Zonulin as pathway, biomarker and measurement

Zonulin occupies an unusual position in intestinal-permeability research.

It is simultaneously the name given to a biological pathway involved in reversible regulation of intercellular tight junctions, a molecular entity identified as pre-haptoglobin 2, a term increasingly applied to a broader family of structurally and functionally related proteins, and the label attached to commercial assays used in hundreds of human studies.^¹–⁴

Those meanings are related.

They are not interchangeable.

This distinction matters because a substantial part of the human literature has moved rapidly from measuring something called zonulin to interpreting that measurement as evidence of increased intestinal permeability.

The inferential distance between those two statements can be considerable.

To interpret zonulin research responsibly, at least five levels need to remain separate:

the pathway, the molecule, the protein family, the assay and the biological conclusion drawn from the measurement.

The pathway came first

The biological story of zonulin emerged from work led by Alessio Fasano and colleagues on the regulation of intestinal tight junctions.

Zonulin was described as an endogenous modulator capable of reversibly altering tight-junction organisation and increasing paracellular permeability.^¹,⁵,⁶ Subsequent mechanistic work identified signalling involving proteinase-activated receptor 2 and epidermal growth factor receptor, linked to cytoskeletal and junctional reorganisation.^²

This work established an important biological principle: intestinal permeability is actively regulated.

The paracellular pathway is not simply sealed in health and structurally destroyed in disease. Tight-junction organisation can change dynamically in response to signalling.

That biological pathway is discussed in detail elsewhere in BIRNA Research.

For measurement, however, another question becomes central:

How do we know that a concentration reported as “zonulin” actually measures that pathway?

The molecular identification added precision

In 2009, Tripathi and colleagues identified human zonulin as the precursor of haptoglobin 2, pre-haptoglobin 2 (pre-HP2).^²

This was a major step because it connected the functional observations surrounding zonulin to a defined molecular entity.

Pre-HP2 is biologically distinct from mature haptoglobin 2. The uncleaved precursor was shown experimentally to influence permeability, whereas proteolytic cleavage into the mature haptoglobin subunits removed this permeability-modulating activity.^²

That distinction immediately creates an analytical challenge.

A test intended to quantify zonulin should ideally distinguish the biologically relevant precursor from mature haptoglobin and from other structurally related proteins.

The existence of a named molecule therefore does not, by itself, validate an assay carrying the same name.

Zonulin may represent more than one protein

The molecular picture subsequently became more complex.

Investigations of antibodies used in zonulin assays suggested that proteins other than pre-HP2 may share structural or immunological characteristics with zonulin.

Scheffler and colleagues examined a widely used commercial ELISA and reported that it did not detect recombinant pre-HP2 as expected. Instead, their experiments identified properdin as a likely target and proposed it as another potential member of the zonulin family.^³

The concept was extended further by work from Takumi Konno and colleagues at Massachusetts General Hospital.

Konno, Martinez, Ji, Miranda-Ribera, Fiorentino and Fasano reported that human coagulation factor X and CD5 antigen-like protein showed structural and functional similarities to previously identified zonulin-family proteins. Both recombinant proteins were recognised by anti-zonulin antibody in their experimental system and reduced epithelial barrier competency in Caco-2 monolayers as assessed by TEER.^⁴

This supports a more nuanced concept of zonulin-family proteins (ZFPs).

That concept is biologically interesting.

It also complicates biomarker interpretation.

If several related proteins can share antibody recognition or functional characteristics, an assay signal cannot automatically be assumed to represent one uniquely defined molecule.

The assay problem is not a minor technical issue

A commercial assay is sometimes treated as a neutral window onto biology.

It is not.

An immunoassay measures what its antibodies recognise under the conditions in which the assay is performed.

If antibody specificity is uncertain, the biological interpretation of the resulting concentration is also uncertain.

Scheffler and colleagues demonstrated this problem directly. Their widely used commercial ELISA did not recognise recombinant pre-HP2, despite being marketed and extensively used as a zonulin assay.^³

Ajamian and colleagues reached a related conclusion from another direction. They investigated proteins detected by a commercial zonulin ELISA and concluded that serum measurements obtained with the assay should not automatically be interpreted as a direct marker of intestinal mucosal barrier function.^⁷

These findings do not demonstrate that every historical zonulin result is biologically meaningless.

They demonstrate something more specific and more important:

the analyte must be known before the assay value can be interpreted as a measurement of a particular molecule or pathway.

Analytical validity and biological validity are different

Several questions are often compressed into one.

Does the assay reliably produce a number?

