Enterotoxin antibodies in relation to diarrhoea in Swedish soldiers in Cyprus

Bulletin of the World Health Organization, 55 (6): 663-668 (1977) Enterotoxin antibodies in relation to diarrhoea in Swedish soldiers in Cyprus A.-M. SVENNERHOLM,l E. BACK,2 & J. HOLMGREN3 The development of antibodies to Escherichia coli heat-labile enterotoxin (LT) and its relationship to diarrhoea was studied in two Swedish battalions (one in the winter, WB, and one in the summer, SB) of the United Nations Peace-keeping Force in Cyprus. Whereas only 6-15 % of the soldiers had detectable levels of anti-LT antibodies in their sera before leaving Sweden, as many as 69 % of the SB and 43 % of the WB soldiers who developed diarrhoea in Cyprus acquired significant titres of enterotoxin antibodies in their sera (acute phase serum). During the summer, many soldiers without diarrhoea also obtained anti-LT antibodies, whereas no such response was observed in healthy soldiers during the winter. Civilian Turkish Cypriots had positive sera in high frequency (60-66 %) during both study periods. Parallel titration of positive sera against E. coli LT and cholera toxin indicated that the enterotoxin-neutralizing antibodies registered were induced by bacteria producing enterotoxin identical or closely related to LT. Recent studies have demonstrated that enterotoxi- nogenic Escherichia coli (ETEC), rather than the classical enteropathogens such as Vibrio cholerae, Salmonella, or Shigella, are responsible for acute diarrhoeal symptoms in North Americans visiting Mexico or Kenya (1). Whether this is true also for the diarrhoeas frequently occurring in North Europeans travelling to southern Europe has not been studied. It has been noticed, however, that most of the ETEC-induced diarrhoeas diagnosed in Sweden have been acquired during or immediately after a visit to southern countries (2). To examine the exposure to enterotoxinogenic bacteria and its relationship to diarrhoea in travel- lers to the Mediterranean area, we conducted a prospective serological study in Swedish soldiers serving for a limited period in the United Nations Peace-keeping Force in Cyprus. The specific aims were to establish the frequency of sera with entero- toxin-neutralizing antibodies from civilian Cypriots, as an indicator of the occurrence of ETEC in the 1 Assistant Professor of Medical Microbiology, Univer- sity of Gothenburg, Sweden. 2 Research Physician, Department of Infectious Diseases, Roslagstull Hospital, Stockholm, Sweden. I Associate Professor of Medical Microbiology, Univer- sity of Gothenburg, Sweden. native population, and to determine the develop- ment of such antibodies in the Swedish soldiers in the presence and absence of diarrhoeal symptoms. MATERIALS AND METHODS Outline of study Two Swedish battalions of the United Nations Peace-keeping Force in Cyprus were studied. The soldiers were recruited from all parts of Sweden for a period of service of 6 months. The first battalion of 560 men (referred to as Winter Battalion, WB) was posted in October 1975 and was studied for the first 3 months in Cyprus. The second battalion of 450 men (Summer Battalion, SB) was studied during the fourth month of duty, in August 1976. The main camp of the battalion was situated outside Famagusta on the east coast of the island, and the soldiers were on duty in the Famagusta district along the border between the Greek and Turkish sectors. Most of the food was imported and distributed by the British base in Cyprus. It was stored in freezers or refrigerators in the camp, prepared in a central kitchen, and either served in the messes or delivered to the observation posts. Fruit and vegetables were obtained from the local market and were disinfected in iodine before being 3653 - 663 - A.