Bioinformatics analysis of homologies between pathogen antigens, autoantigens and the CFTR cystic fibrosis protein: A role for immunoadsorption therapy?

The cystic fibrosis CFTR chloride channel is involved in pathogen entry into epithelial cells, and provides the glutathione and hypochlorous acid necessary for bactericidal and viricidal actions. CFTR mutations block these effects, diminishing pathogen defence and allowing pathogen accumulation in the extracellular space, where antibody encounter is likely. The pathogen antigens observed in cystic fibrosis (including P. Aeruginosa, S.Aureus and S.Maltophilia proteins) are homologous to the autoantigens reported in cystic fibrosis and all are homologous to the CFTR protein itself. Antibodies to pathogens and autoantigens may also target the CFTR protein, acting as antagonists, further compromising its function. The tripartite relationship between pathogen antigens, autoantigens and the CFTR protein creates a feed forward cycle, diminishing the function of the CFTR protein and increasing the probability of pathogen accumulation and further antibody encounters at every turn. Kegg pathway analysis of the CFTR/autoantigen interactome indicates that the CFTR protein is also involved in pathogen entry pathways, diabetes and pancreatic and gastric acid secretion pathways, in pathways related to cardiac myopathy, and in the gonadotrophin signalling network, all which are relevant to cystic fibrosis. Interruption of this cycle by antigen and antibody adsorption, and possibly by immunosuppressant therapy may perhaps be of clinical benefit in cystic fibrosis.

Cystic fibrosis is a devastating condition caused by mutations in the cystic fibrosis transmembrane conductance regulator CFTR chloride channel. The disease affects many organs resulting in general debilitation but especially targets the respiratory system leading to difficulty in breathing . There is no apparent cure or preventive strategy. The disease appears to have an immune and autoimmune component as antibodies to Saccharomyces cerevisiae and Stenotrophomonas Maltophilia{ and to neutrophil cytoplasmic antigens and bactericidal/permeabilityincreasing protein (BPI) and many other proteins (the adrenoreceptor ADRB2, Calgranulin, heat shock proteins, mucins, myeloperoxidase, rheumatoid factor and tumour necrosis factor, inter alia are observed in many patients Bae, Choi, et al. 2010 5301 /id}. The disease is also influenced by infection. For example Burkholderia infection causes severe respiratory infections in cystic fibrosis patients and is often associated with this condition (LiPuma, 1998, LiPuma, 1998, Coutinho, 2007. Stenotrophomonas maltophilia infection has also been reported to worsen pulmonary symptoms while infection with S.Aureus or P.Aeruginosa are known to decrease the lifespan of cystic fibrosis patients . Many bacteria and viruses cause problems by molecular mimicry of human proteins. When homologous to receptors, they may act as decoys, or when homologous to peptide ligands that may act as dummy ligands or decoy substrates. For example the measles virus V protein is a decoy substrate for IkappaB kinase (Pfaller & Conzelmann, 2008). They may also use the host's cognate receptors to gain entry, as is the case with the AIDS virus and the CCR5 or CXCR4 chemokine receptors . When such mimics are antigenic and homologous to host proteins they may cause problems related to autoimmunity. Such mimicry is extensive (Elde & Malik, 2009)and has been observed between Herpes simplex, a risk factor in Alzheimer's disease, and Alzheimer's disease susceptibility gene products (Carter, 2010b), or between the proteins of the Epstein Barr virus or of gut bacterial flora and multiple sclerosis autoantigens (Westall, 2006, Toussirot & Roudier, 2008 . As reported below, proteins from pathogens implicated in cystic fibrosis, and many others (bacteria, fungi and viruses) are homologous to diverse CFTR mutants. Many of these homologous regions are immunogenic, suggesting an important autoimmune component to cystic fibrosis that may be amenable to therapy.

