Fall DAN!^TM 2003 Conference
*** Portland, Oregon *** October 3-5, 2003

Why Use the Gluten-Free and Casein-Free Diet in Autism and What the Results have Shown so Far

Peptides and Autism

Karl Reichelt, MD, PhD¹ and A. M. Knivsberg PhD²

¹ Institute of Pediatric Research, Univ. of Oslo, Rikshospitalet ,N-0027 Oslo, Norway
² Centre for Reading Research, N-4068 Stavanger , Norway

Key words: Autism, Exorphins, Peptide, Serotonin uptake, Peptidase

Abstract:

Objective: The reasons for using diet in autism should also reasonably explain the disorders patho-physiology. Changes ought to be found in rating scales, peptide levels and learning ability after intervention.

Method: Peptides have been separated by HPLC, co-chromatography, immune assay and mass spectrometry and IgA antibodies to food proteins measured. Because a double blind was impossible to carry through (people know what they eat), an open study over 4 years and a single blind controlled pair wise matched intervention with Gluten free/Casein free diet have been used.

Conclusion: An increase in urine peptides and specific peptide like compounds can be demonstrated. Furthermore IgA antibody increases relative to controls point to certain food proteins as being important. Because of the indicated limitations p values lower than 0.01 two tailed were generally chosen as limit in the open dietary experiments. Even so clear-cut behavioral changes can be measured after one and four years.

Abbreviations: 5HT: serotonin. CSF: cerebrospinal fluid. Cm: Casomorphin

Introduction

Several laboratories have found increase in urinary peptides (Fig 1 - 3) in autism (table 1) and that some of these are probably exorphins (1-5). The definite structure of these has been obtained by mass spectrometry and fragmentation mass spectrometry (5, 6). Furthermore co-chromatography and immune assay for peptides have been performed. The presence of rare d-amino acid containing exorphins has also been confirmed (5). Furthermore increase of opioids have been found in serum and CSF (7, 8), some of which are bovine casomorphins (1, 4, 5). IgA antibody increase against gliadin, gluten and casein have been found (4, 9-11). Therefore diet free of these proteins have been used with behavioural testing before and after intervention.

Methods:

Diagnosis: based on ICD 10 or DSM IV (almost the same).

Urine: 10 ml of the first morning urine has been the basis of analysis .It was frozen sent in Styrofoam boxes to our laboratory. The urine had creatinine measured and 600 µ l urine was filtered by centrifugation at 3000 x g for 20 min. using Spin-X Costar centrifuge Tube filters (0.22µlm cellulose acetate) in 2 ml polypropylene tubes. (Corning Inc, Corning, NY 14831, USA). Filtered urine equivalent to 250 nanomoles of creatinine was routinely used for analysis.

Urine analysis: With an automatic sample applicator and thermostatic control set at 30^o C. and HPLC machine from Hewlet Packard (Agilent series 1100) was used and Vydac C-.18 reversed phase column (0.4x 20 cm). Peptide bonds were measured at 215 nano-meters wave length and aromatic side groups at 280 nano-meters. The latter is important because many drugs contain aromatic groups and can thus be found. Flow rate was 1 ml per minute and gradients of trfluoroacetic acid (10mM) against acetonitril was run (for details see 12)

Antibody analysis: was analysed in serum taken from venous blood sample in the cubital vein. IgA and IgG antibodies were measured against gliadin and gluten and IgA against lact-albumin, lacto-globulin, casein and ov-albumin were measured using Elisa as described (13, 14)

Dietary intervention: Because we have not been able to run dietary intervention double blind (people must know what they eat) two strategies for gluten free/casein free diet were used.

Group A: It is exceedingly difficult to have a control group on diet and given gluten and casein blind. The reason for this is that the effort involved excludes willingness to continue if no improvement is registered.

We therefore ran one experiment on 15 participants for 4 years to rule out possible placebo in an open study.

Group B: In another experiment run for 1 year matched pairs were randomly assigned to diet or control group and studied single blind. All other treatments were unchanged. p values lower than 0.01 increases the probability of a likely effect (13 ).