-M. SVENNERHOLM ET AL. served. The water used was chlorinated and was regularly tested for bacteria. When off duty, the soldiers usually ate in local restaurants or simple kebab houses. Sera were collected from most individuals before leaving Sweden (pre-Cyprus sera), as well as on their return from Cyprus (post-Cyprus sera). All soldiers were informed about the investigation and instructed to report, as early as possible, any symptoms of diarrhoea. Diarrhoea was defined as two or more stools in one day, subjectively considered abnor- mally loose. A blood sample was taken from all soldiers with diarrhoea, usually within 24 h and always less than 2 days after the first loose stool (acute phase serum). Whenever possible serum was also collected from a " matched control ", i.e., a soldier occupying an adjacent bed but without diar- rhoea. A convalescent serum sample was taken 2 weeks later from the individuals in the diarrhoea group. During each of the study periods, serum samples were also collected from civilian Turkish Cypriots representing different socioeconomic groups and geographical areas of the Famagusta district (per- sonnel at the main battalion camp, outpatients at the surgical clinic of the Famagusta hospital, and 15-18- year-old pupils at the secondary school in Fama- gusta). In close connexion with the serum sampling, stool specimens for microbiological examination were col- lected. The results of this study will be reported elsewhere (Back et al., unpublished data). Serological tests The sera were analysed for neutralizing antibodies against E. coli heat-labile enterotoxin (LT) by means of the YI adrenal cell system (3). The LT prepara- tion used was kindly supplied by Dr R. Mollby, Stockholm, and was obtained by fermentor cultiva- tion of E. coli 411-5 (originally supplied by Dr R. B. Sack) in a defined medium under controlled condi- tions (Landwall & M6llby, unpublished data). The culture supernatant was dialysed to remove medium components and concentrated 50 times by lyophili- zation and dialysis against polyethylene glycol. The LT preparation was used at a final dilution 1: 160; this toxin concentration, as well as further dilutions varying from 1: 2 to 1: 8 on the various test occa- sions, produced typical rounding of >75% of the adrenal cells. The heat-inactivated sera were tested on two occasions at final dilutions of 1: 2 and 1: 4. When the results were discrepant a third test was performed. Neutralization was defined as an inhibi- tion of toxin activity resulting in less than 25% rounded cells. Sera neutralizing the toxin at a titre of 4 were further tested at least twice at a final dilution of 1: 30. Selected samples were titrated twice in parallel, by two-fold serial dilution against the E. coli LT prepa- ration and against a crude enterotoxin preparation of V. cholerae (Cholera Toxin Lot 4493G, obtained from the National Institutes of Health, Bethesda, MD, USA; at least two such titrations were per- formed with each sample. Rabbit antisera against these toxins, obtained by repeated subcutaneous immunizations, were run for comparison. RESULTS In the WB, 50 individuals (8.9%) developed diar- rhoea during the study period of three months. In the SB, 29 cases of diarrhoea (6.4%) were detected during August, i.e., the incidence per month was about twice that in the WB. The diarrhoeal episodes were generally brief (<3 days) and mild, and in no case was parenteral fluid therapy required. Of the WB soldiers with diarrhoea, 43% had detectable neutralizing antibodies against E. coli LT in the acute phase serum (Table 1). This frequency was significantly higher than that for the matched controls (P <0.01, x2 = 8.61). In the corresponding Cypriot control group as many as 66% of the sera were positive, i.e., significantly more than in both the Swedish matched controls (P <0.001) and in the group with diarrhoea (P <0.02). The proportion of sera with titres > 30 was also higher for the Cypriots (17%) than for the Swedish soldiers (2%) (see Table 1). In the SB, 69% of the soldiers with diarrhoea had antibodies neutralizing E. coli LT in the acute phase serum. This frequency is higher than was observed in the WB (P <0.05). In the matched control group, there was also a significantly higher proportion of positive sera (50 %) in the SB than in the VVB (P <0.01). Such seasonal variations were not found in civilian Cypriots, among whom the number of positive sera and their titre distribution during the summer did not differ from the winter study period (Table 1). The difference in frequency of positive sera between the diarrhoea and matched control groups in the SB was not significant, nor were the differences between any of these groups and the Cypriot civilians. 