Results
The immunity spectrum of the CFTR mutants.
The localisation of the mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) that were examined is depicted in Fig 1. Several of these mutations are in regions of high predicted B-cell antigenicity (R171H, G480C, G551D, S895N, K1250A, N1303K) while others are less so (Fig 2).
The CFTR F508 deletion or point mutations can dramatically change the antigenicity, not only of the amino acid concerned, but also of the surrounding peptide as shown in Fig 2. For example the F508Del mutant markedly increases the predicted B-cell antigenicity over a long stretch of amino acids, not only confined to the deleted amino acid and also generates two T cell epitopes that have a higher affinity (1.5 -3fold) than those of the native protein (Fig 3). For the 19 other mutants, B cell antigenicity can be increased, decreased or little changed by the point mutation ( Fig  2).
The ten T cell epitopes (mutant Delta F508 protein) with the highest affinity are homologous to proteins expressed by a number of bacterial species, including S.Aureus. Other noteworthy species containing CFTR epitope homologues included Clostridial species, and Klebsiellae (Table 2, which are known to colonise CF patients (see Table 3). This type of epitope mapping may be of use in identifying novel pathogen suspects that may pose a problem in cystic fibrosis. For example, proteins from B.cereus and Brachyspira species were well represented as CFTR epitope matches (See Table 2).

F508del CFTR homology with autoantigens and pathogen proteins
The F508del mutant protein is homologous to ten autoantigens and four P.Aeruginosa and S.Maltophilia antigens reported in cystic fibrosis. The autoantigens are in turn homologous to proteins from three major pathogens implicated in cystic fibrosis (S.Aureus, P.Aeruginosa and S.Maltophilia) ( Table 4), suggesting that the autoantigens are likely to have been created by antibodies that initially targeted the pathogen proteins. The Blast results for this exercise are available at http://www.polygenicpathways.co.uk/cftrpathant.htm

CFTR/pathogen protein homologies
Several viral or bacterial pathogens colonise cystic fibrosis patients to a much greater extent than observed with the normal population . Many of these pathogens have been reported to worsen symptomatology, for example S.Maltophilia and even to decrease the lifespan of infected patients (S.Aureus or P.Aeruginosa) . These effects are summarised in Table 3.
The heptapeptides surrounding the 19 point mutations, or the octapeptide surrounding the F508del mutation, are all homologous to proteins expressed by S.Aureus, P.Aeruginosa or S.Maltophilia, (Table 5 ) as well as to many other strains (not shown: see website BLASTs) .

P.Aeruginosa and S.Aureus vatches in the mutant CFTR protein
Vatches (Viral mATCHES are short contiguous amino acid stretches covering the entire human proteome that are identical in human, and viral proteins and also in the proteins of other pathogens ( see http://www.polygenicpathways.co.uk/blasts.htm . They are a probable legacy of our evolutionary decent from microorganisms, and of pathogen mimicry of human proteins: Despite chromosomal shuffling over millions of years, the current human DNA can still encode for quite large peptide stretches that are identical to those expressed by pathogen proteins (Carter, 2010d, Carter, 2010a, Carter, 2010b. The S.Aureus and P.Aeruginosa vatches within the CFTR polymutant are shown in Fig 4. The CFTR polymutant displays extensive homology with proteins expressed by these two pathogens. The homologous regions are often within highly immunogenic regions of the CFTR and pathogen proteins, and also cover the CFTR point mutations.

Homology with the native CFTR protein
As the mutations in cystic fibrosis are point mutations, the native protein too is evidently homologous to these same pathogen proteins. However, the pathogen riddance pathways are intact in these cases, and the immune system is not compromised by CFTR mutations. There is no reason to suppose that high levels of pathogen proteins could be attained, or that the host could not appropriately deal with the pathogens . Whether the CFTR mutations increase or decrease homology to pathogens is also perhaps irrelevant, as the hyper colonisation by pathogens would be an expected consequence of any functional mutation (see discussion); an outcome that would favour antibody production that could target any CFTR matching epitopes. As antibodies are able to enter cells , such targeting could be relevant to domains in both the intracellular and extracellular portions of the CFTR protein.

