Roy S. Jamron holds a B.S. in Physics from the University of Michigan and an M.S. in Engineering Applied Science from the University of California at Davis, and independently investigates the latest research on celiac disease and related disorders.
This article originally appeared in the Spring 2004 edition of Celiac.com's Scott-Free Newsletter.
Celiac.com 05/10/2004 - Identical twins enter life from the same womb sharing the same genetic code, the same family, the same home, largely experiencing the same environment as they develop from infancy through childhood and mature into adults. When celiac disease strikes one identical twin, the odds are the other twin also has celiac disease. Twin studies lead to the conclusion that celiac disease is strongly linked to genetic factors. Yet one identical twin may develop celiac disease while the other twin may remain completely free of celiac disease for decades if not for a lifetime.
One study looked at 20 pairs of identical twins and 27 pairs of fraternal twins where at least one twin of the pair was known to have celiac disease. In 75% of the pairs of identical twins, both twins had celiac disease. In contrast, in only 11% of the pairs of fraternal twins did both twins have celiac disease. However, in 25% of the 20 identical twin pairs studied, one twin of the pair did not have celiac disease1. In another study which followed 5 pairs of female identical twins for 11-23 years (at least one twin of the pair having celiac disease or dermatitis herpetiformis), it was found that two of the twins who began the study with neither celiac disease or dermatitis herpetiformis remained free of the disease throughout the study2. In other words, something beyond genetics, some environmental factor, seems to be responsible for the onset of celiac disease. Exactly what is it that makes one twin intolerant to gluten and not the other?
Looking for Answers
To find an answer, one might start by asking when do signs of an intolerance to gluten first begin to emerge? A recent study in the UK looked at a screened sample of 5,470 children aged 7 years old and found 54 who tested positive for both tTG antibodies and IgA-EMA (tissue transglutaminase and antiendomysial antibodies) indicating celiac disease is likely present. This 1% prevalence in children is comparable to the 1% prevalence of celiac disease in adults in the UK. Since the prevalence of celiac disease is not greater in adults, this suggests that the onset of celiac disease begins in early childhood, even in cases where celiac disease is not diagnosed until later in adulthood. The authors of this study concluded, “The search for the trigger resulting in the breakdown of immune tolerance to gluten therefore needs to focus on infancy and intrauterine life3.”
Breast-feeding has long been thought to delay or reduce the risk of developing celiac disease in children. This effect has been attributed to a number of potentially protective milk components and antibodies passed from the mother. Studies relying on questionnaires have found that the onset of celiac disease in children is significantly delayed if gluten is introduced into the diet while the child is still being breast-fed4-7. The effect of epidermal growth factor (EGF), a component of breast milk, was studied in newborn rats. Interferon-gamma and gliadin, a gluten protein, were administered to rat pups to induce gluten enteropathy. Celiac disease-like villus atrophy was found in rat pups fed an artificial milk diet without EGF but not in breast-fed pups or pups supplemented with EGF8. Recent research shows that breast milk also passes bacterial flora from mother to newborn9. Growth factors found in human milk have been shown to aid in establishing predominant species of commensal bacteria in the gut of breast-fed infants10. The makeup of microflora which colonize the gut in early infancy is dependant on many factors, including whether babies are bottle-fed or exclusively breast-fed, whether or not delivered by caesarean section, on treatment in neonatal intensive care units, hygienic conditions, and antimicrobial procedures. Initially, it is the maternal microflora that is the source of bacteria for the newborn gut. A diet of breast milk induces the development of a flora rich in Bifidobacterium in full-term infants11. The possibility that these microflora play critical symbiotic roles in the development of the intestine and its immunological functions has not yet been considered as a factor in the onset of celiac disease.
