Paul Ashwood, PhD
University of California, Davis, M.I.N.D. Institute
Dr. Ashwood earned his Ph.D. in Immunology at King’s College in London in 1999. He is currently an assistant professor in the Department of Medical Microbiology and Immunology at the M.I.N.D. Institute at the University of California, Davis.
Presentation:
There is a growing awareness that the immune system plays an important role in autism. The interface between the cellular immune and nervous systems is exceedingly complex, with extensive communication occurring between them in both health and disease. Immune cells and neuroimmune molecules such as cytokines and chemokines modulate brain function, affecting cognitive and emotional processing.
For example cytokines, which are chemical mediators that confer messages in the immune system, also have widespread effects on neuronal pathways, and may contribute to common features of autism, such as mood and sleep disturbances. In fact, normal neurodevelopment of the brain is contingent upon successful communication between these two systems.
What do we know about the immune system in children with autism? We know that products of an activated immune system are important in neurodevelopment, and that neurodevelopment relies on a well-ordered, well-balanced immune system. We know that various immune system abnormalities have been reported in children with autism throughout the world by a number of different laboratories. Enhanced autoimmunity has been shown in some children, but reduced immune response has been shown in others.
Moreover, very recently we have seen the development of many animal models of autism that have an immune basis. Importantly, in children with autism and GI symptoms, dietary restrictions have been shown to improve both intestinal inflammation and behavior, suggesting that there is a link between the two. What we also know is that the nervous system and the immune system are in constant communication.
Many products of the nervous system—such as opioid peptides, 5-HT, vasopressin, VIP, and oxytocin—have important immunomodulatory effects. And products of immune cell activation, including inflammatory cytokines IL-1 and TNF-alpha in particular, can affect mood and sleep. Previously published data have shown that there are decreased immune cell numbers, in particular T lymphocytes, diminished responses to T-cell stimulants, decreased NK cell activity, skewed cytokine profiles, and the presence of antibodies directed to central nervous system proteins and to brain proteins. Inflammation may alter the properties and the function of neurons and alter neurotransmission.
Furthermore, activated microglia cells in the brain respond to inflammatory signals from the periphery and can release cytokines in response. However, so far there has not been a general consensus on what the immune system abnormalities are, and so far the findings are not present in all of the patients with autism. Often the literature has shown conflicting results, which may be due to the diversity of the autism patients that have been studied as well as some of the control groups that have been chosen. Larger study cohorts are definitely needed.
What we would really like to know is, is there a signature autoantibody response? Is there a signature cell-mediated immune response, such as a cytokine response, that we can use that will help, first, to identify and distinguish specific children with autism and, second, that will help identify treatment for their diagnosis? We would also like to evaluate the immune response to give us clues to a possible causative agent or agents. I will now split this talk into three main parts: First, talking about mucosal response (inflammation in the GI tract); second, the work presently ongoing at the M.I.N.D. Institute, which is looking at autoimmunity and the presence of ‘self’ reactive antibodies; and finally, the peripheral response, which includes serology and the cytokine profile.
So, why study the GI tract as an immunologist? Primarily because it is the single largest immune organ and has far more immune cells than any other organ in the body. In addition, in autism there have been findings that include reduced glycosylation and defective sulfation of ingested phenolic amines such as Tylenol. There are other findings that are important to an immunologist, such as bacterial overgrowth, increased permeability in the gut, and the importance of dietary intervention.
The estimates of GI problems in autism range from 20% to 76%. The main complaints, as you've already heard, are abdominal pain, constipation, and diarrhea. The group I’m going to talk about includes those who have GI symptoms sufficient to justify invasive colonoscopy and endoscopy. (Points to slide) As you have heard, in the terminal ileum we have seen lymphoid nodular hyperplasia, and in the colon prominent lymphoid follicles. These are indicative of immune responses. By histology we see that there is an increased level of immune cells. Using a technique called immunohistochemistry we can see that in a typically developing control there are very few cells positive for CD3, a marker of T-lymphocytes in the epithelial layer (which is the layer that provides a barrier between the gut lumen and the body proper), and the lamina propria layer (the mucosal layer). In contrast, in children with autism and GI symptoms there is an increase in the number of immune cells in both layers, the outside epithelial layer and within the lamina propria, when compared with typically developing controls and also an increase compared with cerebral palsy controls.
One of the best things you can do in science to confirm a hypothesis is to use a separate technique and a separate study group but find the same thing. Using a separate technique, flow cytometry, we found that, compared with neurotypical controls, there was an increase in the lymphocytes in the autism population in both the lamina propria and the epithelial layers. We performed this work in the colon, the duodenum, and also in the terminal ileum, and found that there were increases in lymphocytes at all three sites that we studied. Increases in both T and B cells (the B cell which can produce antibodies) were seen.
