one level does not carry to another.

🧬 The Molecule — why the naming confusion is not an accident

Chapter 1 §1.8 taught you to read names. Apply that here and something falls out immediately.

Thymosin β4 is a descriptive scientific name. It tells you the family (thymosins, originally isolated from thymus tissue), the class within the family (β), and the specific member (4). It is the name used in the research literature, which is where the animal data lives.

TB-500 is a laboratory code. Per §1.8, a compound still known only by a code, long after its introduction, has told you something about its regulatory history: it never entered the formal development pipeline far enough to be assigned a generic name by an international naming authority. That is strong evidence about regulatory history and no evidence at all about pharmacology. The compound might work. It has simply never been through the process that would tell us.

Now notice the rhetorical effect of using the two names interchangeably. Search the scientific name and you find a substantial peer-reviewed literature. Search the code and you find message boards. A seller who uses both names in the same paragraph — "TB-500 (thymosin beta-4)" is the standard formulation — has, in five characters of parenthesis, attached one molecule's literature to another molecule's product.

This is not necessarily deliberate deception. A great deal of it is people repeating what they read elsewhere. But the effect is the same regardless of intent, and the correction is available to any reader who knows to ask: is the study you are citing about the 43-residue peptide or about the fragment? In the overwhelming majority of cases the honest answer is the former, and in the overwhelming majority of cases nobody asks.

One further complication: what is actually in a given vial sold under this code is a separate question again. Chapter 34 covers analytical testing of gray-market products in detail; the short version is that identity, purity, and concentration are three independent things, none guaranteed. So the chain of assumptions runs: animal data on the parent → assumed to transfer to the fragment → assumed to be the fragment in the vial → assumed to be at the stated concentration → assumed to produce a clinical outcome. Five links. Each is a real question. The market treats all five as settled.

🔍 Check Your Understanding

1. A product page states: "TB-500 (Thymosin Beta-4) has been extensively researched for tissue repair." Identify the two distinct claims fused in that sentence and say which one the research actually supports.
2. Why does the fact that TB-500 has only a laboratory code count as evidence about regulatory history but not as evidence about whether it works?
3. Suppose a fragment did reproduce its parent molecule's biochemical activity exactly. Name two reasons that still would not establish a clinical effect.

18.2 The animal repair data: what it actually shows

Now be fair to the science, because the science is real.

The animal literature on thymosin β4 spans three areas worth naming individually. Dermal wound healing — models in which a wound is created and closure rate, re-epithelialization, and granulation tissue are measured. Cardiac repair after injury — infarction models following cardiac function, scar size, and vessel formation after treatment. Corneal healing — models of corneal injury, where transparency makes healing unusually easy to observe, and where some human work has been undertaken.

This is not fringe work. It has been conducted by academic laboratories, published in peer-reviewed journals, and pursued over a sustained period. The accompanying mechanistic story is coherent: a peptide that modulates actin dynamics would plausibly affect cell migration, and wound closure requires cell migration. Chapter 17 asked you to take the BPC-157 animal literature seriously, and the same discipline applies here. Dismissing this work because of the company its name keeps in the consumer market would be exactly the error Rule 4 forbids: you may not downgrade with distaste.

And human evidence is limited. For the parent molecule, there has been clinical investigation — corneal and dermal indications are the areas where it has gone furthest — and it has not produced an approved product for tissue repair in a major jurisdiction. For the marketed fragment specifically, the situation is thinner still. There are no completed randomized controlled human trials of TB-500 for tissue repair or athletic recovery in the published literature as of this writing.

Both of those paragraphs are true. Chapter 17's central skill was holding them together without letting either collapse into the other, and the reason it is hard is that each paragraph has a constituency that wants to hear only its own.

Why the animal data does not settle it

Chapter 5 §5.3 gave five reasons animal results fail to transfer. All five apply here, and it is worth walking them rather than gesturing at them.

Species differences in physiology. A rodent's healing is not a human's healing. Rodent skin is loose and heals substantially by contraction; human skin is tethered and heals substantially by re-epithelialization and granulation. A compound that accelerates the rodent process is not thereby shown to accelerate the human one, because they are partly different processes.

Dose and exposure scaling. The exposures used in animal models are chosen to demonstrate an effect if one exists, not to represent what a person would experience. Translating between them is a real pharmacological problem with real methods, and "it worked in mice" tells you nothing about whether the required exposure is achievable, tolerable, or safe in a human.

Model artificiality. An acute, sterile, standardized surgical wound in a young healthy animal is a very different object from a chronic tendinopathy in a forty-five-year-old with years of accumulated loading history, or a diabetic foot ulcer with impaired perfusion. The models are designed to be uniform; patients are not. Compounds that work in uniform injury frequently disappoint in messy injury, and that is one of the most reliable patterns in translational medicine.

Endpoint mismatch. Animal studies measure what is measurable in animals: closure rate, histology, biomarker levels, infarct size on a slide. Patients care about pain, function, and time to return to activity. Chapter 6 established that surrogate endpoints are hypotheses about clinical endpoints, not substitutes. An animal study measuring granulation tissue is measuring a surrogate for a surrogate.

Publication filtering. This is the one most people skip, and it is the most important. Animal research is largely unregistered. Human trials, in most jurisdictions and journals, must be registered before they begin, which means that a trial that produced a disappointing result exists in a public record whether or not anyone publishes it. Animal work carries no equivalent requirement. A laboratory that runs a repair model and sees nothing is under no obligation to record that the experiment happened.

The consequence is precise and worth stating carefully: the published animal literature on any compound is a filtered sample of an unknown total. Not a biased sample of a known total — you could correct for that. An unknown total. You cannot compute how much you are missing, because there is no denominator anywhere. This is why "there are many animal studies showing an effect" is a much weaker statement than it sounds, and it is also why this book will not tell you how many studies exist. Any number would imply a completeness that nobody can verify.

