Advertisement

Peptides Derived from the Tight Junction Protein CLDN1 Disrupt the Skin Barrier and Promote Responsiveness to an Epicutaneous Vaccine

Open ArchivePublished:August 02, 2019DOI:https://doi.org/10.1016/j.jid.2019.06.145
      Keratinocytes express many pattern recognition receptors that enhance the skin’s adaptive immune response to epicutaneous antigens. We have shown that these pattern recognition receptors are expressed below tight junctions (TJ), strongly implicating TJ disruption as a critical step in antigen responsiveness. To disrupt TJs, we designed peptides inspired by the first extracellular loop of the TJ transmembrane protein CLDN1. These peptides transiently disrupted TJs in the human lung epithelial cell line 16HBE and delayed TJ formation in primary human keratinocytes. Building on these observations, we tested whether vaccinating mice with an epicutaneous influenza patch containing TJ-disrupting peptides was an effective strategy to elicit an immunogenic response. Application of a TJ-disrupting peptide patch resulted in barrier disruption as measured by increased transepithelial water loss. We observed a significant increase in antigen-specific antibodies when we applied patches with TJ-disrupting peptide plus antigen (influenza hemagglutinin) in either a patch-prime or a patch-boost model. Collectively, these observations demonstrate that our designed peptides perturb TJs in human lung as well as human and murine skin epithelium, enabling epicutaneous vaccine delivery. We anticipate that this approach could obviate currently used needle-based vaccination methods that require administration by health care workers and biohazard waste removal.

      Abbreviations:

      HA (hemagglutinin), IM (intramuscular), PHFK (primary human foreskin keratinocyte), TER (transepithelial electrical resistance), TEWL (transepidermal water loss), TJ (tight junction), TJDP (tight junction–disrupting peptide)

