Molecular Mimicry Revisited: Gut Bacteria and Multiple Sclerosis
http://www.100md.com
微生物临床杂志 2006年第6期
Institute for Disease Research, P.O. Box 890193, Temecula, California 92589
ABSTRACT
Molecular mimicry is a possible explanation for autoimmune side effects of microorganism infections. Protein sequences from a particular microorganism are compared to known autoimmune immunogens. For diseases such as multiple sclerosis (MS), where the infectious agent is unknown, guesses to its identity are made. Mimics are assumed to be rare. This study takes a radically different approach. Reported sequences from all known human bacterial and viral agents were searched for autoimmune immunogen mimics. Three encephalitogenic peptides, whose autoimmune requirements have been studied extensively, were selected for comparison. Mimics were seen in a wide variety of organisms. For each immunogen, the mimics were found predominantly in nonpathogenic gut bacteria. Since the three immunogens used in this study are related to MS, it is suggested that a microorganism responsible for autoimmune activity in MS could be a normally occurring gut bacterium. This would explain many of the peculiar MS epidemiological data and why no infective agent has been identified for MS and supports recently found MS gut metabolism abnormalities.
INTRODUCTION
During the past 20 years molecular mimicry has been proposed as an explanation for autoimmune side effects of microorganism infections (3-5, 13, 15, 19, 41, 53). The process has involved comparing sequences within the proteins of a microorganism, usually a virus, with known human autoimmune immunogens. The probability of finding a particular sequence is 1 in 20n, where 20 refers to the 20 amino acids and "n" is the number of residues within the sequence under consideration. Finding a duplicate sequence in a microorganism which is identical to a human sequence would appear to be impossible. However, it is not that difficult. First, some amino acids, such as leucine and glycine, are used far more than others, such as tryptophan and histidine. Thus, some sequences are more likely to appear than others. Second, amino acid usage within sequences varies with the species. Third, one does not need to find an exact match to a human immunogen. Each peptide immunogen is composed of amino acids that each contribute differently to the overall immunogenicity of the peptide. Therefore, substitutions can be made in the original immunogen without destroying its activity. For example, it can be easily estimated that 10,000 peptides would satisfy the encephalitogenic requirements for the tryptophan peptide, fswgaegqr, of myelin basic protein (43, 44, 47).
Quite often if a sequence, under consideration, contains some amino acids that match the immunogen in type and location, the sequence is considered a possible mimic. Very seldom is there a concern for the location of the microorganism's sequence within the cell or within its protein structure itself, i.e., whether the microorganism's sequence has access to the human immune system. More importantly, the individual contributions to autoimmunity of the individual amino acids of the human immunogen are not considered. Without considering the individual amino acid contributions of the sequence fswgaegqr responsible for autoimmune encephalomyelitis (EAE), for example, which one of the following sequences would be considered a likely encephalitogen (characters in boldface represent changes in the sequence from the original fswgaegqr): fswgaigqr, fswaaegqr, or fswgaeger The correct answer is the first sequence (47). Most would pick the second or the third. The glycine allows an essential bend at the glycine-tryptophan bond (17, 25). The glutamate/glutamine substitution results in negating the positive arginine (50). Therefore, to adequately analyze potential microorganism mimics, one needs to have a totally defined human immunogen. This presents a problem.
In 1970 I synthesized and chemically defined the requirements for EAE induction from the first myelin basic protein encephalitogenic sequence (8, 43, 47). The contribution of each of the nine amino acids to immunity was ascertained. Since then, numerous (Table 1) other short encephalitogenic regions have been discovered within the myelin basic protein and the other myelin proteins. With a few exceptions, the other regions have only been isolated and in some cases synthetically produced. Even though EAE is by far the most studied autoimmune disease and is considered to be the autoimmune model for multiple sclerosis (MS) and viral neuropathies, few of its immunogens have been adequately defined, i.e., the contribution of each of the amino acids to immunity has not been examined.
The molecular mimicry process has centered on examining an individual microorganism for a sequence which fulfills the requirements for a given human immunogen. I report here the results of the opposite approach. The proteins of a large number of human bacterial and viral nonpathogenic and pathogenic organisms are compared to the three most studied encephalitogenic sequences. The following questions are asked: (i) How plentiful are the matches (ii) Are these sequences found more often in bacterial or viral organisms (iii) Is there a particular species that contains these mimics (iv) Do matches favor a particular encephalitogen (v) How are these results relevant to multiple sclerosis (vi) Are these results applicable to searches for other human immunogen mimics
MATERIALS AND METHODS
Selection of encephalitogenic sites. EAE has for years been used as the autoimmune model for multiple sclerosis (18). This disease is induced when whole brain, myelin, one of the myelin proteins, or encephalitogenic peptides in an appropriate adjuvant are injected into test animals. The fact that one autoimmune disease, EAE, is initiated even when whole brain is used as the encephalitogen points to the special immunological nature of the myelin proteins. Table 1 lists the major encephalitogenic regions within each of the encephalitogenic myelin proteins. Most of the encephalitogenic regions found in the myelin proteins only have been sequenced. However, the major three encephalitogenic regions of the myelin basic protein have been extensively studied. The encephalitogenic contribution of each of the amino acids from the tryptophan region, fswgaegqr, has been examined thoroughly in the guinea pig (8, 29, 37, 43, 44, 47, 50). The midpeptide, tthygslpqk, has been studied in the DR rat and rabbit (31, 33). The hyperacute EAE site, pqksqrtqdenpv, has been analyzed in the Lewis rat (20-22, 45, 52). Furthermore, the tryptophan peptide is encephalitogenic in rabbits (9), guinea pigs, and monkeys (7) but not in Lewis rats. The midpeptide has been found active in rabbits and DR rats but not in guinea pigs. Finally, the hyperacute site is active in Lewis rats and monkeys but not in guinea pigs.
The relevance of any of these sequences to human disease, and particularly to MS, is questionable. The only way to determine whether any of these sequences are encephalitogenic in humans is to inject humans. This, of course, is not feasible. Many of the encephalitogenic regions are capable of reacting with activated lymphocytes from MS patients. However, this is no proof that they can induce disease in humans. Without reviewing this topic, five pertinent papers can be cited. First, Field and Caspary (10) showed that the tryptophan peptide could react with activated lymphocytes from a variety of cancer patients. These patients had neither EAE nor MS. The human cancer cells possess a surface protein that sequentially is similar to the tryptophan peptide. Second, Weizman et al. (42) reported a cell-mediated autoimmune response to human myelin basic protein in 76% of the autistic children studied. Here again, these children were not suffering from EAE, MS, or cancer. Finally, Spitler et al. (34, 35) showed that guinea pigs sensitized with encephalitogenic peptides having amino acid sequences different from that in the test protein did not show cellular immunity in vivo or in vitro to myelin basic protein, although the animals developed EAE. Einstein (6a) showed that antigen (MBP) suppression of EAE in guinea pigs did not require an intact tryptophan, and yet the tryptophan is an absolute requirement for disease induction. The specificity for disease induction and activated immune recognition are different.