Pathway analysis of the CFTR interactome (Fig 5)
Pathway analysis of protein interaction networks is a powerful tool for divining the functions of particular proteins. Those proteins shown to interact with the CFTR protein, from pSTIING, are shown in Table 8.
Pathway analysis of the CFTR interactome (Table 9) also included the autoantigens reported in cystic fibrosis, as their function is also likely to be compromised by their respective autoantibodies. This pathway analysis clearly demonstrates an important role for the CFTR protein in the immune system and in pathogen invasion (  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  6 chemokine signalling and in lysosomal function which is also related to antigen processing and pathogen destruction ,as well as in chemokine signalling, while others are involved in bacterial invasion and Vibrio infection or pathogen destruction (endocytosis, junctions, phagosomes and lysosomes).These pathways are illustrated in Fig 5. Interaction with viruses in the CFTR interactome.
The virusMINT and HSV-1 interactions showed that a number of the CFTR interacting proteins also interact with viral proteins from the adenovirus and papillomavirus as well as the Epstein-Barr, Herpes simplex, Hepatitis B and C and HIV-1 viruses ( Table 8) , all of which also express proteins with homology to the CFTR protein (Table 7) . In other words, certain viral proteins with homology to CFTR may bind to the same targets as the CFTR protein and, when present, could form an integral part of the CFTR interactome. With the exception of a replete HIV-1 interaction database, viral/human protein networks are not extensively referenced in online databases, and more interactions are likely to exist.
Certain of the CFTR interactome pathways trace out a route that is used by the Herpes simplex virus, and probably other related viruses, during its life cycle. This involves entry and endocytosis, entry and exit to and from lysosomes , phagosomes and nuclei, and interference with protein processing pathways (see http://www.polygenicpathways.co.uk/herpeshost.html for a detailed view). These pathways suggest that the CFTR protein is involved in both bacterial and viral defence (Fig 5).

The pancreas, cardiac myopathy and the vas deferens in cystic fibrosis
Pancreatic insufficiency and diabetes are common features of cystic fibrosis as are cardiac myopathy and related cardiovascular problems (Moss, 1982). Bilateral loss of the vas deferens in men, or of the uterus and vagina in women are also commonly associated with cystic fibrosis . The CFTR/autoantigen pathway analysis indicates that the CFTR protein is involved in pancreatic and gastric acid secretion pathways, in several pathways related to cardiac myopathy, and in the gonadotrophin signalling network, which latter controls the development of the sexual organs. The autoantigens implicated in cystic fibrosis are also members of a signalling network related to diabetes (Table 8; Fig 5). These pathways relate to all of the coexisting conditions described above. The involvement of the CFTR protein in these signalling networks indicates that these associated conditions are a direct result of defects in CFTR signalling.

