Abstract
This review highlights our current knowledge regarding expression of transient receptor potential (TRP) cation channels in lung endothelium and evidence for their involvement in regulation of lung endothelial permeability. Six mammalian TRP families have been identified and organized on the basis of sequence homology: TRPC (canonical), TRPV (vanilloid), TRPM (melastatin), TRPML (mucolipin), TRPP (polycystin), and TRPA (ankyrin). To date, only TRPC1/4, TRPC6, TRPV4, and TRPM2 have been extensively studied in lung endothelium. Calcium influx through each of these channels has been documented to increase lung endothelial permeability, although their channel-gating mechanisms, downstream signaling mechanisms, and impact on endothelial structure and barrier integrity differ. While other members of the TRPC, TRPV, and TRPM families may be expressed in lung endothelium, we have little or no evidence linking these to regulation of lung endothelial permeability. Further, neither the expression nor functional role(s) of any TRPML, TRPP, and TRPA family members has been studied in lung endothelium. In addition to this assessment organized by TRP channel family, we also discuss TRP channels and lung endothelial permeability from the perspective of lung endothelial heterogeneity, using outcomes of studies focused on TRPC1/4 and TRPV4 channels. The diversity within the TRP channel family and the relative paucity of information regarding roles of a number of these channels in lung endothelium make this field ripe for continued investigation.
INTRODUCTION
The transient receptor potential (TRP) channel superfamily is perhaps the largest known cation channel family contributing to calcium entry in lung endothelium. Since calcium influx through plasmalemmal calcium channels can impair the integrity of the lung endothelial barrier, and thus its permeability to fluid and protein, TRP channels have the potential to be both critical participants in and therapeutic targets for acute lung injury. This review focuses on our current knowledge regarding TRP channel expression in the lung and evidence supporting a role for TRP channels in regulation of lung endothelial barrier integrity and permeability. Changes in lung endothelial permeability due to calcium influx can certainly be modulated by the availability of cyclic nucleotide pools and mediated by myriad downstream signaling pathways. However, these issues have been reviewed at length.1–6 Thus, we focus on TRP channel–mediated calcium entry and loss of endothelial barrier integrity. Given the growing recognition of structural and functional heterogeneity of lung endothelium,7–10 we also discuss the evidence for TRP channels' involvement in regulation of lung endothelial permeability from that perspective as well.
Currently, 27 human TRP isoforms have been identified (28 in mouse), grouped into 6 families based on amino acid homology: TRPC (canonical), TRPV (vanilloid), TRPM (melastatin), TRPML (mucolipin), TRPP (polycystin), and TRPA (ankyrin). A seventh family—TRPN (NO-mechano potential C)—has not been identified in any vertebrate species other than zebra fish. 11 TRP channel proteins contain 6 transmembrane spanning domains, with a pore loop region between transmembrane domains 5 and 6. Domains within the intracellular N- and C-termini contribute to channel assembly and regulation. Individual TRP proteins assemble into homo- or heterotetramers to form functioning cation channels. While Nilius and Owsianik 11 recently suggested that most TRP channels exist as homotetramers, heteromultimeric assembly of several TRP proteins from within or between subfamilies into functional cation channels has been described.12–15 Nonetheless, aside from a recent report using Förster resonance energy transfer (FRET) to document assembly of TRPC1 and TRPC4 into heteromeric channels in pulmonary artery endothelial cells and endothelial caveolar fractions isolated from intact lung, 16 there is very little direct evidence for TRP channel composition and stoichiometry in lung endothelium. To further complicate this issue, protein expression, channel assembly, and trafficking to the plasma membrane may potentially be dictated by developmental stage and/or disease in vivo,17,18 passage of endothelial cells in culture, 19 or the degree of confluency of endothelial monolayers in culture. 20
TRP CHANNELS, CALCIUM ENTRY, AND LUNG ENDOTHELIAL PERMEABILITY
TRPC (canonical)