Three sequences were compared in order that any conclusions presented would reflect all of the sites, not just the characteristic of one region. The tryptophan peptide has been studied by far the most. As part of my Ph.D. thesis (43), I systematically examined in guinea pigs the contribution to encephalitogenicity of each amino acid in the sequence. Further studies in later years also added to this initial work (8, 44, 47, 50).
Of the nine residues of the tryptophan peptide, fswgaegqr, five appear considerably more important. The five are the glutamine, a positively charged amino acid adjacent to the glutamine, the tryptophan located six residues from the glutamine toward the amino-terminal end with the glycine C terminally adjacent to it, and a hydrophobic region (preferably a ring structure) eight residues from the glutamine. The glutamine itself can be replaced by serine or asparagine, but not by glutamic acid. The negative charge of the glutamic acid will negate the positive of the adjacent amino acid. The glutamine appears to be contributing a hydrogen-bonding side chain. The phenylalanine and tryptophan form a hydrophobic pocket, called a molecular sandwich (26, 28, 43). The glycine permits a bend in the molecule at the tryptophan-glycine bond (17, 25).
Using the midpeptide, Smeltz et al. (33) showed by alanine replacement in the DA rat the importance of the "hygslp" sequence. Since only alanine substitutions were used, it is difficult to ascertain the variability acceptable. However, it appears that, unlike the tryptophan-glycine requirement, the tyrosine-glycine shows some flexibility. As stated above, the midpeptide is also active in rabbits. The midpeptide structurally is quite similar to the tryptophan peptide, which is also active in rabbits. Therefore, sequences that satisfy the general "hygslp" requirement and contain a glutamine adjacent to a charged amino acid at the C-terminal end are also noted (46).
The hyperacute sequence, pqksqrtqdenpv, is the least studied of the three encephalitogenic regions. The "tqde" sequence appears to be significant (32). A serine for threonine substitution converts the potency of the molecule from regular EAE to hyperacute EAE (20, 52). The aspartic acid appears to be essential. Although not as specific as the aspartic acid, the serine is important (21, 22). Therefore, small hydrophilic residues were included as acceptable for the replacement of the serine. With respect to the glutamic acid, the "tqde" data indicate some variability. For this position, glutamic acid, aspartic acid, glutamine, and asparagine were included.
Using the information on the requirements for the three encephalitogenic regions, bacterial and viral sequences were selected as potential encephalitogens.
Analytical process. The BLAST (basic local alignment search tool) program of the National Center for Biotechnology Information was used to ascertain potential encephalitogenic mimics (www.ncbi.nlm.nih.gov/BLAST). Four databases—nr, refseq, Swiss-Prot, and month—were used.
The proteins in the following human bacterial groups were examined for potential encephalitogenic mimics: Klebsiella, Morganella, Proteus, Serratia, Enterococcus, Micrococcus, Streptococcus, Bifidobacterium, Lactobacillus, Prevotella, Bacteroides, Fusobacterium, Eubacterium, Burkholderia, Mycobacterium, Salmonella, Chlamydophila, Haemophilus, Bacillus, Pseudomonas, Actinobacillus, Clostridium, and Escherichia.
The proteins in the following human virus groups were examined for potential encephalitogenic mimics: Morbillivirus, Paramyxovirus, Rubulavirus, Pneumovirus, Filoviridae, Influenza virus, Arenaviridae, Bunyaviridae, Rotavirus, Coltivirus, Orthreovirus, Coronavirus, Torovirus, Flaviviridae, Togaviridae, Calicivirus, Astrovirus, Enterovirus, Rhinovirus, Hepatitis A virus, Hepatitis B virus, Hepatitis D virus, Hepatitis E virus, Herpesvirus B, Varicella-Zoster virus, Herpes simplex virus, Herpesvirus, Cytomegalovirus, Epstein-Barr virus, Adenoviridae, Parvovirus, Polyomavirus, Echovirus, Bluetongue virus, and Papillomavirus.
Obviously, only the data that have been accumulated to date can be examined. The entire genomes of some microorganisms, including strains, have been ascertained. Others have not. Therefore, as more data are revealed more potential encephalitogenic mimics will be found. Comparative analysis of 16S rRNA sequences amplified from human feces indicated that less than 25% of the molecular species identified corresponded to known organisms (36). Obviously many bacteria in the gut, at least, have yet to be identified.
RESULTS AND DISCUSSION
Tables 2 (midregion), 3 (tryptophan peptide), and 4 (hyperacute site) present the potential encephalitogenic mimics within proteins from known human bacteria. Table 5 gives the same data for the potential encephalitogenic mimics within proteins from known human viruses. No attempt was made in these tables to separate pathogenic from nonpathogenic organisms. Moreover, no attempt to separate the organisms by location of potential "infection" was made.
The bacteria present 111 potential encephalitogenic mimics: 39 midregion, 23 tryptophan peptide, and 49 hyperacute site. Thus, there are numerous sequences that are potentially encephalitogenic in bacteria. No one immunogen dominates the group. In fact, the differences probably reflect more on the specific requirements imposed on selection. The tryptophan site, which is by far the most studied (and presents the most restrictions), has the fewest mimics but still presents a large selection. The sites are found in most of the bacteria examined. In fact, the number of sites found in a particular bacterium probably reflects more on the amount of data known about the species than any bacterial restriction. Gram stain characteristics seem to have little significance. Finally, the locations of the sites are in proteins located in all parts of the bacterial cell—inside, outside, etc. The locations of the sites within the proteins are also quite variable.
For molecular mimicry to work the immunogen must be "visible" to the immune system. It has therefore been assumed that it must be located on the surface of a protein that is either excreted by the microorganism or is on its surface. This is not necessarily true. The microorganism itself can initiate an immune response. Cells are hydrolyzed, and their contents are released. A great amount of digestive enzyme activity is present. All this can "expose" potential immunogens (23). It certainly happens within active EAE lesions. These "new" immunogens can be processed by the already-present immune cells.
The one feature of all of the diseases that have been related to molecular mimicry is that the incidence of the primary infection is very high, e.g., measles and influenza. However, the incidence of the autoimmune "side effect" is extremely low, e.g., the viral neuropathies incidence is 1/10,000. Any hypothesis that attempts to explain this process must take into account the disparity in incidence rates.
One explanation is taken from experimental autoimmune disease research. In order to induce EAE in animals, both the immunogen and the adjuvant must be injected at the same time. In fact, it has been shown that the two form a complex (27, 28, 48). The need for two separate compounds to be present simultaneously is used to explain the low incidence of viral and postvaccinal neuropathies (49). It has been shown that secondary infections are significantly present during the induction of these neuropathies. The cell wall material of the microorganism responsible for the secondary infection is a source of the adjuvant.
How can this be used as an explanation of multiple sclerosis In this disease it is hard to find one infection (other than the autoimmune reaction at the lesion), much less two. Tables 2 to 5 present more than potential encephalitogenic sites. They show the probability of finding such sites resides in bacteria and not viruses and, in fact, in the bacteria located in the most concentrated bacterial area of the human anatomy—the gut. Most of the gut bacteria are not pathogenic but form a symbiotic relationship with the individual. Therefore, the requirement for the first "infection" is satisfied by the presence of nonpathogenic bacteria.