Does it detect the molecule it claims to detect?

Does the detected molecule reflect activity of the zonulin pathway?

Does pathway activity correspond to altered intestinal permeability at the time of measurement?

Does that permeability alteration have clinical significance?

These are separate validation problems.

An assay may be technically reproducible while lacking molecular specificity.

A molecularly specific assay may measure a protein whose circulating concentration does not correspond closely to intestinal pathway activity.

A biomarker may correlate with a functional permeability test in one population but not another.

And even a valid measure of altered permeability does not establish that the permeability alteration caused the clinical phenotype under investigation.

Reliability, analytical validity, biological validity and clinical interpretation therefore cannot be collapsed into one claim.

Serum zonulin is not synonymous with small-intestinal permeability

This distinction has been tested empirically.

In healthy adults, serum zonulin measured by ELISA has been compared with functional assessment of small-intestinal permeability. One such study found that serum zonulin measured with the commercial assay did not provide a reliable substitute for functional small-intestinal permeability testing.^⁸

This matters because the two approaches measure fundamentally different things.

A differential sugar test assesses the passage and subsequent recovery of defined probe molecules.

A serum immunoassay quantifies antibody-reactive material in blood.

Calling both measurements “intestinal permeability” does not make them equivalent.

If they disagree, the discrepancy should not automatically be treated as failure of one test.

It may reveal that they are interrogating different biological processes.

Serum and faecal zonulin should not be treated as the same measurement

Specimen type adds another layer.

Zonulin has been measured in serum, plasma and faeces. These matrices differ biologically and analytically.

A concentration in blood reflects material present within the systemic circulation.

A faecal measurement reflects material present within a complex luminal matrix containing microbial products, digestive material, host proteins and other compounds.

The same numerical unit or assay label does not make these measurements biologically equivalent.

Recent studies continue to investigate faecal zonulin as a non-invasive marker. For example, a 2025 prospective cohort study reported an association between faecal zonulin and hepatic steatosis, but discriminatory performance was modest and cohort-specific calibration performed differently from the manufacturer’s threshold.^⁹

That kind of finding may be clinically interesting.

It should not be transformed into evidence that faecal zonulin directly quantified intestinal paracellular permeability.

Association, discrimination and physiological measurement are different questions.

Population and disease context can alter biomarker meaning

A biomarker does not necessarily mean the same thing in every population.

This becomes especially important when the biology of the disease can itself alter production, clearance or distribution of the measured protein.

A 2025 systematic review and meta-analysis of serum zonulin in liver cirrhosis illustrates the problem. The included studies showed substantial heterogeneity, and the authors specifically cautioned that zonulin cannot simply be treated as a permeability marker in cirrhosis because hepatic biology itself affects its interpretation.^¹⁰

This is a general biomarker principle.

The meaning of a concentration depends not only on the assay but on the biological system in which it is measured.

Renal function can influence clearance.

Liver function can influence protein synthesis.

Inflammation can influence multiple pathways simultaneously.

Age, metabolic state, medication and disease activity may affect both the proposed biomarker and the outcome with which it is being compared.

A biomarker cannot be interpreted outside its biological context.

Intervention studies create another interpretive trap

Intervention studies frequently report changes in serum or faecal zonulin after dietary, probiotic, prebiotic or other interventions.

A decrease is then sometimes described as an “improvement in intestinal permeability”.

That conclusion may be stronger than the measurement allows.

Recent systematic reviews illustrate both the interest and the problem. Intervention studies using probiotics, synbiotics or prebiotics have reported changes in zonulin and other barrier-associated markers, but substantial heterogeneity remains across interventions, populations and outcome measures.^¹¹,¹²

If an intervention lowers the concentration produced by a commercial zonulin assay, the immediate result is:

the assay signal changed.

To conclude that functional intestinal permeability improved requires additional evidence connecting that assay signal to permeability in that experimental context.

This is not excessive caution.

It is simply keeping the conclusion at the level of the measurement.

My own systematic review exposed this problem early

This distinction became central in my own work.

In our systematic review and meta-analysis of zonulin-dependent intestinal permeability in children diagnosed with mental disorders, we examined observational evidence in children with ADHD, autism spectrum disorder and obsessive-compulsive disorder.^¹³

Five studies representing 402 participants met the inclusion criteria. The available evidence suggested differences in zonulin-related measurements in some groups, but the literature was small and methodologically heterogeneous.

The scientific question was therefore not resolved by observing a difference in a biomarker.

A reported zonulin difference could not establish, by itself, that functional intestinal permeability was altered.