664 ENTEROTOXIN ANTIBODIES Table 1. Neutralizing antibodies against E. coil LT in Swedish soldiers (diarrhoea cases and matched controls) and Cypriot civilians No. of sera with anti-LT activity Battalion Serum sample No. of Titre Total 2 4-29 . 30 No. % Winter Diarrhoea cases 47 9 10 1 20 43 Matched controls 40 1 4 1 6 15 Cypriot controls 103 33 17 18 68 66 Summer Diarrhoea cases 29 8 10 2 20 69 Matched controls 24 6 2 4 12 50 Cypriot controls 25 5 4 6 15 60 Table 2. Development of anti-LT antibodies in relation to diarrhoea cases in Cyprus No. of sera with anti-LT activity Battalion Serum sample No.sof Titre Total samplesTirToa 2 4-29 . 30 No. % Winter Pre-Cyprus 36 2 - - 2 6 Acute phase 47 9 10 1 20 43 (0-2 days) Convalescence 41 1 2 3 6 15 (12-14 days) Post-Cyprus 25 5 1 - 6 24 Summer Pre-Cyprus 20 3 - - 3 15 Acute phase 29 8 10 2 20 69 (0-2 days) Convalescence 16 4 8 1 13 81 (12-14 days) Post-Cyprus 21 1 5 5 11 52 The occurrence of E. coli LT neutralizing anti- bodies in serum of the soldiers seems to reflect enterotoxin exposure in Cyprus. Thus, only a few of the soldiers who later experienced diarrhoea during the study period had anti-LT antibodies in their pre- Cyprus serum sample, and then in low titres (Table 2); the difference between pre-Cyprus and acute phase sera was highly significant (P <0.001) for both battalions. There was a striking difference between the WB and SB with regard to positive convalescence sera. In comparison with the acute phase samples, the frequency decreased in the WB whereas it increased in the SB (Table 2). Nonethe- less, in three instances in the WB, antibodies were demonstrated only in the convalescent samples. Such seroconversion was found in 2 SB cases, and in 1 other with a low titre in acute phase serum there was a more than 15-fold titre increase. Post-Cyprus serum samples of the soldiers of the WB diarrhoea group were positive with about the same frequency as were the convalescent sera. In the SB, the proportion of positive post-Cyprus sera from the diarrhoea group had decreased slightly as com- pared with the convalescent samples (Table 2). On 665 666 A.-M.- SVENNERHOLM ET AL. their return to Sweden, more than twice as many of the SB soldiers had LT antibodies in serum as the WB soldiers. To verify that the neutralizing antibodies were directed against E. coli LT and not against immuno- logically cross-reactive enterotoxin of V. cholerae (4, 5), selected sera were titrated in parallel against these toxins (Table 3). In similarity with rabbit antiserum against the E. coli LT used in this study, the sera of the soldiers as well as those of the Cypriot controls had, with one exception, a higher neutraliz- ing capacity against LT than against cholera toxin. Rabbit antiserum to cholera toxin, on the other hand, had considerably higher titres against the homologous toxin than against LT (Table 3). These findings indicate that the enterotoxin-neutralizing antibodies registered in the human sera were induced by toxin identical or related to E. coli LT rather than by a more cholera-like enterotoxin. DISCUSSION This study shows that a high proportion of the Swedish soldiers who developed diarrhoea in Cyprus had significant levels of serum antibodies capable of neutralizing E. coli heat-labile enterotoxin (LT). During the summer period, many of their healthy comrades also produced detectable anti-LT anti- bodies, whereas there was no increase in antibodies in the absence of diarrhoea during the winter. For the WB, the correlation between the incidence of diar- rhoea and the appearance of enterotoxin antibodies in serum is highly significant. For the SB, no such relationship was noted, due to the high frequency of positive sera in the matched control group. In contrast to the soldiers, the Cypriot controls showed no seasonal variation in enterotoxin anti- bodies, with 60-65% of the population having posi- tive sera during both study periods. This is consider- ably more than in healthy Swedes, since only 6-15% of the serum samples taken from the soldiers before leaving Sweden contained anti-LT antibodies, and then only in very low titres. This