Zonulin as prehaptoglobin2 regulates lung permeability and activates the complement system

Daniel Rittirsch, Michael A. Flierl, Brian A. Nadeau, Danielle E. Day, Markus S. Huber-Lang, Jamison J. Grailer, Firas S. Zetoune, Anuska V. Andjelkovic, Alessio Fasano, Peter A. Ward


Zonulin is a protein involved in the regulation of tight junctions (TJ) in epithelial or endothelial cells. Zonulin is known to affect TJ in gut epithelial cells, but little is known about its influences in other organs. Prehaptoglobin2 has been identified as zonulin and is related to serine proteases (MASPs, C1qrs) that activate the complement system. The current study focused on the role of zonulin in development of acute lung injury (ALI) in C57BL/6 male mice following intrapulmonary deposition of IgG immune complexes. A zonulin antagonist (AT-1001) and a related peptide with permeability agonist activities (AT-1002) were employed and given intratracheally or intravenously. Also, zonulin was blocked in lung with a neutralizing antibody. In a dose-dependent manner, AT-1001 or zonulin neutralizing antibody attenuated the intensity of ALI (as quantitated by albumin leak, neutrophil accumulation, and proinflammatory cytokines). A similar pattern was found using the bacterial lipopolysaccharide model of ALI. Using confocal microscopy on sections of injured lungs, staining patterns for TJ proteins were discontinuous, reduced, and fragmented. As expected, the leak of blood products into the alveolar space confirmed the passage of 3 and 20 kDa dextran, and albumin. In contrast to AT-1001, application of the zonulin agonist AT-1002 intensified ALI. Zonulin both in vitro and in vivo induced generation of complement C3a and C5a. Collectively, these data suggest that zonulin facilitates development of ALI both by enhancing albumin leak and complement activation as well as increased buildup of neutrophils and cytokines during development of ALI.

  • inflammation
  • acute lung injury
  • tight junctions
  • C3a
  • C5a

intercellular tight junctions (TJ), also known as zonula occludens, are junctional complexes between endothelial cells and epithelial cells. TJ regulate paracellular trafficking and fluid flow from plasma to extravascular spaces. Until recently, TJ were thought to be unregulated structures forming an impermeable barrier within the paracellular space (13). However, it has become evident that TJ are complex and dynamic structures involved in several key functions of epithelial and endothelial barriers under physiological conditions as well as in pathological circumstances (22). In addition, recent research has indicated that various signaling proteins and transduction pathways are associated with these junctions (26). The discovery of Vibrio cholerae Zonula occludens toxin (Zot), a toxin that causes opening of TJ, led to the identification of its eukaryotic counterpart, zonulin, as an endogenous regulator of paracellular permeability by modulating intercellular TJ (1). While systemic bacteremia or septicemia by V. cholerae is well known, acute lung injury (ALI) is not part of the pathological picture of the disease. Zonulin is an ∼47-kDa protein that has been shown to increase permeability in the intestinal epithelium and is also involved in intestinal innate immunity (1, 7). Zonulin is overexpressed in patients with autoimmune disorders, including celiac disease and type 1 diabetes in which TJ dysfunction seems to be a key functional defect (5, 8, 32). With respect to its mechanism, zonulin has recently been identified as the precursor of haptoglobin2 (preHP2) that engages a key signalosome involved in the pathogenesis of various immune-mediated diseases (39). Although zonulin has been described as a modulator of intestinal permeability in health and disease, its role and mechanism of action in extraintestinal tissues remains poorly defined. Besides the gastrointestinal tract, the involvement of zonulin in regulating paracellular permeability in the brain has also been suggested (34). In contrast, the role of zonulin in regulating lung permeability is unclear.

ALI and acute respiratory distress syndrome (ARDS) are characterized by a leakage of plasma components into the lungs, compromising the ability of lungs to expand and optimally engage in gas exchange with blood, resulting in respiratory failure (42). ALI/ARDS can be caused by a variety of different insults, and the underlying pathophysiology is highly complex. While the initial mechanisms of the lung inflammatory response may be divergent depending on the causative insult, the common downstream events evolve into increased lung vascular permeability, which is a characteristic feature of ALI (17). In the pathogenesis of ALI, local activation of the complement system plays a central role in the initiation and progression of disease, and products of complement activation can induce production of adhesion molecules on endothelial cells and on leukocytes, the release of oxygen radicals, and the expression of cytokines/chemokines. Finally, the complement activation product C5a can act as a direct chemoattractant for neutrophils (6, 15, 17, 27, 35). As part of the first line of defense, the complement system represents a cascade of plasma serine proteases. Complement is considered to be a phylogenetically ancient part of the innate immune system. Interestingly, preHP2 and certain complement-activating proteins (MASPs and C1qrs) are descendants of a common ancestral protein (21). While numerous clinical and experimental studies have investigated the immunopathogenesis of ALI, little is known about mechanisms leading to a loss of endothelial and epithelial barriers in the lung.