      Introduction

      An intact skin barrier is important for good health, functioning to restrict exposure of environmental toxins, antigens, and pathogens to the immune system (
      • De Benedetto A.
      • Kubo A.
      • Beck L.A.
      Skin barrier disruption: a requirement for allergen sensitization?.
      ,
      • Kubo A.
      • Nagao K.
      • Amagai M.
      Epidermal barrier dysfunction and cutaneous sensitization in atopic diseases.
      ,
      • O'Neill C.A.
      • Garrod D.
      Tight junction proteins and the epidermis.
      ). As such, it impedes transepidermal delivery of therapeutic agents and vaccines. Current methods of vaccination primarily rely on intramuscular (IM), subcutaneous, and intradermal injection of antigens. These vaccination routes, although effective, require medical personnel to deliver, generate biohazards (sharps) requiring disposal, and cause patients pain and anxiety. This has fueled research efforts to identify needle-free methods of immunization.
      A number of epicutaneous vaccine delivery systems have been explored, including electroporation and microneedle-based techniques (
      • Leone M.
      • Mönkäre J.
      • Bouwstra J.A.
      • Kersten G.
      Dissolving microneedle patches for dermal vaccination.
      ,
      • Levin Y.
      • Kochba E.
      • Kenney R.
      Clinical evaluation of a novel microneedle device for intradermal delivery of an influenza vaccine: are all delivery methods the same?.
      ,
      • Todorova B.
      • Adam L.
      • Culina S.
      • Boisgard R.
      • Martinon F.
      • Cosma A.
      • et al.
      Electroporation as a vaccine delivery system and a natural adjuvant to intradermal administration of plasmid DNA in macaques.
      ). Although producing notable successes, each of these methods suffers from complications that make them challenging to implement on a large scale for mass vaccination strategies. Electroporation of antigens into the skin requires expensive machinery, and microneedles suffer from inadequate antigen loading, poor reproducibility, incomplete dissolution of microneedles, and costly manufacturing (
      • Ita K.
      Transdermal delivery of vaccines - Recent progress and critical issues.
      ).
      In the skin, tight junctions (TJs) and the stratum corneum act together to maintain a formidable epidermal barrier (Figure 1). TJs are composed of claudin proteins that control barrier formation by homodimerizing on adjacent cells through extracellular loop domain interaction (
      • Haftek M.
      • Callejon S.
      • Sandjeu Y.
      • Padois K.
      • Falson F.
      • Pirot F.
      • et al.
      Compartmentalization of the human stratum corneum by persistent tight junction-like structures.
      ,
      • Sugawara T.
      • Iwamoto N.
      • Akashi M.
      • Kojima T.
      • Hisatsune J.
      • Sugai M.
      • et al.
      Tight junction dysfunction in the stratum granulosum leads to aberrant stratum corneum barrier function in claudin-1-deficient mice.
      ,
      • Yoshida K.
      • Yokouchi M.
      • Nagao K.
      • Ishii K.
      • Amagai M.
      • Kubo A.
      Functional tight junction barrier localizes in the second layer of the stratum granulosum of human epidermis.
      ). The composition of these extracellular loops determines the claudin function, ranging from a tight seal to a more leaky channel (
      • Günzel D.
      Claudins: vital partners in transcellular and paracellular transport coupling.
      ). Disruption of TJs and, specifically, reduced expression of CLDN1 is a key feature of human and canine atopic dermatitis (
      • De Benedetto A.
      • Rafaels N.M.
      • McGirt L.Y.
      • Ivanov A.I.
      • Georas S.N.
      • Cheadle C.
      • et al.
      Tight junction defects in patients with atopic dermatitis.
      ,
      • De Benedetto A.
      • Slifka M.K.
      • Rafaels N.M.
      • Kuo I.H.
      • Georas S.N.
      • Boguniewicz M.
      • et al.
      Reductions in claudin-1 may enhance susceptibility to herpes simplex virus 1 infections in atopic dermatitis.
      ,
      • Roussel A.J.
      • Bruet V.
      • Marsella R.
      • Knol A.C.
      • Bourdeau P.J.
      Tight junction proteins in the canine epidermis: a pilot study on their distribution in normal and in high IgE-producing canines.
      ,
      • Tokumasu R.
      • Tamura A.
      • Tsukita S.
      Time- and dose-dependent claudin contribution to biological functions: lessons from claudin-1 in skin.
      ). TJ disruption results in increased movement of molecules and viruses via the paracellular route both into and out of the lower levels of the epidermis (
      • De Benedetto A.
      • Slifka M.K.
      • Rafaels N.M.
      • Kuo I.H.
      • Georas S.N.
      • Boguniewicz M.
      • et al.
      Reductions in claudin-1 may enhance susceptibility to herpes simplex virus 1 infections in atopic dermatitis.
      ). This process can be measured by transepithelial water loss (TEWL) or tracer flux through the epidermis, both of which are increased when TJs are disrupted (
      • Furuse M.
      • Hata M.
      • Furuse K.
      • Yoshida Y.
      • Haratake A.
      • Sugitani Y.
      • et al.
      Claudin-based tight junctions are crucial for the mammalian epidermal barrier: a lesson from claudin-1-deficient mice.
      ). Cldn1 knockout mice have extensive TEWL, indicating that the protein is essential for skin barrier integrity (
      • Furuse M.
      • Hata M.
      • Furuse K.
      • Yoshida Y.
      • Haratake A.
      • Sugitani Y.
      • et al.
      Claudin-based tight junctions are crucial for the mammalian epidermal barrier: a lesson from claudin-1-deficient mice.
      ). A less dramatic variation of this is seen in humans with atopic dermatitis; these patients have been shown to have greater TEWL and paracellular permeability than healthy subjects. This is thought to be due to reductions in CLDN1, with expression levels ∼50% lower than in nonatopic controls (
      • De Benedetto A.
      • Rafaels N.M.
      • McGirt L.Y.
      • Ivanov A.I.
      • Georas S.N.
      • Cheadle C.
      • et al.
      Tight junction defects in patients with atopic dermatitis.
      ). Therefore, we hypothesized that by targeting CLDN1, we could disrupt TJ function, which would enhance paracellular permeability and facilitate greater epicutaneous adaptive immune reactivity. Importantly, other groups have shown that synthetic peptides derived from the sequence of the extracellular loops of transmembrane TJ proteins, such as claudins and OCLN, are able to disrupt barrier function at high concentrations (
      • Baumgartner H.K.
      • Beeman N.
      • Hodges R.S.
      • Neville M.C.
      A D-peptide analog of the second extracellular loop of claudin-3 and -4 leads to mislocalized claudin and cellular apoptosis in mammary epithelial cells.
      ,
      • Beeman N.
      • Webb P.G.
      • Baumgartner H.K.
      Occludin is required for apoptosis when claudin-claudin interactions are disrupted.
      ,
      • Mrsny R.J.
      • Brown G.T.
      • Gerner-Smidt K.
      • Buret A.G.
      • Meddings J.B.
      • Quan C.
      • et al.
      A key claudin extracellular loop domain is critical for epithelial barrier integrity.
      ,
      • Wong V.
      • Gumbiner B.M.
      A synthetic peptide corresponding to the extracellular domain of occludin perturbs the tight junction permeability barrier.
      ,
      • Zwanziger D.
      • Hackel D.
      • Staat C.
      • Böcker A.
      • Brack A.
      • Beyermann M.
      • et al.
      A peptidomimetic tight junction modulator to improve regional analgesia.
      ,
      • Zwanziger D.
      • Staat C.
      • Andjelkovic A.V.
      • Blasig I.E.
      Claudin-derived peptides are internalized via specific endocytosis pathways.
      ).
      Figure thumbnail gr1
      Figure 1Schematic of TJs in the epidermis, CLDN1 in TJs, and peptides used in this study. (a) In the epidermis, TJs (magenta) form a paracellular barrier between keratinocytes in the SG. The SC and TJ provide barrier function for the skin. Langerhans cells are shown in orange. (b) CLDN1 self-assembles and interacts with TJ proteins through extracellular loops. Peptide 1 represents half of the first extracellular loop of human CLDN1 (light blue), with a Cys to Ser mutation (hCLDN1 [53–81, C54,64S]). Peptide 2 consists of the same amino acids but altered sequence order. Peptide 3 and 4 are alterations of Peptide 2 that reduce the number of charged residues or remove charge completely, respectively. BM, basement membrane; Cys, cysteine; SB, stratum basale; SC, stratum corneum; Ser, serine; SG, stratum granulosum; SS, stratum spinosum; TJ, tight junction.
      We synthesized a peptide (Peptide 1) derived from amino acid residues 53–81 of the first extracellular loop of human CLDN1, which has extensive homology with mouse CLDN1, containing only one amino acid change (serine to asparagine at position 74). Importantly, we replaced several cysteine residues in the peptide (Figure 1) with serine to prevent complications arising from disulfide bond formation. This domain was chosen as a target for TJ disruption because the first extracellular loop has been shown to facilitate transepithelial electrical resistance (TER) development and determine ion permeability selectivity (
      • Colegio O.R.
      • Van Itallie C.
      • Rahner C.
      • Anderson J.M.
      Claudin extracellular domains determine paracellular charge selectivity and resistance but not tight junction fibril architecture.
      ,
      • Colegio O.R.
      • Van Itallie C.M.
      • McCrea H.J.
      • Rahner C.
      • Anderson J.M.
      Claudins create charge-selective channels in the paracellular pathway between epithelial cells.
      ). As a test of the importance of sequence order versus overall amino acid identity, we also synthesized a peptide with the same amino acid composition as Peptide 1, but with an altered sequence; we call this Peptide 2. Additionally, Peptides 3 and 4 (Figure 1) were designed to test if reducing the number of charged residues while retaining the net charge of the peptide (Peptide 3) or reducing the net charge to zero (Peptide 4) diminished the ability to disrupt TJs.
      When all four of these peptides were used in an epithelial cell model derived from human lung cells (16HBE), they showed robust TJ disruption. Peptide 2, which showed the most robust TJ disruption in 16HBE cells, also was able to significantly delay TJ formation in primary human foreskin keratinocytes (PHFKs). To extend these findings, an in vivo patch-based delivery system was developed to determine if Peptide 2 could enhance an immune response against the viral protein influenza hemagglutinin (HA) in a mouse vaccination model. A perturbed skin barrier was observed by increased TEWL after application of a patch containing TJ-disrupting peptides (TJDPs). Furthermore, serological studies indicated that antigens delivered in tandem with a TJDP had an enhanced humoral immune response. Overall, this work establishes the validity of using TJ disruption as a method to deliver antigens epicutaneously, suggesting that this approach may be useful as a vaccine or drug delivery method.

