Entamoeba Histos>Fca and Entamoeba Dispar

Entamoeba histoS>fcaand Entamoeba dispar: Mechanisms of adherence and implications for virulence

Dylan Ravindran Pillai

A thesis subrnitted in confonnity with the requirements for the degree of Doctor of Philosophy Institute of Medicai Science University of Toronto

O Copyright Dylan R. Pillai, 2000 AcquiSiiand Acquisiins et Bblibgtaphic Services services bibliographques

The author has granteci a non- L'auteur a accordé une licence non exclusive iicence allowing the exclusive pettantà la National Lïbrary of Canada to BïbIioth&pe nationale du Canada de reproduce, loan, disûibute or seil reproduire, prêter, disniiuer ou copies of this thesis in microform, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfichef61m, de reproduction sur papier ou sur fotmat electronique.

The author retams ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts hmit Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimes reproduced without the author's ou autrement reproduits sans son permission. autonsatioa, In 1925, Emile Brumpt had proposed the existence of two morphologically identical species, Entarnoeba hisrolylica (a human pathogen) and Entmoeba dispm (a non-pathogenic human commensal). Recent molecuiar and genetic evidence has supported the two-species concept. The pathogen relies on the galactose/N-Acetyl-O- galactosamine Iecth for adherence that leads to contact-dependent cytolysis of host target ceils, a key step in amebic vinilence. Work presented here explores the structure, fiinction, and expression of the lectin nom E-histoZyticu and E-dispar. We hypothesized that molecular differences between the pathogenic and non-pathogenic lectin homologues would be instructive in understanding the differentid role of each in colonization

(E-dispor) versus invasion (E.histolytica). To acheive this, a gene encoding the lectin in

E-dispar (dhgI2) was cloned and sequenced. Homologous lectin message and proteh levels in E. histolyrica and E-dispur under axenic culture conditions were compared. A heterologous expression system (in COS 7 cells) was constructed to test anti-lectin monoclonal antibody reactivity for E. histoi'ytica and E-dispar lectin homologues. Lech protein, which was generated in COS7 cells, was used for enzyme-linked immunosorbent assays in which antisera from patients infected with pathogen or non-pathogen were tested. Further experiments were also performed in order to localize the carbohydrate recognition domai@) of the amebic lectin using a newly developed in vitro assay. Taken together, evidence is presented supporthg the view that E.histolylica and E-dispar possess highly conserved lectin molecules that are differentially expressed in pathogen vernis non-pathogen. Additional experùnents were Mormed to isolate and characterize a P2 integrin-like molecuie expressed on the amebic dace* Data supporthg the identification of a distinct B2 integrin on the daceof E. histolytica is presented. A putative role for this molecule in amebic virulence is discussed. Finally, epidemioIogîcaI questions regarding E.histolyticu and E.disp4~infections were revisited, now relying on stool antigen immunoassays. These immunoassays, which are able to distinguish

E. histoZyticu fiom E.dispar9demonstrated that the overwhelming majonty of infections in a non-endemic setting are caused by the non-pathogen. These results have broad clinical implications since specific molecdar diagnosis would obviate the need for unnecessary chemotherapy of E-dispar infections, with its attendant costs, nsk of side effects, and danger of drug resistance.

iii This work is dedicated to my parents, Sam and Kamala, whose love and sacrince are the source of my energies. Above d,I would like to th& Kevin Kain for the guidance, mentorship, and enthusiasm in his laboratory dining the course ofthis work. I would also Wre to dia&

Me1 Silverman for having the vision to pursue and foster the MD/PhD programs in

Canada, and for aiiowing me to be a part of it at the University of Toronto in my hometown. 1wouid iike to express my gratitude to members of my advisory cornmittee

James Brunton, Ian Crandall, and Keiiy MacDonald. Not least, 1owe a great deal to Seth

Bitting, Charles Dela Cruz, Anand Ghanekar, Shankar Karunanithi, Rhona Liridsay,

Bryan Lo, Christopher Rickerd, and Mieke va.Zante for their golden companionship in recent years, 1- Introduction: Amebiasis and Entmoeba histolyticu infection...... 1