🔬 Read the Study

```text FIGURE 18.1 — "What a repair-model animal study can and cannot tell you" [Tier 3 — illustrative composite, not a specific published study]

THE STUDY A constructed composite representing the standard design used across the thymosin β4 repair literature. An injury is created under controlled conditions in a rodent model. Treated and control animals are compared on healing measures over a defined follow-up. Tissue is examined at the end. This composite is presented to teach the DESIGN, not to report a result. No number below is a finding from any real paper.

THE QUESTION Does administration of the peptide accelerate healing in this model, relative to vehicle control, on the model's chosen measures?

WHAT IT SHOWS Whether treated animals differ from controls on measures the experimenters selected — closure rate, histological appearance, vessel density, functional measures where the model supports them. In a well-run study with blinded assessment and adequate group sizes, a difference here is a real difference in that model.

WHAT IT DOESN'T • Whether the effect occurs in humans, whose repair biology differs • Whether an exposure producing this effect is achievable or safe in humans • Whether it works in aged, comorbid, chronically injured tissue rather than acute standardized injury • Whether faster closure means better function, less pain, or lower re-injury rate — none of which this design measures • Whether similar studies were run and not published (no registry exists) • Whether the FRAGMENT sold commercially does any of this — the parent molecule is what was studied

THE VERDICT Legitimate preclinical evidence supporting further investigation of the parent molecule. Not evidence of human efficacy, and not evidence about the marketed fragment at all.

THE LESSON The gap between "this experiment was well done" and "this compound works" is not a gap in study quality. A flawlessly executed animal study is still an animal study. Improving the rigor of preclinical work does not convert it into clinical evidence — only a trial does that. ```

That last line addresses a common confusion. People imagine that a sufficiently large, rigorous, replicated animal literature would eventually amount to human evidence. It would not. It would amount to an excellent case for running a human trial. The categories do not convert.

📊 Evidence Rating

Claim: TB-500 improves tissue repair or accelerates recovery from musculoskeletal injury in humans.

Rating: ❌ Hype outpaces evidence (as of this writing)

Reason: There are no completed randomized controlled human trials of this compound for these endpoints in the published literature; the marketed compound is a fragment rather than thymosin β4; and the supporting animal data concerns the parent molecule in models with the standard translation problems.

What would change it: A registered, completed, randomized, controlled human trial of the marketed fragment, in a defined injury population, reporting a clinical endpoint — pain, function, or time to return to activity — with prespecified analysis. A single such trial reporting a positive result would move this to ⚠️ immediately. Publication of the trial protocol and registration alone would be meaningful news.

Note what that ❌ says and does not say. It says the confident version of the claim is unsupported by human data. It does not say the compound does nothing. An ❌ is a statement about the state of evidence, not a verdict on the molecule. Some ❌ compounds in this book will be shown to work; some will be shown not to. The rating tells you which of those futures we are currently in a position to distinguish, which is neither.

18.3 Thymosin alpha-1: approved there, not approved here, and what that tells you

Now a genuinely different molecule, which shares a family name and almost nothing else.

Thymosin alpha-1 (Tα1) is a 28-amino-acid peptide derived from prothymosin alpha, with immunomodulatory activity. It is not a fragment of thymosin β4, it is not structurally related to it in any useful sense, and the shared word "thymosin" reflects a historical accident of isolation — both were originally identified in thymus-derived preparations — rather than a family resemblance that predicts anything. Treating them as siblings because of the name is a version of the same error §18.1 warned about, one taxonomic level up.

Here is the fact that makes this section worth a section:

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

Most writing about this compound uses that split in one of two ways, and both are wrong.

The promotional use: it's approved in dozens of countries, so it obviously works, and the FDA is just slow or captured. The dismissive use: the FDA hasn't approved it, so it obviously doesn't work, and the other regulators are lax.

Both readings share a hidden premise — that one regulator is right and the other is making a mistake. Discard that premise and the split becomes far more informative than either camp's version.

What regulatory divergence actually indicates

Regulators are not oracles. They are institutions applying a decision rule to a body of evidence, under conditions of uncertainty, with different thresholds, different statutory mandates, different standards of what constitutes an adequate endpoint, different populations in mind, and different disease burdens driving their priorities. A jurisdiction with a very high hepatitis B burden faces a different cost-benefit calculation than one where the disease is comparatively rare and existing therapies are widely available.

So consider what a split outcome implies about the underlying evidence.

WHAT REGULATORY AGREEMENT AND DISAGREEMENT IMPLY

  EVERY MAJOR REGULATOR APPROVES
    → the evidence cleared every threshold that was applied to it
    → the effect is large enough and consistent enough that
       reasonable decision rules converge
    → e.g. insulin, semaglutide for the indications each carries

  NO MAJOR REGULATOR APPROVES, AND NONE HAS BEEN ASKED
    → nobody has assembled a dossier
    → says nothing about efficacy; says a great deal about who
       is willing to fund a trial
    → e.g. most of the compounds in Part III

  SOME APPROVE AND SOME DO NOT  ←── thymosin alpha-1 is here
    → the evidence is strong enough that some reasonable decision
       rules clear it, and weak enough that others do not
    → this is the SIGNATURE of a genuinely contestable middle
    → it is not a scandal in either direction; it is what the
       middle looks like from the outside

That third row is the point of this section. A split decision is the characteristic external signature of evidence that sits in the genuinely contestable middle. If the effect were large, consistent, and measured on unambiguous endpoints, you would expect convergence. If there were no credible human evidence at all, you would expect no approvals anywhere. The split tells you that serious people looking at largely the same material reached different conclusions — and the honest inference is that the material genuinely supports different conclusions depending on where you set your threshold.

This is not a comfortable place to be. Readers want a verdict, and "the evidence is genuinely contestable" reads as evasion. It is not. It is an accurate description of a real state of affairs, and refusing to state it accurately in order to sound decisive is the failure mode this book exists to avoid.

The literature underlying Tα1 includes actual randomized human trials — in chronic hepatitis B, in sepsis, in immune support during cancer treatment, and in other settings where immune function is thought to be compromised. That is more human evidence than any other compound in this chapter, by a wide margin. The trials are also heterogeneous in design, endpoint, population, and quality, and the effect sizes reported are contested. Both halves matter, and reporting only one of them is how you get the two wrong readings above.