What about the need for an adjuvant molecule Adjuvant molecules are portions of the bacterial cell walls. The simplest is N-acetylmuramyl dipeptide (MDP) (6). However, MDP does not aid all encephalitogens in producing EAE. Larger portions apparently are needed for some encephalitogens (28). Fox et al. have shown quite convincingly that adjuvant molecules normally are not in human body fluids or tissue, except the gut (11, 12). So, apparently, the two groups of substances, potential immunogens (mimics) and adjuvant molecules, known to be required for an autoimmune response are normally found together in the normal human gut. Although it appears that the adjuvant itself cannot cross the gut, the adjuvant-immunogen complex probably can. In fact the MDP-tryptophan peptide complex forms a spherical ball with one hemisphere hydrophobic and the other hydrophilic. This soap-like structure is ideal for membrane penetration (28, 48).
Why is the incidence of MS so low Although both the immunogen and adjuvant molecules are present, they must be processed in such fashion that the correct units are made in high enough concentration to form the complex. This will depend entirely upon the nature of the gut and its hydrolytic enzymes. The gut is a dynamically evolving biosystem whose bacterial content is changing and therefore the concentrations and types of hydrolytic enzymes continually vary.
There is other evidence that the gut is involved in MS. Serotonin metabolism is greatly altered during multiple sclerosis relapses. This has been seen by measuring the 5-hydroxy indoleacetic acid (5-HIAA)/tryptophan ratio from urine (Table 6). Ninety percent of the serotonin metabolism is derived from the gut and not the nervous system. Furthermore, one of the adjuvant molecules, MDP, is known to mimic serotonin (28).
Rather than study a single organism, the present study examined a large number of nonpathological and pathological human bacteria and viruses in order to ascertain the prevalence of encephalitogenic mimics. There appear to be many potential encephalitogens within bacteria and viral cells. It also seems that the most probable source of these mimics is the normal gut. This information has been applied to MS. However, the procedure used is just as valid for the location of any potential immunogenic mimic. Therefore, it has a broad application to the field of molecular mimicry.
Mailing address: Institute for Disease Research, P.O. Box 890193, Temecula, CA 92589. Phone: (951) 699-5177. Fax: (951) 699-5177. E-mail: fcwestallidr@adelphia.net.
REFERENCES
Amor, S., D. Baker, N. Groome, and J. L. Turk. 1993. Identification of a major encephalitogenic epitope of proteolipid protein (residues 50-70) for the induction of EAE in Biozzi AB/H and nonobese diabetic mice. J. Immunol. 150:5666-5672.
Amor, S., N. Groome, C. Linington, M. M. Morris, K. Dornmair, M. V. Gardinier, J.-M. Matthieu, and D. Baker. 1994. Identification of epitopes of myelin oligodendrocyte glycoprotein for the induction of EAE in SJL and Biozzi AB/H mice. J. Immunol. 153:4339-4356.
Bachmaier, K., N. Neu, L. M. de la Maza, S. Pal, A. Hessel, and J. M. Penninger. 1999. Chlamydia infection and heart disease linked through antigenic mimicry. Science 283:1335-1343.
Benoist, C., and D. Mathis. 2001. Autoimmunity provoked by infection: how good is the case for cell mimicry Nat. Immunol. 2:797-801.
Carnegie, P. R., and R. Weise. 1987. Visna and MBP. Nature 329:294-295.
Chedid, L., F. Audibert, P. Lefrancier, J. Choay, and E. Lelderer. 1976. Modulation of the immune response by a synthetic adjuvant and analogs. Proc. Natl. Acad. Sci. USA 73:2472-2475.
Einstein, R. E. 1972. Suppression of EAF, p. 121-129. In E. J. Field, T. M. Bell, and P. R. Carnegie (ed.), Multiple sclerosis: progress in research. North Holland Publishing Co., Amsterdam, The Netherlands.
Eylar, E. H., S. Brostoff, J. Jackson, and H. Carter. 1972. Allergic encephalomyelitis in monkeys by a peptide from the A1 protein. Proc. Natl. Acad. Sci. USA 69:617-619.
Eylar, E. H., J. Caccam, J. J. Jackson, F. C. Westall, and A. B. Robinson. 1970. EAE: synthesis of disease-inducing site of basic protein. Science 168:1220-1223.
Eylar, E. H., F. C. Westall, and S. Brostoff. 1971. Allergic encephalomyelitis. J. Biol. Chem. 246:3418-3424.
Field, E. J., and E. A. Caspary. 1971. Lymphocyte sensitization in cancer. Lancet i:189-190.
Fox, A., and K. Fox. 1991. Rapid elimination of a synthetic adjuvant peptide from the circulation after systemic administration and absence of detection of natural muramyl peptides in normal serum at current analytical limits. Infect. Immun. 59:1202-1205.
Fox, A., K. Fox, B. Christensson, D. Harrelson, and M. Kramer. 1996. Absolute identification of muramic acid, at trace levels, in human septic synovial fluids in vivo and absence in aseptic fluids. Infect. Immun. 64:3911-3915.
Fujinami, R. S., and M. B. Oldstone. 1985. Amino acid homology between the encephalitogenic site of MBP and virus: mechanism for autoimmunity. Science 230:1043-1045.
Greer, J. M., V. K. Kuchroo, R. A. Sobel, and M. B. Lees. 1992. Identification and characterization of a second encephalitogenic determinant of myelin proteolipid protein (residues 178-191) for SJL mice. J. Immunol. 149:783-788.
Jahnke, U., E. D. Fisher, and E. C. Alvord, Jr. 1985. Sequence homology between certain viral proteins and proteins related to encephalomyelitis and neuritis. Science 229:282-284.
Karkhanis, Y. D., D. J. Carlo, S. W. Brostoff, and E. H. Eylar. 1975. Allergic encephalomyelitis. Isolation of an encephalitogenic peptide activity in the monkey. J. Biol. Chem. 250:1718-1722.
Khanarian, G., S. A. Margeson, W. J. Moore, S. J. Pasaribu, and F. C. Westall. 1979. Biophys. Biochem. Res. Commun. 87:236-243
Kies, M. W., and E. C. Alvord, Jr. 1959. Myelin basic protein, p. 239-299. In Allergic encephalomyelitis. Charles C Thomas, Springfield, Ill.
Lang, H. L. E., H. Jacobson, S. Ikemizo, C. Andersson, K. Harla, L. Madsen, P. Hjorth, L. Sondergaard, D. Svejgard, K. Wucherpfennig, D. I. Stuart, J. L. Bell, E. Y. Jones, and L. Fugger. 2002. A functional and structural basis for the cross-reactivity in multiple sclerosis. Nat. Immunol. 3:940-943.
Lennon, V. A., F. C. Westall, M. Thompson, and E. Ward. 1976. Antigen, host, and adjuvant requirements for induction of hyperacute EAE. Eur. J. Immunol. 6:805-810.
