By Rosemary H. Waring , School of Biosciences, University of Birmingham, Birmingham B15 2TT U.K.
Sulphate synthesis
I have always been interested in the sulphation pathway as addition of a sulphate group can make dramatic changes to the properties of both drugs and tissue components. Our group first started working in the field of autism about 15 years ago, when we were asked to measure the metabolism of paracetamol in an autistic child. At the time, I had only heard the orthodox medical view that autism was ‘all in the mind’ and had no biochemical basis. To our great surprise, we found that children with autism, unlike the age-matched controls, were much less able to form the sulphate conjugate of paracetamol, although the other metabolic pathways were normal. We went on to look at the levels of sulphate in the blood plasma, because sulphation capacity depends on both the amount of sulphate available and also the activity of the enzyme that carries out the reaction. We found that autistic children generally had low sulphate levels, typically about 10-15% of the control values. Sulphate is produced in vivo by oxidation of methionine or cysteine, both sulphur – containing amino acids which are provided from dietary proteins, and this pathway probably provides ~ 80% of the sulphate required in man.
The first stage in this process involves the enzyme cysteine dioxygenase (CDO); cysteine sulphinic acid is formed and undergoes fission to provide sulphite (SO 3 2- ) ions which are then further oxidised to sulphate (SO 4 2- ) ions by the enzyme sulphite oxidase (SOX). Obviously, if CDO or SOX have reduced activity, the provision of sulphate will also be decreased. The human CDO gene is localised to chromosome 5 (5q22-23) and it is interesting that analysis of 110 multiplex families with autism, where one sibling had autism and the other a diagnosis of Asperger’s syndrome or pervasive developmental disorder, suggested linkage on chromosomes 5 and 19 while a study on ADHD (attention deficit/hyperactivity disorder) found a linkage to chromosome 5q33. The CDO protein is found in heart, thyroid and kidney, as well as brain and the liver, localisation in the CNS being particularly found in the cerebellum and the Purkinje neurons; these are known to be abnormal in patients with autistic spectrum disorders. CDO activity is variable in human populations and there are sub-sets with lower activity (~ 30% of the population) or null activity (~ 3% of the population).
The null S-oxidisers are heavily over-represented in chronic disease states with an auto-immune component such as rheumatoid arthritis and primary biliary cirrhosis; in general auto-immune problems are more common in the family background of autistic children. We now know that inflammatory cytokines such as TNF-?, which are at relatively high levels in many autistic children and in auto-immune diseases, can reduce expression of CDO and SOX and therefore reduce the supply of sulphate for conjugation with drugs and biocomponents. Expression of both CDO and SOX was inhibited in vitro at levels of 0.1 ng/ml TNF- a , concentrations which could easily occur in vivo during an infection. This work, however has all been carried out in vitro and it is difficult to deduce from this whether the effects would also occur in vivo . To check this, a small pilot study was carried out in this laboratory when students and staff were offered a vaccination against hepatitis B (the antigen is in fact one of the virus coat proteins). Volunteers were asked to take a therapeutic (1000 mg) dose of paracetamol (acetaminophen) on Day 1 before the vaccination which took place on Day 7 and paracetamol was also ingested on days 8, 10 and 15 afterwards. The vaccination had no ill effects apart from slight reddening at the injection point in some volunteers but it greatly altered the detoxication pathways for paracetamol. In particular, the phase 2 conjugation reaction of sulphation was severely depressed, only reaching control values about a week later. It is clear that even a simple vaccination in healthy volunteers can dramatically affect some metabolic pathways, at least in the short term. It is obvious from these results that the process of sulphate formation and sulphation itself is potentially severely disrupted in inflammation; the in vivo findings appear to correlate with those seen in vitro , suggesting that cell culture systems can act as useful models. It is interesting that autistic children who were challenged with a paediatric dose of paracetamol were less able to form its sulphated derivative than controls of the same age although the glucuronidation pathway was unaffected. Their general metabolic profile was very similar to that found for the adult volunteers with a Hepatitis B vaccination. This seems to be a general finding which has been replicated in UK, Italian and USA populations and probably reflects the fact that raised cytokine levels in autism have secondary effects on sulphation of a range of substrates.