Therefore, we sought to investigate the role of zonulin in ALI to determine if blockade of zonulin would attenuate increased lung permeability in ALI. We further hypothesized that zonulin (also known as preHP2) can activate the complement system, which is known to play an important role in the pathophysiology of ALI/ARDS.


Acute lung injury
Acute respiratory distress syndrome
Synthetic zonulin antagonist
Synthetic peptide with permeability-inducing activity
Trypsinized preHP2
Prehaptoglobin2 (zonulin)
Mannose-binding lectin-associated serine protease
Tight junction
Zonula occludens toxin



Young adult male (22–25 g) specific pathogen-free C57BL/6 mice were used in these studies. All studies were done in accordance with the University of Michigan committee on use and care of animals.

IgG immune complex ALI.

For intrapulmonary immune complex deposition, 125 μg rabbit anti-BSA IgG (ICN Biomedicals, Aurora, OH) was administered intratracheally followed by intravenous injection of BSA (500 μg). Sham-operated animals underwent the same procedure with intratracheal injection of PBS. Permeability index as a quantitative marker for vascular leakage was determined by leak of intravenously injected 125I-labeled BSA in lung and compared with radioactivity in 1.0 ml blood, as described elsewhere (28). For analgesia and anesthesia, mice were injected intraperitoneally with ketamine (100–200 mg/kg body wt) and with xylazine (8–10 mg/kg body wt) immediately before induction of ALI. For bronchoalveolar lavage (BAL) retrieval, airways were flushed with 0.8 ml PBS (pH 7.4). If not otherwise noted, permeability index was determined and BAL fluids were collected 4 h after lung injury induction.

Lipopolysaccharide ALI model.

For lipopolysaccharide (LPS)-induced ALI, the trachea was surgically exposed in anesthetized mice, and 40 μl of PBS containing 50 μg of LPS (from Escherichia coli 0111:B4; Sigma) were slowly instilled intratracheally during inspiration as previously described (3).

Enzyme-linked immunosorbent assay for mouse IL-6, TNF-α.

For measurement of IL-6 and TNF-α in BAL fluids, commercially available enzyme-linked immunosorbent assay (ELISA) kits (“Duo set”; R&D Systems, Minneapolis, MN) were used according to the manufacturer's protocol.

ELISA for mouse C3a and C5a.

To measure the concentration of mouse C3a and C5a in BAL fluids, ELISA plates were coated with purified monoclonal anti-mouse C5a or anti-mouse C3a IgG (capture antibody, 5 μg/ml; BD Pharmingen). After blocking, BAL fluids and recombinant mouse C5a (R&D Systems) or C3a (BD Pharmingen) were applied. For detection, biotinylated monoclonal anti-mouse C5a or anti-mouse C3a antibody was applied (500 ng/ml; BD Pharmingen), followed by incubation with streptavidin-peroxidase. As substrate solution, equal volumes of H2O2 and tetramethylbenzidine (R&D Systems) were mixed and added to the wells. The color reaction was stopped with 2 N sulfuric acid, and the absorbance was read at 450 nm.

Lung myeloperoxidase activity in tissue extracts.

After 4 h, mouse lungs were perfused through the right cardiac ventricle with 2 ml PBS and then snap-frozen in liquid nitrogen and stored at −80°C. To measure myeloperoxidase (MPO) activity, whole lungs were homogenized in 50 mM potassium phosphate buffer containing 0.5% hexadecyltrimethylammonium bromide and 5 mmol/l EDTA. After centrifugation at 12,000 g for 10 min at 4°C, the supernatant fluids were incubated in a 50 mmol/l potassium phosphate buffer containing the substrate H2O2 (1.5 mol/l) and o-dianisidine dihydrochloride (167 μg/ml; Sigma Aldrich). The enzymatic activity was determined spectrophotometrically by measuring the change in absorbance at 460 nm over 3 min (Molecular Devices, Sunnyvale, CA) (36).

Leukocyte count in BAL fluids.