Mannie, M. D., P. Y. Paterson, D. C. U'Prichard, and G. Flouret. 1989. Encephalitogenic and proliferative responses of Lewis rat lymphocytes distinguished by position 75- and 80-substituted peptides of myelin basic protein. J. Immunol. 142:2608-2616.
Mannie, M. D., P. Y. Paterson, D. C. U'Pritchard, and G. Flouret. 1985. Induction of EAE in Lewis rats. Proc. Natl. Acad. Sci. USA 82:5515-5519.
Matsuo, A., G. C. Lee, K. Terai, K. Takami, W. F. Hickey, E. G. McGeer, et al. 1997. Unmasking of an unusual MBP epitope during the process of myelin degeneration in humans: a potential mechanism for the generation of autoantigens. Am. J. Pathol. 150:1253-1266.
Mendel, I., N. Kerlero de Rosbo, and A. Ben-Nun. 1996. Delineation of the minimal encephalitogenic epitope within the immunodominant region of myelin oligodendrocyte glycoprotein. Eur. J. Immunol. 26:2470-2479.
Moore, W. J., B. E. Chapman, G. E. James, G. Khanarian, S. A. Margetson, S. J. Pasaribu, and F. C. Westall. 1981. Conformation of encephalitogenic proteins and peptides. Excerpta Medica 546:198-208.
Root-Bernstein, R. S., and F. C. Westall. 1984. Serotonin binding sites. Brain Res. Bull. 12:425-436.
Root-Bernstein, R. S., and F. C. Westall. 1986. Complementarity between dual antigens in the induction of EAE. J. Int. Dev. Biol. 2:1-10.
Root-Bernstein, R. S., and F. C. Westall. 1990. Serotonin binding sites. II. Muramyl dipeptide binds serotonin binding sites on MBP, LHRH, and MSH-ACTH 4-10. Brain Res. Bull. 25:827-841.
Sasaki, Y. 1974. Studies on encephalitogenic fragments of myelin basic protein. III. Synthesis of H-Arg-Phe-Trp-Gly-Ala-glu-Gly-Asn-Arg-OH as an analog of encephalitogenic decapeptide. Chem. Pharm. Bull. 22:2188-2191.
Schumacher, G. A., G. Beebe, R. F. Kibler, L. T. Kurland, J. F. Kurtzke, F. McDowell, B. Nagler, W. A. Sibley, W. W. Tourtellotte, and T. L. Wilmon. 1965. Problems of experimental trials in multiple sclerosis. Ann. N. Y. Acad. Sci. 122:552-568.
Shapira, R., F. C. H. Chou, S. McKneally, E. Urban, and R. F. Kibler. 1971. Biological activity and synthesis of an encephalitogenic determinant. Science 172:736-738.
Smeltz, R. B., M. H. M. Wauben, N. A. Wolf, and R. H. Swanborg. 1999. Critical requirement for aspartic acid at position 82 of myelin basic protein 73-86. J. Immunol. 162:829-836.
Smeltz, R. B., N. A. Wolf, and R. H. Swanborg. 1998. Delineation of two encephalitogenic myelin basic protein epitopes for DA rats. J. Neuroimmunol. 87:43-48.
Spitler, L. E., C. M. Von Muller, H. H. Fudenberg, and E. H. Eylar. 1972. EAE: dissociation of cellular immunity to brain protein and disease. J. Exp. Med. 136:156-174.
Spitler, L. E., C. M. Von Muller, and J. Young. 1975. Study of cellular immunity to the encephalitogenic determinant. Cell. Immun. 15:143-151.
Suau, A., R. Bonnet, M. Sutren, J. J. Godon, G. R. Gibson, M. D. Collins, and J. Dore. 1999. Direct analysis of genes encoding 16S rRNA from complex communities reveals many novel molecular species within the human gut. Appl. Environ. Microbiol. 65:4799-4807.
Suzuki, K., and Y. Sasaki. 1974. Studies on encephalitogen fragments of myelin basic protein. IV. Synthesis of glycine analogs of tryptophan containing fragment. Chem. Pharm. Bull. 22:2181-2187.
Tuohy, V. K., Z. Lu, R. A. Sobel, R. A. Laursen, and M. B. Lees. 1988. A synthetic peptide from myelin proteolipid induces EAE. J. Immunol. 141:1126-1130.
Tuohy, V. K., Z. Lu, R. A. Sobel, R. A. Laursen, and M. B. Lees. 1989. Identification of an encephalitogenic determinant for myelin proteolipid protein for SJL mice. J. Immunol. 142:1523-1527.
Udenfriend, S., E. Titus, and H. Weissbach. 1955. The identification of 5-hydroxy-3-indoleacetic acid in normal urine and a method for its assay. J. Biol. Chem. 261:499-505.
Weise, M. J., and P. R. Carnegie. 1988. An approach to searching protein sequences for superfamily relationships or chronic similarities relevant to the molecular mimicry hypothesis: applications to the basic protein of myelin. J. Neurochem. 51:1267-1272.
Weizman, A., R. Weizman, G. A. Szekely, H. Wijsenbeak, and E. Livni. 1982. Abnormal immune response to brain tissue antigen in the syndrome of autism. Am. J. Psychiatr. 139:1462-1465.
Westall, F. C. 1970. Applications of solid phase peptide synthesis. Ph.D. thesis. University of California, San Diego.
Westall, F. C. 1972. Solid phase peptide synthesis as applied to EAE, p. 72-77. In E. J. Field, T. M. Bell, and P. R. Carnegie (ed.), Multiple sclerosis: progress in research. North Holland, Amsterdam, The Netherlands.
Westall, F. C. 1977. Hyperacute allergic encephalomyelitis: a single determinant. Immunol. Commun. 6:227-237.
Westall, F. C. 1978. Possible change of a recognition mechanism for determining encephalotogenicity. J. Theor. Biol. 71:465-468.
Westall, F. C., A. B. Robinson, J. Caccam, J. J. Jackson, and E. H. Eylar. 1971. Essential chemical requirement for induction of allergic encephalomyelitis. Nature 229:22-24.
Westall, F. C., and R. S. Root-Bernstein. 1983. An explanation of prevention and suppression of EAE. Mol. Immunol. 20:169-177.
Westall, F. C., and R. S. Root-Bernstein. 1986. Cause and prevention of postinfectious and post-vaccinal neuropathies in light of a new theory of autoimmunity. Lancet ii:251-252.
Westall, F. C., and M. Thompson. 1977. Further definition of the encephalitogenic region in guinea pigs. Immunol. Commun. 6:23-31.
Westall, F. C., and M. Thompson. 1978. An encephalitogenic region for rabbits. Immunochemistry 15:189-191.
Westall, F. C., M. Thompson, and V. A. Lennon. 1977. Hyperacute autoimmune encephalomyelitis-unique determinant conferred by serine in a synthetic autoantigen. Nature 269:425-427.
