<p>The human stomach normally contains high levels of bioactive nitric oxide (NO). This NO derives from salivary nitrate (NO₃^-) that is converted to nitrite (NO₂^-) by oral bacteria and thereafter non-enzymatically reduced in the acidic gastric lumen to NO. Nitrate is a common component in vegetables, and after ingestion it is absorbed in the small intestine. Interestingly, circulating nitrate is then concentrated by the salivary glands. Hence, intake of nitrate-rich vegetables results in high levels of NO in the stomach. The physiological effects of the high concentration of NO gas normally present in the gastric lumen have been hitherto unknown, and the present investigations were therefore conducted to address this issue.</p><p>NO produced in the gastric lumen after nitrate ingestion increased gastric mucosal blood flow and the thickness of the firmly adherent mucus layer in the stomach. The blood flow and mucus layer are essential defense mechanisms that protect the mucosa from luminal acid and noxious agents. Nonsteroidal antiinflammatory drugs (NSAID) are commonly prescribed and effective drugs for treating pain and inflammation, but are associated with severe gastrointestinal side effects. We demonstrated that a nitrate-rich diet protects against NSAID-induced gastric damage, as a result of the increased formation of NO in the stomach. We also showed that the gastroprotective effect attributed to nitrate depended completely on conversion of nitrate to nitrite by the bacterial flora colonizing the tongue, and that the oral microflora is therefore important in regulating physiological conditions in the stomach.</p><p>In summary, this thesis challenges the current dogma that nitrate intake is hazardous, and on the contrary suggests that dietary nitrate plays a direct role in regulating gastric homeostasis. It is likely that a sufficient supply of nitrate in the diet together with the oral microflora is essential for preventing pathological conditions in the gastrointestinal tract.</p>

The human stomach normally contains high levels of bioactive nitric oxide (NO). This NO derives from salivary nitrate (NO3-) that is converted to nitrite (NO2-) by oral bacteria and thereafter non-enzymatically reduced in the acidic gastric lumen to NO. Nitrate is a common component in vegetables, and after ingestion it is absorbed in the small intestine. Interestingly, circulating nitrate is then concentrated by the salivary glands. Hence, intake of nitrate-rich vegetables results in high levels of NO in the stomach. The physiological effects of the high concentration of NO gas normally present in the gastric lumen have been hitherto unknown, and the present investigations were therefore conducted to address this issue. NO produced in the gastric lumen after nitrate ingestion increased gastric mucosal blood flow and the thickness of the firmly adherent mucus layer in the stomach. The blood flow and mucus layer are essential defense mechanisms that protect the mucosa from luminal acid and noxious agents. Nonsteroidal antiinflammatory drugs (NSAID) are commonly prescribed and effective drugs for treating pain and inflammation, but are associated with severe gastrointestinal side effects. We demonstrated that a nitrate-rich diet protects against NSAID-induced gastric damage, as a result of the increased formation of NO in the stomach. We also showed that the gastroprotective effect attributed to nitrate depended completely on conversion of nitrate to nitrite by the bacterial flora colonizing the tongue, and that the oral microflora is therefore important in regulating physiological conditions in the stomach. In summary, this thesis challenges the current dogma that nitrate intake is hazardous, and on the contrary suggests that dietary nitrate plays a direct role in regulating gastric homeostasis. It is likely that a sufficient supply of nitrate in the diet together with the oral microflora is essential for preventing pathological conditions in the gastrointestinal tract.