Immediately after collection of BAL fluids, red blood cells were lysed with 1% acetic acid, and total white cell count of each BAL sample was determined using a Neubauer hemocytometer (Hauser Scientific, Horsham, PA).

Histological analysis.

Lungs were instilled via the trachea with 10% buffered (pH 7.0) formaldehyde and then surgically removed, embedded in paraffin, and sectioned, followed by staining with hematoxylin and eosin.

Modulation of zonulin-induced lung permeability.

For the blockade of zonulin, mice were treated with the zonulin antagonist AT-1001 (1) by intratracheal or intravenous administration in the doses designated or with vehicle solution at the time of lung injury induction. AT-1001 competitively blocks the apical zonulin receptor and prevents the opening of TJ secondary to zonulin (1, 4). Alternatively, 40 μg rabbit anti-Zot IgG (40, 41) or nonspecific rabbit IgG (nsIgG; Jackson ImmunoResearch, West Grove, PA) were given intratracheally together with anti-BSA IgG. In additional experiments, healthy mice or mice with ALI were treated with the zonulin agonist AT-1002 (12) to enhance lung permeability or vehicle solution only (37). Both peptides, AT-1001 (antagonist) and AT-1002 (agonist), have been shown to bind to the same zonulin receptor through a specific binding motif, whereas AT-1002 showed higher affinity than AT-1001 (19).

Molecular markers to define leakage.

BAL fluids were harvested from control mice or mice 2, 4, or 6 h after IgG immune complex (IgGIC) deposition. Mice were previously injected intravenously with a mixture (1 mg/ml of each tracer) of fluorescein isothiocyanate (FITC)-albumin (66 kDa; Sigma), Cascade blue-dextran (3 kDa; Invitrogen), and TRITC-dextran (20 kDa; Sigma) retrieval of BAL fluids 90 min later. BAL fluids were mixed 1:1 with methanol and centrifuged 1,000 g for 30 min. The supernatant fluids were collected, and fluorescence levels for each tracer were measured on a microplate fluorometer (Tecan Group, Mannedorf, Switzerland). Readings were compared with a standard curve to generate values.

Immunostaining for TJ proteins in lung.

Snap-frozen samples of lung were fixed with 4% paraformaldehyde for 15 min and then preincubated in blocking solution containing 5% normal goat serum, 1% BSA, and 0.05% Tween in PBS. Lung sections were then incubated overnight at 4°C with the following primary antibodies: mouse occludin-Alexa Fluor 596, claudin-5-Alexa Fluor 488, zonula occludens (ZO)-1-FITC, and rabbit claudin-3 (all from Invitrogen). Reaction for claudin-3 was visualized by fluorescein-conjugated anti-rabbit FITC-conjugated antibody. All samples were viewed on a confocal laser scanning microscope (LSM 510 Zeiss; Germany).

In vitro incubation of human serum with zonulin or two-chain HP2.

Recombinant human zonulin was cloned and expressed in a baculovirus expression system, as described elsewhere (39). For generation of two-chain HP2, single-chain zonulin was digested in its α2- and β-chains with trypsin-agarose beads (T-1763; Sigma) for 20 min at 25°C. The beads were removed by centrifugation, and removal of trypsin was confirmed (trypsin peptidase activity; Bachem BioScience) (39).

Blood was drawn from healthy probands and incubated overnight at 4°C. After centrifugation, serum was collected and incubated with zonulin or zonulin-derived HP2 secondary to treatment using concentrations of trypsin at 37°C for 30 min. The reaction was stopped by adding EDTA at a final concentration of 0.01 M.

ELISA for human C3a and C5a.

To assess the extent of complement activation in human serum samples after in vitro incubation, the concentration of C3a and C5a was determined using commercially available ELISA kits [C3a (Quidel, San Diego, CA) and C5a (DRG Diagnostics, Marburg, Germany)] according to the manufacturers' protocols.

Statistical analysis.

All values were expressed as means ± SE. Data sets were analyzed by one-way ANOVA followed by Tukey's multiple-comparison test with GraphPad Prism 4 software (GraphPad Software, San Diego, CA). Results were considered statistically significant when P < 0.05.


Blockade of zonulin attenuates experimental ALI.

