Immune response to verotoxin 1 and 2 in children with Escherichia coli O 157 : H 7 hemorrhagic colitis and classic hemolytic uremic syndrome

Presented in part at the annual meeting of the Royal College of Physicians and Surgeons of Canada, Ottawa, Ontario, September 12, 1992 Departments of Pediatrics, Microbiology and Immunology, Hôpital Sainte-Justine, Université de Montréal, Montréal, Québec; and the National Centre for Enteric Bacteriology, Laboratory Centre for Disease Control, Ottawa, Ontario Correspondence and reprints: Dr François Proulx, Department of Pediatrics, Hôpital Sainte-Justine, 3175 chemin Côte Sainte-Catherine, Montréal, Québec H3T 1C5. Telephone (514) 345-4675, Fax (514) 345-4822 Received for publication August 22, 1994. Accepted February 13, 1995 F PROULX, JP TUR GEON, G DEL AGE, H LIOR, L LAF LEUR, L CHI COINE. Im mune re sponse to vero toxin 1 and 2 in chil dren with Escheri chia coli O157:H7 hem or rhagic co li tis and clas sic hemo lytic ure mic syn drome. Can J In fect Dis 1995;6(3):136140.

I N NORTH AMER ICA, VEROTOXIN-PRODUCING ESCHERI CHIA COLI se ro type O157:H7 is the main etio logi cal agent of both hem or rhagic co li tis (HC) (1) and clas sic hemo lytic ure mic syn drome (HUS) (2,3).E coli O157:H7 pro duces two vero toxins (VT).VT-1 is al most iden ti cal to the toxin of Shigella dys ente riae se ro type 1 ex cept for one amino acid sub sti tu tion, while VT-2 DNA is only 50 to 60% ho molo gous (4).Mul ti ple VT vari ants have also been re ported (5).Both VT-1 and VT-2 have di rect cy to toxic ac tiv ity to vero cells (Af ri can green mon key kid ney cells) (6).In HC, in fec tion and ische mia may dis rupt the in tes ti nal mu co sal bar rier, pos si bly al lowing trans lo ca tion of VT and the ini tia tion of re nal en do the lial cell dam age lead ing to HUS.The cel lu lar re cep tor for VT-1 and VT-2 is a gly cosphingolipid glo bo trio syl ce ra mide, which has been iden ti fied on vero cells, hu man re nal tis sue (7) and hu man B lym pho cytes (8).
Se rum an ti bod ies against vari ous E coli O157:H7 an ti gens such as VT (3,9), outer mem brane pro teins (2), fla gella (10) and lipo poly sac cha ride (2,(11)(12)(13) have been shown in humans with HUS.Cy to lytic an ti bod ies against en do the lial cells have also been de scribed in pa tients with HUS (14).How ever, the sig nifi cance of these find ings is not well de ter mined.
The in ci dence of HUS var ies with age (4), and HUS pa thophysi ol ogy may be re lated to the im mune re sponse of the host.We hy pothe sized that there is an in creased im mune response against VT in chil dren with HUS.We com pared neu traliz ing an ti body lev els in se rum against VT-1 and VT-2 in chil dren with clas sic HUS ver sus those with un com pli cated E coli O157:H7 HC.Sec ond, in vi tro stud ies have re ported that trimethoprim-sulfamethoxazole (TMP-SMX) (15) and po lymyxin B (16) may in crease the amount of cy to toxin re leased by E coli O157:H7.We there fore set out to de ter mine whether the mag ni tude of the im mune re sponse to VT-1 or VT-2 was modified by an ti bi otic ther apy in chil dren with proven E coli O157:H7 HC.