Testing before and after:

The following tests were used because they have been standardized against Norwegian/Danish children and were therefore chosen.

A: The group followed for four years, and tested after 1 and 4 years on diet.

1: DIPAB (14) testing especially for social isolation and bizarre behaviour. This was compared with Tafjord test (15), Ravens Progressive Matrices (16) and to compensate for natural maturation: Illinois test of psycholinguistic ability (17) and the peptide levels.

After four years The Illinois test of Psycholinguistic ability and Ravens progressive matrices were checked (18).

B: In the single blind 1-year intervention with a control group on randomly assigned paired children we used DIPAB, as above. Information on autistic traits, cognitive level, language , and motor skills were also obtained using the tests described ( 13). Thus linguistic ability (ITPA) standardized for Norwegian children between 4 and 10 years old (17) or Reynells language test between 1.5 and six years of age (19), non verbal cognitive level (20), and Motor problem registration(21) were used. Profiles of each child were obtained before and after and the children were pair wise matched on severity of autistic symptoms and age (13) and randomly assigned to diet or control group. All other interventions were unchanged.

Results:

In Fig 1 the normal urine pattern and level is shown running urine equivalent to 250 nano-moles of creatinine on HPLC. Fig 2 shows a late onset regressive autistic pattern, while fig 3 shows an infantile autistic .Indolyl acryloyl glycine (IAG) found by the Sunderland group

(22) seems to be the best marker for infantile autism .

In table the level of peptides eluting after hippuric acid in autistic syndromes measured as area under the UV 215nm curve is presented and controls likewise. The difference is statistically highly significant. That the peaks are peptide like has been extensively documented (23) and is based on release of amino acids on hydrolysis, peptidase treatment and isolation of peptides after purification ( 24 ) .

Some of the peptide peaks show opioid activity (1,13), cochromatography on HPLC after purification and on spiking with standards, react with specific antibodies, and by mass spectrometry indicates that some of these are exorphins . (Table 2) (1,10,13), which agrees with data from other groups (4, 5).

Furthermore the peptide levels decrease on diet (4,18), as expected if they had a dietary origin.

Immune assay: As can be seen in table 3 the IgA antibodies are significantly different from controls for gliadin, gluten and casein. This indicates increased permeability for these proteins without celiac disease (negative endomyzium or transglutaminase. Some specificity is present however, because IgA antibodies against lact-albumin, lacto-globulin and ovalbumin ere not statistically different in autistic syndromes and controls.

Results of diet:

Group A followed for 1 and 4 years to compensate for the placebo effect even if no change was induced in training /schooling .

The details can be seen in table 4. There are very statistically significant changes and Haracopos scheme and Tafjord scheme. By matching each other closely this increases the probability that an effect is seen. The improvement in language in the Tafjord test is also reflected in ITPA.

Three children quit the diet for various reasons such as break up of the family. After four years these three show not significant changes. The difference for the three that broke the diet to those that stay on diet has a Mann Whitney U statistic =0.0 and U’= 36 and p= 0.004; or a very significant difference. As can be seen in the original publication those that quit diet do in fact show clear cut regression.(18). This is thus an unplanned control of the changes.

Group B:

In the controlled single blind study the changes measured can be seen in table 5. For the measures taken the diet group improved with decrease in autistic behaviour , increase in non-verbal cognitive level and decrease in motor problems. There was in this 1-year intervention no difference in linguistic age (13) and this may reflect that the children chosen were linguistically relatively high functioning to start with.

The changes noticed are similar to those published ( 4,8,9,25,26) and as outlined (24) the found peptides can indeed explain the disorder. Furthermore the measurable regression in those that quit diet as well as the lack of changes in the control group in the single blind series, clearly indicates that diet is effective. A recent Internet study also points to this conclusion.(27 )

Because neuroactive peptides may show long term effects after early exposure ( 28,29 ) and several of the opioids have trophic effects ,( 30,31) and are epileptogenic ( 32 ) ,we think it is important to get started as early as possible with diet . Given supplements of vitamins , enough other proteins and cod liver oil it is furthermore not hurtful and after many years clearly quite safe. This is not the case for most medications.