Discussion
Nearly 2,000 mutations/polymorphisms have been described in cystic fibrosis patients. The most common is the DeltaF508 deletion which is expressed in almost 70% of patients and the G551D, G542X, and R553X mutations are also relatively common . 20 different mutations were covered by this survey. Several mutations, particularly truncations, result in non-expression of the CFTR protein or compromised delivery to the cell surface (Davidson & Porteous, 1998). The bacterial and viral homology is of less direct relevance to these mutants, although defects in the immune and microbial related functions of the CFTR protein would also favour pathogen colonisation and immune dysfunction. These and other mutant proteins result in malfunction of the chloride channel encoded by the CFTR protein, with the resultant pulmonary pathology associated with cystic fibrosis.
In addition to its actions as a chloride channel, CFTR has a number of other properties that are highly relevant to immunity and microbiology. For example it controls the efflux of glutathione which exerts viricidal and bactericidal properties, including the S.Aureus and P.Aeruginosa targets . Glutathione levels are reduced in cystic fibrosis and glutathione aerosols have been reported to ameliorate lung epithelia oxidative stress in cystic fibrosis patients . Clinical trials with glutathione or its prodrugs are ongoing . CFTR is also important in pathogen defence, providing the chloride for the generation of hypochlorous acid by myeloperoxidase in neutrophil phagosomes. This bactericidal mechanism is defective in cystic fibrosis , likely rendered the more so by the presence of myeloperoxidase autoantibodies in cystic fibrosis .
The CFTR protein is also expressed in lymphocytes and negatively regulates the nuclear factor kappa beta (NFKB) and toll receptor (TLR4) mediated innate immune response . The delta F508 mutation has also been shown to inhibit the antigen presentation pathway (Hampton & Stanton, 2010), and autoantigens and other antigens in cystic fibrosis would therefore not be properly processed. CFTR mutations also increase immune activation in mice . In addition to these effects, CFTR is a pattern recognition receptor that recognises P.Aeruginosa . The CFTR protein appears to be involved in P.Aeruginosa ingestion and destruction, as the delta508 mutation in infected transgenic mice increases the pulmonary P.Aeruginosa burden and decreases its clearance. This mutation-related reduced uptake of the pathogen into epithelial cells favours multiplication of P.Aeruginosa within the lungs , The CFTR protein is also an entry portal for Chlamydia Trachomatis, and Salmonella Typhi, but not the closely related murine S. typhimurium and the delta508 mutation also reduces pathogen entry into epithelial cells . C.Trachomatis binding to CFTR also reduces its chloride channel activity . Not all bacteria use the CFTR protein which may itself thus determine which bacteria are 8 most likely to be present in large quantities. The Kegg pathway analysis of the CFTR binding proteins also revealed a key role in immunity and in pathogen entry and/or elimination.
Thus the CFTR mutations might be expected to compromise not only the chloride channel, but also the ability to kill pathogens via glutathione, or hypochlorous acid. Mutations might also be expected to alter the ability to process antigens to pathogens, or to self. CFTR mutations also activate the immune system.
Many of the mutations in the CFTR protein lie within regions that are highly immunogenic, and such high immunogenicity would be shared by the viral, bacterial and fungal homologues of the protein, of which there are several thousand. The autoantigens reported in cystic fibrosis, as well as P.Aeruginosa antigens are also homologous to the Delta508 mutant protein, again within regions that are highly immunogenic. Given the vast number of pathogen proteins that show homology with various regions of the CFTR protein, and the fact that such species are more abundant in cystic fibrosis patients, cross-reactivity with the CFTR protein would seem inevitable, although to date no antibodies to CFTR have been reported or apparently assessed. Although many of the CFTR mutations are intracellular, antibodies do enter cells , and even if not mounting an intracellular immune response would be expected to bind to the immunogenic regions of the CFTR protein, in effect producing protein knockdown, equivalent to the effects of the truncated mutants that fail to reach the cell surface. It is also clear that the viral homologues of the CFTR protein are capable of binding to CFTR binding partners, potentially modifying the function of CFTR by interactome interference.