TRPC family members 1–7 share an invariant amino acid sequence (EWKFAR) in the C-terminus called the TRP box, as well as 3–4 NH2-terminal ankyrin repeat domains. Generally, TRPCs form nonselective cation channels, with calcium-to-sodium selectivity ratios ranging from 1.1 to 9. 21 Protein expression for TRPC1, 3, 4, 6, and 7 has been variously identified in pulmonary vascular endothelium in the intact lung by Western blot or immunohistochemical approaches.18,22–24 In addition, expression of TRPC1, 2, 4, and 6 messenger RNA (mRNA) and protein has been identified in human, mouse, and rat pulmonary artery endothelial cells.7,23,25–30
In terms of activation mechanisms, TRPC channels can be distinguished as those activated by depletion of endoplasmic reticulum (ER) calcium stores (store-operated channels)—TRPC1/4—and those activated by diacylglycerol (DAG)—TRPC3/6/7. While the latter are often termed receptor-operated channels, this is somewhat of a misnomer, as synthesis of both inositol-1, 4, 5-trisphosphate (IP3, an endogenous initiator of store depletion) and DAG can result from receptor-ligand interactions in lung endothelial cells. However, this functional distinction may not clearly delineated between TRPC subfamily groups because of potential heterologous channel assembly, i.e., interaction between TRPCs typically associated with store-operated channels (TRPC1 and 4) and those associated with receptor-operated channels (TRPC3, 6, and 7). Resolution of this issue is complicated by the relative lack of information regarding stoichiometry for TRPC channel assembly in lung endothelium. The only available information has been provided by Cioffi and colleagues,
16
who used a FRET-based reporter system to identify the composition of the channel responsible for the calcium-selective store-operated current (
Agents that evoke depletion of ER calcium stores, resulting in activation of store-operated channels, include IP3 and ER calcium ATPase (SERCA) inhibitors such as the plant alkyloid thapsigargin and cyclopiazonic acid, as well as the calcium chelators 1,2-bis(o-aminophenoxy) ethane-
Regardless of the tools used to elicit store depletion, it is clear that calcium entry into lung endothelium following store depletion increases lung endothelial permeability, both in vitro and in vivo.18,29,35–37 The mechanisms linking store depletion and activation of store-operated channels are actively debated.38–40 However, the increased permeability response to

Scanning electron microscopy in thapsigargin-treated rat lung. Thapsigargin-mediated activation of store-operated calcium entry promoted interendothelial gap formation in pulmonary arteries (
Unlike store-operated calcium entry, activation of TRPC3, 6, and 7 is triggered by DAG and occurs independent of store depletion.
43
Alvarez et al.
18
found that
TRPV (vanilloid)
The TRPV family members (1–6) are broadly classified into 4 groups based on their structure and function: TRPV1/2, TRPV3, TRPV4, and TRPV5/6. The TRPV proteins contain versions of the TRP box, the TRP domain, and 3–6 N-terminal ankyrin repeat domains. TRPV1 has been described as both nonselective and calcium selective.21,44 TRPV2–4 form nonselective channels, with calcium-to-sodium selectivity ratios ranging from ~1 to 10. In contrast, TRPV5 and 6 have been identified as selective calcium channels, with calcium-to-sodium ratios greater than 100. 21 TRPV proteins tend to form homomeric channels,11–13 although interactions between TRPV1 and 2 and between TRPV5 and 6 have been observed by FRET, co-immunoprecipation, and colocalization studies in heterologous expression systems. 13
Expression of TRPV1 has been identified in human pulmonary artery and human lung microvascular endothelial cells by reverse-transcription polymerase chain reaction (RT-PCR).23,45 TRPV1 is activated by the vanilloid capsaicin, capsaicin analogs (e.g., nonivamide, resiniferatoxin, and olvanil), moderate heat (≥43°C), and low pH (≤5.9).21,44,46 TRPV1 can also be activated by endogenous lipid compounds referred to as “endovanilloids,” such as anandamide, 2-arachidonoyl glycerol, and N-arachidonoyl amino acids.47–51 There is little evidence supporting a role for TRPV1 in regulation of lung endothelial permeability. Neither pharmacological silencing nor genetic deletion of TRPV1 affected the edema response to lipopolysaccharide, or LPS.
45
In support of these findings, treatment of isolated rat lungs with the TRPV1 agonist 4α-phorbol-12,13-didecanoate-20 homovanillate (4αPDDHV) had no impact on
TRPV4 protein is expressed in cultured rat pulmonary arterial and microvascular endothelial cells and in endothelial cells from rat lung.7,25 Further, immunohistochemistry reveals prominent expression of TRPV4 in the alveolar septal wall of human, rat, and mouse lung, while expression in extra-alveolar vessels is more variable.