To investigate the role of zonulin in the regulation of lung permeability, the well-established IgGIC model of ALI in mice was employed (9, 14). In this model, ALI is induced by intrapulmonary immune complex deposition, causing activation of the complement system, neutrophil accumulation and lung macrophage activation, production of proinflammatory cytokines and chemokines, and release of proteases and oxidants. As a hallmark of ALI, lung permeability substantially increases because of a capillary leak, resulting in alveolar edema and hemorrhage (42). In this model, lung injury is defined by the permeability index (albumin leak) as a direct measure for the severity of disease, which peaks 4–6 h after ALI induction (28). When the zonulin antagonist AT-1001 was administered intratracheally in a range of doses together with the anti-BSA IgG, the albumin leak was significantly attenuated as a function of dose, with a maximum reduction of ∼67% (Fig. 1A). Interestingly, ALI was not only diminished after intrapulmonary administration of AT-1001 (25 μg) but also when the compound was given intravenously (50 μg) following lung injury induction (Fig. 1B). In additional experiments, pulmonary polymorphonuclear neutrophil (PMN) accumulation was assessed by determination of MPO activity (whole lung) and leukocyte count in BAL fluids. As shown in Fig. 1C, buildup of lung MPO was modestly reduced (24%, P < 0.05) when the zonulin inhibitor AT-1001 was administered intratracheally. In line with this finding, the presence of AT-1001 also resulted in a significantly reduced number (by 25%) of leukocytes (>95% being neutrophils) in BAL fluids compared with mice with ALI in the absence of AT-1001 treatment (Fig. 1D). These results were further supported by the observation that the local inflammatory response, as determined by the release of the proinflammatory cytokines IL-6 and TNF-α in BAL fluids, was greatly reduced when AT-1001 was administered intratracheally (Fig. 1, E and F). Taken together, these findings suggest that blockade of zonulin reduces the severity of ALI by reducing the capillary leak, diminishing accumulation of PMNs in the interstitial and alveolar compartments, and attenuating levels of proinflammatory mediators.

Fig. 1.

Effects of a blockade of zonulin in acute lung injury (ALI) by administration of the zonulin antagonist larazotide acetate (AT-1001). A: lung permeability as assessed by the leak of 125I-labeled albumin in lung (permeability index) after it administration of AT-1001 as a function of dose (10, 25, or 50 μg). B: comparison of the permeability index in ALI after it (25 μg it) vs. iv (50 μg iv) administration of AT-1001. C: lung myeloperoxidase (MPO) activity as a measure of polymorphonuclear neutrophil (PMN) accumulation (AT-1001, 25 μg it). D: number of leukocytes in bronchoalveolar lavage (BAL) fluids from mice with ALI treated with AT-1001 (25 μg it) or vehicle solution. Effects of zonulin blockade (AT-1001, 25 μg it) on the buildup of IL-6 (E) and TNF-α (F) in BAL fluids. For each bar, n = 5 mice. *P < 0.05.

Reduced ALI by antibody-induced neutralization of zonulin.

Either neutralizing polyclonal antibody against zonulin [anti-Zot (40, 41)] or nsIgG was administered intratracheally (40 μg) together with anti-BSA IgG (125 μg) used for ALI induction and intravenously infused BSA (0.5 mg). In control to normal lung (Fig. 1A), lungs from mice with ALI (positive control; Fig. 2B) or ALI + nsIgG (Fig. 2C) showed typical signs of lung injury, including abundant PMN accumulation, fibrin deposits, hemorrhage, and edema in the alveolar compartment. In contrast, when zonulin was blocked by cross-reacting anti-Zot antibodies, lungs revealed much reduced signs of ALI (Fig. 2D), with alveolar spaces showing scant evidence of PMNs and fibrin, and diminished hemorrhage. Consistent with the findings in Fig. 1, antibody-induced neutralization of zonulin in experimental ALI resulted in a significant reduction (50%) of lung permeability index (Fig. 2E). Similar results were found when mouse albumin concentration in BAL fluids was used as an endpoint measure for lung permeability (Fig. 2F).

Fig. 2.

Histological features in control lung (A) and lungs from mice with IgG immune complex (IgGIC)-induced ALI (B) treated it with either nonspecific IgG (nsIgG) (C) or zonulin neutralizing antibody (anti-Zot, D). Hematoxylin and eosin, ×40. Effects of anti-Zot (it) in ALI on lung permeability index (leak of 125I-albumin in lung) (E) or the leakage of mouse albumin in BAL fluids (F). For each bar, n = 5 mice. NS, not significant. *P < 0.05.

Administration of either zonulin or its agonist AT-1002 intensifies ALI and increases lung permeability.