Master of Science / Department of Clinical Science / Elizabeth Davis / Background: While there is a commercially-available vaccine for Streptococcus equi subsp. equi licensed for the intranasal route of administration, some equine practitioners are administering this vaccine orally despite a lack of evidence for its efficacy by this route of administration. Objectives: To compare systemic and local immune responses following intranasal or oral administration of the USDA-approved, live-attenuated Streptococcus equi subspecies equi vaccine (Pinnacle IN®, Zoetis, Florham Park, New Jersey). Study Design: Experimental, randomized clinical trial Methods: Eight healthy horses with low Streptococcus equi M protein (SeM) titers (<1:1600) were randomly assigned to an intranasal or oral two-vaccine series. SeM-specific serum immunoglobulins G (IgG) and A (IgA) and nasal secretion IgA were assessed using a commercially-available ELISA (Equine Diagnostic Solutions, LLC, Lexington, Kentucky) and a novel magnetic microsphere assay utilizing fluorescence. A general linear mixed models approach was used for statistical data analysis. Results: As expected, intranasal vaccinates showed substantial increases in both serum SeM-specific IgG and IgA levels post-vaccination (P=0.0006 and P=0.007, respectively). Oral vaccinates showed an increase in serum SeM-specific IgG post-vaccination (P=0.0150), though only one-third the magnitude of intranasal vaccinates. Oral vaccinates showed no evidence of change in SeM-specific IgA post-vaccination (P=0.15). Main Limitations: Changes in mucosal antibody responses were not identified in this study which may be related to small change in antibody response, timing of sample collection, or method of nasal secretion collection. Conclusions: Results indicate that intranasal or oral vaccine administration resulted in increased serum SeM-specific IgG, though the magnitude of response differed between routes.

Introduction: Derangement of systemic and mucosal immunity, which are the integral components of the immune system, increases the risk of septic complications in patients postoperatively. The aims of this study were to investigate the integrity of systemic immunity as well as the mucosal immune system in patients with pancreatic cancer (PC), chronic pancreatitis (CP) and obstructive jaundice (OJ).Method: Healthy controls, as well as four groups of patients were studied. These included; jaundiced patients with PC, jaundiced patients secondary to benign disease (choledocholithiasis), non-jaundiced patients with PC and non-jaundiced patients with CP. The study evaluated the nutritional status including anthropometric measurements and the serum proteins: retinal binding protein (RBP), transferrin (TRF) and prealbumin (PALB). This study also evaluated systemic immunity in terms of total lymphocyte count, lymphocyte subsets (CD4+, CD8+, CD25+and CD56+), tumour necrosis factor alpha (TNF- alpha), interleukin-1alpha (IL-1 alpha) and complement components; and mucosal immunity in terms of CD3+, CD4+, CD8+, CD20+, CD57+, CD68+ and mast cells. Results: 78 patients were recruited (including 39 males) as follows: normal controls (n=17), benign OJ (n=9), patients with PC with jaundice (n=23), non-jaundiced patients with PC (n=20) and CP (n=9). Circulating CD25+ and CD4+ were significantly lower in the PC group whereas CD8+ showed increased levels in the same patients with a significant decrease in OJ patients when compared with controls. Circulating CD56+ showed no statistically significant difference between all four groups. In addition, IL-1 and TNF-alpha showed no statistically significant difference in all groups when compared with the control group. Also, C3 and CH50 showed significantly raised levels in PC with jaundice when compared with the control group. On the other hand mucosal lymphocyte subsets showed no statistically significant difference among all groups in comparison with the control group. As for prealbumin and transferrin, both showed significantly low levels in OJ, PC with jaundice and with PC when compared to healthy controls. Survival analysis for both PC groups was carried out and showed no difference in terms of age, however PC patients who survived over 13 months showed increased levels of prealbumin as well as low levels of CH50.Conclusion: Patients with PC both with and without jaundice showed some signs of altered and dysfunctional systemic immunity as well as a reduction in serum proteins. These findings may have implications on the disease progression and postoperative complications. This may warrant therapeutic interventions to restore nutrition and improve immunity before major surgical intervention is planned which could result in improving prognosis.