difference, together with the finding in a parallel study (Back et al., unpublished data) that LT-producing bacteria could be isolated in 7% of faecal specimens taken from healthy Cypriots, but not from any of the more than 400 stools taken from the soldiers before leaving Sweden, indicates a higher risk of being exposed to ETEC in Cyprus than in Sweden. This is in agree- ment with a recent report that ETEC is isolated from less than 5% of individuals with acute diarrhoea in Table 3. Comparison of neutralizing capacity against E. coli LT and cholera toxin (CT) Neutralizing Sera Titre against capacity LT a CT b LT/CT c Rabbit anti- LT 410 d 50 16.4 Rabbit-anti-CT 37 2050 0.04 Human diarrhoea cases Acute phase WB 10 11 2.0 Convalescence WB 24 12 4.0 Acute phase SB 20 16 2.8 Convalescence SB 30 16 3.8 Cypriot controls Winter I 20 8 5.0 11 38 19 4.0 Summer 12 40 0.6 11 16 8 4.0 a The concentration of LT used was twice that minimally required for > 75 % rounded cells. For the Cypriot controls, sera I and If correspond to the acute phase and convalescence sera. b The CT was used in the minimal effective concentration for 275 % rounded cells. c Titre per toxin unit ratio. Titres against LT have been multiplied by 2, since neutralization was performed with 2 toxin units. d Titres were determined by interpolation from the results with serial 2-fold dilutions, as the serum dilution resulting in 25 % rounded cells; the mean values of 2-3 experiments are given. Sweden, and then usually from cases who have acquired the illness during travel in southern Europe or North Africa (2). Few studies of the antibody response in man to E. coli LT have been undertaken. In individuals living in areas where exposure to ETEC is frequent, such as Bengal and Bangladesh (6, 7), high levels of neutralizing antibodies have been observed after ETEC-induced diarrhoea, usually with fourfold or greater titre increases between acute phase and con- valescence. On the other hand, North Americans acquiring enterotoxin diarrhoea during travel in Mexico or Kenya have only rarely shown such titre increases (1). In the present study, although develop- ment of antitoxin antibodies was frequent in relation to diarrhoea, a titre rise between acute phase and convalescence was registered in only a few instances. An increase (although not significant) in frequency of positive sera from acute phase to convalescence, ENTEROTOXIN ANTIBODIES from 69 to 81 %, was noted for the SB, while for the WB there was a significant decrease. A possible explanation for this difference is that the WB sol- diers were studied in relation to their first exposure to immunogenic amounts of enterotoxin, whereas during the summer and at a later stage of service many of the soldiers had probably been immunologi- cally primed prior to their diarrhoeal episode. The higher frequency of positive sera among the healthy controls in the SB, as compared to the WB, is consistent with this hypothesis. Although the development of enterotoxin-neu- tralizing antibodies could be related to diarrhoea in the WB soldiers, the isolation of enterotoxin-pro- ducing bacteria was far from regular (Back et al., unpublished data). This complicates judgment of the bacterial 'species' responsible for induction of the anti-enterotoxin antibody response, since E. coli LT cross-reacts immunologically with the enterotoxin produced by V. cholerae, and possibly also with enterotoxins from other bacteria (4, 5). The con- siderably higher titres found against the homologous than against the heterologous toxin for hyperim- mune sera to E. coli LT and to cholera toxin, indi- cate the presence of strong type-specific determinants on the two toxins in addition to their shared deter- minants. In view of the higher neutralization titres against E. coli LT than against cholera toxin in the soldiers' sera, it appears that in Cyprus the exposure was to bacteria producing enterotoxin identical or related to LT rather than to organisms producing more cholera-like toxin. ACKNOWLEDGEMENTS The authors are indebted to Dr Neelam Khosla for help with the adrenal cell assay and to Ms Britt-Marie Lavstrand for technical assistance. This study was supported by the Swedish National Defence