Infliximab
Infliximab is a tumour necrosis factor -alpha (TNF) monoclonal antibody used to treat autoimmune disorders. TNF antagonism prevents the activation of other inflammatory cytokines and leukocyte activation and this approach is a target in many autoimmune and inflammatory conditions (Hoffman, 2009). A recent case study has reported 2 year remission in a cystic fibrosis patient treated with infliximab . Apart from the use of immunosuppressants in cystic fibrosis lung transplant patients, and limited studies with cyclosporine, the therapeutic potential of this class of drug does not appear to have been widely studied . TNF is one of the autoantigens reported in cystic fibrosis , and shares sequence similarities with the CFTR protein (Table 2). Although certain TNF antibodies would be expected to cross-react with the CFTR protein, such effects would depend upon the epitopes targeted by the antibody, and these details are not available.
A possible scenario for cystic fibrosis (Fig 6) Irrespective of any homology to pathogens, CFTR mutations lead to defects in chloride channel function, but also to a reduction in glutathione levels and defects in hypochlorous acid production, that would compromise viral and bacterial destruction. The channel itself is involved in bacterial entry, and impaired CFTR function reduces bacterial entry into epithelial cells, resulting in increased colonisation of the extracellular milieu. In this space, the likelihood of encountering immunocompetent cells is increased, favouring the production of anti-pathogen antibodies. Pathogen binding to the CFTR channel also impairs its function. Such mutations may also compromise the immune system, rendering it less able to process antigens, but more susceptible to activation. Polymorphisms in immune, inflammation and glutathione related genes fine tune this network, modifying its function, for better or worse.
Upon infection, the surfeit of pathogens triggers an immune response that generates antibodies to the pathogen that also target human proteins that are homologous to the antigenic pathogen proteins, generating the autoantigens observed in cystic fibrosis. As judged by epitope homology, antibodies to pathogen proteins and to autoantigens may also tag the CFTR protein, rendering it incapable of assuming its normal functions. The constant presence of the pathogens and of the autoantigens sustains this immune response. Viral infections, in particular, would also be expected to modify CFTR function via the theft of interactome partners. Thus, antibody knockdown would have the same effect on CFTR function as the mutations that prevent CFTR expression, or its delivery to the cell surface. In these cases, the antibodies are acting as antagonists, rather than as immune activators. In extreme cases, an autoimmune response to the CFTR protein might be expected to damage, or kill the cells in which the protein resides. The bioinformatics analysis suggests that antibodies to the CFTR protein should be detectable in cystic fibrosis. This does not appear to have been assessed, judging from the absence of any mention of CFTR autoantibodies in the literature. However, the high titre of pathogen antibodies, whose antigen targets are homologous to the CFTR protein, suggests that even low affinity T cell epitope binding sites would be saturated.
Taken together, although clearly a genetic disorder, these data suggest that cystic fibrosis has a crucial autoimmune component, triggered by pathogens with homology to the mutant and related proteins.
Antibacterial agents are already used in cystic fibrosis (Wat, 2003). There are no phage or bacterial vaccines as yet, and antiviral agents and vaccination strategies could also perhaps be useful. Unfortunately, the repertoire of pathogens colonising cystic fibrosis patients is so vast that polypharmacy, with its attendant risks, might seem the only plausible option. Clearly the potential benefits of glutathione supplementation appear to be promising . Other methods of enhancing pathogen defence require further research.
It is possible that immunosuppression might be of benefit in cystic fibrosis. This is extremely counter-intuitive, given the problems of multiple infections in these patients, but a carefully controlled and supervised clinical trial may well be warranted. Indeed, the reported benefits of Infliximab (see above) , although only so far reported in a cse study suggest that such approaches may be of more general clinical use.
If the problems in cystic fibrosis stem even partly from autoantigens and autoantibodies, then their riddance can only be beneficial. Immunoadsorption/plasma exchange has been reported to be of benefit in the autoimmune disorder, myasthenia gravis and this type of therapy may be applicable to cystic fibrosis, using targeted antigen and antibody columns to remove the circulating antibodies and antigens. Tryptophan or phenylalanine columns have also been reported to be of use in antibody adsorption .
In summary, CFTR mutations are themselves responsible for bacterial hypercolonisation, and for reduced bactericidal and viricidal effects, creating a situation where antibody generation to a plethora of pathogens in inevitable. These antibodies target other antigens that are homologous to the pathogens' proteins, and these include the various autoantigens that have been recorded in cystic fibrosis. The pathogen antigens and autoantigens are both homologous to the CFTR protein itself, and antibody related CFTR antagonism is a likely consequence of these effects. Interruption of this feed forward cycle may be of clinical benefit in cystic fibrosis.      . ::.** : :.:* .::.::: S100A9 MTCKMSQLERNIETIINTFHQYSVKLGHPDTLNQGEFKELVRK 43 S.Aur *.
. ::.** : :.:* .::.::: LPACDGERPTLAFLQDVMNILLQYVVKSFDRSTKVIDFHYPNELLQ 46      Ryanodine receptor 2 -S100A7 Protein S100-A7 -S100A9 Protein S100-A9 -SEC61A1 Protein transport protein  Gap junction function is impaired in cystic fibrosis pancreatic duct cells Airway epithelial tight junction function is compromised in cystic fibrosis (Godfrey, 1997) Gap junction (4)   A pathogenic feed forward cycle in cystic fibrosis 1: CFTR mutations result in chloride channel deficiency with associated problems in fluid homoeostasis/ They also favour pathogen accumulation in the extracellular mileu. 2 : CFTR mutations also compromise glutathione and hypochlorous acid availability, reducing bactericidal and viricidal effects. 3: Hypercolonistaion by diverse pathogtens results in antibody production. 4: Because of pathogen mimicry these antibodies also target autoantigens, and possible the CFTR protein itself. Epitope sharing between pathogen/autoantigen and the CFTR protein favours the maintenance of antibody production. 5: Reductions in CFTR and autoantigen function, compromises CFTR related pathways, which include those related to the associated pathologies of cystic fibrosis. Repeated reductions in CFTR function continue the cycle, resulting in further pathogen colonisation etc…….