Sir,
        Population-based studies (1), in contrast with molecular and
immunological studies (2-6), have not found an association between MMR
vaccination and autism.  As pointed out by Madsen et al (NEJM
2002;347:1478-1482 (1)) and endorsed by others (7), epidemiological
studies that have examined this relationship have been inadequate. Have
Madsen et al fared better?
        I have no doubt that other correspondents will deal with
a principal limitation of their study, that is, the failure to
disaggregate the relevant autism subset - one which they attempt to
describe in the introduction to their paper - from the overall autism
population. This is equivalent to looking at the totality of hepatitis,
irrespective of aetiology, in a study designed to examine a possible
causal relationship with a single, specific exposure that may account
for a minority hepatitis subtype only.
        My purpose is to try and help clarify the hypothesis of
my group, and to dissociate this from the many proxy hypotheses
generously, if erroneously tested in our name. Our studies have been
concerned with examining the aetiology and pathogenesis of autism in a
subset of children who became encephalopathic after a period of normal
development and suffer an immune-mediated gastrointestinal pathology
(2-4,8-14). Within the relevant subset of children we have observed
frequent atopy (especially food allergy), antibiotic use, ear infection,
multiple concurrent vaccine exposure and a strong family history of
atopic and autoimmune disease, as reported by others (15). Consistent
with these clinical observations, there appears to be, in many affected
children, a TH2 mucosal and systemic immune bias; this is evident in
lymphocyte cytokine profiles (14,16), eosinophil infiltration of the
intestinal mucosa, and up-regulation of class II antigen within the
intestinal lamina propria that is not seen on the adjacent epithelium
(8-10).  Dysregulated mucosal immunity in affected children is
accompanied by an excess of TNFa-positive lymphocytes, to an extent that
distinguishes the autistic lesional mucosa from both inflammatory and
non-inflammatory paediatric controls (14) that is consistent with the
findings of others (17). There is a profound expansion of CD19+
lymphocytes in the lamina propria, mirroring the associated hyperplastic
lymphoid response that, at the macroscopic level, is particularly
evident in the ileum and colon (13). In controlled, systematic studies
intestinal lymphoid hyperplasia of the degree seen in affected children
is clearly not, as anecdotal impression would have it, a normal variant
(9,18). While the TH1-TH2 model is an oversimplification, its serves as
a useful template for our working model.
        Early on in the current debate, in a paper that sought
to articulate the hypothetical relationship between MMR and regressive
autism, we wrote, "At the level of the immune response, the newborn
tends towards a TH2 response to pathogens and gradually shifts towards a
TH1 response with age. If this transition does not take place
appropriately, the infant is likely to be at greater risk of mounting
aberrant immune responses in later life, as seen in patients with
allergies. Given that, under normal circumstances the age of this
transition will be different for different children, it seems inevitable
that a ubiquitous viral exposure [MMR] of all 15-month-old children
could induce an immune response that is consistent with the individual
dynamics of this TH2-TH1 transition." (19).
        A precursor to an adverse reaction to MMR may be a
congential or acquired aberrant TH2 immune programming. This would
increase the likelihood of an inadequate antiviral immune response in
the face of a live viral vaccine and may facilitate viral persistence
and immunopathology, as described for measles virus in affected children
(2,4).
        The key to defining the "child at risk", therefore, is
an examination of the co-factors that may interfere with the appropriate
TH2-TH1 transition prior to, or concomitant with, MMR exposure. One such
factor may be mercury, for which the immuno-toxicity (putting aside for
now the associated neurotoxicity) of organic and inorganic derivatives
is qualitatively similar. Is a synergistic adverse interaction between
mercury and a live viral vaccine biologically plausible? The
immunosuppressive and immunomodulatory effects associated with mercury
exposure are accompanied by increased susceptibility to challenge with
infectious agents. One of the best-characterised examples of T-helper
cell phenotypic polarity in response to infection is the murine model of
Leishmania major. Murine susceptibility to L. major infection is
dependent upon induction of a genetically restricted TH2 response.
Resistant animals, that exhibit a genetically restricted TH1 response to
L.major, are rendered susceptible by prior exposure to mercury (20). In
previously resistant animals, sub-toxic doses of mercuric chloride
induced an autoimmune syndrome characterised by the expansion of TH2
cells, IL-4 production by splenocytes and IgG1 and IgE production. This
was accompanied by a non-healing phenotype with increased footpad
swelling and parasite burden. Methyl mercury enhanced the immune damage
and chronicity of coxsackie B3 myocarditis in mice, compared with mice
infected without prior mercury exposure (21). Similarly, mercuric
chloride exposure significantly impaired macrophage-mediated resistance
to generalised infection with herpes simplex type-2 in a murine model
(22).
        Mercury is only one of several exposures to infants that
may potentially influence the immune response to live viral vaccines. In
testing the correct hypothesis at the population level, these factors
will need to be taken into account and appropriate adjustments made. It
may be, for example, that the rapidly changing pattern of infant mercury
exposure - as thimerosal in bacterial and subunit vaccines - will with
the necessary adjustments, reduce statistical power to the extent that
such studies fail to offer any convincing evidence either way. It is my
personal opinion that the answer will be found in the detailed analysis
of each individual child - from clinical history to molecular
idiosyncrasy.
