946 in neurotransmitters was demonstrated by Fischer and James10 in experimental acute hepatic failure in rats, the increase in brain octopamine being closely related to the level of consciousness. Precursors of false neurochemical transmitters such as phenylalanine and tyrosine are produced in the gut from protein by the action of bacterial aminoacid decarboxylases. Normally such precursors are cleared from the portal blood by monoamine-oxidase activity, notably in the liver, but with impairment of hepatic function
extrahepatic portal shunting
these aromatic amines may enter the nervous system in excessive amounts and be &bgr;-hydroxylated locally by a relatively non-specific enzyme.18 They are taken up by sympathetic nerve endings and stored in granules, replacing dopamine and noradrenaline.19 20 The accumulation of false neurotransmitter amines in the brain preceded the rise in plasma-octopamine concentration.21 Increased concentrations of plasma phenylalanine and tyrosine have been demonstrated in human cirrhosiszz and are available for the formation of false neurotransmitters ; furthermore, increased brain concentrations of these aminoacids have been demonstrated in experimental liver failure in dogs.23 We have found statistically significant increases in urinary and serum octopamine concentrations between patients with and without encephalopathy, between those in hepatic precoma and in coma, and between grade-0 and grade-1 encephalopathy. Urinary octopamine excretion was significantly greater in grade-4 encephalopathy than in those with grade-3 encephalopathy. Serum-octopamine concentrations in these two grades were not, however, significantly different. There was no significant difference between grades 1 and 2, and grades 2 and 3. Lam et al.11 also demonstrated a close relation between the grade of encephalopathy and serum-octopamine concentrations, although these were significantly less than those given in this paper, probably because of differences in methodology. Levels of urinary octopamine excretion overlapped for the different grades of encephalopathy. None of the findings noted was of
prognostic significance. Lam et al. 11 suggested that impaired renal function may contribute to raised serum-octopamine concentrations. In this study, there was no correlation between blood, urea, and serum-octopamine concentrations; indeed, in the patient with the highest serum-octopamine concentration (67 ng/ml), blood-urea was 17 mg/dl and urinary output was 2 litre/24 h. We suggest that the measurement of octopamine, particularly in the serum of patients with liver disease, may be used as another marker of hepatic encephalopathy. Whether the finding of increased urinary and serum concentrations of octopamine in hepatic encephalopathy that these substances are involved in the of genesis portal-systemic encephalopathy has still to be established. in
We thank Dr Josef E. Fischer for his advice, Mr P. Matkin and Miss A. Lane for technical assistance, and Dr J. Kennedy for the F.F.G, mterpretations. K.M. was in receipt of a Commonwealth Tropical Medicine Research Fellowship and M.L. was a Watson Smith Fellow of the Royal College of Physicians during this study. We thank the Burton Trust for support.
Requests for reprints should be addressed to B.
MUCOSAL ADHERENCE OF HUMAN
ENTEROPATHOGENIC ESCHERICHIA COLI P. TURNER N. EVANS
A. S. MCNEISH JEAN FLEMING Institute
of Child Health, University of Birmingham
An in-vitro assay system has been developed to measure bacterial adhesion to the mucosa of human fetal small intestine. Two strains of Escherichia coli that are proven human enteropathogens (E.P.E.C.) have been shown to adhere in large numbers, compared with control organisms. The attachment mechanism is species specific and is not caused by common fimbriæ. Mucosal adhesion may be as important as enterotoxin production or invasiveness in determining the virulence of some strains of human E.P.E.C.
Introduction CERTAIN strains of Escherichia coli are enteropathogenic and cause infectious diarrhoea.’ These enteropathogenic E. coli (E.p.E.c.) elaborate soluble enterotoxins that have a similar effect upon enterocyte function in several animal models,2 yet an individual strain of E.P.F.C. will generally cause clinical disease only in one particular species.3 Clinical disease is associated with a proliferation of E.P.E.C. in the upper small intestine,3 and it is likely that bacteria proliferate more readily if they can adhere to the mucosal surface.4 Adhesion may also allow enterotoxin to be released close to the enterocytes. In E. coli that are enteropathogenic for newborn pigs, a surface protein antigen (K88) allows K88-positive bacteria to adhere to pig intestinal mucosal Adhesion is essential for the virulence of these E. coli in conventionally reared piglets.6 K88-associated mucosal adhesiveness shows some species specificity, because K88-positive E. coli do not adhere to calf intestine.5
DR MANGHANI AND OTHERS: REFERENCES
1. Zieve, L. Archs intern. Med. 1966, 118, 211. 2. Walker, C. O., Schenker, S. Am. J. clin. Nutr. 1970, 23, 619. 3. Muto, Y., Takahashi, Y., Kawamura, H. Brain Nerve, 1964, 16, 608. 4. Walker, C. O., McCandless, D. W., McGarry, J. D., Schenker, S. J. Lab clin. Med. 1970, 76, 569. 5. Chen, S., Mahadevan, V., Zieve, L. ibid. 1970, 75, 622. 6. Challenger, F., Walshe, J. M. Lancet, 1955, 1, 1239 7. Phear, E. A., Ruebner, B., Sherlock, S., Summerskill, W. H. J. Clin Sci 1956, 15, 93. 8. Walshe, J. M., De Carh, T., Davidson, C. S. ibid. 1958, 17, 11. 9. Fischer, J. E., Baldessarini, R. J. Lancet, 1971, ii, 75. 10. Fischer, J. E., James, J. H. Am. J. Surg. 1972, 123, 222. 11, Lam, K. C., Tall, A. R., Goldstein, G. B., Mistilis, S. P. Scand. J Gastroent.