Research Institute (506 H563), the World Health Organization, and the Swedish Medical Research Council. R1tSUMIt ANTICORPS CONTRE L ENTIROTOXINE ET DIARRHEE CHEZ LES SOLDATS SUEDOIS A CHYPRE La relation entre l'apparition dans le serum d'anticorps contre l'enterotoxine thermolabile (LT) d'Escherichia coli et les manifestations diarrheiques a ete etudiee sur envi- ron 1000 soldats composant deux bataillons suedois de la Force de Maintien de la Paix des Nations Unies ia Chypre au cours de leur sejour de 6 mois dans l'ile. Un bataillon d'hiver (WB) a ete contr6le pendant ses trois premiers mois dans l'ile et un bataillon d'et6 (SB) pendant son quatrieme mois de service, en aoiut. Un groupe de 128 civils cypriotes turcs a ete soumis 'a la meme etude aux fins de comparaison. La presence d'anticorps neutralisants contre LT a ete determinee au moyen d'un test sur des cellules de la glande surrenale. Un niveau significatif de ces anticorps a t observe dans le serum de 6% seulement des soldats du WB et de 15% de ceux du SB 'a leur depart de Suede. Dans le WB, l'incidence de la diarrhee a ete de 3% par mois. Parmi les soldats atteints, la proportion de ceux dont le serum, preleve en phase aigue, contenait des anticorps contre LT, etait passee a 43% - proportion beaucoup plus elevee (P < 0,01, x2 = 8,61) que celle de 15% etablie pour les soldats exempts de diarrhee dont le serum a 6te analyse au meme moment. L'incidence de la diarrhee dans le SB a 6te 'a peu pres le double de celle constat6e dans le WB. Des anticorps seriques contre LT ont ete trouves en quantite importante chez 69% des soldats du SB presentant ce symptome. Chez les sodats (appaires# non atteints de diarrhee, le taux de sero- positivite (50%) ne presentait cependant pas de diff&e rence significative. Les analyses de serum pr6lev6 en phase de convalescence ont egalement montr6 un ecart frappant entre le WB et le SB: Par rapport au serum pr6leve en phase aigue, la frequence de seropositivite est passee chez les convalescents du SB 'a 81 %, alors qu'elle a decline i 15% dans le WB. Au retour en Suede, le nombre de soldats du SB dont le serum contenait des anticorps contre LT etait de 52%, soit plus du double de celui des soldats du WB a serum positif (24%). Les variations saisonnieres du taux d'anticorps contre l'enterotoxine observees chez les soldats suedois n'ont pas e't constatees chez les civils cypriotes, chez qui ce taux est demeure d'environ 65% durant les deux periodes d'etude. Le titrage parallele sur des serums choisis des anticorps contre LT d'E. coli et contre l'enterotoxine brute de Vibrio cholerae a permis de conclure que les anticorps 667 668 A.-M. SVENNERHOLM ET AL. presents avaient ete induits par une toxine identique A LT d'E. coli ou apparentee plutot que par une toxine plus voisine de celle liee au cholera. Les resultats obtenus indiquent que l'exposition a E. coli enterotoxinogene est, A Chypre, aussi commune chez les visiteurs etrangers que chez les civils de l'ile et qu'elle suscite souvent un taux eleve d'anticorps speci- fiques. Ils suggerent egalement que cette exposition est une cause frequente de diarrhee du voyageur, bien qu'il existe de nombreux cas asymptomatiques. REFERENCES 1. MERSON, M. H. ET AL. Travelers' diarrhea in Mexico. New England journal of medicine, 294: 1299-1305 (1976). 2. BACK, E. ET AL. Enterotoxigenic Escherichia coli in Sweden. Infection, 5: 3-6 (1977). 3. SACK, D. A. & SACK, R. B. Test for enterotoxigenic Escherichia coli using YI adrenal cells in minicultures. Infection and immunity, 11: 334-336 (1975). 4. SACK, R. B. Human diarrheal disease caused by enterotoxigenic Escherichia coli. Annual review of microbiology, 29: 333-353 (1975). 5. HOLMGREN, J. ET AL. Cross-reactivity between heat- labile enterotoxins of Vibrio cholerae and Escherichia coli in neutralization tests in rabbit ileum and skin. Acta pathologica et microbiologica scandinavica, Sec- tion B, 81: 757-762 (1973). 6. SACK, R. B. ET AL. Antitoxin responses to infections with enterotoxigenic Escherichia coli. Journal of in- fectious diseases, 129: 330-335 (1974). 7. RYDER, R. W. ET AL. Enterotoxigenic Escherichia coli and reovirus-like agents in rural Bangladesh. Lancet, 1: 659-662 (1976).