Conclusion:

Peptides are excreted in increased amounts in autism and some of these are opioids.

The excreted peptides may explain the social isolation as expected form opioids (33). Also trophic changes may be due to opioid peptides, as such compounds to have effects on brain maturation (30,31 ). Furthermore Fos antigen induction is found in key nuclei such as N Accumbens , and the superior and inferior colliculus after IV casomorphin 1-7(bovine) (34). Because immunocompetent cells have receptors for opioids ( 35 ) immunological changes are to be expected as found. Although not double blind the result of our dietary intervention strongly indicates that gluten free, gliadin free and casein free diet is an effective treatment. This is furthermore supported by the effect of oral peptidase treatment to increase protein and peptide break down ( 36).

The anatomical changes found in autistic children in brain (37,38) lends some urgency to the treatment as early as possible , the more so because of their behavioural effects(39).

References:

1. Reichelt KL, Knivsberg A-M, Lind G, Nødland M. Probable etiology and possible treatment of childhood autism. Brain Dysfunct 1991; 4: 308-19
2. Shattock P, Kennedy A, Rowell F, Berney T. Role of neuropeptides in autism and their relationship with classical neurotransmitters. Brain Dysfunct 1990; 3: 328-45
3. Israngkun P, Newman HA, Patel ST, Duruibe VA, Abou-Issa, H. Potential biochemical markers for infantile autism. Neurochem Pathol 1986; 5: 51-70.
4. Cade RJ, Privette RM, Fregly M, Rowland N, Sun Z, Zele V, Wagemaker H, Edelstrein C. Autism and schizophrenia: Intestinal disorders. Nutritional Neuroscience 1999; 2: 57-72
5. Shanahan MR, Venturini AJ, Daiss JL and Friedman AE. Peptide diagnostic markers for human disorders. European Patent application. 2000. EP 0 969 015 A2 1-44:1-44.
6. Gillberg C, Terenius L, Lønnerholm G. Endorphin activity in childhood psychosis. Arch Gen Psychiat 1985; 42: 780-3
7. LeBoyer M, Bouard MP, Racasens C, Philippe A., Guillod-Bataille M, Bondeaux D. TabuteauF, Fugas M, Panksepp J, Lauoy J-M. Difference between plasma N-and C-terminally directed beta-endorphin immunoreactivity in infantile autism. Am J Psychiat 1994; 151: 1797-801
8. Reichelt KL, Ekrem J, Scott H. Gluten, milk proteins and autism: dietary intervention effects on behavior and peptide secretion. J Appl Nutrition 1990; 42: 1-11
9. Lucarelli S, Frediani T, Zingoni AM, Ferruzzi F, Giardini O, Quintieri F, Barbato M, D’Eufemia P, Cardi E. Food allergy and infantile autism. Panminerva Med 1995; 37: 137-141
10. Reichelt WH and Reichelt KL. The possible role of peptides derived from food proteins in diseases of the nervous system. In Gobbi G et al. (Eds) Epilepsy and other Neurological Disorders in coeliac Disease. London: John Libbey & Comp. 1997; pp225-235
11. Scott H, Rognum TO, Midtvedt TJ, Brandtzaeg P. Age related changes in human serum antibodies to dietary and colonic bacteria antigens measured by enzyme-linked immunoabsorbant assa. Acta Pathol Microbiol Immunol Scand. 1985a, 93: 65-70.
12. Scott H, Rognum TO, Brandtzaeg P. Performance testing of antigen coated polystyrene micro-plates for ELISA measurements of serum antibodies to bacterial and dietary antigens. Acta Pathol Microbiol Immunol Scand. 1985b,93:117-123.
13. Knivsberg A-M, Reichelt Kl, Høien T, Nødland M. A Randomized, controlled study of dietary intervention in Autistic syndromes. Nutr.neuroscience 2002; 5:251-261
14. Haracopos D , Kelstrup A .DIPAB observationsskema .Special-pædagogisk forlag a/s,Herning ,Denmark.
15. Tafjord M. Observasjon av forutsetninger for lek og aktivitet; Observasjonsskjema . Statens Spesiallærerhøgskole, Oslo , Norway. 1952