The Beneficial Roles of Gut Bacteria
Over 500 species of bacteria may be present in the human gut in concentrations of between 100 billion to 1 trillion microbes per gram adding up to about 95% of the total number of cells in the human body12,13. For many years it has been known that gut bacteria play an important and beneficial role in one’s health. Extraordinary new findings on how commensal microflora participate in early gut development and in the development of the immune system have been uncovered by recent research. Here is sampling of some of these discoveries:
A study of 64 healthy formula and breast-fed infants, aged 0-6 months, examined fecal samples for intestinal colonization of Bacteriodes fragilis, Bifidobacterium-like, and Lactobacillus-like bacteria, and compared these results with counts of IgA, IgM, and IgG antibody-secreting cells in blood fluids drawn from the infants. The result was that infants colonized with B. fragilis at one month of age had significantly higher counts of IgA- and IgM-secreting cells at the age of two months than infants not colonized with B. fragilis. It was concluded that colonization timing and the type of bacteria colonizing the gut of newborns may influence the maturation of the naive immune system14.
Bacteriodes thetaiotaomicron, a species abundant in the guts of humans and mice, has been the focus of much research, chosen because of its predominance in the microflora and ability to be genetically manipulated. Studies of this microbe introduced into the developing guts of gnotobiotic (germ-free) laboratory mice have found B. thetaiotaomicron seems to communicate with host cells in the intestine, altering and influencing gut development and function.
One study has shown gene activity in the host is affected by B. thetaiotaomicron colonization. Using sophisticated DNA microarray devices, a comparison of gene expression of some 25,000 mouse genes was made between germ-free and B. thetaiotaomicron colonized mice. The activity of 118 genes was found to be increased or reduced by colonization. These genes are involved in several important intestinal functions, including nutrient absorption, intestinal permeability, toxin neutralization, intestinal blood vessel development, and postnatal gut maturation suggesting that these functions should be examined further in future studies15.
An influence on fructose production in the gut by B. thetaiotaomicron was the first finding uncovered by researchers. Pre-weaned mice produce fructose sugar on the surface of cells lining the intestine providing a food source helping to establish commensal bacteria. B. thetaiotaomicron colonizing the gut of germ-free mice causes intestinal cells to continue fructose production after weaning. If B. thetaiotaomicron is not present after weaning, fructose synthesis stops. B. thetaiotaomicron actually senses when its supply of fructose is low and instructs the host to produce more fructose in response16.
Gene activity findings led researchers to look at the development of the intricate network of intestinal blood vessels in mice raised germ-free and in mice raised colonized with B. thetaiotaomicron or normal gut flora. When the mice reached adulthood, capillary development in the intestines was examined. Capillary development in mice colonized with B. thetaiotaomicron or normal flora was normal and complex, but capillary development in the germ-free mice was immature and arrested. Further, it was, found for blood vessel development to occur, these microbes must interact with Paneth cells (epithelial cells located at the base of the “crypts” in the small intestine)17.
The relationship of B. thetaiotaomicron with Paneth cells was further studied. It was discovered that Paneth cells produce a protein called angiogenin 4 or Ang4 and that Paneth cells are induced to express Ang4 by B. thetaiotaomicron. Ang4 and other angiogenins were found to exhibit bactericidal and fungicidal activities against certain known pathogens. It appears that B. thetaiotaomicron and other commensal microbes, which are themselves resistant to Ang4, take part in shaping the microbial ecology of the gut and innate immunity18.
Another study found a relationship between commensal bacteria and the development of gut-associated lymphoid tissue (GALT) in rabbits. GALT consists of lymphocytes and organized tissues called Peyer’s patches and mesenteric lymph nodes (MLNs) located within the intestinal mucosa, which are involved in the induction of immunity and tolerance. During the first few months after birth, newborn animals and humans rely on antibodies passed maternally to fend off infections until the immune system can mature. After those first few months, a diversification of antibody repertoire normally takes place within the GALT. When, shortly after birth, the appendices of rabbits are tied-off and isolated to prevent colonization by microflora, GALT development within the appendices is arrested. Rabbit pups delivered sterilely, isolated and hand-reared on a sterile diet exhibited underdeveloped GALT and antibody repertoires. In further experimentation, a number of different bacteria species were introduced into surgically-rendered, germ-free appendices of rabbits. No bacteria species alone promoted GALT development. However, the combination of Bacteroides fragilis and Bacillus subtilis consistently resulted in the development of GALT and antibody repertoire. The conclusion is that specific combinations of microflora are required for GALT development19,20.