Zhao, Z.-S., F. Granucci, L. Yeh, P. Schaffer, and H. Cantor. 1998. Molecular mimicry by herpes simplex virus-type I: autoimmune disease after viral infection. Science 230:1043-1051.(Fred C. Westall)
ABSTRACT
Molecular mimicry is a possible explanation for autoimmune side effects of microorganism infections. Protein sequences from a particular microorganism are compared to known autoimmune immunogens. For diseases such as multiple sclerosis (MS), where the infectious agent is unknown, guesses to its identity are made. Mimics are assumed to be rare. This study takes a radically different approach. Reported sequences from all known human bacterial and viral agents were searched for autoimmune immunogen mimics. Three encephalitogenic peptides, whose autoimmune requirements have been studied extensively, were selected for comparison. Mimics were seen in a wide variety of organisms. For each immunogen, the mimics were found predominantly in nonpathogenic gut bacteria. Since the three immunogens used in this study are related to MS, it is suggested that a microorganism responsible for autoimmune activity in MS could be a normally occurring gut bacterium. This would explain many of the peculiar MS epidemiological data and why no infective agent has been identified for MS and supports recently found MS gut metabolism abnormalities.
INTRODUCTION
During the past 20 years molecular mimicry has been proposed as an explanation for autoimmune side effects of microorganism infections (3-5, 13, 15, 19, 41, 53). The process has involved comparing sequences within the proteins of a microorganism, usually a virus, with known human autoimmune immunogens. The probability of finding a particular sequence is 1 in 20n, where 20 refers to the 20 amino acids and "n" is the number of residues within the sequence under consideration. Finding a duplicate sequence in a microorganism which is identical to a human sequence would appear to be impossible. However, it is not that difficult. First, some amino acids, such as leucine and glycine, are used far more than others, such as tryptophan and histidine. Thus, some sequences are more likely to appear than others. Second, amino acid usage within sequences varies with the species. Third, one does not need to find an exact match to a human immunogen. Each peptide immunogen is composed of amino acids that each contribute differently to the overall immunogenicity of the peptide. Therefore, substitutions can be made in the original immunogen without destroying its activity. For example, it can be easily estimated that 10,000 peptides would satisfy the encephalitogenic requirements for the tryptophan peptide, fswgaegqr, of myelin basic protein (43, 44, 47).
Quite often if a sequence, under consideration, contains some amino acids that match the immunogen in type and location, the sequence is considered a possible mimic. Very seldom is there a concern for the location of the microorganism's sequence within the cell or within its protein structure itself, i.e., whether the microorganism's sequence has access to the human immune system. More importantly, the individual contributions to autoimmunity of the individual amino acids of the human immunogen are not considered. Without considering the individual amino acid contributions of the sequence fswgaegqr responsible for autoimmune encephalomyelitis (EAE), for example, which one of the following sequences would be considered a likely encephalitogen (characters in boldface represent changes in the sequence from the original fswgaegqr): fswgaigqr, fswaaegqr, or fswgaeger The correct answer is the first sequence (47). Most would pick the second or the third. The glycine allows an essential bend at the glycine-tryptophan bond (17, 25). The glutamate/glutamine substitution results in negating the positive arginine (50). Therefore, to adequately analyze potential microorganism mimics, one needs to have a totally defined human immunogen. This presents a problem.
In 1970 I synthesized and chemically defined the requirements for EAE induction from the first myelin basic protein encephalitogenic sequence (8, 43, 47). The contribution of each of the nine amino acids to immunity was ascertained. Since then, numerous (Table 1) other short encephalitogenic regions have been discovered within the myelin basic protein and the other myelin proteins. With a few exceptions, the other regions have only been isolated and in some cases synthetically produced. Even though EAE is by far the most studied autoimmune disease and is considered to be the autoimmune model for multiple sclerosis (MS) and viral neuropathies, few of its immunogens have been adequately defined, i.e., the contribution of each of the amino acids to immunity has not been examined.
The molecular mimicry process has centered on examining an individual microorganism for a sequence which fulfills the requirements for a given human immunogen. I report here the results of the opposite approach. The proteins of a large number of human bacterial and viral nonpathogenic and pathogenic organisms are compared to the three most studied encephalitogenic sequences. The following questions are asked: (i) How plentiful are the matches (ii) Are these sequences found more often in bacterial or viral organisms (iii) Is there a particular species that contains these mimics (iv) Do matches favor a particular encephalitogen (v) How are these results relevant to multiple sclerosis (vi) Are these results applicable to searches for other human immunogen mimics
MATERIALS AND METHODS
Selection of encephalitogenic sites. EAE has for years been used as the autoimmune model for multiple sclerosis (18). This disease is induced when whole brain, myelin, one of the myelin proteins, or encephalitogenic peptides in an appropriate adjuvant are injected into test animals. The fact that one autoimmune disease, EAE, is initiated even when whole brain is used as the encephalitogen points to the special immunological nature of the myelin proteins. Table 1 lists the major encephalitogenic regions within each of the encephalitogenic myelin proteins. Most of the encephalitogenic regions found in the myelin proteins only have been sequenced. However, the major three encephalitogenic regions of the myelin basic protein have been extensively studied. The encephalitogenic contribution of each of the amino acids from the tryptophan region, fswgaegqr, has been examined thoroughly in the guinea pig (8, 29, 37, 43, 44, 47, 50). The midpeptide, tthygslpqk, has been studied in the DR rat and rabbit (31, 33). The hyperacute EAE site, pqksqrtqdenpv, has been analyzed in the Lewis rat (20-22, 45, 52). Furthermore, the tryptophan peptide is encephalitogenic in rabbits (9), guinea pigs, and monkeys (7) but not in Lewis rats. The midpeptide has been found active in rabbits and DR rats but not in guinea pigs. Finally, the hyperacute site is active in Lewis rats and monkeys but not in guinea pigs.
The relevance of any of these sequences to human disease, and particularly to MS, is questionable. The only way to determine whether any of these sequences are encephalitogenic in humans is to inject humans. This, of course, is not feasible. Many of the encephalitogenic regions are capable of reacting with activated lymphocytes from MS patients. However, this is no proof that they can induce disease in humans. Without reviewing this topic, five pertinent papers can be cited. First, Field and Caspary (10) showed that the tryptophan peptide could react with activated lymphocytes from a variety of cancer patients. These patients had neither EAE nor MS. The human cancer cells possess a surface protein that sequentially is similar to the tryptophan peptide. Second, Weizman et al. (42) reported a cell-mediated autoimmune response to human myelin basic protein in 76% of the autistic children studied. Here again, these children were not suffering from EAE, MS, or cancer. Finally, Spitler et al. (34, 35) showed that guinea pigs sensitized with encephalitogenic peptides having amino acid sequences different from that in the test protein did not show cellular immunity in vivo or in vitro to myelin basic protein, although the animals developed EAE. Einstein (6a) showed that antigen (MBP) suppression of EAE in guinea pigs did not require an intact tryptophan, and yet the tryptophan is an absolute requirement for disease induction. The specificity for disease induction and activated immune recognition are different.