16. Raven JC . Raven progressive Matrices H.K. Lewis, London, UK, 1958
17. Gjessing HJ, Nygaard HD. ITPA; Håndbok. Norsk utgave .Universitetsforlaget, Oslo , 1975
18. Knivsberg A-M., Reichelt KL , Nødland M, Høien T . Autistic syndromes and diet: a follow –up study. Scand J educat. Res. 1995; 39: 223-236
19. Hagtvet B, Lillestølen R. Håndbok. Reynells språktest(Reynell developmental language scale ). Universitetsforlaget, Oslo, Norway, 1985.
20. Leiter R . Leiter International performance Scale. Stoelting Co., Wood dale ., Il.USA 1979.
21. Henderson S and Sugden D , Movement assesment battery for children. The psychological Corp Ltd. Sidcup, Kent, UK 1992
22. Anderson RJ, Bendell DJ, Garnet I. Groundwater PW, Lough J, Mills MJ, Savery D , Shattock EG . identification of indolyl-3-acryloylglycine in the urine of people with autism. J. Pharamacy and Pharmacol. 2002; 54: 295-298.
23. Reichelt WH, Stensrud MB, Reichelt KL . Peptide excretion in celiac Disease. J Pediatr. Gastroenterol. Nutr. 1998; 26: 305-309
24. Reichelt Kl , Knivsberg A-M.. Can the pathophysiology of Autism be explained by the nature and the discovered urine peptides? Nutr. Neuroscience 2003; 6: 19-28
25. Whiteley P, Rodgers J, Savery D, Shattock P. A gluten-free diet as an intervention for autism and associated spectrum disorders: preliminary findings. Autism 1999; 3: 45-65.
26. Kniker WT, Andrews A, Hundley A, Garver C. The Possible role of Intolerance to Milk/dairy and wheat/gluten foods in older children and adults with autism spectrum disorder. In ‘2001: An Autism odyssey’. (The autism Res. Unit, Sunderland Univ: edit) .2001:pp 183-191.
27. Klaveness J, Bigam J. The GFCF Kids diet survey. In ‘Building bridges.’ (The autism Res Unit, Sunder land Univ.: edit) 2002; pp77-84
28. Meyerson BJ.Höglund U. Neuropeptides and behaviour: The behavioural effects of neonatal exposure to peptides and peptide antagonists. In ‘Hormones, Brain and Behaviour in Vertebrates 1’ (J Balthazart: ed) Karger, Basel.1990; 8:pp205-214
29. Gschanes A, Windisch M. Early postnatal treatment of peptide preparations influences spatial navigation of young and adult rats. Behav Brain Res 1999; 100: 161-6
30. Zagon IS, McLaughlin PJ. Endogenous opioid systems regulate cell proliferation in the developing rat brain. Brain Res 1987; 412: 68-72
31. Hauser KF, McLaughlin PJ, Zagon IS. Endogenous opioid systems and the regulation of dendritic growth and spine formation. J Compar Neurobiol 1989; 281, 13-22
32. Siggins GR, Henriksen SJ, Chavkin C and Gruol D . Opioid peptides and epileptogenesis in the limbic system: cellular mechanisms. Adv. Neuro. 501-512
33. Panksepp J, Normansell L, Sivily S, Rossi J, Zolovick AJ. Casomorphins reduce separation distress in chickens. Peptides 1978; 5: 829-31
34. Sun Z, Cade RJ, Fregly MJ, Privette RM. Beta-casomorphin induces Fos-like immunoreactivity in discrete brain regions relevant to schizophrenia and autism. Autism 1999; 3: 67-83
35. Singh VK. Neuropeptides as native immune modulators. Progress in Drug Res 1995; 45: 10-31
36. Brudnak MA, Rimland B, Kerry RE, Dailey M, Taylor R, Stayton B, Walckman F, W, Walckman M, Pangborn J, Buchholz I.Enzyme-based therapy for autism spectrum disorders-is it worth another look? Meed Hypothesis 2002; 58: 422-428.
37. Hashimoto T, Tayama M, Masahito, M, Sakorama N, Yoshimoto T, Murakawa K. Reduced brain stem size in children with autism. Brain and Development 1992; 14: 94-7
38. Courchesne E, Townsend J, Saitoh, O. The brain in infantile autism: Posterior fossa structures are abnormal. Neurology 1994; 44: 214-23
39. Sun Z and Cade JR. A peptide found in schizophrenia and autism causes behavioral changes in rats. Autism 1999; 3: 85-95