PA TIENTS AND METH ODS
Sainte-Justine Hos pi tal is a ter ti ary care pe di at ric cen tre in Mont real.In formed con sent was ob tained from par ents of all pa tients.The study was ap proved by the eth ics com mit tee of Sainte-Justine Hos pi tal.
From June 1, 1989 to June 1, 1990 paired se rum sam ples were col lected from 12 chil dren with full-blown clas sic HUS and from 41 chil dren with culture-proven, un com pli cated E coli O157:H7 HC.Of these, 18 were ran dom ized to re ceive TMP-SMX (4 mg/kg, twice daily for five days) and 23 to re ceive no an ti bi otic ther apy.E coli O157:H7 HC was de fined as the oc cur rence of an en teri tis with bloody di ar rhea and the iden tifi ca tion of sorbitol-negative colo nies on Mac Con key sor bi tol agar with an E coli bio chemi cal pro file and a posi tive slide agglu ti na tion (Difco Labo ra to ries, Inc, Michi gan) to se ro type O157:H7.All strains of E coli O157:H7 were con firmed by the Labo ra toire de Santé Pub lique du Qué bec.Clas sic HUS was de fined as the oc cur rence of ane mia with a he mo glo bin value be low the third per cen tile for age, throm bo cy topenia (plate let count less than 100x10 9 /L), pres ence of schis to cytes on blood smear and acute re nal fail ure with a cre ati nine value greater than the 90th per cen tile for age af ter a pro drome of en teri tis.Atypi cal forms of HUS were ex cluded.Treated HC, un treated HC and HUS were es tab lished as in de pend ent di agnos tic cate go ries.Age, sex, hos pi tali za tion, time of on set of HC and time of se rum col lec tion were re corded for all pa tients.Se rum speci mens: The acute phase se rum was ob tained at ran dom from all pa tients with HC and from six pa tients who presented with HC but who sub se quently de vel oped HUS.The acute phase se rum was taken at di ag no sis in the re main ing six chil dren who pre sented with es tab lished HUS.The con va lescent se rum was ob tained 10 days later from most pa tients.Sera were fro zen at -80°C un til analy sis.All sam ples were tested for both VT-1 and VT-2 an ti body by quan ti ta tive neu traliza tion by a stan dard ized method (17) modi fied as fol lows: Réponse immunitaire à la vérotoxine 1 et 2 chez des enfants atteints de colite hémorragique à Escherichia coli 0157:H7 et de syndrome urémique et hémolytique classique OBJEC TIFS : Com parer les ti tres d'an ti corps neu trali sants con tre la vé ro tox ine (VT-1 et VT-2) en tre des en fants at te ints de co lite hémor ragique non com pli quée (CH) et des en fants at te ints d'un syn drome hémo lytique et urémique clas sique (SHU).Les ti tres d'an ti corps anti-VT ont égale ment été com parés chez des en fants at te ints de CH qui re ce vaient du triméthoprim-sulfaméthoxazole et ceux qui n'en re ce vaient pas.MODÈLE : Étude pro spec tive.CON TEXTE : Hôpi tal pédia tri que de soins ter ti aires.POPU LA TION ÉTU DIÉE : En fants at te ints de CH (n=41) ou de SHU clas sique (n=12).INTER VEN TIONS : Les anti corps sé ri ques développés con tre VT-1 et VT-2 ont été me su rés à l'aide d'une tech nique de neutrali sa tion quan ti ta tive.PRIN CI PAUX RÉSUL TATS : Les anti corps ont été dé ce lés chez 40 % (21 sur 53) des échan til lons sé ri ques con tre la VT-1 et chez 100 % (53 sur 53) des échan til lons con tre la VT-2.Une réac tion im mu ni taire posi tive, définie comme une aug menta tion du quad ru ple des ti tres d'an ti corps anti-VT ou comme un seul ti tre de 1/64 ou plus a été ob servée chez 0 % (0 sur 12) des pa tients at te ints de SHU en com parai son avec 7 % (3 sur 41) des pa tients at te ints de CH pour la VT-1 (P=0,4) et chez 17 % (2 sur 12) des pa tients at te ints de SHU con tre 22 % (9 sur 41) des pa tients at te ints de CH pour la VT-2 (P=0,3).Les taux de sé ro con ver sion à l'é gard de la VT-1 ou de la VT-2 étaient com pa rables chez les pa tients trai tés et non trai tés at te ints de CH non com pli quée.CON CLU SIONS : Rien n'in dique que les taux d'an ti corps neu trali sants diri gés con tre la VT-1 ou la VT-2 dans la SHU clas sique ou après un traite ment anti bio tique ne soi ent sub stan tiel le ment diffé rents de ceux de pa tients qui pré sen tent une CH non com pli quée.20 mL of each of VT-1 and VT-2 was added sepa rately to 0.2 mL of vero cell mono lay ers to de ter mine one unit of toxin; one unit of toxin ac tiv ity was de fined as the amount pres ent in the highest di lu tion of a toxin prepa ra tion that caused 50% cell death (CD 50 ) af ter 48 to 72 h in cu ba tion.Toxin prepa ra tions were obtained from bacteria-free fil trates af ter over night in cu ba tion with Evans Me dium (18) of ref er ence strains H19 for VT-1 and the authors' ref er ence strain 90-2380 for VT-2.For neu trali zation, 30 mL of 2.5 units of each toxin was added sepa rately to 30 mL of each se rial di lu tion of pa tients' se rum (1:2 to 1:4096).The toxin/an tise rum mix tures were in cu bated at 37°C for 3 h then re frig er ated at 4°C for 18 h; 20 mL ali quots of each toxin/an tise rum mix ture were then dis pensed into cor re sponding wells of vero cell mono lay ers in mi cro ti tre trays.The mi croti tre trays were in cu bated at 37°C in a 5% car bon di ox ide in cu ba tor and the end-point was the high est se rum di lu tion caus ing in hi bi tion of CD 50 af ter 48 to 72 h of in cu ba tion.Included in the as say were ap pro pri ate tox ins, se rum and cell con trols.Each se rum pair (VT-1 and VT-2) was run on the same mi cro ti tre tray.To the authors' knowl edge, there is no im muno logi cal cross-reaction be tween VT-1 and VT-2.A posi tive immune re sponse was de fined as a four fold in crease in VT an ti body ti tres or as a sin gle ti tre of 1/64 or greater (3).