Three sequences were compared in order that any conclusions presented would reflect all of the sites, not just the characteristic of one region. The tryptophan peptide has been studied by far the most. As part of my Ph.D. thesis (43), I systematically examined in guinea pigs the contribution to encephalitogenicity of each amino acid in the sequence. Further studies in later years also added to this initial work (8, 44, 47, 50).
Of the nine residues of the tryptophan peptide, fswgaegqr, five appear considerably more important. The five are the glutamine, a positively charged amino acid adjacent to the glutamine, the tryptophan located six residues from the glutamine toward the amino-terminal end with the glycine C terminally adjacent to it, and a hydrophobic region (preferably a ring structure) eight residues from the glutamine. The glutamine itself can be replaced by serine or asparagine, but not by glutamic acid. The negative charge of the glutamic acid will negate the positive of the adjacent amino acid. The glutamine appears to be contributing a hydrogen-bonding side chain. The phenylalanine and tryptophan form a hydrophobic pocket, called a molecular sandwich (26, 28, 43). The glycine permits a bend in the molecule at the tryptophan-glycine bond (17, 25).
Using the midpeptide, Smeltz et al. (33) showed by alanine replacement in the DA rat the importance of the "hygslp" sequence. Since only alanine substitutions were used, it is difficult to ascertain the variability acceptable. However, it appears that, unlike the tryptophan-glycine requirement, the tyrosine-glycine shows some flexibility. As stated above, the midpeptide is also active in rabbits. The midpeptide structurally is quite similar to the tryptophan peptide, which is also active in rabbits. Therefore, sequences that satisfy the general "hygslp" requirement and contain a glutamine adjacent to a charged amino acid at the C-terminal end are also noted (46).
The hyperacute sequence, pqksqrtqdenpv, is the least studied of the three encephalitogenic regions. The "tqde" sequence appears to be significant (32). A serine for threonine substitution converts the potency of the molecule from regular EAE to hyperacute EAE (20, 52). The aspartic acid appears to be essential. Although not as specific as the aspartic acid, the serine is important (21, 22). Therefore, small hydrophilic residues were included as acceptable for the replacement of the serine. With respect to the glutamic acid, the "tqde" data indicate some variability. For this position, glutamic acid, aspartic acid, glutamine, and asparagine were included.
Using the information on the requirements for the three encephalitogenic regions, bacterial and viral sequences were selected as potential encephalitogens.
Analytical process. The BLAST (basic local alignment search tool) program of the National Center for Biotechnology Information was used to ascertain potential encephalitogenic mimics (www.ncbi.nlm.nih.gov/BLAST). Four databases—nr, refseq, Swiss-Prot, and month—were used.
The proteins in the following human bacterial groups were examined for potential encephalitogenic mimics: Klebsiella, Morganella, Proteus, Serratia, Enterococcus, Micrococcus, Streptococcus, Bifidobacterium, Lactobacillus, Prevotella, Bacteroides, Fusobacterium, Eubacterium, Burkholderia, Mycobacterium, Salmonella, Chlamydophila, Haemophilus, Bacillus, Pseudomonas, Actinobacillus, Clostridium, and Escherichia.
The proteins in the following human virus groups were examined for potential encephalitogenic mimics: Morbillivirus, Paramyxovirus, Rubulavirus, Pneumovirus, Filoviridae, Influenza virus, Arenaviridae, Bunyaviridae, Rotavirus, Coltivirus, Orthreovirus, Coronavirus, Torovirus, Flaviviridae, Togaviridae, Calicivirus, Astrovirus, Enterovirus, Rhinovirus, Hepatitis A virus, Hepatitis B virus, Hepatitis D virus, Hepatitis E virus, Herpesvirus B, Varicella-Zoster virus, Herpes simplex virus, Herpesvirus, Cytomegalovirus, Epstein-Barr virus, Adenoviridae, Parvovirus, Polyomavirus, Echovirus, Bluetongue virus, and Papillomavirus.
Obviously, only the data that have been accumulated to date can be examined. The entire genomes of some microorganisms, including strains, have been ascertained. Others have not. Therefore, as more data are revealed more potential encephalitogenic mimics will be found. Comparative analysis of 16S rRNA sequences amplified from human feces indicated that less than 25% of the molecular species identified corresponded to known organisms (36). Obviously many bacteria in the gut, at least, have yet to be identified.
RESULTS AND DISCUSSION
Tables 2 (midregion), 3 (tryptophan peptide), and 4 (hyperacute site) present the potential encephalitogenic mimics within proteins from known human bacteria. Table 5 gives the same data for the potential encephalitogenic mimics within proteins from known human viruses. No attempt was made in these tables to separate pathogenic from nonpathogenic organisms. Moreover, no attempt to separate the organisms by location of potential "infection" was made.
The bacteria present 111 potential encephalitogenic mimics: 39 midregion, 23 tryptophan peptide, and 49 hyperacute site. Thus, there are numerous sequences that are potentially encephalitogenic in bacteria. No one immunogen dominates the group. In fact, the differences probably reflect more on the specific requirements imposed on selection. The tryptophan site, which is by far the most studied (and presents the most restrictions), has the fewest mimics but still presents a large selection. The sites are found in most of the bacteria examined. In fact, the number of sites found in a particular bacterium probably reflects more on the amount of data known about the species than any bacterial restriction. Gram stain characteristics seem to have little significance. Finally, the locations of the sites are in proteins located in all parts of the bacterial cell—inside, outside, etc. The locations of the sites within the proteins are also quite variable.
For molecular mimicry to work the immunogen must be "visible" to the immune system. It has therefore been assumed that it must be located on the surface of a protein that is either excreted by the microorganism or is on its surface. This is not necessarily true. The microorganism itself can initiate an immune response. Cells are hydrolyzed, and their contents are released. A great amount of digestive enzyme activity is present. All this can "expose" potential immunogens (23). It certainly happens within active EAE lesions. These "new" immunogens can be processed by the already-present immune cells.
The one feature of all of the diseases that have been related to molecular mimicry is that the incidence of the primary infection is very high, e.g., measles and influenza. However, the incidence of the autoimmune "side effect" is extremely low, e.g., the viral neuropathies incidence is 1/10,000. Any hypothesis that attempts to explain this process must take into account the disparity in incidence rates.
One explanation is taken from experimental autoimmune disease research. In order to induce EAE in animals, both the immunogen and the adjuvant must be injected at the same time. In fact, it has been shown that the two form a complex (27, 28, 48). The need for two separate compounds to be present simultaneously is used to explain the low incidence of viral and postvaccinal neuropathies (49). It has been shown that secondary infections are significantly present during the induction of these neuropathies. The cell wall material of the microorganism responsible for the secondary infection is a source of the adjuvant.
How can this be used as an explanation of multiple sclerosis In this disease it is hard to find one infection (other than the autoimmune reaction at the lesion), much less two. Tables 2 to 5 present more than potential encephalitogenic sites. They show the probability of finding such sites resides in bacteria and not viruses and, in fact, in the bacteria located in the most concentrated bacterial area of the human anatomy—the gut. Most of the gut bacteria are not pathogenic but form a symbiotic relationship with the individual. Therefore, the requirement for the first "infection" is satisfied by the presence of nonpathogenic bacteria.