In other research, the composition of commensal flora in rats was shown to alter intestinal permeability. Colonization with Escherichia coli, Klebsiella pneumoniae, and Streptococcus viridans significantly increased colonic wall permeability while colonization with the common probiotic strain, Lactobacillus brevis, significantly reduced permeability of the colon wall. Bacteroides fragilis induced only a slight permeability reduction21.
Gut pathogens in combination with stimulation by cytokines such as TNF-alpha (tumor necrosis factor) can cause cells of the intestinal epithelium to respond by releasing proinflammatory cytokines like interleukin-8 (IL-8). A study found that probiotic strains, Bifidobacterium longum and Lactobacillus bulgaricus, can suppress IL-8 secretion in intestinal epithelia when stimulated by proinflammatory cytokines. Hence, some probiotic strains of bacteria may be able to down-regulate inflammation in the gut22.
Other beneficial functions of microflora include the fermentation and removal of non-digestible dietary residue and the mucus residue produced by the epithelia; the derivation of energy as short-chain fatty acids by fermentation of carbohydrates in the colon; the production of vitamins, particularly those of the B group and vitamin K; the absorption of minerals and ions including calcium, magnesium and iron; and the formation of a protective functional barrier against pathogens23,24.
A Role for Bacteria in Celiac Disease?
As can been seen, commensal microflora play a myriad of complex, diverse and important roles in normal health and development. Much remains to be investigated, and new roles and functions microflora play are waiting to be discovered. The possibility that commensal bacteria are involved in the pathogenesis of celiac disease cannot be overlooked. Certainly, differences in the mix of microflora could account for why one identical twin may develop celiac disease while the other does not. Could the mix of commensal bacteria in newborn infants set the stage for the development of celiac disease? Could the onset of celiac disease be triggered by an event such as illness, use of antibiotics, stress, or pregnancy which alters the mix of microflora opening the door to a pathogenic interaction with gluten? One recent study has already found an association between antibiotic use and the development of Crohn’s disease25.
Over the course of the last few years, much new understanding of the pathogenesis of celiac disease has come to light, but a fundamental question remains unanswered: Why does the immune system fail to tolerate gluten in some people? A possible mechanism involving one or more unidentified species of commensal bacteria possibly explaining why tolerance to gluten fails will be proposed and discussed here.
Tolerance and Immunity
The subject of tolerance and immunity is involved and complex, and science remains far from fully comprehending its workings. At heart, is how the immune system decides to react when an antigen is first presented to a naive T cell. The response of the immune system to an antigen is mediated and regulated by cell secretions of numerous proteins called “cytokines” sensed by a multitude of receptors on the various specialized cells of the immune system. Structural components of pathogens are also sensed by immune cell receptors called “Toll-like receptors”. Antigens may be any substance foreign to the body and may or may not actually be harmful. They could be components of food, or could be components of either friendly or pathogenic organisms.