      Results

      Peptides 1 and 2 strongly aggregated when prepared directly in buffer or cell culture media. Screening a series of surfactants to facilitate formation of stable peptide structure to improve handling revealed that the inclusion of 0.12% Pluronic F127 allowed for solubilization and subsequent dilution of Peptides 1 to 4 into cell culture media without precipitation (
      • Khattak S.F.
      • Bhatia S.R.
      • Roberts S.C.
      Pluronic F127 as a cell encapsulation material: utilization of membrane-stabilizing agents.
      ).
      Initial experiments employed the human bronchial epithelial cell line 16HBE, because this cell line is known for its robust TJ formation (
      • Saatian B.
      • Rezaee F.
      • Desando S.
      • Emo J.
      • Chapman T.
      • Knowlden S.
      • et al.
      Interleukin-4 and interleukin-13 cause barrier dysfunction in human airway epithelial cells.
      ). In the media used for peptide exposure, 16HBE cells achieved a stable TER of 800–1200 Ω/cm2 (Supplementary Table S1). Immunofluorescence staining revealed reactivity for the TJ molecules CLDN1, CLDN4, OCLN, and ZO-1 at the periphery of epidermal cells with a honeycomb-like appearance (Supplementary Figure S1). The honeycomb fluorescence pattern of these proteins on the cell surface is a hallmark of a barrier-competent TJ (
      • Furuse M.
      • Fujita K.
      • Hiiragi T.
      • Fujimoto K.
      • Tsukita S.
      Claudin-1 and -2: novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin.
      ,
      • Furuse M.
      • Hirase T.
      • Itoh M.
      • Nagafuchi A.
      • Yonemura S.
      • Tsukita S.
      • et al.
      Occludin : a novel integral membrane-protein localizing at tight junctions.
      ). Following TJ formation, cells were exposed to peptides (0.4 to 50 μM) or vehicle control. Peptides 1 and 2 significantly decreased TER (P < 0.01 and P < 0.0001, respectively) (Figure 2a). The disruption was dose-dependent, with both peptides eliciting minimal disruption at 0.4 μM (2–3%) and increasing reductions at higher concentrations. Peak disruption was observed at day 3 for Peptide 1 (72% decrease at 50 μM), with even greater disruption observed with Peptide 2 at day 2 (90% decrease at 50 μM). Peptides 3 and 4 were less effective at disrupting the barrier, with maximum TER reductions of 68% for Peptide 3 on day 3 and 74% for Peptide 4 on day 2 when used at 30 μM (Supplementary Figure S2). Importantly, TJs recovered after peptide washout except at the highest dose (50 μM), and the Pluronic F127 surfactant vehicle control did not disrupt TJs. A control peptide, (FKFE)2, tested in the same concentration range (12–96 μM) also did not affect TER values, suggesting that our TJDPs did not nonspecifically affect barrier integrity (Supplementary Figure S3). Using a water-soluble tetrazolium-1 assay, we observed no loss of viability after treatment of 16HBE with 50 μM of Peptides 1 or 2, the highest concentration tested (Figure 2b). We observed enhanced TJ penetration of a labeled monoclonal antibody (palivizumab, 150 kDa) in peptide-treated cells (Figure 2c) (
      • Anderson E.J.
      • Carosone-Link P.
      • Yogev R.
      • Yi J.
      • Simões E.A.F.
      Effectiveness of palivizumab in high-risk infants and children: A propensity score weighted regression analysis.
      ). Antibody penetration was enhanced 2.5 ± 0.2-fold and 3.2 ± 0.4-fold by 2.4 and 12 μM of peptide 1, respectively. We observed greater permeability after Peptide 2 exposure, with 4.9 ± 0.4-fold and 5.6 ± 0.9-fold enhancement using 2.4 and 12 μM, respectively. The antibody diffusion was measured at 30 minutes to reduce the likelihood that we were measuring active transcellular transport.
      Figure thumbnail gr2
      Figure 2TJDPs decrease barrier function in lung epithelial cells in the absence of cytotoxicity and enable protein diffusion. (a) Peptides 1 and 2 were used to disrupt 16HBE cells after they had formed TJs. TER was measured daily to observe both disruption and recovery (Peptide 1, n = 3; Peptide 2, n = 10). (b) Viability changes as a result of peptide exposure were measured at day 1, 2, and 4 (recovery) using the WST-1 assay (Peptide 1, n = 3; Peptide 2, n = 6). (c) 24 hours after disruption, a mAb was applied, and diffusion through the monolayer was determined 30 minutes or 18 hours later (n = 4). Error bars represent SD. Significance was calculated compared to vehicle control using (a, b) the Kruskal-Wallis analysis and (c) Mann-Whitney t-test with Prism software v8.0. Asterisk color signifies peptide concentration that is significant. mAb, monoclonal antibody; SD, standard deviation; TER, transepithelial electrical resistance; TJ, tight junction; TJDP, tight junction–disrupting peptide; WST-1, water-soluble tetrazolium-1.
      To validate peptide-mediated TJ disruption in skin, PHFKs were isolated and propagated from neonatal human foreskins. PHFKs form TJs as measured by TER following differentiation in high calcium media (day 3, 140–450 Ω/cm2, Supplementary Table S1). To determine if TJs could be perturbed during differentiation, PHFKs were treated with Peptide 2 upon addition of high calcium media. Peptide 2 was chosen for all further studies given its robust phenotype in 16HBE. Notably, when PHFKs were differentiated (three days after high calcium media), they became refractory to TJ disruption by peptide treatment (data not shown). In contrast, cells treated with Peptide 2 at initiation of differentiation experienced a significant delay in TJ formation for five days (P < 0.001) (Figure 3a). We observed recovery of barrier equivalent to media-treated controls after peptide removal at day 3. Modest toxicity was observed at higher concentrations by water-soluble tetrazolium-1 assay (30 μM, P < 0.05), but at lower concentrations, where robust inhibition of TJ formation was observed, no appreciable cell death was detected (Figure 3b).
      Figure thumbnail gr3
      Figure 3TJDPs delay barrier formation in PHFKs without eliciting cytotoxicity. (a) Peptide 2 was used to disrupt TJs in PHFKs during differentiation. Cells were differentiated (media containing 1.8 mM calcium) in the presence of TJDPs for 3 days after which media was replaced. TER was measured over six days to observe disruption and recovery kinetics. Data were normalized to the media controls (n = 7–13). (b) Viability changes resulting from Peptide 2 exposure were measured at days 1, 2, and 4 (recovery) using the WST-1 assay (n = 5–9). Error bars represent SD. Significance was calculated compared to vehicle control using the Kruskal-Wallis analysis with Prism software v8.0. Asterisk color signifies peptide concentration that is significant. PHFK, primary human foreskin keratinocyte; SD, standard deviation; TER, transepithelial electrical resistance; TJ, tight junction; TJDP, tight junction–disrupting peptide; WST-1, water-soluble tetrazolium-1.
