Wallace P. Rowe, M. D. 
April 2, 1979 
PAGE TWO 
do produce glomerulonephritis. Further, the suggestion made by King that influenza viral 
infection may predispose to autoimmune responses against glomerular basement membrane is 
based on very little evidence since in our series of some 300 cases of this disease, reasoncble 
evidence for an associated influenza viral infection was obtained only twice, an association 
which might well be due to chance alone. Finally, Dr. King's paper tends to imply that 
autoimmune immunologic phenomena are of necessity pathologic. Some autoreactive anti- 
body responses for example, rheumatoid factor, may follow a variety of immunizations and 
exposure to microbes, yet there is little evidence these autoreactive immunoglobulins 
necessarily produce disease. A similar view may be taken of many of the incidental auto- 
reactive anti-tissue antibodies seen in a wide variety of diseases. 
The hypothetical question raised by Dr. Beckwith in his letter to Dr. Gartland is of course 
impossible to answer definitively. It is entirely possible that a human protein combined with 
a bacterial protein presented to the human body in a particularly immunogenic manner (not in 
the gastrointestinal tract) might stimulate an immune response against the human protein. 
There are, however, a number of considerations which tend to play down the likelihood of 
such an event occurring and leading to pathological consequences. First of all, those con- 
stituents of E. Coli which are particularly immunogenic are the lipopolysaccharides and 
according to Dr. Beckwith these are not the bacterial components he is concerned about in 
combination with human protein. Thus, bacterial proteins in combination with human proteins 
would probably not present an unusually antigenic stimulus. Second, as mentioned above, the 
production of autoreactive antibodies is not synonymous with autoimmune disease. Judging on 
the basis of past experience, it is difficult indeed to induce such diseases in experimental 
animals. For example, one may immunize with a variety of homologous tissues such as brain, 
and develop circulating antibodies against brain in the complete absence of any demonstrable 
disease. In order to develop an allergic encephalomyelitis the brain antigen must be injected 
in a particular manner along with materials such as mycobacteria with extremely strong adjuvant 
properties. If this particular combination of parenteral injection and use of adjuvant is not 
followed, an autoimmune response does occur but no disease develops. The same situation has 
been observed with experimentally induced anti-heart antibodies in animals which were 
associated with little or no cardiac damage. 
From the above it would seem that available information would downplay any deleterious effect 
of cross-reacting antigens in the gastrointestinal tract. Further, a likelihood of pathological 
consequences from exposure to hypothetical hybrid cross-reacting antigens would seem remote 
indeed. However, if this possibility is considered significant enough, it could be tested in 
animals with appropriate hybrid molecules in a matter of a few months. The host-E. Coli hybrid 
molecules or the bacteria themselves could be injected or presented in whatever manner is deemed 
appropriate to experimental animals and any autoimmune responses and/or pathological conse- 
quences should be observable within a reasonably short period of time. If under such "worst 
situation" exposures, no autoimmune disease results, I would think these concerns could be 
layed aside. I hope these comments are of help to you. 
With all good wishes. 
Sincere ly yours, 
' '■ -> . 
Frank J. Dixon, M.D. 
[ 246 ] 