What about the need for an adjuvant molecule Adjuvant molecules are portions of the bacterial cell walls. The simplest is N-acetylmuramyl dipeptide (MDP) (6). However, MDP does not aid all encephalitogens in producing EAE. Larger portions apparently are needed for some encephalitogens (28). Fox et al. have shown quite convincingly that adjuvant molecules normally are not in human body fluids or tissue, except the gut (11, 12). So, apparently, the two groups of substances, potential immunogens (mimics) and adjuvant molecules, known to be required for an autoimmune response are normally found together in the normal human gut. Although it appears that the adjuvant itself cannot cross the gut, the adjuvant-immunogen complex probably can. In fact the MDP-tryptophan peptide complex forms a spherical ball with one hemisphere hydrophobic and the other hydrophilic. This soap-like structure is ideal for membrane penetration (28, 48).
Why is the incidence of MS so low Although both the immunogen and adjuvant molecules are present, they must be processed in such fashion that the correct units are made in high enough concentration to form the complex. This will depend entirely upon the nature of the gut and its hydrolytic enzymes. The gut is a dynamically evolving biosystem whose bacterial content is changing and therefore the concentrations and types of hydrolytic enzymes continually vary.
There is other evidence that the gut is involved in MS. Serotonin metabolism is greatly altered during multiple sclerosis relapses. This has been seen by measuring the 5-hydroxy indoleacetic acid (5-HIAA)/tryptophan ratio from urine (Table 6). Ninety percent of the serotonin metabolism is derived from the gut and not the nervous system. Furthermore, one of the adjuvant molecules, MDP, is known to mimic serotonin (28).
Rather than study a single organism, the present study examined a large number of nonpathological and pathological human bacteria and viruses in order to ascertain the prevalence of encephalitogenic mimics. There appear to be many potential encephalitogens within bacteria and viral cells. It also seems that the most probable source of these mimics is the normal gut. This information has been applied to MS. However, the procedure used is just as valid for the location of any potential immunogenic mimic. Therefore, it has a broad application to the field of molecular mimicry.
Mailing address: Institute for Disease Research, P.O. Box 890193, Temecula, CA 92589. Phone: (951) 699-5177. Fax: (951) 699-5177. E-mail: fcwestallidr@adelphia.net.
REFERENCES
Amor, S., D. Baker, N. Groome, and J. L. Turk. 1993. Identification of a major encephalitogenic epitope of proteolipid protein (residues 50-70) for the induction of EAE in Biozzi AB/H and nonobese diabetic mice. J. Immunol. 150:5666-5672.
Amor, S., N. Groome, C. Linington, M. M. Morris, K. Dornmair, M. V. Gardinier, J.-M. Matthieu, and D. Baker. 1994. Identification of epitopes of myelin oligodendrocyte glycoprotein for the induction of EAE in SJL and Biozzi AB/H mice. J. Immunol. 153:4339-4356.
Bachmaier, K., N. Neu, L. M. de la Maza, S. Pal, A. Hessel, and J. M. Penninger. 1999. Chlamydia infection and heart disease linked through antigenic mimicry. Science 283:1335-1343.
Benoist, C., and D. Mathis. 2001. Autoimmunity provoked by infection: how good is the case for cell mimicry Nat. Immunol. 2:797-801.
Carnegie, P. R., and R. Weise. 1987. Visna and MBP. Nature 329:294-295.
Chedid, L., F. Audibert, P. Lefrancier, J. Choay, and E. Lelderer. 1976. Modulation of the immune response by a synthetic adjuvant and analogs. Proc. Natl. Acad. Sci. USA 73:2472-2475.
Einstein, R. E. 1972. Suppression of EAF, p. 121-129. In E. J. Field, T. M. Bell, and P. R. Carnegie (ed.), Multiple sclerosis: progress in research. North Holland Publishing Co., Amsterdam, The Netherlands.
Eylar, E. H., S. Brostoff, J. Jackson, and H. Carter. 1972. Allergic encephalomyelitis in monkeys by a peptide from the A1 protein. Proc. Natl. Acad. Sci. USA 69:617-619.
Eylar, E. H., J. Caccam, J. J. Jackson, F. C. Westall, and A. B. Robinson. 1970. EAE: synthesis of disease-inducing site of basic protein. Science 168:1220-1223.
Eylar, E. H., F. C. Westall, and S. Brostoff. 1971. Allergic encephalomyelitis. J. Biol. Chem. 246:3418-3424.
Field, E. J., and E. A. Caspary. 1971. Lymphocyte sensitization in cancer. Lancet i:189-190.
Fox, A., and K. Fox. 1991. Rapid elimination of a synthetic adjuvant peptide from the circulation after systemic administration and absence of detection of natural muramyl peptides in normal serum at current analytical limits. Infect. Immun. 59:1202-1205.
Fox, A., K. Fox, B. Christensson, D. Harrelson, and M. Kramer. 1996. Absolute identification of muramic acid, at trace levels, in human septic synovial fluids in vivo and absence in aseptic fluids. Infect. Immun. 64:3911-3915.
Fujinami, R. S., and M. B. Oldstone. 1985. Amino acid homology between the encephalitogenic site of MBP and virus: mechanism for autoimmunity. Science 230:1043-1045.
Greer, J. M., V. K. Kuchroo, R. A. Sobel, and M. B. Lees. 1992. Identification and characterization of a second encephalitogenic determinant of myelin proteolipid protein (residues 178-191) for SJL mice. J. Immunol. 149:783-788.
Jahnke, U., E. D. Fisher, and E. C. Alvord, Jr. 1985. Sequence homology between certain viral proteins and proteins related to encephalomyelitis and neuritis. Science 229:282-284.
Karkhanis, Y. D., D. J. Carlo, S. W. Brostoff, and E. H. Eylar. 1975. Allergic encephalomyelitis. Isolation of an encephalitogenic peptide activity in the monkey. J. Biol. Chem. 250:1718-1722.
Khanarian, G., S. A. Margeson, W. J. Moore, S. J. Pasaribu, and F. C. Westall. 1979. Biophys. Biochem. Res. Commun. 87:236-243
Kies, M. W., and E. C. Alvord, Jr. 1959. Myelin basic protein, p. 239-299. In Allergic encephalomyelitis. Charles C Thomas, Springfield, Ill.
Lang, H. L. E., H. Jacobson, S. Ikemizo, C. Andersson, K. Harla, L. Madsen, P. Hjorth, L. Sondergaard, D. Svejgard, K. Wucherpfennig, D. I. Stuart, J. L. Bell, E. Y. Jones, and L. Fugger. 2002. A functional and structural basis for the cross-reactivity in multiple sclerosis. Nat. Immunol. 3:940-943.
Lennon, V. A., F. C. Westall, M. Thompson, and E. Ward. 1976. Antigen, host, and adjuvant requirements for induction of hyperacute EAE. Eur. J. Immunol. 6:805-810.