Au cours des dernières décennies, la technologie des nanoparticules pour la délivrance des vaccins au niveau de muqueuses a reçu un intérêt croissant. L’administration intranasale possède de grands avantages pour la stimulation du système immunitaire, telles que la stimulation d’une immunité protectrice locale et systémique. Cependant des systèmes de délivrance et des adjuvants sont souvent nécessaires pour déclencher efficacement la réponse immunitaire. Nous avons appliqué la technologie des nanoparticules en tant que système de délivrance d'un vaccin universel nasal contre la grippe dans un projet européen FP7 appelé UniVacFlu. Nous avons formulé un antigène adjuvé CTA1-3M2e-DD avec les NPL. Cet antigène est composé de la sous-unité A1 de la toxine du choléra et d’un épitope conservé du virus de la grippe A (M2e), ainsi que du dimère de l’analogue synthétique de la protéine A de Staphylococcus aureus (DD). Les nanoparticules utilisées sont poreuses et constituées de maltodextrines réticulées ayant un coeur lipidique (NPL). L’association de cet antigène avec les NPL est quantitative et la formulation est stable pendant au moins six mois à 4°C. Les NPL permettent également de délivrer d’une manière accrue cet antigène dans les cellules épithéliales des voies respiratoires et les macrophages. Actuellement ces formulations sont évaluées chez la souris par le consortium UniVacFlu.L'un des principaux problèmes des vaccins nasal est la toxicité qui peut être provoquée par le passage nez-cerveau de l'un de ses composants. Le but de ce travail est d'évaluer le potentiel des NPL, en tant que vecteurs pour la délivrance des vaccins nasal. Ainsi, nous avons étudié le chargement d’un antigène dans les NPL et sa délivrance dans les cellules épithéliales des voies respiratoires. Notre étude révèle que les NPL interagissent fortement avec les muqueuses et délivrent d’une manière accrue les antigènes dans les cellules. Nous avons également montré l'absence de transcytose et de passage paracellulaire des NPL ou des antigènes délivrés dans un modèle de barrière épithéliale in vitro. Les résultats in vivo confirment l'absence de passage nez-cerveau des NPL et montrent qu’elles prolongent fortement le temps de résidence nasale des antigènes qui sont ensuite éliminés par le tractus gastro-intestinal.Ces résultats mettent en évidence l'intérêt des NPL comme vecteurs pour la prochaine génération de médicaments et de vaccins. / Nanoparticles technology for mucosal delivery of vaccines received a growing interest in the last decades. Intranasal administration owns great advantages for immune system stimulation, such as local and systemic protection against infectious diseases. However delivery systems and adjuvants are often required to efficiently trigger mucosal and systemic immune responses. In this thesis, nanoparticles (NP) have been evaluated as delivery system for a nasal universal influenza vaccine in a People Program of the European Union Seventh Framework Program FP7 called UniVacFlu. The aim of the UniVacFlu network is to develop a universal influenza vaccine administered through the mucosal route. We used porous maltodextrin nanoparticles with a lipidic core (NPL). We loaded an adjuvanted antigen named CTA1-3M2e-DD in the NPL. CTA1-3M2e-DD is composed of the A1 subunit of the cholera toxin and a conserved epitope of influenza A virus (M2e), while DD, dimer of the synthetic analogue of the Staphyloccous aureus protein A, targets B cells. Interestingly the antigen loading in NPL was quantitative for the antigen: NPL 1:5 mass ratio and the formulation was stable for at least six months at 4°C. We assessed the successful delivery of the antigen by NPL in airway epithelial cells and macrophages. These formulations are currently evaluated by the UniVacFlu consortium in mice.One of the main issues of intranasal vaccines is the toxicity that can be elicited by the nose-brain passage of one of their components. We investigated the loading of antigens in NPL and their delivery in airway mucosa. We observed a high endocytosis of NPL and an increased protein delivery into the cells. On a transwell model of the airway mucosa we assessed the absence of transcytosis and paracellular passage of the NPL. In vivo results confirmed the lack of nose-brain passage of the NPL, as NPL were found not to cross the mucosa. Interestingly, we observed an increased nasal residence time of the protein targeted by NPL. The particles after having delivered their payload are totally eliminated through the gastrointestinal tract, making these nanoparticles good candidates for mucosal delivery system. These results highlight the interest of NPL as vectors for mucosal delivery of drugs.