1.1 Epidemiology of amebiasis: Entmoeba hisrolyh'cu and E. dispar ...... -2 1.2 Current diagnostic ..approach to E.histoZytica Xection...... 10 1 -3 Treatment of amebiasis...... 17 1.4 Pathogenesis of E-hi~olyrcainfection: the key vinilence factors...... 28 1.5 E. histo&tica infection and immunobiology...... O...... 35 1.6 Lectins: an ove~ew...... 41 1.7 The GaVGdNAc Iectin: adhesin and vinilence factor...... 47 1 -8 Rationaie and Hypotheses...... 61 1.9 References ...... -64

II . The cysteine-rich domain of Entamoeba histolytca adherence lectin is sunicient for hi&-aff?nity GaVGaiNAc-specinc binding in vitro ...... -87

2.1 Abstract ...... -88 2.2 Introduction ...... -89 2.3 .Materials & Methods...... -90 2.4 Results ...... -99 2.5 Discussion...... 109 2.6 References ...... -113

III. A gene homologous to hgU of Entmoebu histolytica is present and expressed in E. dispar ...... 1 18

3.1 Abstract ...... 119 3-2 Introduction...... 120 3 -3. Materials & Methods ...... 121 3 -4 Results ...... 126 3.5 Discussion ...... 131 3 -6 References ...... -132 N .Molecula. r characteruaton of the WGaINAc lectin of E.dispm ...... -135

4. f Abstract-...... 136 4.2 Introduction...... 137 4.3 .Materials & Methods...... 139 4-4 Redts...... 146 4.5 Discussion ...... 158 4.6 References ...... 163

V-A novel PZintegrin (CD 18)-Wre molecule participates in Entmoeba histolylica adherence to endothelid celis...... -168

5.1 Abstract...... 169 5.2 Introduction...... 170 5.3 .Materials & Methods ...... 172 5 -4 Results ...... 1 76 5.5 Discussion ...... 185 5.6 References...... 189

VI. Entamoeba histolytica and E.dirpar: rnolecular diagnosis. epidemiology. and treatment implications in a non-endemic setting...... 195

6.1 Absmct...... 196 6.2 Introduction...... 197 6.3 .Materials & Methods ...... 199 6.4 Results ...... ,...... 201 6.5 Discussion ...... 216 6.6 References ...... 220

VII .Irnmunochromatographic sîrip based detection of Entamoeba histolytica/ E-dispur and Giardia lamblia coproantigen ...... 226

7.1 Abstract...... 227

7.2 Introduction~ ...... ,...... 228 73.Materials & Methods ...... ,,... 230 7.4 Results...... 231 7.5 Discussion...... ,235 7.6 References ...... 237

vii 8-1 Fine-mapping the carbohydrate recognition domains (CRDs) of the

E-histolytica GaVGaNAc Iectin heavy subunit,-,....,. - .,,. t.t.ttttt.tttt...tt. .. - .- 240 82ImpIicatÏons of structure-hction similarities and differences between the E, hislolyrica and E-dispar GaVGaINAc lectin- - .. . - ...... ,...... ,. ,-. , . . , . . .., 246 8.3 .Tissue tropism and ''vinilence factors"...... 251 8 -4 Vaccinology :the utility of a mucosal DNA vaccine to pavent E. hisrolytica infection.. .. , , . , . , . , ...... ,. .,...... ,, ., , , , , . ,, ,-- .. . , . - - .. . . - - .- .. . , . . -. .- -.. . -255 8.5 Specific.. diagnosis of E.histo&tica and not E-dispur: clinical implications...... - - - ...... - ... . - .-.. . - .. . . - .- -.,...... - ...... -256 8-6References.. ,-,,- - ,. - - - - -..- .- ... - - .- .., , --. .. .. -... - -.-. . . - --...-...... - 260