7
TRPV4 can be activated by heat,
53
cell swelling,54,55 mechanical and shear stress,53,55,56 epoxyeicosatrienoic acids (EETs),57,58 and synthetic phorbol esters such as 4α-phorbol-12,13-didecanoate (4αPDD).
59
The impact of TRPV4 activation on

Activation of either TRPV4 or store-operated channels increases endothelial permeability in isolated, perfused mouse lung. Challenge of lungs with either 4α-phorbol-12,13-didecanoate-20 (4αPDD) or thapsigargin (TG) to activate TRPV4 or store-operated transient receptor potential (TRP) channels, respectively, similarly increased endothelial permeability, as measured by the filtration coefficient

Activation of calcium entry via TRPV4, but not that via store-operated channels, elicits lung endothelial barrier disruption via distinct targeting of the alveolar septal barrier. Activation of TRPV4 by 4α-phorbol-12,13-didecanoate-20 (4αPDD) significantly increased the frequency of blebs and breaks in the alveolar septal wall in both rat and mouse lung (not shown), which led to significant increases in the alveolar fluid volume fraction (

TRPV4 activation preferentially targets the alveolar septal network. Isolated rat and mouse lungs treated with 4α-phorbol-12,13-didecanoate-20 to activate TRPV4 were visualized by transmission electron microscopy. The increase in
TRPV4 activation by mechanical stress induces lung injury as well. Jian et al. 61 found that high venous pressure–induced increases in lung permeability were mediated by cytochrome P450 expoxygenase–dependent activation of TRPV4 in the mouse lung. P450 epoxygenases are enzymes that synthesize EETs from arachidonic acid. 62 Thus, these data are consistent with EET-induced activation of TRPV4 noted in vitro.57,58,63 In a model of ventilator-induced lung injury, calcium-dependent increases in lung permeability were exacerbated by heat in TRPV4+/+ mice but were absent in TRPV4−/− mice. This mechanical stress–induced lung injury was also attenuated by use of cytochrome P450 epoxygenase inhibitors. 64 However, the permeability response to ventilator-induced injury in mouse lung appears to specifically require TRPV4 expression in alveolar macrophages, because repletion of wild-type macrophages to lungs of TRPV4−/− mice restored the injury response. 65 Collectively, these studies support a role for EETs in initiating TRPV4 activation and the permeability response to mechanical stress in the intact lung, but they also suggest differences in the cellular repertoire recruited to elicit lung injury. The role of TRPV4 in the permeability response to high airway and vascular pressure in mouse lung has been well documented by us and others.61,64–66 In both experimental paradigms, the permeability response is significantly attenuated by pretreatment with the nonspecific TRPV antagonist ruthenium red and by use of TRPV4−/− mice. However, until recently, extension of these observations to rats and larger mammals has been hampered by the lack of pharmacologic and genetic tools to specifically target TRPV4 in these species. New studies using small-molecule antagonists—GSK2193874 and GSK2263095—that are selective for TRPV4 have confirmed a TRPV4-mediated permeability response to high vascular pressure in rat and canine lung. 67 Further, use of these specific TRPV4 antagonists has provided support for involvement of TRPV4 in the development of pulmonary edema in acute pulmonary venous hypertension induced by aortic banding in rats, as well as the development and resolution of pulmonary edema in a murine model of chronic myocardial infarction. 67
The downstream targets of TRPV4-mediated calcium entry that mediate permeability increases in the lung are unknown. Our observation that interendothelial junctions of capillary endothelial cells in mouse or rat lung appear to be intact after TRPV4 activation (Fig. 4) 7 challenges the established paradigm that increases in lung endothelial permeability are invariably due to endothelial cell retraction and subsequent gap formation. In cultured human umbilical vein endothelial cells, TRPV4 activation leads to dose-dependent cell detachment, 60 leading us to consider activation of proteases to mediate basement membrane degradation, loss of cell-matrix tethering, and/or initiation of cell death processes as potential mechanisms for injury. In preliminary work, we have identified increased active matrix metalloproteinase (MMP) 2 and 9 expression in mouse lungs in concert with TRPV4-induced lung injury, 68 although whether these MMPs actually cause the TRPV4-induced increase in lung permeability remains to be determined. Increases in intracellular calcium, if sufficient, can orchestrate cell injury and death,69,70 and TRPV4 activation is capable of initiating cell death in several tissues.71–73 However, we do not observe activated caspase-3 or positive TUNEL (terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling) staining in mouse lungs where TRPV4-mediated lung injury occurred (P. C. Villalta and M. I. Townsley, unpublished observations). As an alternative, the cell swelling, blebbing, and vacuolation induced by TRPV4 activation resemble structural defects induced by another form of programmed cell death, oncosis. 74 Oncosis, implicated as a mode of cell death in a few models of acute lung injury,75–77 can be initiated by calcium-dependent activation of calpains, a family of calcium-dependent cysteine proteases. 78 Although there are no reports linking calpains and oncosis in TRPV4-induced lung injury, calpain inhibition does attenuate lung interstitial and air space edema during inflammation, 79 suggesting a potential role for deformation of the septal barrier. In vitro, separate reports have demonstrated that cyclic stretch both activates TRPV4 80 and induces calpain-mediated alveolar epithelial barrier disruption. 81 Mechanisms by which TRPV4 activation leads to endothelial barrier disruption in the lung should be further resolved.