      TJ-associated proteins were then examined by immunofluorescence staining of PHFKs to determine whether peptide exposure resulted in changes to the appearance or distribution of the key TJ transmembrane proteins CLDN1 and OCLN (Figure 4a). At two days after peptide treatment, PHFKs appeared to have a higher intensity of CLDN1 distributed throughout the monolayer, with some of the staining observed within the cytoplasm (Figure 4a top panels, Supplementary Figure S4). At four days after peptide treatment, there was a notable lack of honeycomb OCLN staining compared with control cells, consistent with functional changes in TJs (Figure 4a bottom panels). Four days after Peptide 2 treatment, PHFKs appeared to have CLDN1 distribution similar to control cells at two days, demonstrating a peptide-induced delay in TJ organization during barrier formation (Figure 4a bottom panels). Quantification of DAPI-stained nuclei showed minimal changes in cellular density with mean values of 557, 623, and 527 cells per image for media, vehicle, and Peptide 2 images, respectively, supporting the conclusion that peptide treatment does not affect barrier function by altering proliferation of cells in the monolayer (Figure 4b). Significant reduction in the magnitude of OCLN staining was observed in cells treated with Peptide 2 at day 2 and 4 (Figure 4b, P < 0.05 and P < 0.01, respectively). Interestingly, greater CLDN1 intensity per number of DAPI+ PHFKs was observed at two and four days after peptide treatment, suggesting aberrant accumulation of TJ proteins not typically seen during PHFK differentiation.
      Figure thumbnail gr4
      Figure 4TJDPs alter staining of TJ proteins (OCLN and CLDN1) critical for the establishment of skin barrier function. (a) Cells were exposed to Peptide 2 (10 μM), vehicle, or media alone containing calcium (1.8 mM), which initiates differentiation. At two and four (recovery) days post-differentiation, cells were stained for CLDN1 and OCLN (TJ proteins) and nuclei (DAPI). (b) Ten images from each condition were quantified for the number of DAPI+ cells (left), amount of OCLN-covered area (center), and level of CLDN1 intensity per cell using ImageJ from a representative donor (n = 3). The white bar indicates a 50-μm distance. Error bars represent SD. Significance was calculated using the Kruskal-Wallis analysis with Prism software v8.0. SD, standard deviation; TJ, tight junction; TJDP, tight junction–disrupting peptide.
      We developed an epicutaneous patch delivery system to determine whether our TJDPs could impair barrier function in murine skin. To accomplish this, patches containing either Peptide 2 or vehicle control were applied to the flanks of animals three days following removal of fur. Patches were removed 18 hours later, and TEWL was measured at 1, 3, and 24 hours as an indicator of barrier disruption. Patches containing Peptide 2 significantly (P < 0.01) increased TEWL at 1 and 3 hours after patch removal (1.8- and 1.6-fold, respectively) compared with patches containing vehicle control, with barrier function returning to baseline after 24 hours (Figure 5, Supplementary Figure S5).
      Figure thumbnail gr5
      Figure 5TJDPs reduce barrier function of murine skin. Female Balb/c mice, 8–10 weeks old, were shaved and treated with a depilatory cream. Animals were then rested for three days before TJDP treatment. To disrupt the skin barrier, Peptide 2 (7.8 nmol/cm2) was added to a filter paper (patch) and then applied to mouse skin using a Tegaderm dressing on the right flank (n = 10). A vehicle-laden control patch was attached to the left flank of the same animal. After 18 hours, the patch was removed and TEWL was measured 1, 3, and 24 hours later. Lines connect TEWL measurements from a single mouse on either the vehicle- or peptide-treated flank. Significance was calculated using the paired Wilcoxon t-test with Prism software v8.0. TEWL, transepidermal water loss; TJDP, tight junction–disrupting peptide; n.s., not significant.
      Building on this result, we investigated whether TJ disruption achieved with our peptides was sufficient to promote immunologic responsiveness to an epicutaneously applied antigen. We used influenza A HA as a model antigen in two models, a patch-based prime followed by IM boost or an IM prime followed by a patch-based boost. These two models were chosen to simulate a naïve response (patch-prime) to an antigen or, as occurs with influenza, seasonal boosting of a pre-existing response (patch-boost). No detectable response was measured after a patch-prime containing either vehicle control or TJDP with HA (Figure 6a). Upon boosting mice with an IM delivery of HA, animals that received a TJDP-containing patch during primary exposure to the antigen had significantly (P < 0.05) increased HA-specific IgG antibody titers compared with vehicle control, with a mean antibody endpoint titer of 33,000 compared with 6,400 (day 38 post-boost; Figure 6a). To determine whether patch-based delivery could boost pre-existing immunity to an antigen (as is done during annual influenza vaccine campaigns), we primed animals with an IM immunization of influenza and followed with patch delivery of HA. Animals boosted with HA in a patch containing Peptide 2 had significant increases in the antibody response observed as early as 14 days post-boost. This response was delayed compared with animals that received the boost by IM delivery of antigen, but the same level of antibody titers was achieved by day 28. Additionally, the capability of antibodies elicited in our patch-boost experiments to neutralize the virus was measured at day 35 by hemagglutination inhibition assay, which showed comparable titers to IM control animals (Figure 6b). Because all animals began with similar antibody titers, this strongly implicates TJ disruption as a key event promoting robust and protective antigen-specific responses comparable to IM immunization.
      Figure thumbnail gr6
      Figure 6TJDPs can prime and boost the immune system to epicutaneously delivered influenza HA. (a) Mice were primed with a patch containing 2 μg HA and 7.8 or 0.78 nmol/cm2 of Peptide 2 or vehicle at day 0, 1, and 2 (n = 3). HA was delivered by IM injection as a positive control (n = 1). Animals were boosted intramuscularly with 1 μg of inactivated influenza virus 21 days later. (b) Animals were primed by IM injection and then patch-boosted (n = 4) or IM-injected (n = 3) as a positive control. Anti-HA serum antibodies were measured before and after boost. HAI titers were measured to determine whether protective antibodies were elicited. Error bars represent SD. Significance was calculated using the Mann-Whitney t-test with Prism software v8.0. HA, hemagglutinin; HAI, hemagglutination inhibition; IM, intramuscular; SD, standard deviation; TJDP, tight junction–disrupting peptide.