viii 1 .2 .1 Computed tomography scan of amebic iiver abscess...... -12 1 .7.1 Subunit structure of the E-histolyticaWGalNAc lectin ...... -52 2.1 Immunoprecipitationof En vitro synthesized cysteine-rich Hg12 ...... 100 2.2 Immunoreactivity of 3F4 with the hololectin and native IectÏn heavy subunit... 103 2.3 Iïnmunoprecipitation of PDI-treated 35~-labelledcysteine nch Hg12 ...... 104 2.4 Binding of 35S-Lectin CR1 Hg12 to CHO cells...... O...... 1 05 2.5 Binding of "s-~etand 35~-~yslabelled CR1 Hg2 to GaiNAcci9BSA...... 106 2.6 Bindhg of 35~-1abelled~~12 to GaINAc19BSA ...... 107 2.7 ~indin~of 3s~-labelled CRI Hg12 to GalNAclgBSA...... 108 3.1 The complete predicted amino acid sequence of the GaUGaiNAc-inhihitable lectin gene dhgl2 of E.dispm ...... 123 3 -2Southern blot of E-hisrolyrica and E.dispm genomic DNA ...... -127 3 -3 RT-PCR of E-histolyticaand E-dispur total RNA ...... 129 3 -4PCR-RFLP analysis diffierentiates E-dispar fiom E.histolyrica ...... 130 4.1 Relative expression of lectin genes by RT.PCR ...... 147 4.2 Relative expression of hololectin on amebic sudace by flow cytometry ...... 149 4.3 Lectin heavy subunit fiom E-dispar @hgl2) is sdücient for high-afnnty GaVGalNAc-specific binding in vitro ...... -151 4.4 hunoreactivity of axenic E-histolyticaand E-dispar under native conditions 152 4.5 Immunoreactivity of axenic E. histolytica and E-dispar after fixation ...... 154 4.6 E-dispar lectin retains the lG7 epitope when expressed on COS-7 ceiis...... 156 4.7 Seroreactivity of cysteine-rich region in human amebic infection...... 157 5.1 Amebic f32 integrin daceexpression by flow cytometry ...... 178 5.2 Detection ofamebic GaVGalNAc ho1okcti.n by ELISA ...... 179 5.3 Flow cytometric analysis of $72 l insect cells expressing Hg12 ...... 180 5.4 Immunoprecipitation of the putative amebic B2 integrin ...... 181 5.5 Adherence of activated endothefiai ceils to amebic trophozoites ....,...... 182 5.6 Adherence of CHO cells expressing human ICAM-1 with amebic trophozoites 183 6.1 Demographic and epidemiological data on study population ...... 202 6.2 Distribution of results fiom ELISA kits performed on stool specimens ...... 207 6.3 Prevalence of E-histoZytica and Edispar among patients ...... -208 8.1. Schematic representation of in vitro binding assay ...... 241 8.2 Secondary structure prediction for hg12 and dhg12 ...... 247 8.3 Working hypothesis of E- histolyrica and E. dispar GaüGINAc lectin ...... , 250 8 -4 Putative role for f32 integrin-mediated adherence at intestinal epithelium ...... 253 1.1.1 Estimated global prevaience and incidence ofamebiasis ...... S...... -...... 8 1 .2.1 Results of ELISA performed on amebic Ever abscess aspirate ...... 14 1 .3.1 Treatment of the asymptomatic cyst passer...... 19 1.3 -2Treatment of invasive disease (amebic colitis)...... 22 1.3.3 Treatment of invasive disease (arnebic liver abscess) ...... -25 1.3 -4 Mications for aspiration of amebic iiver abscess...... 27 1.6.1 Cornparison of features which distinguÏshC- and S-type lech...... 42 1.7.1 Chhese Hamster Ovary ceIl mutants ...... 4 . 1.7.2 Carbohydrate specificity of the E-kistolyticaGaVGaINAc lectïn ...... 30 1.7.3 Features ascribed to anti-Iectin monoclonal antibodies ...... 37 6.1 Cornparison of ELISA redts: ProSpecT, Entmoeba test, and mAb 3 18.28 ....210 6.2 Specincity and sensitivity for detection of E.histoZjticafE.dispar complex .....-212 6.3 Predicitve value for a positive result by micmscopy ...... 213 6.4 Results of ELISA kits for patients with positive amebic serology...... 214 6.5 Results of indirect hemagglutination test ...... 215 7.1 Resdts of Triage versus ProSpecT for the detection of E. histolyricafdispar ..... 232 7.2 Limit of detection of E~hi~toi~cafdîspmby Triage and ProSpecT ...... -233 7.3 Results of Triage versus standard O & P ~croscopy...... 234 A - adenosine ADCC- antibody dependent ceUuiar cytotoxïcity Ala - danine ALA - amebic Liver abscess AM0 - abscess-derived macrophage Arg - arginine Asn - asparagine bp - base pair C - cytidine Cl - comptement factor 1 CD - cluster of Herentiation CHEF - contour-cIamped homogeneous electric fieId CHO - Chinese Hamster Ovary ConA - concanavaiin A CR - cysteine-rich CRD - carbohydrate recognition domain CRE - core regdatory element CT - cholera toxin CTL - cytotoxic T lymphocyte CTLD - C-type Iectin-like domain Cys - cysteine DAF - decay accelerating factor DNA - deoxyribonucleic acid DNase - deoxyribonuclease dNTP - deoxynucïeotide triphosphate EGF-R - epidsrmd growth factor receptor ELISA - enzyme-Iinked immunosorbent assay F-actin - filamentous actin Fab - antigen binding gagment FITC - fluoroscein isothiocyanate G - guanosine Ga1 - GaIactose GaINAc - N-acetyl-D-galactosamine Glu - glutamic acid GM-CSF - granulocyte macrophage colony stimulatuig factor HA - hemagglutinin hr - hour FA- indirect immunofluorescence assay IgA - immunogiobuIin A IgG - immunogIobulin G IL - UiterleuIàa NOS- inducible nitric oxide synthase Iso - isoleucine kb - kilobase kDa - kiloclalton LT - heat-labile enteroto* mAb - monoclonal antibody MAC - membrane attack complex Met - methionine MHC - major histocompatibility complex min - minute mLIF - murine leukemia inhibitory factor NO - nitric oxide O & P - ova and parasite PCR - polymerase chah reaction PGE - prostagland.E Phe - phenyIalanine PHA - phytohemagglutiain PT - pertussin toxin PVA - polyvinyl akohoi RNA -ribonucleic acid SAA - soluble amebic dgen sIg - secretory immmogiobulin T - thymidine Th - T helper TM - transmembrane TNF - tumor necrosis factor Tyr - Tyrosine URE - upstream regdatory element