Dr. Wakefield also alluded to some epithelial changes that we saw which can be important and indicate that there is a disruption in the barrier function of the intestinal lining. This could allow things to get into the body which could exacerbate the inflammation. Also, we saw that there were increases in NK cells and monocytes. These are important in the innate immune response. We know these cells are increased in the intestinal lining, but are they doing anything? This is an important question. Are they functionally active? Do they produce cytokines? Broadly speaking, what we can do is split up cells into two different groups: Those that produce an inflammatory response and lead to inflammation-inducing cytokine such as TNF alpha, and those that are there to balance out such a response and counter the inflammatory signals, such as the regulatory T-cells product IL10.
Jyonouchi, looking at a group very similar to ours with GI symptoms and a regressive phenotype, showed that there was a skew towards inflammation with an increase in TNF alpha. She used an ELISA technique ("Proinflammatory and regulatory cytokine production associated with innate and adaptive immune responses in children with autism spectrum disorders and developmental regression" J Neuroimmunol. 2001 Nov 1;120(1-2):170-9). In an independent study with a separate technique we were able to extend these findings, and were able to say that TNF-alpha had been produced by CD3 lymphocytes—the cell group that is the major infiltrator in the gut.
Here (points to slide) you see that there is an increase in TNF alpha and an increase in interferon gamma and IL2—the proinflammatory cytokines in CD3 lamina propria cells in the duodenum of children with autism. What's interesting is that the balance towards inflammation has shifted because there is also a decrease in IL10, the regulatory cytokine. We saw the same thing in the colon, the duodenum, and the terminal ileum. What is also very interesting is that the level of TNF alpha was increased in children with autism to levels similar to those seen for established inflammatory bowel diseases, but what was different was the level of a balancing IL10 signal, which was much further reduced in the children with autism. This may in part be due to effective treatments in the IBD group that are able to address the balance between pro-inflammatory and regulatory signals.
To summarize this first part of my talk, within the mucosa there is a cytokine milieu that is directed towards a pro-inflammatory response. Overall, there is evidence of a diffuse, subtle panenteric pathology, with increased inflammatory cell infiltrate. There is increased production of pro-inflammatory cytokines, and there is a decreased production of the regulatory cytokine IL10. There is also this possibility that an autoimmune response occurs that affects the epithelial cells and may lead to a disrupted intestinal barrier.
Moving along to the second part of my talk, at the M.I.N.D. Institute we are investigating whether there is any evidence of autoimmunity—the reaction of antibodies toward "self" proteins—in children with autism. In the literature, there have been a number of papers that have shown that there are antibodies that are directed to brain proteins present in children with autism. In addition there is also significant history of autoimmunity in families who have children with autism.
(Describing the slide) In this experiment we looked for the evidence of antibodies directed towards brain extracts. I won't go into the details of this particular test too much, except to say that we took commercially available human brain extracts and we ran sera—that is the component of the blood that contains antibodies—from patients with autism and from controls and found that there were distinct patterns of reactivity to brain proteins in the autism group that were not there in the controls. Not all the autism patients had the same patterns of reactivity. Some had two; some just had one, but they were present in an increased number of the autism patients and not the neurotypical controls.
We found this for proteins that were present in different brain structures such as the hypothalamus, the thalamus, and the cerebral cortex. Overall what this tells us is that antibodies are generated—maybe as a result of inflammation in the brain—in children with autism that can bind to proteins present in brain tissue. Potentially these antibodies may be able to block or impair the function of neurons in the brain that contain these proteins. To confirm the study we used a different technique—we used immunohistochemistry on monkey cerebellum sections—and here we covered the sections with serum either from autism patients or serum from controls. (Points to slide) Again what we saw were distinct patterns of staining/reactivity on very specific cells within the granular layer of the cerebellum. The nature of these particular cells needs confirmation, but their importance may be in the control of stimulatory signals to purkinje cells.
So for the autism group, using two separate techniques, we have been able to demonstrate the presence of autoantibodies directed to brain proteins in the children with autism, which were not present in the controls. These patterns are differential, i.e. not the same in all autism subjects, but the majority of autistic children have at least one pattern. What these autoantibodies do, unfortunately, at present we do not quite know. We would like to know if they cause neuronal damage, if they block or impair neuronal function, or whether they are generated as a consequence of an inflammatory process that occurs within the brain in autistic children.
Moving on to the final part of my talk, the peripheral response: First, are there different antibody responses to infectious agents? Second, are peripheral T cells in autism patients activated to produce cytokines? There has previously been much anecdotal data that suggests that in autism the immune response is dysregulated. Often parents have reported that their child has had prolonged and recurrent infections. This may indicate that the immune response is not acting properly and that there would be differences in the levels of antibodies to infectious agents and differences in cytokine levels within the blood.
If we look at antibody levels, we can check to see if there have been responses against foreign agents such as bacteria or viruses. In this study for the large part we chose antigens (agents on bacteria/viruses that elicit a specific immune response) that are present in vaccines, primarily because we know the vaccination history of the children and that they have been exposed to these antigens. We also looked at total immunoglobulin (antibody) levels in autistic and control populations, and specific subclasses of these immunoglobulins. We found that the total levels of immunoglobulins (IgG, IgM, and IgA) were decreased in the autistic group compared with age-matched neurotypical healthy controls and children that were age-matched who had other developmental disabilities, but that the levels were similar to age-matched sibling controls. So far this has been seen for more than 100 patients in the autistic group and 140 in the neurotypical control group, so we are looking at large numbers.