To further support the concept that zonulin is involved in regulating lung permeability, zonulin (20 μg) was administered intratracheally to sham mice. As shown in Fig. 3A, intratracheal administration of zonulin (20 μg) in healthy mouse lungs increased the permeability index by 3.2-fold compared with the value in normal (control) lung. In parallel, when 20 μg HP2 were injected intratracheally, the permeability index rose by 2.6-fold, suggesting that both zonulin (preHP2) and HP2 (trypsinized zonulin) increased vascular permeability in mouse lung. In line with an earlier report (39), the permeability-increasing activity of zonulin was slightly reduced after pretreatment of zonulin with trypsin (Fig. 3A), which converted zonulin to HP2. As further support for the ability of zonulin to increase vascular permeability in mouse lung, we used AT-1002, which represents a peptide agonist of zonulin with permeability-inducing activity in the small intestine or in monolayers of intestinal epithelial cells (12). Similar to the application of AT-1001 in the previous experiments (Fig. 1A), AT-1002 was given together with anti-BSA IgG intratracheally at the time of induction of ALI. Under these conditions, lung permeability was substantially intensified by 1.5-fold (Fig. 3B) compared with mice that received anti-BSA in the absence of AT-1002. When AT-1002 was administered intratracheally in healthy mice, lung permeability, as determined by the leak of I125-labeled BSA in lungs, did not change compared with negative control mice (data not shown). In summary, ALI was not only attenuated by antibody neutralization of zonulin (Fig. 2) but, in turn, was intensified in the presence of its small peptide permeability-inducing agonist AT-1002 (Fig. 3B). Moreover, zonulin in the setting of healthy lungs was capable of increasing the permeability in lungs (Fig. 3A), further suggesting a role of zonulin in the regulation of lung permeability. Combined, these data demonstrate for the first time that zonulin participates in the regulation of lung permeability and in the intensity of ALI.

Fig. 3.

A: effect of the it injection of zonulin or trypsin-treated zonulin [haptoglobin2 (HP2)] in healthy lungs on the permeability index (leak of 125I-albumin in lung). B: effects of the permeability enhancer AT-1002 (50 μg it) in the ALI model on the permeability index. For each bar, n = 5 mice. *P < 0.05.

Defining the nature of the lung leak in IgGIC-induced ALI.

As shown in Fig. 4A, we used fluorescent-labeled molecular markers that were injected intravenously 1.5 h before lung lavage after induction of ALI. BAL fluids were then assessed for the amount of fluorescence 2, 4, and 6 h after induction of ALI. Clearly, 3 kDa, 20 kDa and albumin markers all escaped from the vascular into the alveolar compartment, implying a substantial amount of leak. This was expected, since alveolar hemorrhage develops in this type of ALI (Fig. 2). In Fig. 4B, confocal microcopy was carried out on fixed frozen sections of control (uninjured) and 6 h IgGIC ALI mouse lungs. The following TJ proteins were assessed: ZO-1, claudin-5, claudin-3, and occludin. As is apparent, in normal lungs, TJ proteins showed linear patterns consistent with capillary walls. In the ALI lungs, the fluorescence patterns were very much less intense, being discontinuous and fragmented.

Fig. 4.

A: molecular weight characterization of vascular leak in the alveolar space during ALI. Fluorescent molecular tracers (albumin, 66 kDa; dextran, 3 or 20 kDa) were administered iv 90 min before harvest of BAL fluids from healthy lungs (control) or mice 2, 4, or 6 h after onset of IgGIC ALI. The fluorescence level for each tracer in BAL fluids was quantified. Enhanced presence of tracers in BAL fluids indicates vascular leakage into the alveolar compartment. B: immunofluorescent localization of tight junction proteins occludin, claudin-3, claudin-5, and zonnula occludens (ZO)-1 in uninjured lung (control) and in ALI tissue sections. In control samples, the tight junction proteins claudin-3 (expressed only on the epithelial cell-cell junction), claudin-5, ZO-1, and occludin (expressed on both epithelial and endothelial cells-cell junction) showed a continuous linear staining pattern. ALI samples showed discontinuous and fragmented staining patterns for all tight junction proteins, indicating disruption of epithelial and alveolar epithelial cell junctional complexes and increased permeability. Scale bar 100 μm.

Zonulin (preHP2) and HP2 activate the complement system in vitro.