xii

The epidemiology of amebiasis has a tortwus history owing in large part to co~on over the naming of the causative agent, Early on, complications arose because the life cycle of the organism is comprised of infective cysts (rounded and non-mode) that are acquired orally and MenttrophoPtes (arneboid and motïie) that are responsible for symptomatic disease.

Entamoeba histoiytr'ca resembles several amoebae that were subsequently identified as distinct species, and many of which are Eiannless cornmensais in the human host-

1.1.1 Genesis of a disease

The name of the causative agent of amebiasis, Entamoebu histoiyticu, was first introduced by Fritz Schaudinn in 1903 (21 1)'. However, the idenacation of the organism was made much eariier by Fedor Losch in 1875 (137)'. Losch had identined motile trophozoites, amoeboid in nature, in the dysenteric stool of a labourer in St. Petersburg, Russia. Over the next

16 years, there were many physicians and învestigators who reported the presence of bactena and amoebae in dysenteric stool, but no association was made between the presence of amoebae in the stool and disease. In 189 1, Councilrnan and Lafleur reported the presence of ccAm~ebacoli" in bacteria-fiee berabscess pus, demonstrating the pathogenlc potential of the organisrn independent of bacteria (50)'. Schaudinn suggested that "Entamoebahistolyrca" should be reserved for the dysentery-associated species, and sirniiar organisms harboured in heaithy inviduais (which had also been observed) shouid be termed "Enrmnoeba coi?'.

' Reprinted in "'ïropical Medicine and Parasitology: Classic Investigations" Volume 1, eds. Kean BH, Mott KE, and Russell AJ. Cornell University Press, Ithaca, NY, 1978. The first clear fbifïhent of Koch's postdate was pedionned by Emest Walker and

Andrew Sellards with human experîmental subjects in Mda, The Philippines (244)'. In 19 13, these authors reported that ifseveral individuals were infected wïth the same isolate, some but not a11 would develop dysentesr, However, cysts isolated fiom the faeces of these primary infected individuais @respective of whether they were symptomatic or not) were fed oraiiy to secondary infected individuais and dysentery was observed. Aithough these experiments were ethicaiiy tenuous in the absence of curative chemotherapy, they were definitive in three ways: (i)

Koch's postulate was fuifiiled for amebic dysentery caused by E-histolyrica. (ii) Not ail individuals infected with E- histolyria succumbed, suggesting E-histoiytica was not always invasive. (iii) The mode of transmission of amebic dysentery was estabiïshed as faecai-oral. By

19 19, Clifford Dobeli had anaiyzed the vanous descriptions of this protozoal parasite of humam, and concluded there was only one species of Entamoeba with quadrinucleate cysts ingcsted orally, and that released infective trophozoites in the bowel, namely E..histo[ytica (62)'.