This may explain why some children with autism are reported as responding badly to dosage with paracetamol and with other drugs; obviously toxic effects are more likely if clearance is impaired by reduced metabolism to water-soluble derivatives.
Sulphate in the brain
Principally, sulphation is a major inactivation pathway for catecholamines such as the neurotransmitter dopamine, about 80% of which is sulphated in man. Usually, when chemical neurotransmitters are released in the central nervous system, they act at receptor proteins and are then inactivated by sulphation or by FAD-linked mono-oxygenases or alternatively are carried by transporter proteins back into the initiating neurone. Failure of a major pathway such as sulphation will lead to a neurotransmitter imbalance and raised serum and CSF levels of dopamine and elevated urinary levels of dopamine metabolites have been found in autistic children. In rats, high dopamine concentrations like this are associated with stereotyped and repetitive behaviour, not unlike that sometimes seen in autism. Other catecholamines, such as noradrenalin, also control behaviour and affect mood so that changes in their levels can have obvious effects.
Sulphation also affects the synthesis of brain tissue. Sulphated polysaccharides and glycosaminoglycans are so important in the development of the foetal and neonatal brain that any alteration in their structure may have serious consequences – it is currently thought that these compounds act as ‘scaffolding’ to direct the direction and ‘wiring’ of brain neurons. Sulphate transport across the placenta increases dramatically around the time of birth when most of the glial cells are being formed and these increased levels of sulphate are associated with formation of astrocytes and oligodendrocytes from progenitor cells.
Children have higher levels of plasma sulphate than adults (0.47 nmol/l at birth decreasing to 0.33 nmol/l at 36 months; adult levels are around 0.27 nmol/l although there can be a wide range). This relatively high level of sulphate, as compared with the adult state and with, for instance, laboratory rats, seems to show that humans have a definite requirement for sulphation in neuronal development both before and after birth and that reduced levels could affect brain structure and function,. Recent research suggests that brain development relies on particular patterns of sulphation occurring in the right sequence; rather like the fairy tale of ‘Goldilocks and the three bears’ we need ‘not too little, not too much but just right’! But are infections in pregnancy, which would be expected to give raised cytokine levels and potentially reduce sulphate formation, actually linked with altered brain development or function in the baby? In a small pilot survey in this laboratory which looked at 200 mothers of autistic children, it was found that they were eight times more likely to have received antibiotic treatment for an infection in pregnancy than age-matched controls and 5 times more likely to have had long-term therapy for recurrent infections. Certainly studies using rats as a model have shown that increased cytokine levels in pregnancy affect the development of neural integration in the neonate; it seems probable that one of the many factors in autism may be raised levels of cytokines or other factors, possibly released in infections, affecting sulphation and neurodevelopment.
Sulphation and the gastrointestinal tract
The process of sulphation also affects the functioning of peptides and proteins. Mucin proteins, which line the gastrointestinal tract, are sulphated glyco-proteins which control adhesion and absorption of nutrients. They have long peptide backbones with repeating sub-units and also peptide side-chains, rather similar to a ‘bottle-brush’. These amino acid sequences also have strings of attached sugars which are sulphated like the peptides themselves. As the addition of sulphate residues (SO 4 2- ) sticks on net negative charges, the proteins spread out since the negative charges repel each other (Figure 1). If the sulphate residues are lost, this leads to a protein which has a more globular structure and provides less protection for the tissues from the intestinal contents as there are ‘gaps’ between the proteins. Reduced sulphation has been linked with gut dysfunction in irritable bowel disease and Andrew Wakefield’s group showed that lower levels of sulphation of the ileal mucins occured in children with autism which probably explains why gut permeability is increased in many autistic children. Sulphation of mucins increases their resistance to colonisation by pathogenic bacteria (and viruses). It is interesting that Helicobacter pylori , which can colonise the stomach, only does so when it has produced a sulphatase enzyme to de-sulphate the gastric mucins. This reduced sulphation of gut proteins may make Candida infections more likely in autistic children, since the slight negative charges on Candida cells would lead to their repulsion by negatively charged sulphate groups on the mucins.