In celiac disease, the antigens are those gluten peptides which survive the process of digestion. In the current understanding of celiac disease, these peptides are transported across the mucosal epithelium as polypeptides. In mainly the subepithelial region, gluten peptides undergo a process called deamidation by an enzyme called tissue transglutaminase (tTG). A peptide is a chain of amino acids. Deamidation is a process that converts glutamine amino acid components of a gluten peptide into glutamic acid components. In the lamina propria region of the intestines, deamidated gluten peptides are taken up by antigen presenting cells called dendritic cells and presented by HLA-DQ2 or -DQ8 molecules on the surface of dendritic cells to receptors of gluten-sensitive naive CD4+ T cells (Note CD here refers to a “cluster of differentiation” number, a numbering system for the cell-surface molecules which identify T cell type). Activated CD4+ T cells then differentiate and proliferate. Some T cells interact with B cells which, in turn, then differentiate into plasma cells producing antigliadin, antiendomysial and anti-tTG antibodies. Other T cells become natural killer or cytotoxic T cells, secreting cytokines which cause inflammation and damage to the enterocytes in the epithelium. Connective tissue cells called “fibroblasts” increase their output of matrix metalloproteinase enzymes which may play an active role in villus atrophy. Intraepithelial lymphocytes also increase, but their role is not clear26-29.
Human leukocyte antigen (HLA) genes encode the class II molecules DQ2 and DQ8, the key genetic risk factors in celiac disease. The HLA system is the human version of the major histocompatibility complex (MHC). HLA class II molecules are expressed on the surface of antigen presenting cells such as dendritic cells. Virtually all celiac disease patients carry DQ2 or DQ8, but carrying DQ2 or DQ8 alone does not confer celiac disease. DQ2 and DQ8 molecules may be encoded by several different haplotypes. Haplotypes are combinations of alternative genes for the same trait (alleles) occupying different locations on a chromosome which tend to be inherited as a group. These DQ2 and DQ8 molecules play a central role in the pathogenesis of celiac disease. The function of HLA class II molecules is to bind peptide antigens and present them to CD4+ T-cell receptors. The pattern of amino acids in the makeup of the chain that forms the peptide antigen is called an epitope, and that pattern is crucial to the binding between HLA molecule and peptide. It is the misfortune of celiac disease patients that epitopes of deamidated gluten peptides just happen to match up and firmly anchor into the binding grooves of DQ2 and DQ8 molecules. This strong binding results in the activation of CD4+ T cells and the subsequent processes which damage the intestinal epithelia. But why is it that CD4+ T cells are not activated in everyone who possesses the appropriate HLA-DQ2 and -DQ8 haplotypes? The question arises again. Why is one identical twin tolerant to gluten and not the other?26-30
Whether an outcome of tolerance or intolerance results when a dendritic cell presents an antigen to a naive T cell depends on many factors. A dendritic cell is a special type of white blood cell (leukocyte) which circulates throughout the body looking to acquire antigens. Dendritic cells engulf and internalize antigens through a process called endocytosis. In receptor-mediated endocytosis, dendritic cells express a variety of surface receptors to capture protein antigens. In macropinocytosis, dendritic cells surround and “drink up” soluble antigens. In phagocytosis, dendritic cells engulf pathogenic bacteria, viruses, fungi, dead or infected cells, or their products. After digestion and processing, the antigens are bound to HLA (or MHC) molecules and expressed on the surface of dendritic cells for presentation to T cells. Antigen presentation occurs after dendritic cells migrate to the lymph nodes which are rich with T cells. T cell activation also requires secondary stimulation by costimulatory molecules expressed on the dendritic cell surface. Dendritic cells have three stages in their life cycle: Precursor, immature and mature. Precursor dendritic cells arise from the bone marrow. Subsets of precursor dendritic cells have been identified that grow and differ with regard to observable characteristics (phenotype), function and anatomical location. Studies have linked dendritic cell subsets with particular functions such as T cell differentiation or tolerance induction. Immature dendritic cells spread throughout tissues seeking antigens. Dendritic cells enter the mature stage when they reach the lymph nodes after antigen capture and having become primed and ready to activate T cells with antigens and costimulatory molecules. The processing of antigens produces roughly 100,000 to 300,000 peptide-laden HLA molecules on the dendritic cell surface, most peptides represented by about 100 copies. A single mature dendritic cell is capable of stimulating 100–3,000 T cells31-34.