Involvement of other TRPV channels in regulation of lung endothelium permeability is uncertain. Although TRPV2 mRNA and protein are indeed expressed in rat lung endothelium,23,25 there is no evidence to date linking this TRP channel to regulation of endothelial permeability. In human pulmonary artery endothelium, mRNA for TRPV3, 5, and 6 was not identified with RT-PCR. 23 No studies to date have assessed expression of these TRPV channels in lung microvascular endothelium.
TRPM (melastatin)
The TRPM (1–8) family members, divided into the TRPM1/3, 2/8, 4/5, and 6/7 groups, lack N-terminal ankyrin repeat domains. 21 TRPM channels vary greatly in calcium selectivity ratio, ranging from those that are not calcium permeable (TRPM4/5) to those with selectivity ratios greater than 3.
Expression of all TRPMs, with the exception of TRPM5, has been verified by RT-PCR in human pulmonary artery endothelial cells.23,82 In addition, protein expression of the alternate splice form TRPM2-s and the full-length protein TRPM2-l, as well as TRPM4, has been confirmed by Western blot in human pulmonary artery endothelial cells.23,82 TRPM2 can be activated in the setting of oxidant stress, although this effect is now recognized to be indirect. Direct activation of TRPM2 is elicited by intracellular adenosine diphosphate ribose (ADPR),83,84 cyclic ADPR, H2O2, 85 nicotinic acid adenine dinucleotide phosphate, 86 and arachidonic acid. 87 Nonetheless, the role of TRPMs in lung endothelial permeability is unclear. Hecquet et al. 82 provided evidence for involvement of TRPM2 in oxidative stress–induced damage to human pulmonary artery endothelium. They determined that treatment of human pulmonary artery endothelial monolayers with sublytic concentrations of H2O2 led to calcium influx and decreased monolayer transendothelial resistance in a dose-dependent fashion. Inhibition of calcium entry by blockade of intracellular ADPR generation or siRNA-mediated suppression of TRPM2-l expression diminished the decrease in monolayer barrier integrity. The authors also confirmed a regulatory mechanism—noted in other cell lines88,89—by which TRPM2-s isoform overexpression modulates the activity of the TRPM2 channel and the increased monolayer permeability response to H2O2. The involvement of TRPM2 in the intact lung is more complex. In a model of endotoxin-induced lung injury, where oxidative stress is prevalent, Di and colleagues 90 found that lung edema increased and survival decreased in TRPM2−/− mice bred into a C57BL/6 background, compared to wild-type controls. However, Hardaker et al. 91 reported that TRPM2 deficiency in a BALB/c mouse strain had no effect on bronchoalveolar-lavage albumin levels following LPS challenge, compared to that in wild-type controls. These discrepancies between the permeability effects of TRPM2 activation in human pulmonary artery endothelial cells and those in mouse lung could be due a number of factors. However, the varied sensitivity to lung injury among different mouse strains may be an important consideration. Notably, BALB/c mice display a lower sensitivity to lung injury induced by LPS or ischemia/reperfusion than do C57BL/6 mice.92,93 Thus, no definitive conclusions can be reached yet regarding the role of TRPM2 in mediating increases in lung endothelial permeability.