Sta tis ti cal analy sis:
The Fisher ex act test was used to compare the num bers of pa tients with a posi tive se ro logi cal response in the group with HUS ver sus the group with un com pli cated HC.The rate of se ron con ver sion was also compared in treated ver sus un treated chil dren with un com pli cated HC.Sta tis ti cal sig nifi cance was es tab lished at 0.05.

RE SULTS
There were 53 pa tients (29 males and 24 fe males), of whom 41 (77%) were hos pi tal ized.Me dian age was 62 months (range seven to 213) in pa tients with treated HC, 48 months (range three to 166) in those with un treated HC and 32 months (range 15 to 156) in those with HUS.Me dian time from on set of HC to acute se rum sam ple col lec tion was six days (range four to 11) in pa tients with treated HC, eight days (range three to 17) in pa tients with un treated HC and 6.5 days (range three to 16) in those with HUS.Me dian time be tween acute and con va les cent se rum col lec tion was 11 days (range eight to 13) in pa tients with treated HC, 10 days (range seven to 21) in pa tients with un treated HC and 10 days (range three to 16) in those with HUS.Both age and time of speci men collec tion af ter on set of HC were com pa ra ble among groups.The time of ran domi za tion af ter on set of HC was simi lar in treated and un treated pa tients (mean ± SD): 7.4±5.0and 7.2±2.7 days, re spec tively.Five pa tients with HUS re quired peri to neal di aly sis and one was he mo dia lyzed.
Fig ure 1 shows acute and con va les cent neu tral iz ing an tibody ti tres against VT-1 in chil dren with un com pli cated HC (treated and un treated) and in those with HUS.Se rum ti tres against VT-2 are pre sented in Fig ure 2.An ti bod ies were detected in 40% (21 of 53) of se rum sam ples for VT-1 and in 100% (53 of 53) of sam ples for VT-2.A posi tive im mune response against VT-1 was found in 0% (0 of 12) of pa tients with HUS com pared with 7% (three of 41) of those with HC (P=0.4).Among the lat ter, two chil dren were treated and one did not re ceive an ti bi otic ther apy (P=0.3).A posi tive im mune response against VT-2 was found in 17% (two of 12) of pa tients with HUS com pared with 22% (nine of 41) of those with un compli cated HC (P=0.3).Among the lat ter, 28% (five of 18) of treated pa tients and 17% (four of 23) of those who did not receive an ti bi otic ther apy showed a posi tive im mune re sponse (P=0.2).