Text to Figures

Fig 1: Shows the HPLC separation of peptides from a normal boy, based on the principles published ( ). C –18 reverse phase column is used and absorption read at 215 and 280 nm giving us a purity index. That most of the material after hippuric acid is peptidic has been discussed extensively ( ), and is based on amino acid release by hydrolysis and also peptidase treatment followed by amino acid analysis and mow also Mass spectrometry.

Fig 2: The urinary profile from a 6-year-old boy with autism. Comparing this with fig 1 the difference is striking. Both runs are based on urine volumes equivalent to 250 nanomoles creatinin.

Fig 3: Another boy with the same diagnosis and CARS (Childhood autism rating scale) score as in fig 2. Notice the different pattern in spite of having the same diagnosis, sex and age. This indicates that different enzyme defects are probably present in different families, but with overlaps.

Table 1: The quantitative aspects of autism in prepubertal children found from 1995 to 1998 from 8 different countries in our lab.

The units are Abs units in µm² under the UV 215 nm trace of the peaks eluting after hippuric acid, and based on urine = 250 nanomoles creatinin. Only patients that have been diagnosed by certified psychiatrists are included. Ratio of UV 215/ UV 280 was used to check of puritiy of individual peaks. Many drugs have high UV 280 nm peaks due to aromatic ring structures. The autistic children are different from controls with a p= 0.001. Samples from 8 different countries and MD’s did not differ statistically indicating that diagnosis has become quite standardized through the use of DSM III to IV.

Table 2: Different compounds found in autism and the frequency of appearance.

The preliminary opioid receptor assay was carried out by Prof L Terenius, Stockholm and the immune-assays by professor Dr H Teschemacher from Giessen, who kindly provided us with antibodies for later work. Only peaks that are statistically larger than in controls are included. CM = Casomorphin , GluM : Glutemorphin. Gli M: Gliadinomorphin. The mass spectra (MS) and the fragmentation (MS/MS) spectra have been identified for urinary components and published (5). We identified these by purification as described (1,9), amino acid analysis after hydrolysis and sequencing (9). Antibody binding to antibodies against casomorphine 1-8 was used (1) as guideline for 1-8 and 1-7. Finally cochromatography with synthetic samples in at least two different HPLC systems was routinely used. Compounds run on MS are marked *.

Table 3. Immune data in autistic syndromes.

CI = confidence interval – Antibodies measured by ELISA in serum as described.

Table 4: 1 and 4 year intervention (18)

Notice that Raven C reaches a maximum level after 1 year while Illinois test of psycholinguistc ability (ITPA) improves even more after 4 years, probably reflecting the more complex nature of advanced learning (language). DIPAB = Haracopos scheme: ” Diagnosis of Psychotic Behaviour in Children”). Peptide levels decreased and followed the decrease in symptoms. Tafjord is a Norwegian scheme for registering play, interaction, structural ability and sensory motor behaviour during play. For details see (18). Mann Whitney U test used throughout.

Table 5. Single blind paired and controlled dietary intervention. (13)

Non parametric MannWhitney U test after 1 year, two tailed.

Corresponding author:

KL Reichelt, MD, PhD
Inst. of Pediatric Res. Univ of Oslo
Rikshospitalet, N-0027 Oslo, Norway
Phone:+ 47 23 07 29 85
Fax:+ 47 23 07 27 80
Email: K.L.Reichelt@klinmed.uio.no