💊 In the Clinic — what "approved elsewhere" does and does not mean for a patient

This comes up constantly, and it deserves a careful answer rather than a reflex.

What it means. A regulator with statutory authority reviewed a dossier and concluded the benefit-risk balance favored approval for a defined indication in a defined population — often with access to unpublished data submitted as part of the application. That is a real institutional act. It is not nothing.

What it does not mean.

It is not an approval for the indication you have in mind. An approval for chronic hepatitis B is an approval for chronic hepatitis B. It is not an approval for general immune support, for recovery from training, or for prevention of infection in a healthy person. The indication is part of the approval, not a detail attached to it — Chapter 12 made the same point about semaglutide's brands.

It does not transfer across borders. A compound approved elsewhere and not approved in the United States is, in the United States, not an approved medicine. What that implies practically — how it could be obtained, and what happens to product quality outside an approved supply chain — is Chapter 35's territory, and the quality question is not minor.

It does not tell you the effect size. Approval means a regulator judged benefit to outweigh risk. It does not tell you whether the benefit was large or marginal, and that answer lives in the trials.

The useful move is not to argue about which regulator is right. It is to ask a clinician who knows your history: for my situation, is there an approved therapy with better-characterized evidence? That question has an answer, and it is usually more useful than the one being debated.

📊 Evidence Rating

Claim: Thymosin alpha-1 produces clinically meaningful immunomodulatory effects in defined patient populations with compromised immune function (for example, chronic hepatitis B).

Rating: ⚠️ Promising but preliminary (as of this writing)

Reason: This compound has genuine randomized human trial evidence and regulatory approval in multiple jurisdictions for specific indications — more human evidence than anything else in this chapter — and simultaneously lacks FDA approval, with heterogeneous trial designs and contested effect sizes across the literature.

What would change it: Toward ✅ — large, well-powered, consistently designed trials in a defined population converging on a clinically meaningful effect, and approval by regulators that have so far declined. Toward ❌ — well-powered trials in the indications where it is currently approved failing to reproduce the effect, or systematic review finding the positive signal concentrated in small or methodologically weak studies.

Defend that ⚠️ rather than treating it as a hedge. This book's ⚠️ is not a shrug or a diplomatic compromise between people who like a compound and people who do not. It is a positive claim with content: there is real human evidence here, and it does not settle the question. Thymosin alpha-1 is the clearest example in this book of a compound that genuinely occupies that position. Most Part III compounds are ❌ because there is nothing to weigh. This one is ⚠️ because there is something to weigh and the scale has not tipped.

If you take one transferable idea from this section, take this: a rating system that never produces a middle rating is not a rating system, it is a sorting hat. The middle has to be real and it has to be defended, otherwise every compound gets pushed toward the pole its advocates or detractors prefer, and the reader loses exactly the information that mattered.

18.4 LL-37: a bridge, not a destination

LL-37 is a human cathelicidin — an antimicrobial peptide produced by our own cells as part of innate immune defense. It is a genuinely fascinating molecule. It kills bacteria directly by disrupting membranes, and it has additional immunomodulatory properties: it interacts with immune cells, influences chemotaxis, and participates in the coordination of inflammatory responses in ways that go well beyond simple bacterial killing.

Chapter 25 owns antimicrobial peptides in full, and this section is a bridge. Stating that explicitly is not throat-clearing. One of the ways textbooks mislead is by covering a topic twice at different depths, so a reader who stops at the shallower treatment walks away thinking they have the whole picture. You do not have the whole picture on LL-37 from this section. What you should take from here is the reason it appears in a chapter about recovery and immune peptides at all.

LL-37 appears here because it is the cleanest natural example of a molecule with genuinely dual function. It is antimicrobial and immunomodulatory. Those are not the same activity, they are measured differently, and they can dissociate — it is entirely possible for a modification to preserve one and destroy the other. That duality makes it a good teaching case for §18.6, which is coming next.

When someone says LL-37 "supports the immune system," which of those two activities do they mean? Direct bacterial killing has a clear endpoint. Immunomodulation is not one property at all.

LL-37 — TWO ACTIVITIES, ONE NAME

  ANTIMICROBIAL ACTIVITY               IMMUNOMODULATORY ACTIVITY
  ────────────────────────             ─────────────────────────
  disrupts bacterial membranes         interacts with immune cell
                                         receptors
  measurable: growth inhibition        measurable: only if you first
    at a stated concentration            specify WHICH cell, WHICH
    against a stated organism            mediator, WHICH direction
  clear endpoint                       no single endpoint
  a claim here is falsifiable          a claim here is falsifiable ONLY
                                         once the specification above
                                         is supplied

  Same molecule. One activity yields testable claims easily.
  The other yields testable claims only with work that is
  usually not done.

That is not a criticism of LL-37 research, which does the specification work routinely. It is a criticism of how the molecule is described when it leaves the research literature and enters a marketing context, where the two activities are merged into "immune support" and the specification disappears.

One further wrinkle, which Chapter 25 develops properly: antimicrobial peptides that work by disrupting membranes use a mechanism that does not perfectly discriminate between bacterial membranes and yours. Selectivity is a matter of degree, driven by differences in membrane composition and charge, and that degree is a central design problem in the field rather than a solved one. A molecule that punches holes in membranes has a safety profile that must be established empirically rather than assumed from its naturalness — Chapter 1's ⚠️ Hype Check in a different costume.

No rating is issued here. LL-37's claims are rated in Chapter 25, where the evidence is presented in full. Rating a compound in a section that explicitly does not present its evidence would violate Rule 5: a rating must be falsifiable, and a rating issued without its evidence base is not something a reader can check.

18.5 KPV, larazotide, and the gut peptides

Two more compounds, both early, and one of them illustrates a principle so cleanly that it is worth the whole section on its own.

KPV