Fig ure 2) Acute (A) and con va les cent (C) neu tral iz ing an ti body lev els in se rum against verotoxin-2 in 41 pa tients with un com pli cated hem orrhagic co li tis, among whom 18 were treated and 23 were un treated, and in 12 chil dren with hemo lytic ure mic syn drome (HUS). Num bers to the right of C points in di cate the number of pa tients with a spe cific an ti body pro file be tween acute and con va les cent sera. The fig ure shows that, al though a large number of pa tients had sta ble se rum an tibody lev els, se rum ti tre in creased mainly in chil dren with hem or rhagic co li tis. Ex cept for one pa tient, se rum ti tres among chil dren with HUS ei ther were sta ble or de creased by one se rum di lu tion
Fig ure 1) Acute (A) and con va les cent (C) neu tral iz ing an ti body lev els in se rum against verotoxin-1 in 41 pa tients with un com pli cated hem orrhagic co li tis, among whom 18 were treated and 23 were un treated, and in 12 chil dren with hemo lytic ure mic syn drome (HUS).Num bers to the right of C points in di cate the number of pa tients with a spe cific an ti body pro file be tween acute and con va les cent sera.The fig ure shows that most pa tients had low an ti body ti tres against verotoxin-1, which were sta ble over time in all groups

DIS CUS SION
A four fold in crease be tween acute and con va les cent serum ti tres is gen er ally con sid ered di ag nos tic of a re cent in fection.We also con sid ered as posi tive any se rum ti tre of 1/64 or greater be cause Kar mali et al (3) showed the me dian con vales cent se rum ti tre to be 1/64 among 16 pa tients with HUS with se ro con ver sion against VTEC.Time in ter vals of se rum sam ple col lec tion from pa tients with HUS and from those with HC and the time of sam ple col lec tion af ter on set of HC were both similar to those re ported by oth ers (3,11).Di aly sis could not have low ered an ti body lev els in the pres ent study, ex cept in the patient who was he mo dia lyzed.
The im mune re sponse to VT in chil dren with E coli O157:H7 HC has not been com pletely de scribed (6,19).Siddons and Chap man (20) re ported com pa ra ble neu tral iz ing an ti body ti tres against VT-1 and VT-2 in pa tients with E coli O157:H7 HC and in healthy con trols with a sin gle se rum sample.Chart et al (21), us ing ELISA, failed to dif fer en ti ate between sin gle se rum speci mens from pa tients with HUS and sam ples from ap par ently healthy con trols; they also suggested that an ti bod ies to VT-1 and VT-2 are of lit tle value in the se ro di ag no sis of HUS caused by E coli O157:H7 (21).We detected an ti bod ies against VT-2 more fre quently than an ti bodies against VT-1; se ro con ver sion against VT-2 also oc curred more fre quently than against VT-1.How ever, we can not de termine whether VT-2 is more im mu no genic than VT-1 be cause we did not as sess the ex cre tion of free fe cal VT-1 and VT-2 in vivo.It has been pre vi ously re ported that strains of E coli O157:H7 may pro duce VT-2 more fre quently than VT-1 (22,23).
The role of an ti bi otic ther apy in E coli O157:H7 HC is unclear (24).Pre vi ous in vi tro stud ies re ported that TMP-SMX (15) and po lymyxin B (16) may in crease the re lease of VT produced by E coli O157:H7.If TMP-SMX does in crease the release of VT in vivo, it was not re flected by an in creased im mune re sponse to ei ther VT-1 or VT-2 in our popu la tion.We rec og nize, how ever, that the power to de tect such a dif ference be tween treat ment and con trol groups was small (24).Con trary to our pri mary hy pothe sis, there is no evi dence that lev els of neu tral iz ing an ti body against VT-1 and VT-2 are dif ferent in clas sic HUS or af ter an ti bi otic ther apy from those in patients with un com pli cated HC.
It was re cently pro posed that VT may be toxic to the immune sys tem.Co hen et al (8) re ported that the vast ma jor ity of shiga-like toxin-sensitive B lym pho cytes are of the im muno globu lin (Ig) G and IgA com mit ted sub set, whereas most IgM and IgM/D pro duc ing cells are re sis tant.Ab nor mal subpopu la tions of T and B lym pho cytes have also been described in HUS (25); how ever, the sig nifi cance of these find ings has not been de ter mined.Lipo poly sac cha ride may be an im por tant modu la tor of the in flam ma tory re sponse of the host in HUS (26,27).More stud ies on the im mune and the in flam ma tory re sponse of the host are needed to help de fine the pa tho physi ol ogy of clas sic HUS.Fur ther analy sis of variables in the clini cal spec trum of E coli O157:H7 in fec tions may iden tify risk fac tors for de vel op ment of HUS.