There was also a decreased response when we looked at specific responses to antigens contained within the vaccine DTaP (diptheria, tetanus, and bordetella), we noticed that there was a decreased response in autistic patients compared with neurotypical controls, for bordetella, tetanus, and diphtheria. For infectious agents that are present in the environment, and that these children would potentially be exposed to, such as CMV (cytomegalovirus) or influenza A and B, we also saw a similar pattern of reduced immunoglobulin responses in autistic patients compared with age-matched neurotypical controls.
However, when we looked at measles, mumps and rubella, we saw no statistical difference between the groups, which was a pattern that was different considering the background of decreased or down-regulated antibody responses. These patterns of antibody levels were similar for autistic patients and sibling controls. Overall, we can see that there’s a suboptimal antibody response to some antigens from infectious agents that may indicate that there is a dysfunctional immune response in children with autism. Incidentally, when we tested for antibody responses to food antigens we did not notice any differences in this population.
We now know that there is a subtle intestinal pathology with increased inflammatory T cell infiltrate. There are increased peripheral and mucosal T cells that produce increased levels of pro-inflammatory cytokines but produce reduced amounts of regulatory cytokines. There is a possible autoimmune response. What we want to know is whether or not there is any evidence of a chronic antigenic or viral stimulation that we can look at at the lymphocyte level. We can do this a number of ways.
One particular way is by using flow cytometry. Activated T cells of the subset which have a CD8+ phenotype can be differentiated based upon their level of activation and stimulation, through the expression of markers on their cell surface such as the co-stimulatory molecules CD28 and CD27. This has been shown for cytomegalovirus, HIV, Epstein-Barr Virus, and Hepatitis C. The paradigm is that a naive cell (one that has not experienced antigen) will express CD28 and CD27 on its surface, if the cell encounters a virus, an antigen, or a bacteria, it will then proliferate and become a functional effector cell and lose CD28 and CD27 expression. In children with autism and GI symptoms there is evidence that shows that a population of CD8+ cells that have lost their CD28 and CD27 expression is increased when compared with age-matched controls.
This data adds to a growing picture, where we have immune activation, which is consistent with a persistent chronic immunological stimulation. We previously asked the question of mucosal cells, and we'll now ask the question of peripheral response: Are the infiltrating T cells activated to produce cytokines? Can we measure mediators in the plasma that will tell us whether there's an activated immune system?
Once we have discovered these biomarkers for an active immune system, can we identify different groups within the autistic population? Can we then use this profile of biomarkers to help for screening or treatment? And then, once kids are being treated, can we use this pattern of biomarkers to see if they are getting better?
In conjunction with a company based here in Austin (Rules Based Medicine), we are looking at a number of different biomarkers. Through this work we have seen that there is a difference in the immune response in children with autism compared with age-matched controls. Not only are a number of biomarkers present in children with autism that are different from neurotypical controls, but the pattern of biomarkers may differ between children with different behavioral phenotypes. This work is being confirmed in a larger study.
As we have seen earlier, the immune response is all about balance, for example the balance between an immune response that eliminates infections from the body and one that is too aggressive and produces injury to "self." There is also the balance between cytokine responses such that pro-inflammatory responses are balanced by regulatory cytokine responses.
In children with autism I have presented evidence to suggest that, somehow, these balances have been disrupted. In the autistic children that have GI symptoms, there is an increase in TNF alpha but no corresponding rise in the counter regulatory cytokine IL-10, suggesting a shift in the immune balance. The activation of the immune system may affect the function of both afferent nerves and the CNS. As a result of activation of the peripheral immune system, various immune factors, including activated lymphocytes, antibodies, and cytokines, could cross the blood-brain barrier and induce inflammation in the brain and CNS.
Various stimulators may be involved, such as pathogenic agents and molecules that are not usually present under physiological conditions (e.g, viruses, bacteria, neuroactive molecules from the diet). Immune defects might also be present, such as altered T cell populations, decreased regulatory cells, altered antigen presenting cells/dendritic cell populations, altered or inappropriate cytokine production upon stimulation, or loss of "self" tolerance.
I will finish the talk with this diagram of what may potentially be going on: There is primary systemic disease such as you may see in the GI tract; this produces cytokines from macrophages and T cells that can cross through the blood-brain barrier and activate microglial cells, which could then affect and impair the function of other neurons. In addition, there could be a release of antibodies that bind to the neurons and impair function or cause damage.
In future studies we need to verify previously reported immune function/dysfunction in autism. This will include the identification of activated immunological cells and their response to immune stimulation. There is clear evidence that the immune system is involved in autism, but we now need to perform larger cohort studies with well defined age-matched controls. Thank you.
Return to Top |