      Discussion

      The results of this study suggest an alternative method of vaccine delivery. The method proposed here, TJ disruption, avoids all complications of needle-based delivery, because a patch-based delivery is painless, can be dried (avoids refrigeration), and is easily applied. We have demonstrated that TJDPs based on the first loop of CLDN1 (Figure 1) disrupt barrier function in a lung epithelial cell line in the absence of cytotoxicity, and this disruption is significant enough to allow the diffusion of large molecular weight proteins (150 kDa) (Figure 2). Using primary epidermal cells, TJDPs were able to delay barrier formation without impacting cell viability (Figure 3). To further characterize this barrier disruption, PHFK monolayers were treated with Peptide 2 and visualized for TJ protein immunoreactivity during differentiation. In peptide-treated PHFKs, OCLN staining was substantially delayed and CLDN1 staining was mainly detectable in the cytoplasm (Figure 4). These observations suggest that TJDPs perturb barrier function, at least in part, by altering the expression and/or localization of key TJ transmembrane proteins.
      Our murine studies confirmed that TJDPs do, in fact, disrupt the skin barrier, as measured by increased TEWL. Importantly, this effect was transient, with TEWL recovering to near baseline values within 24 hours (Figure 5). To determine the biological consequences of epidermal disruption as a noninvasive vaccination method, we tested whether a patch with a viral antigen and our TJDPs could (i) prime the naïve immune system and/or (ii) boost pre-existing immunity to a protein. Studies aimed at priming the naïve immune system to HA antigen failed to demonstrate detectable levels of antigen-specific antibodies but still established memory, as was observed by enhanced antibody responses after an IM boost. Of note, even vehicle delivery of protein in a patch elicited a boost response, suggesting that skin occlusion may be sufficient to deliver an antigen to the murine immune system, even in the context of minimal changes in TEWL (Figures 5 and 6a). Importantly, in all of our mouse studies, we observed no physical changes in the skin over the 3-month period the mice were observed. This observation suggests that TJ disruption in mouse skin does not promote a diseased state and/or increase infection risks, highlighting the safety of this transepidermal antigen delivery system.
      Humans are exposed to influenza as young as 6 months of age, and as a result most individuals have pre-existing immunity to the virus (
      • Zhou H.
      • Thompson W.W.
      • Viboud C.G.
      • Ringholz C.M.
      • Cheng P.Y.
      • Steiner C.
      • et al.
      Hospitalizations associated with influenza and respiratory syncytial virus in the United States, 1993–2008.
      ). Therefore, the function of seasonal flu shots is to stimulate the expansion of influenza-specific, memory B cells to the likely seasonal strains. To model whether an epicutaneous patch with a TJDP could boost the immune response to influenza, we used an IM injection of inactivated virus to establish a memory, or pre-existing immunity, state. To boost, HA was then delivered by a patch containing a TJDP to skin or by IM injection as positive control. Animals receiving a patch containing TJDP and antigen were observed to have enhanced levels of antigen-specific antibodies similar to the IM control (Figure 6b). Importantly, patch-based delivery of HA stimulated increased hemagglutination inhibition titers, which are a known correlate of protection against influenza (
      • Plotkin S.A.
      Correlates of protection induced by vaccination.
      ). This observation suggests that TJ disruption–based antigen delivery through the skin can elicit antibodies that are biologically significant in protection from influenza.
      The data presented here demonstrate that our TJDP transiently disrupts epithelial barrier function at doses that do not affect viability. Their incorporation into an epicutaneous patch may provide a noninvasive, painless method to administer vaccines quickly and cheaply to a large population. Importantly, multiple groups are attempting to establish a universal flu vaccine that would increase the effectiveness of the current vaccine and possibly negate yearly booster immunizations (
      • Erbelding E.J.
      • Post D.J.
      • Stemmy E.J.
      • Roberts P.C.
      • Augustine A.D.
      • Ferguson S.
      • et al.
      A universal influenza vaccine: the strategic plan for the National Institute of Allergy and Infectious Diseases.
      ,
      • Nachbagauer R.
      • Liu W.C.
      • Choi A.
      • Wohlbold T.J.
      • Atlas T.
      • Rajendran M.
      • et al.
      A universal influenza virus vaccine candidate confers protection against pandemic H1N1 infection in preclinical ferret studies.
      ). One possible method to accomplish this is to stimulate cytotoxic CD8+ T cells specific for conserved epitopes in the virus. Typical IM vaccination is extremely poor at eliciting cellular immunity, but skin-based delivery of antigen has been shown to initiate robust T-cell responses that home to other organs in the body (
      • Liu L.
      • Fuhlbrigge R.C.
      • Karibian K.
      • Tian T.
      • Kupper T.S.
      Dynamic programming of CD8+ T cell trafficking after live viral immunization.
      ,
      • Schmidt S.T.
      • Khadke S.
      • Korsholm K.S.
      • Perrie Y.
      • Rades T.
      • Andersen P.
      • et al.
      The administration route is decisive for the ability of the vaccine adjuvant CAF09 to induce antigen-specific CD8(+) T-cell responses: the immunological consequences of the biodistribution profile.
      ,
      • Zaric M.
      • Becker P.D.
      • Hervouet C.
      • Kalcheva P.
      • Ibarzo Yus B.
      • Cocita C.
      • et al.
      Long-lived tissue resident HIV-1 specific memory CD8+ T cells are generated by skin immunization with live virus vectored microneedle arrays.
      ). Therefore, future studies on the cellular component elicited after TJ disruption–based delivery of antigen could highlight that this method can initiate a universal response to influenza, addressing an important public health concern.

      Materials and Methods

      PHFKs and 16HBE cell culture

      PHFKs were isolated from discarded foreskin tissue. Patient consent for experiments was not required because human tissue left over from surgery was de-identified and considered discarded material. Isolation and propagation procedures for both PHFK and 16HBE cells were done as previously described (
      • Poumay Y.
      • Roland I.H.
      • Leclercq-Smekens M.
      • Leloup R.
      Basal detachment of the epidermis using dispase: tissue spatial organization and fate of integrin alpha 6 beta 4 and hemidesmosomes.
      ,
      • Saatian B.
      • Rezaee F.
      • Desando S.
      • Emo J.
      • Chapman T.
      • Knowlden S.
      • et al.
      Interleukin-4 and interleukin-13 cause barrier dysfunction in human airway epithelial cells.
      ).

      TJDP formulation

      Synthesized TJDP (10 mg, RS Synthesis, Louisville, KY) was dissolved in 100 μl of DMSO (Sigma-Aldrich, St. Louis, MO). Dissolved peptide was then diluted in prewarmed (55 °C) DMSO or Pluronic F127 (Spectrum, New Brunswick, NJ) solution formulated in phosphate buffered saline (0.6% and 0.12%, respectively). This solution was heated at 55 °C for 30 minutes then homogenized into detergent using a water bath sonicator (Branson 2200) for 10 minutes and vortexed for 1 minute. Peptide was stored at 4 °C until use.

      TER measurements and paracellular flux

      TER and paracellular flux were done as previously published (
      • De Benedetto A.
      • Rafaels N.M.
      • McGirt L.Y.
      • Ivanov A.I.
      • Georas S.N.
      • Cheadle C.
      • et al.
      Tight junction defects in patients with atopic dermatitis.
      ). Measurements of TER were taken for up to 6 days following exposure to TJ-disrupting peptides or vehicle control. Fluorescently-labeled antibody (palivizumab, 2 μM) was added to cells after 24 hours of exposure to TJDP, and paracellular flux was measured 30 minutes and 18 hours later.

      Cell viability measurement

      For cytotoxicity measurements, cells were plated at a density of 75,000 cells/well in a 96-well plate and grown to confluence (2 days). Cells were then exposed to vehicle control or TJDP, and viability was measured at 24, 48, and 96 hours. Water-soluble tetrazolium-1 reagent (Roche, Basel, Switzerland) was diluted 20-fold into each well, and cells were incubated at 37 °C. Duplicate readings were taken at 0.5 and 1 hour after addition using a Thermo Multiskan EX plate reader (A450 − background A620). Media-only wells were subtracted and values were normalized to media-treated controls.