Photocopy of typescript. / Digitized by Kansas Correctional Industries

Pseudomonas aeruginosa (P. aeruginosa) is a frequently isolated pathogen that causes septicaemia and chronic respiratory infection. It exhibits a higher mortality rate than other gram-negative bacteria and the need for effective immunotherapy is emphasised by the frequency of antibiotic resistance associated with this organism. Mucosal immunisation with a whole killed cell P. aeruginosa vaccine has previously demonstrated a significant immune response in both rodent studies and human trials. This study is a continuation of that research, with the major goal being the identification of a purified protein antigen that could form the basis of a mucosal vaccine against P. aeruginosa. Specifically, the aims of this study were the development of purification protocols for the isolation of previously untested protein antigens, assessment of the efficacy of these antigens to enhance bacterial clearance in an animal model of acute respiratory infection, determination of the immune parameters that are associated with the resolution of P. aeruginosa respiratory infection and finally, cloning of an identified antigen which demonstrated vaccine efficacy. Protocols were established to isolate proteins for use as antigens in immune response studies. The proteins purified in this study were Pa 13, Azurin, acyl carrier protein (ACP), Amidase, Aminopeptidase, KatA and Pa70. These proteins were used to immunise rats by intestinal intra-Peyer's patch (IPP) inoculation and intratracheal (IT) boost. The immunisation protocol employed was designed to target mucosal antigen-specific immune responses where the route of immunisation, Peyer's patch (PP) intestinal inoculation, is akin to the oral delivery of antigens to the gut-associated lymphoid tissue (96). Investigations of a previously uncharacterised antigen, Pa60, later identified this protein as the P. aeruginosa catalase, KatA. This study demonstrated enhanced bacterial clearance of both homologous and heterologous challenge following immunisation with KatA. The level of clearance demonstrated by KatA was promising when compared to that of killed whole cell immunisation. KatA was cloned and studies with the recombinant protein showed enhanced bacterial clearance commensurate with that of the native protein. Immunisations with other proteins identified four additional antigens which enhanced bacterial clearance; Pa13, Pa40, Pa45 and Pa70. Amino acid sequence analysis indicated that Pa13 may be a novel protein, whereas Pa40 was determined to be amidase and Pa45, aminopeptidase. Pa70 was not successfully sequenced. These proteins were effective in significantly enhancing bacterial clearance of homologous P. aeruginosa challenge. For KatA, Pa13 and Pa70, clearance was associated with a marked phagocytic cell recruitment. In contrast, amidase and aminopeptidase demonstrated clearance with a minimal cellular response. Proteins; azurin and ACP were non-protective, failing to clear a live P aeruginosa challenge. Analysis of the antigen-specific responses of these nonprotective proteins and comparison with those antigens which enhanced bacterial clearance were used to determine factors that may contribute to the resolution of an acute pulmonary infection. The study has demonstrated that mucosal immunisation using purified protein antigens can enhance the clearance of pulmonary infection with P. aeruginosa. It has also contributed to the understanding of immune responses to newfound antigens of P. aeruginosa and identified antigen-specific responses which confirm their potential as vaccine candidates.

<p>The lungs and upper airways are mucosal surfaces that are common site for infection with an enormous variety of inhaled pathogens. Therefore, induction of immune responses in the respiratory tract is crucial for protection against respiratory diseases.</p><p>One of the pathogens infecting the host via the respiratory tract is <i>Mycobacterium Tuberculosis</i>. The reported efficacy of the currently used Bacillus Calmette-Guérin (BCG) vaccine against tuberculosis is highly variable, ranging from 50% against pulmonary tuberculosis to 80% against disseminated tuberculosis. Recently, the current route of vaccination (intradermal) has been considered as a possible factor influencing the protective capacity of the BCG vaccine. In this regard, intradermal route most likely induces protective systemic responses while it fails to induce optimal responses in the lungs. Therefore, our working hypothesis is that vaccination should be directed towards the respiratory mucosal immunity in order to improve the degree of host protection in the lungs.</p><p>In this thesis we studied the effect of the route of immunization as well as of different mucosal adjuvants on the induction of mucosal immune responses against the mycobacterial surface antigen PstS-1. We found that, the intranasal (i.n.) route of immunization was a more favorable route inducing strong local immune responses, than intraperitoneal (i.p.) route. Indeed, i.n. route immunization, unlike the i.p. route, elicited strong IgA responses in the lungs accompanied by a major influx of CD4⁺ T cells and a significant local production of IFN-gamma.