Peptides can also be sulphated, usually on tyrosine residues, and the gastric hormones gastrin and cholecystokinin are good examples of this pathway. Both are involved in the digestive process and both are activated by sulphation. In a complex cascade, gastrin is sulphated and, with hydrochloric acid from the stomach, causes release of cholecystokinin, which also requires sulphation. Together with peptide fragments released from proteins by hydrochloric acid in the stomach, this acts with the peptide hormone secretin on pancreatic tissue to induce the secretion of a range of proteolytic enzymes and also amylase and lipases. (Figure 2). Without the sulphation process to trigger the release of pancreatic proteases such as trypsin and chymotrypsin, the complete digestion of proteins to their amino acid building blocks (proteolysis) cannot take place, so that peptides, rather than amino acids, are found in the gastrointestinal tract.
As reduced sulphation of mucins may have made the gut more permeable, the stage is set to allow peptides to penetrate into the blood stream. At the same time, the reduced levels of pancreatic amylase alter the digestibility of starch-based foods and allow increased fermentation of pathogenic bacteria while the decreased pancreatic lipase activity promotes formation of foul-smelling fatty stools which contain undigested triglycerides. Some peptides which cross the gut wall, particularly those derived from casein and gluten, have been found to be neuroactive with effects on the brain where they act at opioid receptors, affecting behaviour, mood and responses to physical stimuli such as pain. This ‘leaky gut’ hypothesis therefore links with the opioid theory to explain why there are peptides in the circulation rather than amino acids and why they have such ready access to the central nervous system. Although the blood-brain barrier is usually seen as being non-permeable to many compounds, it may, like the gut, be ‘leaky’ in autism. Several studies have reported the presence of brain-derived proteins and antibodies, such as those from myelin, within the peripheral circulation. If relatively large proteins can cross from the brain, it seems possible that peptides and proteins could potentially be transported into the brain, although the mechanisms involved are not known. Simple diffusion across ‘leaky’ gap junctions may be all that is necessary.
Sulphotransferase enzymes in autism
Not only is there an impaired level of sulphate in many cases of autism, there is also often a corresponding lack of sulphotransferase activity. These are the enzymes which carry out sulphation of a wide range of substrates. They belong to a super-family which uses PAPS (3′-phosphoadenosine-5′-phosphosulphate) as a co-factor and are widely distributed throughout the body, sulphating tissue components and signal molecules such as steroids, thyroid hormones and neurotransmitters. The major enzymes responsible for the sulphation of phenols and catecholamines are called SULT1A1 and 1A3 respectively. Sulphotransferase activity is known to be altered in some dysfunctional states, for example most patients with migraine have low SULT1A1 and sometimes low SULT1A3 activity. They are therefore less able to sulphate dietary phenols and catecholamines. Sulphation inactivates amines – many migraine patients are susceptible to foods which contain brain-active amines (cheese/tyramine, chocolate/phenylethylamine, bananas/serotonin) or inhibit the SULT enzymes. The increased blood levels of amines/phenols with neurotransmitter activity are thought to ‘trigger’ migraine headaches in those who are already susceptible. It has been shown that individuals who are susceptible to migraine are metabolically unstable (with raised excitotoxic amino acid levels) so that very small changes in blood and brain catecholamine levels are sufficient to provoke a migraine attack . SULT1A1 and 1A3 can be inhibited by flavonoids and by foods containing these compounds which typically occur in fruit and vegetables. Eating citrus fruit, especially oranges, is often reported as being a migraine ‘trigger’ and the component flavonoids (especially naringin) are inhibitors of both SULT1A1 and SULT1A3. The inhibitory effects of flavonoids on SULT1A1 can be partially overcome by the presence of magnesium ions which enhance the enzyme activity.