Mannie, M. D., P. Y. Paterson, D. C. U'Prichard, and G. Flouret. 1989. Encephalitogenic and proliferative responses of Lewis rat lymphocytes distinguished by position 75- and 80-substituted peptides of myelin basic protein. J. Immunol. 142:2608-2616.
Mannie, M. D., P. Y. Paterson, D. C. U'Pritchard, and G. Flouret. 1985. Induction of EAE in Lewis rats. Proc. Natl. Acad. Sci. USA 82:5515-5519.
Matsuo, A., G. C. Lee, K. Terai, K. Takami, W. F. Hickey, E. G. McGeer, et al. 1997. Unmasking of an unusual MBP epitope during the process of myelin degeneration in humans: a potential mechanism for the generation of autoantigens. Am. J. Pathol. 150:1253-1266.
Mendel, I., N. Kerlero de Rosbo, and A. Ben-Nun. 1996. Delineation of the minimal encephalitogenic epitope within the immunodominant region of myelin oligodendrocyte glycoprotein. Eur. J. Immunol. 26:2470-2479.
Moore, W. J., B. E. Chapman, G. E. James, G. Khanarian, S. A. Margetson, S. J. Pasaribu, and F. C. Westall. 1981. Conformation of encephalitogenic proteins and peptides. Excerpta Medica 546:198-208.
Root-Bernstein, R. S., and F. C. Westall. 1984. Serotonin binding sites. Brain Res. Bull. 12:425-436.
Root-Bernstein, R. S., and F. C. Westall. 1986. Complementarity between dual antigens in the induction of EAE. J. Int. Dev. Biol. 2:1-10.
Root-Bernstein, R. S., and F. C. Westall. 1990. Serotonin binding sites. II. Muramyl dipeptide binds serotonin binding sites on MBP, LHRH, and MSH-ACTH 4-10. Brain Res. Bull. 25:827-841.
Sasaki, Y. 1974. Studies on encephalitogenic fragments of myelin basic protein. III. Synthesis of H-Arg-Phe-Trp-Gly-Ala-glu-Gly-Asn-Arg-OH as an analog of encephalitogenic decapeptide. Chem. Pharm. Bull. 22:2188-2191.
Schumacher, G. A., G. Beebe, R. F. Kibler, L. T. Kurland, J. F. Kurtzke, F. McDowell, B. Nagler, W. A. Sibley, W. W. Tourtellotte, and T. L. Wilmon. 1965. Problems of experimental trials in multiple sclerosis. Ann. N. Y. Acad. Sci. 122:552-568.
Shapira, R., F. C. H. Chou, S. McKneally, E. Urban, and R. F. Kibler. 1971. Biological activity and synthesis of an encephalitogenic determinant. Science 172:736-738.
Smeltz, R. B., M. H. M. Wauben, N. A. Wolf, and R. H. Swanborg. 1999. Critical requirement for aspartic acid at position 82 of myelin basic protein 73-86. J. Immunol. 162:829-836.
Smeltz, R. B., N. A. Wolf, and R. H. Swanborg. 1998. Delineation of two encephalitogenic myelin basic protein epitopes for DA rats. J. Neuroimmunol. 87:43-48.
Spitler, L. E., C. M. Von Muller, H. H. Fudenberg, and E. H. Eylar. 1972. EAE: dissociation of cellular immunity to brain protein and disease. J. Exp. Med. 136:156-174.
Spitler, L. E., C. M. Von Muller, and J. Young. 1975. Study of cellular immunity to the encephalitogenic determinant. Cell. Immun. 15:143-151.
Suau, A., R. Bonnet, M. Sutren, J. J. Godon, G. R. Gibson, M. D. Collins, and J. Dore. 1999. Direct analysis of genes encoding 16S rRNA from complex communities reveals many novel molecular species within the human gut. Appl. Environ. Microbiol. 65:4799-4807.
Suzuki, K., and Y. Sasaki. 1974. Studies on encephalitogen fragments of myelin basic protein. IV. Synthesis of glycine analogs of tryptophan containing fragment. Chem. Pharm. Bull. 22:2181-2187.
Tuohy, V. K., Z. Lu, R. A. Sobel, R. A. Laursen, and M. B. Lees. 1988. A synthetic peptide from myelin proteolipid induces EAE. J. Immunol. 141:1126-1130.
Tuohy, V. K., Z. Lu, R. A. Sobel, R. A. Laursen, and M. B. Lees. 1989. Identification of an encephalitogenic determinant for myelin proteolipid protein for SJL mice. J. Immunol. 142:1523-1527.
Udenfriend, S., E. Titus, and H. Weissbach. 1955. The identification of 5-hydroxy-3-indoleacetic acid in normal urine and a method for its assay. J. Biol. Chem. 261:499-505.
Weise, M. J., and P. R. Carnegie. 1988. An approach to searching protein sequences for superfamily relationships or chronic similarities relevant to the molecular mimicry hypothesis: applications to the basic protein of myelin. J. Neurochem. 51:1267-1272.
Weizman, A., R. Weizman, G. A. Szekely, H. Wijsenbeak, and E. Livni. 1982. Abnormal immune response to brain tissue antigen in the syndrome of autism. Am. J. Psychiatr. 139:1462-1465.
Westall, F. C. 1970. Applications of solid phase peptide synthesis. Ph.D. thesis. University of California, San Diego.
Westall, F. C. 1972. Solid phase peptide synthesis as applied to EAE, p. 72-77. In E. J. Field, T. M. Bell, and P. R. Carnegie (ed.), Multiple sclerosis: progress in research. North Holland, Amsterdam, The Netherlands.
Westall, F. C. 1977. Hyperacute allergic encephalomyelitis: a single determinant. Immunol. Commun. 6:227-237.
Westall, F. C. 1978. Possible change of a recognition mechanism for determining encephalotogenicity. J. Theor. Biol. 71:465-468.
Westall, F. C., A. B. Robinson, J. Caccam, J. J. Jackson, and E. H. Eylar. 1971. Essential chemical requirement for induction of allergic encephalomyelitis. Nature 229:22-24.
Westall, F. C., and R. S. Root-Bernstein. 1983. An explanation of prevention and suppression of EAE. Mol. Immunol. 20:169-177.
Westall, F. C., and R. S. Root-Bernstein. 1986. Cause and prevention of postinfectious and post-vaccinal neuropathies in light of a new theory of autoimmunity. Lancet ii:251-252.
Westall, F. C., and M. Thompson. 1977. Further definition of the encephalitogenic region in guinea pigs. Immunol. Commun. 6:23-31.
Westall, F. C., and M. Thompson. 1978. An encephalitogenic region for rabbits. Immunochemistry 15:189-191.
Westall, F. C., M. Thompson, and V. A. Lennon. 1977. Hyperacute autoimmune encephalomyelitis-unique determinant conferred by serine in a synthetic autoantigen. Nature 269:425-427.
Zhao, Z.-S., F. Granucci, L. Yeh, P. Schaffer, and H. Cantor. 1998. Molecular mimicry by herpes simplex virus-type I: autoimmune disease after viral infection. Science 230:1043-1051.(Fred C. Westall)