</p><p>IgA, being the predominant Ig isotype at mucosal tissues, is considered a major effector molecule involved in defense mechanisms against viral and bacterial pathogens at these sites. Therefore, we investigated the possible role of IgA in the protection of the respiratory mucosa against mycobacterial infections, using mice deficient in IgA and in the polymeric Ig receptor. We show that, deficient mice are more susceptible to mycobacterial infections than wild type mice, thereby demonstrating a role for IgA in protection against mycobacteria. Importantly, our studies revealed a reduced production of protective factors, such as INF-gamma and TNF-alpha in the lungs of deficient mice that was associated with the higher susceptibility seen in these mice compared to wild-type mice. We also conducted challenge experiments against another respiratory pathogen, <i>Chlamydia pneumoniae</i>, using IgA deficient mice. Likewise to mycobacteria, our data support a role for IgA in the protection of the respiratory tract against <i>C. pneumoniae</i> infection.</p><p>Finally, we investigated the possible mechanisms explaining the reduced pro-inflammatory responses in IgA deficient mice. Our data indicated that IgA deficient mice present a defective response to stimulation with LPS or 19kDa which appears to be both, essentially due to suboptimal stimulation of macrophages and restricted to the lungs.</p>

This thesis aimed to the identification of immune biomarkers of mycobacterial infection for better diagnosis of tuberculosis (TB) and also focused on new vaccination strategies with a particular emphasis on the immune responses in the respiratory tract using murine models. Since the lung is the natural habitat for the M. tuberculosis, we reasoned that immune responses detected locally in the lungs would be good correlates of infection (Paper I). Likewise, immune responses induced in the respiratory tract following immunization would be more effective against mycobacterial infection. We showed that cytokines (IL-12, TNF, and IFN-γ) and cytokine receptors (sTNFR1 and sTNFR2) together with specific antibodies in the respiratory tract correlated better with the bacterial burden in the organs. In Paper II, we investigated the role of the BCG vaccination as a priming vaccine in a heterologous prime-boost immunization protocol. The results showed that the neonatal BCG vaccination primed the immune system for a relevant antigen and showed a generalized adjuvant effect. Using this immunization protocol, protective immune responses in the lungs were generated independently of the route used for the booster immunization. In Paper III, We showed that exposure to mycobacterial antigens during the gestational period led to antigen transportation from the mother to the fetus and this resulted in an early priming of the fetal immune system. Immunization with the same antigen during the postnatal life increased antigen-specific recall IFN-γ responses and protection against infection. We examined the role of innate immunity for the induction of acquired immune responses upon immunization with mycobacterial antigens using TLR2 deficient mice (Paper IV). Our data indicated that suboptimal innate immune responses in the TLR2-/- mice might compromise the induction of acquired immune responses. Overall, the current findings suggested that a better understanding of the mucosal immunity would be useful for the improvement of diagnostic procedures and the development of efficient vaccines against TB. / At the time of the doctoral defense, the following paper was unpublished and had a status as follows: Paper 4: Manuscript

One of the fundamental questions in mucosal immunology is how the intestine maintains tolerance to food antigens and commensal flora, and yet it is capable of mounting immune responses to pathogens. Peyers patches (PP) are lymphoid aggregates that are found in the small intestine and are the primary sites where adaptive immune responses are initiated in the intestine. An understanding of how PP cells regulate innate immune responses may provide information on how immune responses are regulated in the intestine. The toll-like receptors (TLRs) are a family of pattern recognition receptors (PRR) which provide a sensory mechanism for the detection of infectious threats. TLR9 recognizes bacterial DNA or synthetic CpG oligodeoxynucleotides (ODN). Cells that express TLR9 when stimulated with CpG ODN proliferate and produce Th1-like pro-inflammatory cytokines and upregulate co-stimulatory molecules. Because the intestine is constantly exposed to bacterial DNA from commensal flora, immune cells from the gut must have evolved mechanisms to modulate responses to TLR9 stimulation to prevent responses to harmless bacteria. Our hypothesis is that innate immune responses to the TLR9 agonist CpG ODN in Peyers patches (PP) are attenuated compared to other tissues such as blood or lymph nodes. This is due to local regulatory mechanisms unique to the intestinal microenvironment.