KPV is a tripeptide — three residues, lysine-proline-valine — corresponding to the C-terminal fragment of α-melanocyte-stimulating hormone (α-MSH). α-MSH is a peptide hormone with several established roles, including pigmentation, and it has reported anti-inflammatory activity. KPV is the short fragment of it that has attracted interest specifically for the anti-inflammatory part.

You should already be reaching for §18.1's warning, and you should be right to. KPV is a fragment of a parent molecule, and the same question applies: does the fragment reproduce the parent's activity? The difference from the TB-500 situation is that with KPV the fragment itself is what has been studied — the work is on KPV, under that name, rather than borrowed from work on α-MSH. That distinction is meaningful and works in KPV's favor. It does not change the fact that the literature is preclinical.

Reported anti-inflammatory activity in preclinical models — principally models of intestinal inflammation — is where KPV's evidence currently sits. Chapter 10 §10.9 told you what to do with the phrase "anti-inflammatory": ask which inflammatory process, in which tissue, measured how. In KPV's preclinical work that specification exists, because researchers must specify endpoints to publish. In KPV's consumer marketing it does not.

Larazotide, and the peptide that is designed not to be absorbed

Larazotide is an octapeptide — eight residues — studied for celiac disease. It has been through human clinical trials, which already distinguishes it from most of Part III. And it has a design feature so instructive that it is worth understanding even if you never encounter the compound again.

Larazotide is taken orally and is designed not to be absorbed.

Read that against Chapter 4, and against Chapter 1 §1.6's ⚠️ point about oral peptides. The default problem with an oral peptide is that your digestive system destroys it and whatever survives is too large to cross the intestinal wall. Every oral peptide drug is an engineering answer to that problem — except the ones for which the problem is not a problem, because the target is inside the gut lumen.

Larazotide is proposed to act as a tight junction regulator. Tight junctions are the protein complexes sealing the spaces between adjacent intestinal epithelial cells, and they are a major determinant of what can pass between the gut lumen and the bloodstream. The hypothesis under investigation in celiac disease is that regulating tight junction permeability could reduce the passage of gluten-derived peptides across the epithelium, thereby reducing the immune response they provoke in susceptible people. Whether the compound achieves clinically meaningful benefit on patient-relevant endpoints is what the trials have been testing, and this book does not report a verdict it cannot substantiate.

But note the design logic, which is beautiful regardless of the outcome. If your target is the apical surface of the intestinal epithelium — the side facing the lumen — then systemic absorption is not merely unnecessary, it is undesirable. Absorption would produce systemic exposure with no therapeutic purpose and whatever risk that exposure carries. A peptide that stays in the gut is a peptide with a restricted safety surface.

THE ORAL PEPTIDE PROBLEM — AND THE CASE WHERE IT ISN'T ONE

  TARGET IS SYSTEMIC                     TARGET IS IN THE GUT LUMEN
  (most peptide drugs)                   (larazotide; also linaclotide)
  ────────────────────                   ──────────────────────────────
  must survive digestion   ←problem→     must survive digestion
                                            (still a real problem)
  must cross the gut wall  ←problem→     must NOT cross the gut wall
                                            (a design GOAL)
  systemic exposure is                   systemic exposure is
    the point                              an unwanted side effect
  bioavailability is the                 bioavailability near zero is
    key metric to maximize                 the specification being met

  The lesson generalizes: "poorly absorbed" is a criticism only if
  absorption was required. Ask what the target is before you ask
  whether the delivery is adequate.

That generalization is worth carrying. When you evaluate any oral peptide claim, the first question is not "can it be absorbed?" It is "where does it need to get to?" If the answer is the bloodstream, absorption is the whole ballgame and Chapter 4's constraints apply in full force. If the answer is the gut lumen, absorption is beside the point.

Do not let this become a universal escape hatch, though, because it is used as one. A seller of an oral peptide marketed for tendon repair, muscle growth, or systemic recovery cannot invoke larazotide logic. Those targets are not in the gut lumen. The "it acts locally in the gut" defense is available only to compounds whose claimed effect is local to the gut, and the moment the claimed effect is systemic, the absorption problem returns undiminished.

📊 Evidence Rating

Claim: KPV and larazotide produce clinically meaningful benefit in the inflammatory and intestinal-barrier conditions for which they are being investigated.

Rating: 🔬 Frontier (as of this writing)

Reason: Both are at an early stage of investigation, proceeding through the ordinary sequence — preclinical characterization for KPV, human clinical trials for larazotide — with results that do not yet support a confident statement in either direction. It is too soon to rate, and the correct response to "too soon" is to say so rather than to guess.

What would change it: Completed, adequately powered, randomized trials reporting patient-relevant endpoints in a defined population, with results consistent enough to support a conclusion. For larazotide specifically, the trial infrastructure already exists, so this is a question of results rather than of whether anyone will look.

🔬 is the least exciting rating in the system and, for a reader learning to think about this field, one of the most important. It means this is what a compound looks like when it is being developed properly. No one is selling larazotide as a recovery peptide. It is being tested in patients with a defined disease, on endpoints that matter to those patients, by people who will publish whichever way it comes out. Compare that with the compounds in §18.1 and the contrast is the entire lesson of Part III.

18.6 Why "immune modulation" is the vaguest claim in the field

This is the core of the chapter, and it is not really about peptides.

Here is a claim of the type you will encounter constantly:

"This peptide modulates and balances the immune system, optimizing immune function and helping your body regulate its own defenses."

Ask yourself what result would show that claim to be false.

Take your time. The answer is that no result would, and the reason is that the claim, as stated, specifies nothing that could come out one way rather than another.

The four missing specifications

No direction. "Modulate" is directionless by construction. If a measurement goes up, that is modulation. If it goes down, that is also modulation. If some measurements go up and others go down, that is modulation with extra credit — you can call it "balancing." The word is compatible with every possible outcome, which is exactly what makes it useless. A claim that cannot fail is not a strong claim. It is not a claim at all.

No arm. "The immune system" is not one thing. It is at minimum two large systems — innate and adaptive — comprising many cell populations that can be up- or down-regulated independently and often in opposite directions at once. Neutrophils, macrophages, dendritic cells, natural killer cells, several T cell subsets with genuinely opposed functions, B cells, and a large cast of cytokines and chemokines that are themselves regulated separately. A compound can raise one and suppress another simultaneously. Which one does the claim mean?