Evolutionary studies have demonstrated that zonulin (preHP2) and complement-activating MASP proteins are descendants of a common ancestor (21). On the other hand, activation of the complement system is known to play a crucial role in the pathogenesis of ALI and in the IgGIC model in particular (6, 15, 17, 27, 35). Therefore, we investigated whether zonulin had complement-activating capabilities. When human serum from healthy probands was incubated with zonulin in vitro, complement was activated in a dose-dependent manner, as indicated by the generation of the complement activation products C3a (Fig. 5A) and C5a (Fig. 5B). The generation of these anaphylatoxins was dependent on the concentrations of zonulin and HP2 (trypsin-treated zonulin) employed. When normal human serum was incubated at 37°C for 30 min in the absence of a complement activator, very small amounts of C3a (Fig. 4A) and C5a (Fig. 5B) were generated. Much higher levels of these anaphylatoxins were found following addition of zonulin. Proteolytic cleavage of zonulin into mature double-chain HP2 by pretreatment with trypsin resulted in lesser complement-activating activity (Fig. 5, A and B). However, using high doses of HP2 (trypsinized preHP2), complement activation still occurred, but to a lesser extent compared with untreated zonulin (Fig. 5, A and B). In the next step, human serum was coincubated with zonulin and with anti-Zot antibodies (which neutralize zonulin) or with nsIgG. In the presence of anti-Zot, zonulin-induced complement activation (generation of C3a and C5a) was suppressed (Fig. 5, C and D). These data suggest that zonulin functions as an activator of the complement system. Because pretreatment of zonulin with trypsin results in a decrease of complement-activating activity (converting preHP2 to HP2), this novel mechanism of complement activation appears to be optimized by the spatial conformation that the zonulin α2- and β-subunits acquire when they remain in a single peptide molecular arrangement while their cleavage into two distinct subunits decreases their complement-activating activity.

Fig. 5.

In vitro incubation of human serum with either zonulin or trypsin-digested zonulin (HP2). Effect of zonulin and HP2 as a function of dose on activation of the complement system, as determined by the in vitro generation of the complement activation products C3a (A) and C5a (B). Generation of C3a (C) and C5a (D) after in vitro incubation of human sera with zonulin or HP2 in the presence of nsIgG or anti-Zot; n = 4 probands sera, each condition being run in triplicate. Ctrl, Control. *P < 0.05.

Neutralization of zonulin reduces complement activation in lungs.

Local activation of the complement system is known to play a crucial role in the pathophysiology of ALI (15, 27, 35). In the IgGIC model of ALI, the generation of C5a is known to amplify the production of proinflammatory cytokines, leading to intrapulmonary accumulation and activation of neutrophils and macrophages (27, 33, 36, 43). In additional in vivo experiments employing the IgGIC model, zonulin was blocked by intratracheal administration of anti-Zot antibodies, followed by assessment of complement activation products in BAL fluids. The use of anti-Zot antibodies in the IgGIC model of ALI resulted in significantly reduced levels of the C3a (Fig. 6A) and C5a (Fig. 6B) in BAL fluids compared with mice with ALI that received the same amount of corresponding nsIgG instead. In the case of C3a, generation was suppressed by 30% in the presence of anti-Zot antibody, whereas the levels of C5a in BAL fluids were reduced to baseline levels when zonulin was blocked by anti-Zot antibodies (Fig. 6, A and B). The fact that neutralization of zonulin resulted in attenuated complement activation during ALI suggests that zonulin may contribute to intrapulmonary complement activation in vivo and that development of ALI causes buildup of plasma complement proteins in lung as vascular permeability develops.

Fig. 6.

Assessment of activation of complement in vivo by measurement of the concentrations of C3a (A) and C5a (B) in BAL fluids from mice with IgGIC-induced ALI treated it with either anti-Zot or nonspecific IgG. For each bar, n = 5 mice. *P < 0.05.

Similar observations using the LPS model of ALI.

As shown in Fig. 7, in the LPS model of ALI, the use of a neutralizing antibody to Zot diminished the leak of albumin into the alveolar compartment (Fig. 7A), as did the presence of the synthetic zonulin antagonist AT-1001 (Fig. 7B). The data suggest that, in the LPS model of ALI, the use of anti-Zot or use of the zonulin antagonist AT-1001 have protective effects similar to those seen in the IgGIC model of ALI.

Fig. 7.

A: control and ALI induced by LPS (10) in mice treated with either nsIgG it or anti-Zot. B: LPS induced ALI in mice that were given Dulbecco's phosphate-buffered saline (DPBS) or the zonulin antagonist AT-1001. For each bar, n = 5 mice. *P < 0.05.