      Immunofluorescent staining of TJ formation in PHFKs

      150,000 PHFK or 16HBE cells were plated onto glass coverslips. Cells were grown to confluence over 3 days and treated with TJDP (10 μM), vehicle (0.0015% DMSO or 0.0003% Pluronic F127), or media alone for 48 and 96 hours. Cells were fixed in 4% paraformaldehyde for 10 minutes and washed three times in phosphate buffered saline. Following this, cells were permeabilized with 100% ice cold methanol for 15 minutes at -20 °C, then washed in phosphate buffered saline three times and left overnight at 4 °C. The next day, cells were blocked in 1% BSA dissolved in phosphate buffered saline for 1 hour and stained with anti-CLDN1 and anti-OCLN (Invitrogen, Carlsbad, CA; 500- or 300-fold dilution, respectively) for 2.5 hours at room temperature. Primary antibodies were detected with anti-mouse Alexa Fluor568 and anti-rabbit Alexa Fluor488 (Life Technologies, Carlsbad, CA) and nuclei with DAPI (Invitrogen), all at a 1,000-fold dilution. Coverslips were mounted onto glass slides with 15 μl of ProLong Gold Antifade Reagent (Invitrogen).

      Quantification of TJ protein levels and DAPI Staining by fluorescence microscopy

      PHFK and 16HBE slides were imaged on an Olympus BX60 fluorescent microscope equipped with SPOT RT3 (Diagnostic Instruments, Sterling Heights, MI). Images were processed with ImageJ software. To quantify DAPI+ foci, background was first removed using the threshold function and then converted to black and white using the binary tool. Individual cells were highlighted with the watershed function. Nuclei-sized pixels were then counted with the analyze particles function set to a pixel2 range of 500–infinity to exclude signals too small to be nuclei. To quantify CLDN1 intensity, images were processed as described above. The measure function was then used to give a mean and standard deviation of CLDN1 intensity. To account for variation in the monolayer, each image was divided into quadrants using the rectangle tool and analyzed to confirm homogeneity across a single image. To quantify area covered by OCLN, the image was processed as described above. OCLN signal was selected using the create selection tool, and OCLN-positive area was measured with the analyze (measure) function. The entire area of the image was measured using the same process without creating a selection and used to calculate area of a single image covered by OCLN. Composite images were generated with ImageJ software. Background fluorescence was minimized with the threshold function to enhance signal-to-noise, and DAPI, CLDN1, and OCLN channels were overlaid and pseudocolored blue, green, and red, respectively.

      Patch treatment and TEWL measurements of mouse skin

      All animal studies were approved by the University of Rochester’s committee on animal resources (protocol 2017-017) in accordance with the NIH Guide for the Care and Use of Laboratory Animals. Female Balb/c mice (8–10 weeks old) were anesthetized with an intraperitoneal injection of ketamine (100 mg/kg, Mylan, Canonsburg, PA) and xylazine (20 mg/kg solution, Akorn, Lake Forest, IL) in saline solution (Hospira, Lake Forest, IL). Hair was removed from both flanks by shaving (Oster, Boca Raton, FL) and application of a depilatory cream (VEET, Reckitt Benckiser, Slough, United Kingdom). Animals were rested for three days and then anesthetized again for patch application. Patches were created using a square 0.64 cm2 piece of filter paper (IQ Chamber, Chemotechnique Diagnostics, Vellinge, Sweden) applied to a 2 cm2 piece of Tegaderm dressing (3M, Maplewood, MN). TJDPs (7.8 or 0.78 nmol/cm2) or vehicle control were then applied to filter paper and allowed to absorb before application. Peptide-treated patches were affixed to the mouse’s right flank, whereas vehicle-treated patches were placed on the left flank. After 18 hours, patches were removed and the skin was allowed to dry before TEWL was measured at 1, 3, and 24 hours after patch removal using a Tewameter TM 300 (Courage & Khazaka Electronic, Koln, Germany).

      Animal immunizations

      Mice received either 2 μg of recombinant A/Cal/07/2009 HA (provided by Dr. Florian Krammer, Mount Sinai, NY) or 1 μg of beta propiolactone–inactivated influenza virus. Immunization was delivered either by patch application as stated previously or a 50 μl IM injection into the flank muscle. Animals were boosted either by patch or injection 3–4 weeks later and then killed at 5–6 weeks post-boost.

      HA-specific ELISA and hemagglutination inhibition analysis of mouse sera

      HA antigen–specific antibodies and hemagglutination inhibition titer were measured from immunized mouse serum as previously published (
      • Brewer M.G.
      • DiPiazza A.
      • Acklin J.
      • Feng C.
      • Sant A.J.
      • Dewhurst S.
      Nanoparticles decorated with viral antigens are more immunogenic at low surface density.
      ,
      • Nogales A.
      • Piepenbrink M.S.
      • Wang J.
      • Ortega S.
      • Basu M.
      • Fucile C.F.
      • et al.
      A highly potent and broadly neutralizing H1 Influenza-Specific Human Monoclonal Antibody.
      ).

      Statistical analysis

      All statistical tests were done using GraphPad Prism v8.0. Statistical significance was calculated using either a nonparametric t-test or analysis of variance (specified in figure legends). Values of P < 0.05 were considered statistically significant.

      Data availability statement

      No datasets were generated or analyzed during this study.

      Conflict of Interest

      Peptides 1 and 2 are the subject of US Patent 9,757,428, on which EAA, ADB, LAB, and BLM are listed as inventors. Peptides 3 and 4 are the subject of a US Provisional Patent application. LAB received honoraria as a consultant for Abbvie, Astra-Zeneca, Allakos, Boehringer-Ingelheim, Celgene, Eli Lilly, GI Innovation, GSK, Incyte, Leo Pharma, Medimmune, Meiji Seika Pharma, Novan, Novartis, Realm Therapeutics, Regeneron, Sanofi, and UCB; received grants for clinical trials from Abbvie, Pfizer, Realm Therapeutics and Regeneron; and has Pfizer and Medtronics stock. All other authors state no conflict of interest.

      Acknowledgments

      The authors would like to acknowledge Florian Krammer at the Mount Sinai School of Medicine for providing recombinant hemagglutinin and Stephen Dewhurst at the University of Rochester for inactivated influenza virus. Funding was partially provided by NHLBI T32 grant number HL66988 (TAH) . Funding was partially provided by NIAMS R21 grant number AR062357 and  Atopic Dermatitis Research Network (ADRN) grant numbers HHSN272201000020C and HHSN272201000017C (LAB) . Funding was also provided by a UR (University of Rochester) Ventures technology transfer award (LAB, BLM).