Even evidence of altered permeability would not establish that permeability caused the psychiatric phenotype.

And neither would identify the microbiota, immune system or another pathway as the causal origin of the alteration.

This is the inferential chain that must remain visible:

assay signal → molecular interpretation → pathway activity → functional permeability → systemic consequence → clinical phenotype.

Evidence at one point in that chain cannot simply be used to fill the others.

Meta-analysis cannot repair measurement uncertainty

Systematic review and meta-analysis are often placed near the top of evidence hierarchies.

But synthesis is only as interpretable as the measurements being synthesised.

If several studies use the same assay, meta-analysis can estimate the consistency and magnitude of differences produced by that assay across populations.

It cannot retrospectively establish that the assay detected the molecule originally assumed.

Nor can statistical pooling transform a surrogate biomarker into a direct functional measurement.

This is especially important in zonulin research because apparent consistency across studies may partly reflect use of the same commercial assay rather than independent confirmation of the underlying biological process.

Heterogeneity can also be informative.

Differences among populations, matrices, assay platforms, clinical phenotypes and study designs may indicate that a supposedly single biomarker is capturing multiple biological processes.

Meta-analysis should quantify that uncertainty.

It cannot remove it.

Newer biomarker reviews reinforce the distinction

The broader permeability literature is increasingly moving toward this more differentiated interpretation.

A 2025 review of body-fluid biomarkers for human intestinal-permeability assessment considered sugar probes alongside endotoxin, D-lactate, I-FABP, diamine oxidase, citrulline, zonulin and LBP.^¹⁴

The importance of such a framework is not that all these measures are equivalent markers of the same construct.

It is almost the opposite.

They arise from different biological processes.

Some relate to molecular passage.

Some to epithelial injury or epithelial mass.

Some to microbial products or host responses to them.

Some to proposed regulation of tight junctions.

A stronger measurement framework therefore asks which biological dimension each marker represents before deciding what it says about permeability.

Zonulin should be interpreted within that same discipline.

The zonulin pathway remains biologically important

Measurement controversy should not be confused with evidence that the zonulin pathway does not exist.

The mechanistic work identifying zonulin-related regulation of tight junctions, the identification of pre-HP2, and subsequent work on zonulin-family proteins address biological questions that are distinct from the validity of a particular commercial ELISA.^¹–⁶

This separation is essential.

It is possible for all of the following statements to be true:

• zonulin-related signalling participates in regulation of epithelial permeability;
• pre-HP2 has been identified as zonulin;
• structurally and functionally related zonulin-family proteins may exist;
• some commercial assays labelled as “zonulin” do not specifically measure pre-HP2;
• concentrations generated by those assays may nevertheless associate with biological or clinical phenotypes;
• those associations cannot automatically be interpreted as direct measurements of intestinal permeability.

There is no contradiction in this.

The contradiction appears only when pathway, molecule and assay are treated as the same thing.

From number to claim

When reading a human zonulin study, I therefore ask a sequence of questions.

What specimen was measured?

Serum, plasma, faeces or another matrix?

What assay was used?

Which manufacturer, antibody system and analytical platform?

What is the assay known to detect?

Has molecular specificity been demonstrated?

What biological construct is being inferred?

Pre-HP2 concentration? A broader zonulin-family signal? Pathway activity? Intestinal permeability?

Was permeability measured independently?

For example, with a functional probe test or another validated physiological method?

What is the clinical claim?

Association, diagnostic discrimination, treatment response, mechanism or causation?

Each question moves one step further from the laboratory measurement.

Each step requires evidence.

The useful conclusion is the precise one

Zonulin remains scientifically valuable precisely because it exposes the difference between biological plausibility and measurement certainty.

The pathway contributed substantially to our understanding of intestinal-barrier regulation.^¹,⁵,⁶

The identification of pre-HP2 provided molecular specificity.^²

The recognition of additional zonulin-family proteins broadened the molecular model.^³,⁴

The assay literature then demonstrated that a commercial test carrying the name zonulin cannot automatically be assumed to quantify pre-HP2 or functional intestinal permeability.^³,⁷,⁸

Human observational and intervention studies add another layer, but they do not erase these analytical boundaries.^⁹–¹⁴

The appropriate response is therefore neither to accept every zonulin measurement as a permeability measurement nor to discard the entire field because some assays are imperfect.

It is to ask a more precise question:

What exactly was measured, and what has that measurement earned the right to say?

That is the distinction between using zonulin as a scientific tool and using the word zonulin as a shortcut for a biological conclusion.