TRPML (mucolipin)
The mammalian members of the TRPML family, TRPML1–3, have been identified largely as endolysosomal nonselective cation channels that mediate intraendosomal calcium release in the endocytic pathway.94–96 There is no evidence of TRPML expression or activity in the plasma membrane, and only recently have technological advances allowed for assessment of the channel in lysosomal membranes. 97 TRPML1 mRNA is ubiquitously expressed in all major tissues, and TRPML3 mRNA has been detected in mouse lung. 98 Neither the expression nor the functional role of TRPML channels in lung endothelium is known. Studies are needed to determine whether these channels influence endothelial barrier properties.
TRPP (polycystin)
The TRPP family members TRPP2, 3, and 5 are more homologous to the TRPML family than to any other TRP family. 99 These calcium-regulated, nonselective channels have been extensively studied in renal epithelium, where their role in mechanosensation and involvement in polycystic kidney disease have been reported.99–102 Berrout et al. 103 implicated TRPP2 in a model of traumatic brain injury, where mechanical stretch of brain microvascular endothelial cells promoted calcium influx and actin stress-fiber formation independent of store depletion. These investigators reported protein expression of TRPP2, along with TRPC1, TRPV4, and other TRPs, using immunocytochemistry and Western blot in mouse brain microvascular endothelial cells. Knockdown of TRPP2, TRPC1, or both by siRNA, as well as nonselective inhibition of TRPPs and TRPCs by amiloride and LOE908, respectively, resulted in a decreased calcium response and stress-fiber formation, suggesting that at least TRPP2 and TRPC1 may mediate calcium-dependent rearrangement of the cytoskeleton that could promote increases in endothelial permeability. Indeed, inhibition of calcium-permeable channels in the plasma membrane of brain microvascular endothelial cells, which prevents rises in cytosolic calcium, results in protection against increases in blood-brain barrier permeability.104,105
The channels that may influence the barrier properties of the microvascular component of the blood-brain barrier are highlighted here because there are many parallels between brain and lung microvasculature. In both brain and lung, the microvascular barrier segments are very tight, compared to their arterial counterparts.106–108 Further, cytosolic calcium influx via TRPC 1/4 and TRPV4 in the lung5,7,22,24,35,37,109 and via TRPC and TRPV in brain110–113 leads to a compromise of barrier integrity and increases in endothelial permeability. This may be particularly relevant because TRPP2 has been described as part of a functional, mechanosensitive complex with TRPC1 and TRPV4.114–116 Thus, studies are needed to determine whether TRPP2 is expressed in lung endothelium, whether it interacts with other members of the TRPC and TRPV families known to mediate increases in lung endothelial permeability, and whether its activation per se compromises lung endothelial barrier integrity.
TRPA (ankyrin)
The lone member of the TRPA family, TRPA1, is unique in that it possesses 14–19 ankyrin repeats on its amino terminus. 117 In the lung, TRPA1 mRNA and protein expression has been verified in human A549 cells and human lung fibroblasts by RT-PCR and Western blot, respectively, 118 but an expression profile in lung endothelium is lacking. In cerebral arteries, TRPA1 activation evokes dynamic cytosolic calcium transients and vasodilation.129,120 While TRPA1 appears to play a role in mediating cough and airway inflammation,121–125 no studies to date have explored a functional role for TRPA1 in lung endothelium.
IMPACT OF LUNG ENDOTHELIAL-HETEROGENEITY
Our understanding of the molecular players and mechanisms that mediate lung endothelial permeability responses to TRP channel activation is emerging with a key message: not all endothelial cells, endothelial calcium pools, and functional endothelial responses to calcium entry are cut from the same cloth.
First, the notion that calcium-mediated loss of barrier integrity is solely a consequence of endothelial cell retraction is now being challenged. As discussed above, we have documented increases in lung endothelial permeability in the alveolar septal compartment without the formation of interendothelial gaps, 7 suggesting that other mechanisms are responsible for the loss in barrier integrity.