      Author Contributions

      Conceptualization: MGB, EAA, ADB, LAB, BLM; Data Curation: MGB, RPP, EAA, TAH; Formal Analysis: MGB, EAA, RPP, LAB, BLM; Funding Acquisition: LAB, BLM; Resources: TAH, LMS; Writing - Original Draft Preparation: MGB, EAA, LAB, BLM; Writing - Review and Editing: MGB, LAB, BLM

      Supplementary Material

      Figure thumbnail fx1
      Supplementary Figure S1Immunofluorescence microscopy of TJ proteins in 16HBE cells. Untreated 16HBE cells stained for (a) CLDN1 (green) and ZO-1 (red), and (b) CLDN4 (green) and OCLN (red). The white bar indicates a 25-μm distance. TJ, tight junction.
      Figure thumbnail fx2
      Supplementary Figure S2TJDPs decrease barrier function in lung epithelial cells with varying efficiency. (a) Peptides 3 and 4 (derived from the CLDN1 sequence) were used to disrupt 16HBE cells after they had formed TJs. Cells were exposed to multiple concentrations of TJDPs for two days and then new media was added. TER was measured over the course of four days to observe both disruption and recovery kinetics. Data were normalized to the media control (n = 3). Error bars represent SD. Significance was calculated compared to vehicle control using the Kruskal-Wallis analysis with Prism software v8.0. Asterisk color signifies peptide concentration that is significant. SD, standard deviation; TER, transepidermal electrical resistance; TJ, tight junction; TJDP, tight junction–disrupting peptide.
      Figure thumbnail fx3
      Supplementary Figure S3TJDPs have specificity for disrupting the skin barrier. Peptides 1 and 2 and an (FKFE)2 peptide were used to disrupt 16HBE cells after they had formed TJs. Cells were exposed to different concentrations of peptide (12 or 96 μM) and TER was measured over the course of 24 hours to observe the kinetics of disruption. Data is presented as average Ω/cm2 measured (n = 2). TER, transepidermal electrical resistance; TJ, tight junction; TJDP, tight junction–disrupting peptide.
      Figure thumbnail fx4
      Supplementary Figure S4TJDPs alter staining of TJ proteins (OCLN and CLDN1) critical for the establishment of skin barrier function. Representative images are higher magnification pictures of the images from to better show distribution of TJ proteins. The white bar indicates a 50-μm distance. TJ, tight junction; TJDP, tight junction–disrupting peptide.
      Figure thumbnail fx5
      Supplementary Figure S5Change in TEWL after mice are treated with either a patch containing Peptide 2 or vehicle. Shown are the raw TEWL values in grams/hour/meter2 from . The gray line indicates the baseline average of both sites on the mouse (TEWL of ∼9 g/h/m2). TEWL, transepidermal water loss.