1973, 8, 465. B., Weinshilboum, R., Axelrod, J.J. Pharmac. exp. Ther 1971, 178, 425. 13. Lowry, O. H., Rosebrough, N. J., Farr, A. L., Randall, R. J.J. Biol. Chem 1951, 193, 265. 14. Kakimoto, Y., Armstrong, M. D. ibid. 1962, 237, 422. 15. Molinoff, P. B., Landsberg, L., Axelrod, J. J. Pharmac. exp Ther. 1969. 170, 253. 16 Kennedy, J., Parbhoo, S. P., MacGillivray, B., Sherlock, S. Q. Jl. Med 1973, 167, 549. 17. Lands, A. M., Grant, J.I.J. Pharmac. 1952, 106, 341. 18. Kaufman, S., Friedman, S. Pharmac. Rev. 1965, 17, 71 19. Cohen, R. A., Kopin, I. J., Creveling, C. R., Musacchio, J. M., Fischer. E., Crout, J. R., Gill, J. R. Ann. intern. Med. 1966, 65, 347. 20. Haldessarini, R. J., Vogt, M. J. Neurochem. 1971, 18, 2519. 21. Dodsworth, J. M., Cummings, M. C., James, J. H., Fischer, J. E Surgery. 1974, 75, 811. 22. Levine, R. J., Conn, H. O.J. clin. Invest. 1967, 46, 2012. 23. Mattson, W. J., Coon, W. W., Child, C. G. Surgery Gynec. Obstet 1970 130, 263. 12. Molinoff, P.
possess K88, but these human intestinal mucosa by a organisms may adhere surface antigen that has not yet been identified. Drucker et al.7 noted E.p.E.c. that were adherent to the mucosal surface of the small intestine in infants who had died of
gastroenteritis. developed an assay for adherence of human the mucosa of human fetal small intestine, modified from the technique of Jones and Rutter.5 The results indicate that some strains show mucosal attachment that is species specific and non-fimbrial. We have
Materials and Methods Strains
Human E.P.E.C. -Strains 026 K60 Hll and 078 H12 kindly supplied by Dr H. Williams Smith (Houghton) and 0148 K? H28 by Dr B. Rowe (Colindale). Each strain consistently produces heat-labile enterotoxin in the rabbit ileal-loop test,8 and each has caused clinical disease.
Control strains.-Strains 01 Kl H7, 015 K14’ H4, and 075 K? H5 were obtained from the Central Public Health Laboratories, Colindale. None of these strains produces enterotoxin, and none has been implicated in clinical gastroenteritis.
Bacterial Culture The organisms were incubated aerobically at 370C for 18h on nutrient agar slopes (Oxoid). Bacteria were washed from the slopes into phosphate-buffered saline (pH 7-2) and their concentrations were adjusted to optical density 0.09 at 600 nm. This corresponds to approximately 5 x 108 organisms/m 1.
Number of bacteria of each strain that adhered to fetal intestinal
Fetal Intestine The recommendations and requirements of the Peel committee9 were followed in obtaining and handling the tissue, which was of 16-20 weeks’ maturity. Pieces of intestine, approximately 15 mm2, were cut from the first few centimetres of jejunum. After careful washing in phosphate-buffered saline, the tissue was either used immediately or was cultured for 6-12 h by a roller-tube organ-culture technique. 10 In-vitro Assay of Adhesion
of E. coli to Fetal Intestine
The technique was modified from that of Jones and Rutter.5 Each experiment was done in duplicate. The fragments of small intestine were incubated aerobically with gentle rotation for 30 min or 1 s at 37°C in 1 ml volumes of a suspension of the test organism that contained 0.5% D-mannose. The tissue was then removed, washed with rotation for 15 min in three changes of P.B.s., and homogenised in 1 ml of P.B.S. in a ground-glass homogeniser (Bellco). The number of viable organisms released from-each intestinal fragment was determined. The increase in the number of adhering organisms between 1 s and 30 min of incubation was the measure of their mucosal adhesiveness.
Species Specificity of Mucosal Attachment Brush-border preparations from the small calves, rabbits, and guineapigs were used of the test elsewhere." ness
intestines of pigs,
Fimbrial Hcemagglutination Common bacterial fimbriae were tested for by the technique of Duguid and colleagues.12 13 Tests were done on each strain of E. coli after overnight culture on nutrient agar slopes, and after 72 h culture with two passages in peptone water.
Preliminary experience indicated that bacteria would
non-specifically to non-viable tissue, so that the viability of samples of test mucosa was confirmed by assay of disaccharidases, alkaline phosphatase, and cyclic adenosine monophosphate. The results of duplicate tests are shown in the figure, using intestine from twenty-eight fetuses. The mean increases in adhering organisms for 026 K60 Hll and 078 H12 were 415 and 3.82 loglo units, respectively. than for These results are significantly greater (P