No endpoint. Nothing in the claim can be measured. "Optimized immune function" has no assay. There is no laboratory in the world that can return a result on it, because there is no defined quantity being asserted. Compare with a claim that does specify: "reduces the rate of laboratory-confirmed influenza in adults over 65 during a single season." You know exactly what to measure, and you know in advance what result would falsify it.

No population. Immune function in a healthy twenty-five-year-old, in a person with a chronic viral infection, in someone on immunosuppressive therapy after transplant, and in someone with an autoimmune condition are four different situations in which the same intervention might be beneficial, harmful, or irrelevant — and might be beneficial in one and actively harmful in another. Suppressing an overactive response is a therapeutic goal in autoimmunity and a catastrophe in someone fighting an infection. A claim that does not name the population has not said which of those it means.

FOUR THINGS AN IMMUNE CLAIM MUST SPECIFY — AND ONE MORE

  1. WHICH ARM / CELL POPULATION / MEDIATOR
       "NK cell cytotoxicity" — not "the immune system"

  2. MEASURED HOW
       a named assay with a known reference range and known
       reproducibility — not "improved markers"

  3. IN WHICH DIRECTION
       increase or decrease, stated in advance — not "modulate,"
       "balance," "optimize," "support," or "regulate"

  4. IN WHAT POPULATION
       healthy adults / chronic hepatitis B / post-transplant /
       autoimmune — these have OPPOSITE therapeutic goals

  5. CONNECTED TO WHAT CLINICAL OUTCOME
       fewer infections, lower viral load, less hospitalization,
       longer survival — because a shifted laboratory value is a
       SURROGATE (Ch. 6), and surrogates have failed before

  A claim missing #1–4 cannot be tested at all.
  A claim with #1–4 but missing #5 can be tested, but tests
  something that may not matter to a patient.

That fifth line is the one people forget even when they get the first four right. It is entirely possible to run a rigorous study showing that a compound raises a specific cell count in a specific population by a specific amount, and for that finding to be of no clinical consequence whatsoever. Chapter 6 gave you the general form of this problem and the graveyard of surrogate endpoints that motivated it. Immune markers are a particularly treacherous case, because so many of them shift in response to almost anything — sleep, exercise, stress, a recent meal, the time of day the blood was drawn — and because the relationship between most of them and any outcome a patient cares about is loose at best.

The parallel with "anti-inflammatory"

Chapter 10 §10.9 did this analysis for "anti-inflammatory," and the failure is structurally identical. Inflammation is not one process. It is a coordinated set of responses with many mediators, several phases, and both protective and destructive aspects — it is how you contain an infection and begin repair, and also how you damage a joint over decades. A compound that reduces one mediator in one tissue is not thereby "anti-inflammatory" in any general sense, and calling it that smuggles in a suggestion of broad benefit the underlying finding does not support.

Same shape here. "Immune modulation" and "anti-inflammatory" are umbrella terms that collapse a heterogeneous, multi-directional biological system into a single word with a positive valence. Once collapsed, the word attaches to any finding, because the system is large enough that a compound will affect something in it, and whatever it affects can be described in the umbrella's terms.

Name the pattern and you can spot it in unfamiliar territory. Watch for any claim that a compound "supports," "balances," "optimizes," or "regulates" a large biological system. The system is named; the specification is not; the reader supplies the specification from their own hopes.

⚠️ Hype Check — "supports and modulates immune function"

The claim, in its usual form:

"This peptide is a powerful immunomodulator. It doesn't suppress or stimulate — it balances your immune system, bringing it back to optimal function so your body can defend itself the way it was designed to."

What's true in it. Immunomodulation is a real pharmacological category. There are approved immunomodulatory drugs, they do real things, and some of them are peptides. The immune system is genuinely capable of both excessive and insufficient responses, so the intuition behind "balance" is pointing at something real. Thymosin alpha-1, in this very chapter, is a compound with real trials in immune-relevant indications. This is not a category error.

Where it fails. In the specification, entirely.

Notice the structure of "it doesn't suppress or stimulate — it balances." That sentence is doing something remarkable: it is explicitly declining to name a direction, and presenting the refusal as a superior property. The claim has been built so that no measurement can contradict it, and the lack of a testable direction is being marketed as sophistication.

Notice also what "back to optimal" assumes: that there is a known optimum, that it is the same for everyone, that we can measure distance from it, and that the compound moves you toward it rather than past it. Four assumptions, none stated, none measurable as written.

And notice what the claim is compatible with. If a user gets fewer colds, the compound worked. If a user gets more, it is "the immune system recalibrating." If nothing happens, "it works at the cellular level." Every outcome confirms it. That is the definition of an unfalsifiable claim.

Verdict: the problem is not that the claim is exaggerated. Exaggerated claims can be corrected by better data. This claim cannot be corrected by any data, because it does not assert anything a datum could contradict. The correct response is not disagreement. It is a request for specification: which cells, measured how, moving which way, in whom, connected to what outcome. Ask those five questions and the conversation either becomes productive or ends.

📊 Evidence Rating

Claim: "Immune modulation" — as a general claim, for any compound in this chapter, in the form in which it is typically stated.

Rating: ❌ Hype outpaces evidence (as of this writing)

Reason: The claim form is unfalsifiable as usually stated. It specifies no direction, no immune arm, no measurable endpoint, and no population, and is therefore compatible with every possible result — which means no result can support it either. This ❌ attaches to the form of the claim, not to any molecule.

What would change it: Nothing could change this rating, because the claim as stated is not the kind of thing evidence bears on. What can change is the claim: specify the arm, the assay, the direction, the population, and the clinical outcome, and you have a different claim that this book would rate on its own evidence — as it does for thymosin alpha-1 in §18.3, which is a specified version of the same underlying idea and earns ⚠️.

This is a deliberate extension of the rating format, and it deserves a note. Rule 1 says a rating attaches to a claim, with a population and an endpoint. Here we are rating a claim that has neither, and the rating is precisely a statement that it has neither. Chapter 9 did the same thing for a methodological inference rather than a compound. The system is a tool for evaluating claims, and sometimes the most useful thing it can tell you about a claim is that it is not built to be evaluated.