Second, we have determined that TRP channels and other calcium channels may be coexpressed in the same lung endothelial compartment but still elicit unique downstream signals. Specifically, calcium entry via TRPV4 and that via the voltage-gated T-type channel, both expressed in lung septal capillary endothelium, have divergent effects. Only calcium transients resulting from TRPV4 activation increase lung endothelial permeability, and only those resulting from activation of the T-type channel lead to surface expression of P-selectin. 10
Third, work from our group and others suggests that calcium entry–dependent alterations in lung endothelial barrier integrity can have distinct outcomes for overall lung function based on the targeted endothelial compartment (Fig. 5). Specifically, activation of store-operated TRP channels in extra-alveolar vessels leads to perivascular/peribronchial cuffing and decreased lung compliance. 126 In contrast, activation of TRPV4 leads to alveolar flooding and depressed gas exchange.7,126

Activation of store-operated channels and TRPV4 leads to distinctive leak sites in the extra-alveolar and alveolar septal networks, respectively. Corrosion casting revealed no leak sites in the normal rat lung (
Finally, heterogeneity should be interpreted with recognition of potential signaling- or remodeling-induced endothelial plasticity. This concept can be evidenced acutely with manipulation of signaling pathways and/or chronically with modulation of TRP channel expression. For example, Wu et al.
22
found that acute rolipram pretreatment of rat pulmonary artery and microvascular endothelial cells abolished or unmasked thapsigargin-induced ISOC activation in these cells, respectively. In the intact rat lung, rolipram pretreatment shifted the thapsigargin-induced leak site from the extra-alveolar endothelial compartment to the alveolar septal compartment. Close association of the ER and the plasma membrane has been proposed as a requirement for activation of store-operated TRP channels,
40
and the distance between the ER and the plasma membrane is more than 2-fold greater in pulmonary microvascular endothelial cells than in pulmonary artery endothelium.
9
Interestingly, Wu et al.
109
also found that rolipram treatment leads to translocation of the ER closer to the plasma membrane in pulmonary microvascular endothelial cells. They further demonstrated that dynein, a microtubule motor protein that directs organelle movement away from the plasma membrane, predominates in microvascular endothelial cells, while kinesin, a motor protein that directs organelle movement toward the plasma membrane, predominates in pulmonary artery endothelial cells. Disruption of dynein expression in pulmonary microvascular endothelial cells unmasked thapsigargin-induced
Collectively, we interpret these observations as reflecting endothelial heterogeneity, not only between extra-alveolar and microvascular compartments in lung but also between unique calcium microdomains in any one lung endothelial compartment. Calcium channels expressed in lung endothelium may be coupled to distinct downstream signaling pathways and thus linked to distinct functional outcomes. As a corollary, some TRP channels expressed in lung endothelium are likely not coupled to regulation of lung endothelial permeability.
SUMMARY AND FUTURE DIRECTIONS
From the large library of mammalian TRP channels now identified—28 members to be exact—evidence links a number of TRP channels to regulation of lung endothelial barrier integrity. We are just beginning to appreciate the complex nature by which specific endothelial responses to calcium entry via TRP channels are orchestrated. Thus far, the TRPC1/4 (

Heterogeneity of injury model. The sequential-filling model of pulmonary edema accumulation initially described by Staub and others,128–131 shown at left, is applicable when lungs are challenged with nonspecific agents to increase endothelial permeability. However, given newer evidence on endothelial heterogeneity, we should appreciate that extra-alveolar and alveolar septal endothelium can be differentially targeted. We, and others, have provided evidence for such a differential distribution of effects following activation of store-operated channels and TRPV4.7,35,127 If lung endothelial phenotypic heterogeneity extends to differential, compartment-specific expression of other transient receptor potential (TRP) channels, then their activation may similarly elicit direct edema formation in the extra-alveolar or septal compartments of the lung, thus bypassing the sequential filling model. A: artery; B: bronchiole. Adapted from Townsley et al. 127 A color version of this figure is available online.
Increasing evidence for TRP channel involvement in lung endothelial permeability makes this an exciting opportunity for discovery, and much research is still needed. The set of TRP channels that mediate increases in lung endothelial permeability must be completely identified. Once identified, the physiological and signaling mechanisms involved in mediating the permeability increases should be resolved. In addition, we have yet to fully understand whether permeability responses are modulated by the influx of other ions, such as magnesium and sodium, through activated but nonselective TRP channels. In the context of therapeutic value, a global understanding of the way these channels and their downstream effectors influence alterations in endothelial barrier properties could open up a new array of targets for drug development.