Immature dendritic cells are capable of phagocytosis of bacteria. Dendritic cell phagocytosis of Salmonella and Borrelia burgdorferi has been observed and studied. Immature dendritic cells roaming the lamina propria below the epithelial cells of the intestine not only capture bacteria which invade and cross the epithelial barrier, but have been observed reaching through the tight junctions between epithelial cells with their dendrite arms to directly sample non-invasive bacteria in the gut lumen and mucosa tissues outside the epithelium34-37.
Immature dendritic cells express a variety of surface receptors which when stimulated cause dendritic cells to mature and respond in specific ways which can result in tolerance or immune activity. These receptors include Toll-like receptors (TLR), cytokine receptors, TNF (tumor necrosis factor) receptor, immunoglobulin (antibody) receptors, and sensors for cell death. TNF and other cykotine inflammatory mediators signal infections. In particular, interleukin-1 (IL-1) can prevent oral tolerance in mice by altering the response of normally tolerogenic dendritic cells into an active immune response32,34.
Toll-like receptors are known as pattern recognition receptors which identify structural components found only on the surface of bacteria and other pathogens. These components are referred to as pathogen-associated molecular patterns (PAMPs). At least 10 types of TLR have been identified in humans and given the designations, TLR1-TLR10. Examples of PAMP include microbial carbohydrates like the toxin lipopolysaccharides (LPS), flagellin, products from bacterial cell walls, bacterial RNA and DNA. Signaling through different TLR evokes distinct biological responses. TLR expressed differently by different dendritic cell subsets may determine the manner in which dendritic cell subsets respond to particular microbial structures34,39.
Mature dendritic cells can produce cytokines while activating CD4+ T cells which may influence T cell differentiation and function. Activated T cells divide and proliferate and differentiate into a variety of types. Tolerance and immunity induction are influenced most by differentiation into type 1 and type 2 helper T cells (Th1 and Th2) and regulatory T cells. The type of cytokines produced by the T cells determine their classification. Th2 responses favor tolerance. Th1 responses favor immunity and inflammation. Regulatory T cells suppress immune responses. IL-10 produced by dendritic cells appears to contribute to Th2 and regulatory T cell responses. Dendritic cell production of IL-12, IL-18, and IL-23 contribute to a Th1 response34,40.
Why Does Tolerance to Gluten Fail?
Okay. So why does the immune system fail to tolerate gluten in celiac disease? The immune system receives and responds to all kinds of signals from a pathogen, but how can a simple gluten peptide turn this complex immune machinery into a force against itself? Thinking about this leads to a very provocative question:
What if instead of responding to gluten peptides alone, the immune system responds to a pathogenic gut bacteria which routinely ingests gluten peptides?
A 33 amino acid gluten peptide has been identified as the primary initiator of the inflammatory response in celiac disease. This peptide contains a number of amino acid sequences which correspond to epitopes known to activate T cells and initiate celiac disease response. In particular, this 33-mer peptide was identified because it remained intact in the residue of a solution of gliadin mixed with gastric and pancreatic enzymes. This demonstrates some gluten peptides are difficult to breakdown by normal digestive processes. Another experiment identified a 17 amino acid gluten peptide which also contained epitopes associated with celiac disease41,42.
Bacteria do not ingest nutrients in the normal sense. Nutrients are transported across cell membranes via several different mechanisms. Transported nutrients are necessarily limited in size. Nutrients are broken down externally by enzymes and by processes such as fermentation, an oxidation process resulting from acids produced by bacteria. Growth factors consisting of purines, pyrimidines, vitamins and amino acids are required by some bacteria in order to grow. Other bacteria are able to synthesize these essential growth factors. Researchers have found that some bacteria can transport and internalize amino acids in the form of peptides. Studies so far have found peptides up to 18 amino acids in length can be internalized by bacteria43-46.