Contrast the chapter's two superficially similar ratings. "Immune modulation" gets ❌ because it cannot be tested. TB-500 for tissue repair gets ❌ because it has not been tested. Different failures, different remedies: rewrite the claim, or run the trial.

18.7 Cancer and angiogenesis: the symmetric risk of pro-healing signals

Now the question the recovery-peptide category does not ask about itself.

Every compound in this chapter that is promoted for tissue repair is promoted on the basis of some combination of three mechanisms: angiogenesis — the formation of new blood vessels; cell proliferation — making more cells; and cell survival — keeping cells that would otherwise die. Those three appear in essentially every pro-healing mechanistic story, and for good reason. They are genuinely what repair requires. Damaged tissue needs blood supply, needs new cells, and needs those cells to survive long enough to do their work.

Here is the observation that follows, and it is not subtle once stated:

Angiogenesis, proliferation, and survival are also the three processes a tumor requires.

A tumor beyond a very small size cannot grow without recruiting a blood supply. Uncontrolled proliferation is the defining feature of malignancy. Evading programmed cell death is among the best-characterized features of cancer biology. A compound whose promotional story is "it promotes angiogenesis, proliferation, and cell survival" has described, in the same words, the requirements of both wound healing and tumor growth. This is not a rhetorical trick. It is the same biology.

State this correctly or not at all

The temptation here runs in two directions, and both must be resisted.

Do not assert a link. No causal link to human cancer has been established for any compound in this chapter. This book will not claim one, imply one, or use suggestive phrasing to let a reader infer one the evidence does not support. Writing "peptides may cause cancer" would be exactly the overclaiming this book criticizes, run in the opposite direction, and Rule 4 forbids downgrading with distaste as firmly as Rule 3 forbids upgrading with mechanism. A theoretical concern is not a finding.

Do not dismiss it either. The concern is not manufactured, and it is not resolved by pointing out that no link has been shown. It has not been shown and it has not been studied — and Chapter 5's whole point about the difference between those two is the point here. Absence of evidence of harm, from a literature that has not looked for harm, is not evidence of absence of harm. It is silence.

What is actually the case: the theoretical concern is real, it follows directly from the mechanisms these compounds are promoted for, it is unstudied in humans for the compounds in this chapter, and a person deciding what to do should know all four of those things.

🩺 Safety and Risk — pro-healing and pro-growth are not separable claims

A reader deciding about any compound in this chapter deserves this laid out plainly.

The mechanistic point. At the level of cell biology, "promotes healing" and "promotes growth" are not two claims. They are one claim with two framings, and which framing applies depends entirely on which cells are receiving the signal. A signal that says proliferate, survive, build vessels does not carry an address label distinguishing injured tendon from an occult lesion somewhere else in the body.

What is not known. Whether any compound in this chapter, at exposures people actually use, meaningfully affects the growth of an existing malignancy in a human. This has not been studied. Not "studied and found reassuring" — not studied. The distinction is Chapter 5's central lesson and §18.8's Dossier exercise is built on it.

Why the precedent chapters matter. Chapter 16's treatment of IGF-1 established that a signal promoting growth promotes growth in tissue you did not intend, and that the epidemiology connecting growth signaling to cancer risk in humans is real, complicated, and not reducible to a slogan. Chapter 14's material on acromegaly showed what sustained excess growth signaling does to a human body over years — the closest thing we have to a natural experiment in chronic exposure. Neither chapter establishes that the compounds here carry the same risk. Both establish that the question is not paranoid.

The asymmetry that matters clinically. For a person with no cancer history, this is a theoretical concern of unknown magnitude, weighed against a benefit that — for the compounds in §18.1 — has no human evidence supporting it. For a person with a current or prior malignancy, the calculation differs in kind rather than degree, and this is the single most important thing to raise with an oncologist before rather than after. Chapter 39 is about making that conversation useful.

What a clinician adds here is not a warning you could have read. It is knowledge of your specific history, your specific surveillance status, and your specific other treatments, plus the ability to notice something you would not think to mention. A book cannot do any of that.

There is a broader lesson buried here, and it generalizes past peptides entirely.

A mechanism cited as a benefit is a mechanism. It keeps all of its consequences. When a promotional claim says a compound works by promoting angiogenesis, it has not stopped at the flattering half. It has asserted a biological activity, and biological activities do not restrict themselves to the applications a marketer had in mind. You are entitled — obligated, really — to ask what else that activity does.

This is why Rule 3 takes the form it does. Mechanism cannot upgrade a rating, because mechanism is a story about what should happen, and stories are cheap. But mechanism can generate hypotheses about risk, and those deserve investigation rather than dismissal. The asymmetry is not a double standard. It follows from what each kind of claim requires: an efficacy claim asserts that a specific benefit occurs, which requires demonstration; a safety concern asserts that a question is open, which requires only that the question be well-formed.

🔍 Check Your Understanding

1. A product describes itself as "stimulating angiogenesis and cellular proliferation to accelerate tissue repair." Restate the mechanistic content of that sentence in a way that makes the symmetric concern visible, without asserting any link to disease.
2. Why is "no link to cancer has been demonstrated" a weaker reassurance for these compounds than it would be for an approved drug that has been on the market for twenty years?
3. Explain why Rule 3 (no upgrading with mechanism) does not imply that mechanistic concerns about safety should also be ignored.

18.8 Ratings across the group

Collect the chapter.

Four observations about that table, each of which is a transferable skill rather than a fact about these compounds.

First: same rating, different reasons. TB-500 and "immune modulation" both carry ❌, and the similarity ends at the symbol. One has not been tested; the other cannot be. A trial fixes the first; nothing fixes the second except rewriting the claim. If you find yourself treating two ❌ ratings as equivalent, go read the one-sentence reason, which is the field carrying the actual information.

Second: the parent/fragment split changes what a rating can attach to. The thymosin β4 row has no rating, and that is not an oversight. The animal literature supports statements about animals, and this book's ratings attach to claims about humans, with a population and an endpoint. There is no honest way to compress "real animal work on a molecule that is not the one being sold" into a symbol.

Third: ⚠️ is the hardest rating to hold. Thymosin alpha-1 will be quoted by advocates as "acknowledged to have real evidence" and by skeptics as "not FDA-approved even in a peptide book." Both quotations are accurate and both are incomplete, and that is what a genuinely contestable middle feels like from the inside.