      Supplementary Data

      References

        • Anderson E.J.
        • Carosone-Link P.
        • Yogev R.
        • Yi J.
        • Simões E.A.F.
        Effectiveness of palivizumab in high-risk infants and children: A propensity score weighted regression analysis.
        Pediatr Infect Dis J. 2017; 36: 699-704
        • Baumgartner H.K.
        • Beeman N.
        • Hodges R.S.
        • Neville M.C.
        A D-peptide analog of the second extracellular loop of claudin-3 and -4 leads to mislocalized claudin and cellular apoptosis in mammary epithelial cells.
        Chem Biol Drug Des. 2011; 77: 124-136
        • Beeman N.
        • Webb P.G.
        • Baumgartner H.K.
        Occludin is required for apoptosis when claudin-claudin interactions are disrupted.
        Cell Death Dis. 2012; 3: e273
        • Brewer M.G.
        • DiPiazza A.
        • Acklin J.
        • Feng C.
        • Sant A.J.
        • Dewhurst S.
        Nanoparticles decorated with viral antigens are more immunogenic at low surface density.
        Vaccine. 2017; 35: 774-781
        • Colegio O.R.
        • Van Itallie C.
        • Rahner C.
        • Anderson J.M.
        Claudin extracellular domains determine paracellular charge selectivity and resistance but not tight junction fibril architecture.
        Am J Physiol Cell Physiol. 2003; 284: C1346-C1354
        • Colegio O.R.
        • Van Itallie C.M.
        • McCrea H.J.
        • Rahner C.
        • Anderson J.M.
        Claudins create charge-selective channels in the paracellular pathway between epithelial cells.
        Am J Physiol Cell Physiol. 2002; 283: C142-C147
        • De Benedetto A.
        • Kubo A.
        • Beck L.A.
        Skin barrier disruption: a requirement for allergen sensitization?.
        J Invest Dermatol. 2012; 132: 949-963
        • De Benedetto A.
        • Rafaels N.M.
        • McGirt L.Y.
        • Ivanov A.I.
        • Georas S.N.
        • Cheadle C.
        • et al.
        Tight junction defects in patients with atopic dermatitis.
        J Allergy Clin Immunol. 2011; (127-786:e771–777)
        • De Benedetto A.
        • Slifka M.K.
        • Rafaels N.M.
        • Kuo I.H.
        • Georas S.N.
        • Boguniewicz M.
        • et al.
        Reductions in claudin-1 may enhance susceptibility to herpes simplex virus 1 infections in atopic dermatitis.
        J Allergy Clin Immunol. 2011; 128: 242-246.e5
        • Erbelding E.J.
        • Post D.J.
        • Stemmy E.J.
        • Roberts P.C.
        • Augustine A.D.
        • Ferguson S.
        • et al.
        A universal influenza vaccine: the strategic plan for the National Institute of Allergy and Infectious Diseases.
        J Infect Dis. 2018; 218: 347-354
        • Furuse M.
        • Fujita K.
        • Hiiragi T.
        • Fujimoto K.
        • Tsukita S.
        Claudin-1 and -2: novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin.
        J Cell Biol. 1998; 141: 1539-1550
        • Furuse M.
        • Hata M.
        • Furuse K.
        • Yoshida Y.
        • Haratake A.
        • Sugitani Y.
        • et al.
        Claudin-based tight junctions are crucial for the mammalian epidermal barrier: a lesson from claudin-1-deficient mice.
        J Cell Biol. 2002; 156: 1099-1111
        • Furuse M.
        • Hirase T.
        • Itoh M.
        • Nagafuchi A.
        • Yonemura S.
        • Tsukita S.
        • et al.
        Occludin : a novel integral membrane-protein localizing at tight junctions.
        J Cell Biol. 1993; 123: 1777-1788
        • Günzel D.
        Claudins: vital partners in transcellular and paracellular transport coupling.
        Pflugers Arch. 2017; 469: 35-44
        • Haftek M.
        • Callejon S.
        • Sandjeu Y.
        • Padois K.
        • Falson F.
        • Pirot F.
        • et al.
        Compartmentalization of the human stratum corneum by persistent tight junction-like structures.
        Exp Dermatol. 2011; 20: 617-621
        • Ita K.
        Transdermal delivery of vaccines - Recent progress and critical issues.
        Biomed Pharmacother. 2016; 83: 1080-1088
        • Khattak S.F.
        • Bhatia S.R.
        • Roberts S.C.
        Pluronic F127 as a cell encapsulation material: utilization of membrane-stabilizing agents.
        Tissue Eng. 2005; 11: 974-983
        • Kubo A.
        • Nagao K.
        • Amagai M.
        Epidermal barrier dysfunction and cutaneous sensitization in atopic diseases.
        J Clin Invest. 2012; 122: 440-447
        • Leone M.
        • Mönkäre J.
        • Bouwstra J.A.
        • Kersten G.
        Dissolving microneedle patches for dermal vaccination.
        Pharm Res. 2017; 34: 2223-2240
        • Levin Y.
        • Kochba E.
        • Kenney R.
        Clinical evaluation of a novel microneedle device for intradermal delivery of an influenza vaccine: are all delivery methods the same?.
        Vaccine. 2014; 32: 4249-4252
        • Liu L.
        • Fuhlbrigge R.C.
        • Karibian K.
        • Tian T.
        • Kupper T.S.
        Dynamic programming of CD8+ T cell trafficking after live viral immunization.
        Immunity. 2006; 25: 511-520
        • Mrsny R.J.
        • Brown G.T.
        • Gerner-Smidt K.
        • Buret A.G.
        • Meddings J.B.
        • Quan C.
        • et al.
        A key claudin extracellular loop domain is critical for epithelial barrier integrity.
        Am J Pathol. 2008; 172: 905-915
        • Nachbagauer R.
        • Liu W.C.
        • Choi A.
        • Wohlbold T.J.
        • Atlas T.
        • Rajendran M.
        • et al.
        A universal influenza virus vaccine candidate confers protection against pandemic H1N1 infection in preclinical ferret studies.
        NPJ Vaccines. 2017; 2: 26
        • Nogales A.
        • Piepenbrink M.S.
        • Wang J.
        • Ortega S.
        • Basu M.
        • Fucile C.F.
        • et al.
        A highly potent and broadly neutralizing H1 Influenza-Specific Human Monoclonal Antibody.
        Sci Rep. 2018; 8: 4374
        • O'Neill C.A.
        • Garrod D.
        Tight junction proteins and the epidermis.
        Exp Dermatol. 2011; 20: 88-91
        • Plotkin S.A.
        Correlates of protection induced by vaccination.
        Clin Vaccine Immunol. 2010; 17: 1055-1065
        • Poumay Y.
        • Roland I.H.
        • Leclercq-Smekens M.
        • Leloup R.
        Basal detachment of the epidermis using dispase: tissue spatial organization and fate of integrin alpha 6 beta 4 and hemidesmosomes.
        J Invest Dermatol. 1994; 102: 111-117
        • Roussel A.J.
        • Bruet V.
        • Marsella R.
        • Knol A.C.
        • Bourdeau P.J.
        Tight junction proteins in the canine epidermis: a pilot study on their distribution in normal and in high IgE-producing canines.
        Can J Vet Res. 2015; 79: 46-51
        • Saatian B.
        • Rezaee F.
        • Desando S.
        • Emo J.
        • Chapman T.
        • Knowlden S.
        • et al.
        Interleukin-4 and interleukin-13 cause barrier dysfunction in human airway epithelial cells.
        Tissue Barriers. 2013; 1: e24333
        • Schmidt S.T.
        • Khadke S.
        • Korsholm K.S.
        • Perrie Y.
        • Rades T.
        • Andersen P.
        • et al.
        The administration route is decisive for the ability of the vaccine adjuvant CAF09 to induce antigen-specific CD8(+) T-cell responses: the immunological consequences of the biodistribution profile.
        J Control Release. 2016; 239: 107-117
        • Sugawara T.
        • Iwamoto N.
        • Akashi M.
        • Kojima T.
        • Hisatsune J.
        • Sugai M.
        • et al.
        Tight junction dysfunction in the stratum granulosum leads to aberrant stratum corneum barrier function in claudin-1-deficient mice.
        J Dermatol Sci. 2013; 70: 12-18
        • Todorova B.
        • Adam L.
        • Culina S.
        • Boisgard R.
        • Martinon F.
        • Cosma A.
        • et al.
        Electroporation as a vaccine delivery system and a natural adjuvant to intradermal administration of plasmid DNA in macaques.
        Sci Rep. 2017; 7: 4122
        • Tokumasu R.
        • Tamura A.
        • Tsukita S.
        Time- and dose-dependent claudin contribution to biological functions: lessons from claudin-1 in skin.
        Tissue Barriers. 2017; 5: e1336194
        • Wong V.
        • Gumbiner B.M.
        A synthetic peptide corresponding to the extracellular domain of occludin perturbs the tight junction permeability barrier.
        J Cell Biol. 1997; 136: 399-409
        • Yoshida K.
        • Yokouchi M.
        • Nagao K.
        • Ishii K.
        • Amagai M.
        • Kubo A.
        Functional tight junction barrier localizes in the second layer of the stratum granulosum of human epidermis.
        J Dermatol Sci. 2013; 71: 89-99
        • Zaric M.
        • Becker P.D.
        • Hervouet C.
        • Kalcheva P.
        • Ibarzo Yus B.
        • Cocita C.
        • et al.
        Long-lived tissue resident HIV-1 specific memory CD8+ T cells are generated by skin immunization with live virus vectored microneedle arrays.
        J Control Release. 2017; 268: 166-175
        • Zhou H.
        • Thompson W.W.
        • Viboud C.G.
        • Ringholz C.M.
        • Cheng P.Y.
        • Steiner C.
        • et al.
        Hospitalizations associated with influenza and respiratory syncytial virus in the United States, 1993–2008.
        Clin Infect Dis. 2012; 54: 1427-1436
        • Zwanziger D.
        • Hackel D.
        • Staat C.
        • Böcker A.
        • Brack A.
        • Beyermann M.
        • et al.
        A peptidomimetic tight junction modulator to improve regional analgesia.
        Mol Pharm. 2012; 9: 1785-1794
        • Zwanziger D.
        • Staat C.
        • Andjelkovic A.V.
        • Blasig I.E.
        Claudin-derived peptides are internalized via specific endocytosis pathways.
        Ann N Y Acad Sci. 2012; 1257: 29-37