The discovery of zonulin a decade ago has shed new light on the concept of TJ-mediated epithelial barrier function (8). It is now evident that TJ underlie a dynamic regulation of barrier function for both epithelial and endothelial cells and that TJ dysregulation may play a central role in the pathogenesis of various inflammatory and autoimmune diseases (5, 8, 22, 26, 32). While the majority of studies investigating the mechanisms of TJ regulation have focused on intestinal epithelia, the present study demonstrates that zonulin may also be involved in the regulation of epithelial and/or endothelial barriers in the lung.

Although capillary leak and loss of epithelial barrier integrity represent a hallmark of ALI and ARDS, little is known about the role of TJ and how TJ may affect inflammatory pathways. As a matter of fact, the medical literature database search using the Medical Subject headings terms “acute lung injury (ALI)” or “acute respiratory distress syndrome (ARDS)” and “tight junctions” retrieves only 33 total results. Among these publications, only a few have investigated the underlying mechanisms of barrier dysfunction involving TJ. It is clear from the data in the current report that, in the IgGIC model of ALI, there is extensive destruction of TJ proteins (ZO-1, claudin-5, claudin-3, and occludin) leading to substantial lung leak (Fig. 4B). However, it is likely that proteases and oxidants derived from both activated neutrophils and lung macrophages also contribute to this injury. While claudin-3 is only expressed in TJ of epithelial cells, claudin-5, ZO-1, and occludin are expressed in both TJ between epithelial cells and between endothelial cells (11, 18). In a recent study, the expression of claudin-4 was found to be upregulated in a murine model of ventilator-induced lung injury using gene expression microarrays (44). Claudins are a family of transmembrane proteins involved in the regulation of paracellular permeability by restricting the bulk of fluid movement between cells and selective development of paracellular ion diffusion (20). Claudins are part of the TJ complex that is composed of several different components, including transmembrane, peripheral, and cytoskeletal proteins, the coordinated activity of which is required for TJ integrity (20). Wray et al. proposed a model that claudin-4 could restrict paracellular backflow of sodium while permitting paracellular chloride diffusion and thus helping to maintain fluid efflux from the alveolus (20, 44). The authors concluded that upregulation of claudin-4 might represent an approach to treat pulmonary edema. In another report, claudin-4 levels were associated with alveolar fluid clearance in human lungs, whereas this link was not found for other TJ proteins (30). In line with the study mentioned above, these data suggest that claudin-4 might promote alveolar fluid clearance.

As a hallmark of ALI and ARDS, lung permeability is substantially increased because of inflammatory, infectious, traumatic, or toxic insults, resulting in capillary leak, pulmonary edema, and subsequent respiratory failure (42). The well-characterized IgGIC model of ALI resembles and shares many characteristic features of ALI appearing clinically, including activation of the complement system and generation of C5a and PMNs in BAL fluids (9, 14). The IgGIC model and the LPS model represent useful tools to investigate mechanisms of the inflammatory response in the lung (10, 17, 29). In the present study, blockade of zonulin resulted in an attenuation not only of lung permeability but also in the extent of PMN accumulation and appearance of inflammatory cytokines in BAL fluids, indicating that zonulin may be involved in early events during the pathogenesis of ALI (Fig. 1, 2). In line with this, zonulin has recently been described as an endogenous regulator of the human respiratory epithelium in vitro (31). The fact that ALI was not entirely suppressed when zonulin was blocked may be because of the specific pharmacokinetics of the compounds (AT-1001) and antibody (anti-Zot) used or might indicate that mechanisms in addition to those involving zonulin are also involved in directly or indirectly increasing lung permeability in ALI. Interestingly, when the zonulin agonist (AT-1002) was given intratracheally, lung permeability increased during ALI (Fig. 3B), supporting the conclusion that zonulin plays an important role in the regulation of lung permeability in health and disease. The permeability-enhancing effects of AT-1002 in intestinal and bronchial epithelial cells have been shown to be protease-activated receptor 2-dependent (31). However, AT-1002, in contrast to zonulin, is not known to activate pathways of the inflammatory response, such as the complement system. As an explanation as to why AT-1002 significantly intensifies ALI in the IgGIC model, but does not increase lung permeability when administered in lungs of healthy animals, it is conceivable that the combination of both local inflammation and increased paracellular TJ permeability is required for the robust (detectable) increase in lung permeability.