Fourth: 🔬 is not a weak ❌. KPV and larazotide are not "❌ with better manners." 🔬 means the question is open and is being addressed through the ordinary process — the rating a compound should have early in its life. So few Part III compounds carry it not because they are worse molecules, but because a compound being sold instead of studied never passes through the frontier stage at all.

The chapter's shape, restated

Notice what actually distinguished these compounds. Not molecular class — all peptides. Not mechanism — all plausible. Not how confidently they are marketed, which is roughly uniform and roughly maximal. What distinguished them was what has been done to find out, and that is the only axis that moved.

Chapter 17 made this point with a single compound. This chapter makes it with five, and the spread is the argument: one category word covers a compound with approvals in multiple countries and a compound that has never been in a registered human trial. If the category told you anything, that spread would be impossible.

📋 Your Evidence Dossier

Field 5 — the human evidence — when the literature is thin or hard to reach.

Previous chapters had you fill Field 5 for compounds with something to fill it with. This chapter teaches the harder version: what to write when there is almost nothing, and how to make "almost nothing" informative rather than blank.

The instinct is to write "no evidence" and move on. Resist it, because "no evidence" collapses three completely different situations into one phrase, and the differences between them are exactly what should drive your thinking.

The three-way distinction

FIELD 5 — WHEN THE LITERATURE IS THIN
  Three states that "no evidence" wrongly merges:

  (A) STUDIED AND DISAPPOINTING
      Trials were run. Results were negative, null, or too small
      to matter. The question was asked and answered unfavorably.
      → This is the most informative state. Someone looked.
      → Future trials are less likely, because someone already looked.

  (B) NOT STUDIED
      No trials exist. Nobody has asked. The question is open in
      the fullest sense.
      → Tells you about FUNDING, INCENTIVES, and REGULATORY STATUS,
        and nothing whatsoever about efficacy.
      → A compound sold without patent protection has no one with
        an economic reason to spend on a trial (Ch. 19).

  (C) STUDIED IN A DIFFERENT MOLECULE
      Evidence exists, and it is about something else — a parent
      molecule, a different salt, a different route, a different
      species, a different indication.
      → The MOST DANGEROUS state, because the literature LOOKS full.
      → A search returns results. The results are about a neighbor.

  Write which one you are in. Then write how you know.

State (C) is the one this chapter exists to teach, and it is the state almost nobody notices, because it does not feel like an absence. You search, you find papers, the papers are real, and every instinct says the box is filled. The discipline is to check, for each paper, what molecule it studied and in what species — and to notice when the answer is consistently different from the thing you are holding.

Demonstration: two entries that differ sharply on exactly this axis

FIELD 5 — TB-500                                  [worked demonstration]

  State of the literature      (C) primarily, with (B) underneath

  Human RCTs of THIS compound  None found. Not "few" — none identified
                                 in the published literature for tissue
                                 repair or recovery endpoints.
  Registered trials            None identified. Note: registration is
                                 the check that catches unpublished
                                 negative results. Its absence means
                                 there is no unpublished record either.
  What the literature IS about Thymosin β4 — the 43-residue parent
                                 molecule — in animal models of dermal,
                                 cardiac, and corneal repair. Real work,
                                 different molecule.
  Human work on the parent     Limited; corneal and dermal indications
                                 investigated; no approved product for
                                 tissue repair in a major jurisdiction.
  Non-human uses cited         Veterinary contexts, notably racehorses.
                                 Ch. 31: a use is not a credential.
  What I would need to see     One registered, completed RCT of the
                                 marketed fragment, clinical endpoint,
                                 defined injury population.
  Confidence in this entry     HIGH that the evidence is absent.
                                 LOW-to-none about whether the compound
                                 works. Those are different confidences
                                 and I am recording both.

FIELD 5 — THYMOSIN ALPHA-1                        [worked demonstration]

  State of the literature      Neither (A), (B), nor (C) — this is
                                 the ordinary state of "studied, and
                                 the answer is contested"
  Human RCTs of THIS compound  Yes. Randomized human trials exist, in
                                 chronic hepatitis B, in sepsis, and in
                                 immune support during cancer treatment.
                                 THIS COMPOUND, not a relative.
  Regulatory status            Approved in a number of countries for
                                 defined indications; NOT FDA-approved.
  Quality of the evidence      Heterogeneous — designs, populations,
                                 and endpoints vary across the literature;
                                 effect sizes are contested. Note this
                                 explicitly: "trials exist" and "trials
                                 agree" are separate findings.
  What I would need to see     Toward ✅: large consistent trials in a
                                 defined population converging on a
                                 clinically meaningful effect.
                                 Toward ❌: well-powered trials in the
                                 approved indications failing to
                                 reproduce it.
  Confidence in this entry     MODERATE about the compound. HIGH about
                                 the shape of the uncertainty — I know
                                 what kind of question is open.

Notice what those entries have in common: neither says "it doesn't work." One says the question has not been asked about this molecule; the other says it has been asked repeatedly and the answers disagree. Those are opposite epistemic situations that a careless dossier renders identically as "insufficient evidence."

Your turn

For each compound in your dossier where Field 5 reads thin, do three things.

One: classify. Which state — (A), (B), or (C)? If (C), name in writing the molecule the literature is actually about, next to the molecule you are holding.

Two: separate your confidences. Two lines, not one. How confident are you that you have characterized the evidence correctly? How confident are you about the compound itself? These come apart constantly, and merging them is how "I'm sure there's no good evidence" becomes "I'm sure it doesn't work" without anyone noticing the step.