Epitopes of gluten peptides deamidated by tissue transglutaminase (tTG) are believed central to celiac disease pathogenesis. However, a study of gluten response in children with celiac disease found that T cells can respond to native gluten peptides independent of deamidation47. Celiac disease may begin its course without deamidation. As the disease progresses, inflammation may cause an increase in expression of tTG. An increase in tTG expression has been shown during wound healing, in liver injury, and in response to an inflammatory stimulus by lipopolysaccharide48-50. Through a process called epitope spreading and with the increase in tTG expression, deamidation of gluten peptides is more likely to occur and T cell response to deamidated gluten peptides likely develops. tTG is expressed in the epithelial brush border and extracellularly in the subepithelial region26 (The brush border is composed of the microvilli found on each individual epithelial cell).
In the course of evolution of bacteria in the gut, it would seem highly plausible that at least one or more bacteria species have evolved and adapted in some way to transport, internalize and utilize gluten peptides as a source of amino acids. Since tTG is expressed in the epithelial brush border, deamidated gluten peptides are available to such bacteria (though in the early stage of celiac disease deamidation may not be required). If these bacteria colonize the gut and exhibit some pathogenic characteristic, such as expressing lipopolysaccharide, dendritic cells may be signaled to reach through the epithelial barrier into the lumen to sample and phagocytize the bacteria. When this bacteria is digested and processed by the dendritic cells, the antigens bound to HLA molecules and expressed on the dendritic cell surface are likely to include the difficult to breakdown, intact gluten peptides that have been internalized by the bacteria. As far as the immune system is concerned, these gluten peptides are indistinguishable from the other bacterial peptides bound to HLA molecules expressed on the dendritic cell surface. When these gluten peptide antigens are bound to HLA-DQ2 or -DQ8 molecules and presented to CD4+ T cells, the T cells simultaneously receive all the signals telling them that the gluten peptide is an antigen from a pathogenic bacteria. The result is that the immune system responds to the presence of gluten as though pathogenic bacteria were present. Such gluten-ingesting bacteria may be the missing link in the pathogenesis of Celiac Disease.
If these bacteria exist, there is now a clear explanation as to why one identical twin may develop celiac disease and not the other. Of course, the presence of such a bacteria in the gut of one twin and not the other would fully explain the discordance. It is also possible that such a bacteria may exist in both twins, but is kept under control by the mix of commensal bacteria colonizing the gut of one twin. Some disturbance to this mix, such as an infection or use of antibiotics, might provide an opportunity for this gluten-ingesting bacteria to colonize and proliferate to a level where its pathogenic properties, such as production of endotoxins, are sensed by the immune system initiating the onset of celiac disease. The existence of such bacteria could also explain why there may be varying degrees of gluten sensitivity, even in individuals without DQ2 and DQ8 molecules.
The possibility that these gluten-ingesting bacteria may exist raises another intriguing question: If these gluten-ingesting bacteria are controlled or eliminated from the gut, could tolerance to gluten be restored? There could be a very real possibility that celiac disease might be cured by eliminating these bacteria. After all, peptic ulcers can be cured by eliminating Helicobacter pylori.
So where should research go from here? The most obvious path would be first to try to find and identify any gut bacteria that has gluten peptides present within its cell membranes. From there, the possible link to celiac disease could be studied. Additionally, it would be quite valuable to initiate a large long-term study of the makeup of commensal bacteria in identical twins beginning at birth via fecal samples. By comparing the differences in microflora and the onset and discordance of diseases in identical twins over many years, the relationships of specific species of bacteria to specific diseases, including celiac disease, could be established. And if it proves to be true that gluten-ingesting bacteria cause celiac disease, a similar mechanism involving bacteria and peptides from other proteins may be the root cause for many other autoimmune diseases. A whole class of autoimmune diseases might be cured by eliminating specific species of bacteria.
Roy Jamron holds degrees in physics and engineering
from the University of Michigan and the University of California at Davis
and actively pursues and investigates research on celiac disease and related