Many children with autism, particularly those with g.i. tract problems, have a family background of migraine and in a small pilot study we found that some children with autism also had reduced sulphotransferase activity. They would be expected to react badly to foods containing phenols, catecholamines or flavonoids and this response may underlie the benefits of the Feingold (low-phenol) diet and provide an explanation for the dietary intolerances which can be found in autism. Anecdotally many parents of autistic children report that their condition is made worse by the same foods which affect migraine patients and that chocolate and bananas exacerbate behavioural problems. In children ‘migraine’ often affects the gastro-intestinal tract, causing colic and cramping. This disappears around puberty to become the classic headache syndrome and it is possible that some of the gut dysfunction seen in autism may be a version of this juvenile presentation of migraine.
Other sulphotransferases can also be affected. The enzyme tyrosylprotein sulphotransferase (TPST) is membrane-bound and found in most tissues of the body, including the platelet and the gastrointestinal tract. TPST is the enzyme responsible for sulphation of gastrin and cholecystokinin as well as the sulphation of mucins so it is obviously important in g.i.tract function. It is interesting that sulphated cholecystokinin (CS) has receptors in the brain as well as the gut and is required for release of the peptide hormone oxytocin. CS levels may be low in mothers of autistic children as studies have shown that they are more likely to require pitocin (a synthetic oxytocin analogue) during the birth process . Children with autism have lower levels of oxytocin themselves and as this hormone elicits social behaviour any deficiency may contribute to the social deficits in autistic spectrum disorders. The digestive and neurological systems are therefore dependent on adequate supplies of sulphate (usually low in autism) and on a sufficiently active form of TPST being present to catalyse the digestive ‘cascade’ process and also activate the body’s defences against infections. Relatively little work has been done on this enzyme, but a pilot study in this laboratory with autistic children showed a mean TPST activity which was 33% of the control value. There was a wide range of activity, some children having almost no detectable platelet TPST values while a small number (3/20) had nearly normal levels. None of this latter group had g.i. tract dysfunction while those children who did have gut problems, including diarrhoea and constipation, all fell into the ‘low TPST activity’ category.
Conclusion
The evidence so far is incomplete but is certainly in accordance with the view that while defects in sulphation may not be the prime cause of autism, they are responsible for much of the dysregulation of biochemical and physiological processes. Perhaps the full answer may lie somewhere in the complex interactions of the autoimmune system with neuronal development. Autism may then reflect either in utero damage from maternal cytokines or perinatal damage caused by the actions of infections and vaccinations on a child with faulty autoimmune responses. The aetiology of the condition might then be similar to that for eczema, asthma and allergic responses, all of which seem to be increasingly common in children. Autism is of course heterogeneous but improved understanding of the biochemistry involved must eventually lead to novel therapeutic approaches. Potentially, too, we should be able to identify children ‘at risk’( perhaps measuring cytokine levels at birth or before vaccinations?). They could then follow a controlled diet and possibly a different schedule of vaccinations, while infections would be avoided where possible.
Figure 1
Schematic diagram showing the structure of mucin. The thick horizontal line represents the polypeptide “backbone” and the short vertical lines the polysaccharide side-chains which are studded with sulphate ( ˜ ) and sialic acid ( ™ ) residues.
Figure 2
The role of sulphation in digestion. Mechanical stretching of the stomach walls coupled with chemical stimuli from the food cause the polypeptide hormone gastrin to be secreted by the pyloric glands in the stomach. Gastrin is activated by sulphation and triggers the production of hydrochloric acid and the proteolytic enzyme pepsin. These, in combination with digestive products, prompt the release of two more hormones: the polypeptide cholecystokinin (which also requires sulphation prior to activation) and secretin which stimulates the pancreas to produce bicarbonate which neutralizes the acid from the stomach and the enzymes necessary for digestion to continue in the small intestine.