Three: write the trigger. One sentence naming the finding that would change your assessment. If you cannot write that sentence, your position is not a conclusion about evidence — it is a preference, and §18.6 just explained what happens to claims no result could contradict. The same standard applies to yours.

Conclusion

Five compounds, one marketing category, and almost nothing else in common.

Thymosin β4 is a 43-residue human peptide with a real animal repair literature. "TB-500" is a marketed fragment of it. They are not the same molecule, the evidence generated on one is routinely cited for the other, and that transfer is an unstated assumption sitting under an entire consumer market. The claim that TB-500 improves human tissue repair rates ❌ — no completed randomized human trials, a fragment rather than the parent, and animal data carrying all five of Chapter 5's translation problems.

Thymosin alpha-1 is a different molecule with a different story. Real randomized human trials, approval in a number of countries, no FDA approval, contested effect sizes. That split is not a scandal in either direction — it is what the genuinely contestable middle looks like from outside, and it earns a ⚠️ that should be defended rather than apologized for.

LL-37 is Chapter 25's, and appears here only because it demonstrates, in one molecule, the difference between a specifiable activity and an unspecifiable one. KPV and larazotide are 🔬, and larazotide carries the most elegant lesson here: a peptide designed not to be absorbed, because its target is in the gut lumen, which turns Chapter 4's central problem into a design specification.

Two ideas are worth carrying out of this chapter into everything else you read.

"Immune modulation," as usually stated, is not a weak claim. It is not a claim. No direction, no arm, no endpoint, no population — compatible with every result, and therefore supportable by none. It gets ❌ as a claim form, and the remedy is not more evidence but a rewritten claim. Once you can see the shape you will find it in places this book never mentions.

And pro-healing is pro-growth. Angiogenesis, proliferation, and survival are what repair requires and what a tumor requires, and a promotional story built on those three mechanisms has described both at once. No causal link to human cancer is established for anything here. It also has not been studied, and those are different sentences. A mechanism cited as a benefit keeps all of its consequences.

Chapter 19 asks the question that hangs over this entire part: if these compounds are so promising, why has nobody run the trial? The answer is less about science than about who pays for evidence, and it explains more about the shape of this field than any amount of molecular biology.

Key Terms

Thymosin β4 (Tβ4) — a 43-amino-acid endogenous human peptide that binds G-actin and participates in regulating actin polymerization, cell motility, and tissue repair. Subject of a substantial animal repair literature.

TB-500 — a laboratory code for a compound marketed as a short synthetic fragment of thymosin β4, commonly described as corresponding to the actin-binding region. Not the same molecule as thymosin β4. Never assigned a generic drug name.

G-actin — the free, unpolymerized ("globular") form of actin. Thymosin β4's best-characterized biochemical activity is sequestering it, buffering the pool available for polymerization.

Fragment — a short peptide corresponding to part of a larger parent molecule. Whether a fragment reproduces the parent's activity is a hypothesis requiring test, not an inference from structure.

Thymosin alpha-1 (Tα1) — a 28-amino-acid peptide derived from prothymosin alpha, with immunomodulatory activity. Approved in a number of countries for defined indications including hepatitis B; not FDA-approved in the United States. Unrelated to thymosin β4 despite the shared name.

Prothymosin alpha — the larger precursor protein from which thymosin alpha-1 derives.

Immunomodulator — a compound that alters immune function. A legitimate pharmacological category when the alteration is specified by arm, direction, magnitude, and population; an empty term when not.

Cathelicidin — a family of host-defense peptides produced as part of innate immunity. LL-37 is the principal human member.

LL-37 — the human cathelicidin antimicrobial peptide, with direct membrane-disrupting antibacterial activity and separate immunomodulatory properties. Covered in full in Chapter 25.

KPV — the tripeptide lysine-proline-valine, the C-terminal fragment of α-MSH, with reported anti-inflammatory activity in preclinical models.

α-MSH (alpha-melanocyte-stimulating hormone) — a peptide hormone with roles including pigmentation and reported anti-inflammatory activity; KPV is a fragment of it.

Larazotide — an octapeptide studied for celiac disease as a tight junction regulator. Taken orally and deliberately designed not to be absorbed, because its target is within the gut lumen.

Tight junction — the protein complex sealing the space between adjacent epithelial cells; a major determinant of what passes between the gut lumen and the bloodstream.

Angiogenesis — the formation of new blood vessels. Required for tissue repair and also for tumor growth beyond a minimal size, which is the source of §18.7's symmetric concern.

Unfalsifiable claim — a claim compatible with every possible observation, and therefore incapable of being supported by any of them. "Immune modulation," as usually stated, is this field's canonical example.

Regulatory divergence — regulators in different jurisdictions reaching different conclusions on largely the same evidence. Characteristically indicates evidence in the genuinely contestable middle rather than error by either party.

Publication filtering — the process by which the published literature becomes an unrepresentative sample of the work performed. Especially severe in animal research, which is largely unregistered, leaving no denominator against which to assess completeness.

Surrogate endpoint — a measurement standing in for an outcome a patient cares about. Immune marker shifts are surrogates; Chapter 6 explains why they mislead.

Spaced Review

1. A product page cites six published papers as evidence for a recovery peptide. All six are real, peer-reviewed, and about the compound's parent molecule in rodents. Using §18.1 and the Dossier's three-way distinction, name which state of the literature you are in, and explain why this state is more misleading than finding no papers at all.

2. (Ch. 17, Ch. 18) BPC-157 and TB-500 both rate ❌ for human repair claims, and the reasons are not identical. Name the feature of the TB-500 situation that has no counterpart in the BPC-157 situation, and say why it makes the evidentiary gap wider rather than merely equal.

3. (Ch. 5, Ch. 18) Chapter 5 §5.3 lists five reasons animal results fail to transfer. One of them — publication filtering — is the reason this book refuses to state how many animal studies exist for any compound. Explain the reasoning, and say what would have to change about animal research for a study count to become meaningful.

4. (Ch. 6, Ch. 18) A trial reports that a peptide significantly increased a circulating immune cell population in healthy adults over eight weeks. Using §18.6's five specifications and Chapter 6's account of surrogate endpoints, state which specifications this claim satisfies, which it does not, and what the trial would need to add before the result could support a claim about health.

5. Someone with a prior cancer history asks whether a repair peptide is safe for them. Write the answer you would actually give in three or four sentences: it must state the mechanistic concern accurately, must not assert a link that has not been established, must not offer false reassurance, and must end somewhere more useful than "ask your doctor."