        The foundations of our hypothesis have not shifted.
Failure to take it into account has served merely to polarise the
debate, confuse the consumer, and allow the polemic of Public Health to
soar a little closer to the sun.
References
1.  Madsen MK., Hviid A., Vestergaard M., Schendel D., Wohlfarht J.,
Thorsen P., Olsen J., Melbeye M. A population-based study of measles
mumps rubella vaccination and autism. NEJM 2002;347:1478-1482
2.  Uhlmann V., Martin CM., Shiels O., Pilkington L., Silva I.,
Lillalea A. Murch SH., Wakefield AJ., O'Leary JJ. Potential viral
pathogenic mechanism for new variant inflammatory bowel disease.
Molecular Pathology. 2002;55:1-6
3.  Wakefield AJ. Enterocolitis, autism and measles virus. Molecular
Psychiatry. 2002;7 Suppl 2:S44-46
4.  Shiels O., Smyth P., Martin C., O'Leary JJ. Development of an
allelic discrimination type assay to differentiate between strain
origins of measles virus detected in intestinal tissue of children with
ileocolonic lymphonodular hyperplasia and concomitant developmental
disorder. Journal of Pathology. 2002 .A20
5.  Singh V., Lin S., Yang V. Serological association of measles
virus and human herpesvirus-6 with brain autoantibodies in autism.
Clinical Immunology and Immunopathology. 1998:89;105-108
6.  Singh VK, Lin SX., Newell E., Nelson C. Abnormal
measles-mumps-rubella antibodies and CNS autoimmunity in children with
autism. J Biomed. Sci. 2002;9:359-364
7.  Spitzer WO., Aitken KJ., Dell'Aniello S., Davis MW The natural
history of autistic syndrome in British children exposed to MMR. Adverse
Drug reactions and Toxicol. Rev. 2001:20;160-163
8.  Wakefield AJ, Murch SH, Anthony A, Linnell J, Casson DM, Malik
M, et al. Ileal LNH, non-specific colitis and pervasive developmental
disorder in children. Lancet 1997; 351: 637-641
9.  Wakefield AJ, Anthony A, Murch SH, Thomson M, Montgomery SM,
Davies S, et al. Enterocolitis in children with developmental disorder.
American Journal of Gastroenterology 2000; 95:2285-2295
10. Furlano RI, Anthony A, Day R, Brown A, McGavery L, Thomson MA,
et al. Colonic CD8 and ?d T cell infiltration with epithelial damage in
children with autism. Journal of Pediatrics  2001;138:366-372
11. Torrente F, Machado N, Ashwood P, et al. Enteropathy with T cell
infiltration and epithelial IgG deposition in autism. Molecular
Psychiatry 2002;7:375-382
12. Wakefield AJ, Puleston J., Montgomery SM., Anthony A., O'Leary
JJ., Murch SH. Review article: the concept of entero-colonic
encephalopathy, autism and opioid receptor ligands. Alimentary
Pharmacology and Therapeutics 2002; 16: 663-674
13. Ashwood P., Murch SH., Anthony A., Pellicer AA., Torrente F.,
Thomson M., Walker-Smith JA., Wakefield AJ.  Intestinal lymphocyte
populations in children with regressive autistic spectrum disorder and
entero-colitis. Gastroenterology 2002;122: Suppl. A1004
14. Ashwood P., Walker-Smith J., Murch S., Wakefield A.
Pro-inflammatory cytokine production in the duodenal and colonic mucosa
of children with autistic spectrum disorder (ASD) and a novel
entero-colitis; Gastroenterology 2002;122: Suppl. A617
15. Comi AM, Zimmerman AW., Frye VH., Law PA., Peeden JH. Familial
clustering of autoimmune disorders and evaluation of medical risks in
autism. J. Child Neurol 1999; 14;388-394
16. Gupta S., Aggarwal S., Rashanravan B., Lee T. Th1- and Th2-like
cytokines in CD4+ and CD8+ T cells in autism. J Neuroimmunol 1998;
85:106-109
17. Jyonouchi H., Sun S., Le H. Pro-inflammatory and regulatory
cytokine production associated with innate and adaptive immune responses
in children with autism spectrum disorders and developmental regression.
18. Kokkonen J., Ruuska T., Kartunen TJ., Maki M. Lymphonodular
hyperplasia of the terminal ileum associated with colitis shows an
increased gd+ T-cell density in children. Am J Gastroenterol.
2002;97:667-672
19. Wakefield AJ.and Montgomery SM. Autism, viral infection,
measles-mumps-rubella vaccination. Israeli Med Assn J. 1999;1:183-187
20. Bagenstose LM., Mentink-Kane MM., Britingham A., Mosser DM.,
Monestier M. Mercury enhances susceptibility to murine Leishaniasis.
Parastite Immunology 2001;23:633-640
21. Ilback NG., Wesslen L., Fohlman Friman G. Effects of methyl
mercury on cytokines, inflammation and virus clearance in a common
infection (Coxsackie B3 myocarditis) Toxicol. Lett. 1996;89:19-28
22. Christensen MM., Ellermann-Eriksen S., Rungby J., Mogensen SC.
Influence of mercuric chloride on resistance to generalized infection
with herpes simplex virus type 2 in mice. Toxicology 1996;114:57-66