In a recent report, zonulin has been identified as preHP2, the precursor of haptoglobin2 (39). Previously, the only function assigned to HPs was the binding of free hemoglobin (Hb) to prevent Hb-induced oxidative tissue damage (2). HPs are tetrameric plasma glycoproteins composed of α- and β-polypeptide chains linked by disulfide bonds (16). In contrast to the β-chain, two different forms of the α-chain exist (α1- and α2-chains). The presence of one or both of the α-chains determines the three different HP phenotypes HP1–1, HP2–1, and HP2–2. From a phylogenetical perspective, all HP variants evolved from the MASPs and are also linked to immunoglobulins (4, 23). MASPs (MASP-1, -2, -3) are known to initiate the lectin pathway of the complement system by cleavage of C4 and C2 after binding to lectins, such as mannose-binding lectin and ficolin (24, 25). As an important component of innate immunity, the complement system plays a pivotal role in defending against invading pathogenic microorganisms. This system can be activated through three different pathways and is known be a phylogenetically ancient part of the immune system. A primitive form of the complement system can be traced back to invertebrates, such as sea urchins, suggesting an evolutionary history of the complement system of more than 500 million years, whereas the antibody-mediated classical activation branch and the lytic terminal pathway arose in the jawed vertebrates with the development of adaptive immunity much later (45). C1r and C1s of the classical pathway and the MASPs of the lectin pathway are crucial for initiation of the proteolytic activation cascades of each of the three pathways. C1r and C1s not only belong to the same serine-protease family of chymotrypsins as MASPs but are also closely related structurally to haptoglobin (38). Other members of the MASP family include a series of plasminogen-related growth factors, such as epidermal growth factor (EGF) and hepatocyte growth factor. Interestingly, preHP2 (but not its cleaved mature form HP2) has recently been demonstrated to transactivate the EGF receptor (39). However, with respect to the mechanism of action, there is some discrepancy between our findings (Fig. 3A) and previously published data by Tripathi et al. (39) showing that tryspin cleavage of preHP2 (zonulin) into HP2 results in a clear reduction of its barrier-deteriorating effect. In the present study, administration of preHP2 in healthy lungs resulted in a significant increase of lung permeability, whereas, in contrast to the above-mentioned study (39), tryspin cleavage of preHP2 (zonulin) into HP2 only leads to a slight loss of this functional activity in vivo. While this implies that secondary effects might also contribute to permeability regulation by zonulin in the lung, it is also conceivable that, in this setting, additional direct but currently unknown mechanisms besides EGFR activation might be involved. As indicated in Figs. 5 and 6, indirect mechanisms of increasing lung permeability by zonulin may comprise activation of the inflammatory response.

In the present study, zonulin was found to be an activator of the complement system in vivo and in vitro (Figs. 5 and 6). The observation that trypsin-treated zonulin partially lost its complement activation activity further supports the notion that zonulin and mature α2-chain HP2 exert distinct biological functions (39). Based on the evolutionary relationship between HPs, MASPs, C1r, and C1s, zonulin may activate the complement system via the lectin and possibly through the classical pathways, but the exact mechanisms are not currently known. However, it is also conceivable that zonulin-induced complement activation may represent a novel pathway of complement activation.

The present study corroborates the in vivo role of TJ in the development of ALI. Using a unique approach that allows modulation of permeability, our findings indicate for the first time that zonulin is likely involved in the regulation of lung permeability in the ALI lung. As another novel aspect, the role of zonulin in ALI links the regulation of permeability with the inflammatory response through direct activation of the complement system.


This study was supported by National Institutes of Health Grants GM-029507, GM-061656 (P. A. Ward), HL-007517-29 (J. J. Grailer), NS-075757 (A. V. Andjelkovic), and DK-048373 (A. Fasano).


The authors have no potential or actual financial conflicts of interest to disclose. No conflicts of interest, financial or otherwise are declared by the authors.


Author contributions: D.R., A.F., and P.A.W. conception and design of research; D.R., M.A.F., B.A.N., D.E.D., J.J.G., F.S.Z., and A.V.A. performed experiments; D.R., B.A.N., D.E.D., M.S.H.-L., A.F., and P.A.W. analyzed data; D.R., M.S.H.-L., A.F., and P.A.W. interpreted results of experiments; D.R., B.A.N., D.E.D., and A.V.A. prepared figures; D.R. drafted manuscript; D.R., M.A.F., M.S.H.-L., J.J.G., F.S.Z., A.V.A., A.F., and P.A.W. edited and revised manuscript; D.R., M.A.F., B.A.N., D.E.D., M.S.H.-L., J.J.G., F.S.Z., A.V.A., A.F., and P.A.W. approved final version of manuscript.


We thank Beverly Schumann and Sue Scott for assistance in the preparation of the manuscript and Robin Kunkel for assistance in the preparation of the figures.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
View Abstract