<p> We conducted a number of experiments to test this hypothesis. We initially assessed the immunostimulatory activity of three available classes of CpG ODN in lymph nodes (LN), peripheral blood mononuclear cells (PBMC) and PP since this had not been done in ruminants. We found that CpG ODN induced strong IFNá, IFN-gamma, IL-12, lymphocyte proliferation and NK-like activity in LN and PBMC. In contrast, these responses were significantly less in PP stimulated with CpG ODN. We wondered whether the reduced responses of PP cells to CpG ODN were unique to the TLR9 agonist. For this reason we tested responses of cells from these tissues to poly (I:C), LPS, and single-stranded RNA, which are agonists for TLR3, TLR4, and TLR7/8 respectively. Additionally, we tested combinations of TLRs since others have reported that multiple TLR agonists may induce synergistic responses. All TLR agonists or their combinations either failed to induce detectable responses or the responses were significantly less in PP compared to other tissues. Thus we concluded that PP cells responses to TLR stimulation were attenuated. In all tissues tested, there were no synergistic responses (IFN-alpha, IFN-gamma and lymphocyte proliferation) following stimulation with combinations of agonists. However, there was inhibition of PBMC responses when TLR7/8 agonists were combined with CpG ODN (TLR9 agonist). Importantly, TLR7/8 agonists reduced the CpG-induced proliferative responses in purified blood B cells. Interestingly, ovine B cells constitutively expressed TLR7/8 and TLR9 mRNA, suggesting the potential for cross-talk between the receptors.<p> Interestingly, cell from all isolated tissues [ileal PP (IPP), jejunal PP (JPP), mesenteric LN (mLN) and PBMC] expressed similar levels of TLR9 mRNA, suggesting that the reduced responsiveness to CpG ODN stimulation in PP was not due to a lack of TLR9 expression.<p> Surprisingly, we observed that PP cells spontaneously secreted significant amounts of the immunoregulatory cytokine IL-10. Furthermore, we confirmed that CD21+ B cells were the source of the IL-10. We then examined the role of IL-10 in regulating IFN and IL-12 responses in PP. Neutralization of IL-10 resulted in a significant increase in the numbers of CpG-induced IFNá-secreting cells detected and in IFN-gamma and IL-12 production by PP cells (both follicular and interfollicular lymphocytes). Similarly, depletion of the CD21+ B cells resulted in significant increases in IFNá, IFN-gamma and IL-12 responses. These observations support the conclusion that IL-10-secreting PP CD21+ B cells suppress innate immune responses in PP. Further characterization by flow cytometry revealed that these cells were CD1b-CD5-CD11c-CD72+CD21+ IgM+ B cells. We have proposed that these IL-10-secreting PP CD21+ B cells are a novel subset of regulatory B cells (Bregs).<p> Finally, we examined the capacity of IL-10 secreting B cells (Bregs) to respond to CpG ODN. To achieve this, we compared CD21+ B cells from blood and JPP. Unlike blood CD21+ B cells, CD21+ B cells from JPP proliferated poorly in response to CpG ODN. Moreover, PP CD21+ B cells, unlike blood CD21+ B cells, do not secrete IgM or IL-12 in response to CpG stimulation, although both PP and blood CD21+ B cells express similar level of TLR9 mRNA. Neutralization of IL-10 did not enhance CpG-induced proliferative responses in PP CD21+ B cells. Thus IL-10 does not play a direct role in the hyporesponsiveness of PP CD21+ B cells to CpG ODN. To further explore the mechanism by which PP Bregs fail to respond to CpG ODN stimulation, we used a kinome analysis to determine whether the TLR9 pathway was functional in PP Bregs compared to blood CD21+ B cells. We observed that peptides representing critical adaptor molecules downstream of TLR9 such as IRAK1, TAK1, Casp8, p-38 MAPK, JNK, FOS, IKKá, NF-KB-p65 were not phosphorylated in JPP CD21+ B cells following CpG ODN stimulation. However, in blood CD21+ B cells stimulated with CpG ODN, the same peptides on the array were all highly phosphorylated leading to a functional TLR9 signaling pathway. Thus PP Bregs have evolved mechanisms by which the TLR9 signaling pathway is not activated following exposure to the TLR9 agonist, CpG ODN.<p> In conclusion, we clearly demonstrated that TLR9-induced responses in cells from PP are significantly attenuated. This is a consequence of PP CD21+ B cells (Bregs) that spontaneously secrete IL-10, which in turn conditions an anti-inflammatory environment in this tissue leading to poor cytokine responses to the TLR9 agonist, CpG ODN. Additionally, we show that Bregs are unresponsiveness to TLR9 stimulation. This unresponsiveness is due to regulatory mechanisms in Bregs leading to a dysfunctional TLR9 signaling pathway. These may represent strategies by which PP dampen innate responses to pathogen-associated molecular patterns (PAMPs) in intestinal immune tissues to maintain intestinal immune homeostasis. These conclusions are consistent with our hypothesis that TLR responses in PP cells are attenuated, and this is due to B cell-mediated regulatory mechanisms that are unique to the intestinal microenvironment.