Confusion arose, however, in 1925 when Ernile Brumpt suggested a morphological1y identical, non-pathogenic organism with the same Mie cycle existed, which he termed Entmoeba dispar (33)". Using a feline model of amebiasis, Brumpt demonstrated that ccE.di~ar''-idected animals failed to develop symptoms. Brumpt had also prospectively followed an individual infected with ccE.dispa?',but who never became symptomatic for over 8 years. Large epidemiological surveys performed by Brumpt and colleagues in temperate countries suggested that invasive disease was unheard of. To Brumpt, these data were consistent with the presence of a non-pathogenic counterpart to E. histolytica in temperate regions He also asserted that

- Reprinted in "Tropical Medicine and Parasitology: Classic investigations" Volume 1, eds- Kean BH, Mott KE, and Russell AJ, Corne11 University Press, Ithaca, NY, 1978. mm Translated by CG Clark, London School of Hygiene and Tropical Medicine (personal communication), E-histclytica was more common in tropical countries where invasive disease was more readily observed.

1.1.2 Evolution of virulence: water-borne transmission

The lgmcentury saw a shift in paradigm for evolutionary biologists who studied the evolution of vinilence. EarLy on, the entrenched view was that ailparasites (hmvirus to bacteria to protozoan) evolved towards balance4 stable associations with their host- However, a cfearer understanding of the genetic basis of evolutionary fitness promoted the following view: if improved Survival of the parasite due to optimd exploitation of the host resulted in extreme harm to the host, then the host-parasite interaction vdiadapt to a situation of increased harm (73).

A coroliary to this view is that a reduction in host mobility would be especiaiiy debilitating to parasites that rely heavily on host mobility for transmission. Reliance on host mobility would dictate that in the developed world, where water supplies are subject to purifcation, Entamoeba would be more reliant on host mobility for transmission. Hence, adaptation to a more benign relationship with the definitive human host would be selected for in

Europe, North Amenca, and Japan. However, in developing nations, where access to safe drinkuig water is not routinely available, there is less reliance of E-oeba on host mobility for transmission. These conditions would not apply selection pressure for a more benign interaction.

Based on this theory, Entamoeba would be more vident in developing nations.

The accumulating epidemiological evidence has supported this view. Harmfulness and pathologie profile of infections in develophg countries is far worse than in weaithier nations (4,

45). Whether this has resulted in a speciation event where E. histoiyiica has differentiated into a Wulent and an avinilent organism has been subject to debate for some the (dlscussed Merin section 1.1 -4).

1.1.3 Estimating the morbidity and mortality

Before an estimate of global morbidity and mortality of amebiasis can be made, it is necessq to acknowledge certain limitations with epidemiologic surveys associated with this disease.

A key criticism of existing studies is that many sample populations descnbed are not representative of the entire population (5 1,52,57,22 1,49,90,116, 1SO), 88, 120,53, 239, 19,22y40,41,56,220,232 as reviewed In 245)- Most studies rely on diagnostic Iaboratory results of hospitalized individuds or outpatients, rather than a randornized cross-sectional sunrey (245). Rapid and sensitive diagnostic methods are not available in many endemic areas, precluding large surveys in these countries (95). Sporadic shedding of cysts in stool seen with E. histolyrca infiection requires several samples to be taken for rnicroscopic andysis in order to reduce the number of falsely negative results (145). The majority of surveys measwe prevalence at a &en tirne, derthan long-tenn follow up. This is problematic since seasonal changes of prevalence cmoccur (28). Serologic assays Vary fiom study to study (at least 9 dinerent ones have been documented), and dinering sensitivities and specincities may skew estimates from region to region (95,245). The clinical spectmm of amebic infection varies considerably. Few studies dehe ccdy~enterf',c'colitis", "rectocoIitis~~,"subacute colonic amebiasis", "hepatic amebiasis", "ameboma" and so on. As a result of these limitations, dobai estimates of prevalence, incidence, morbidity, and mortaiity are inaccurate. Nevertheless, several concIusions can be gieaned fkom these studies.

Entamoeba infection appears to be widespread, but the prevaience varies considerably nom one geographic area to another (244). These studies used stool (cysts and trophozoites observed by microscopy following appropriate staüüng) muor serology (antibodies measured by any one of several immunoassays) diagnosis. Diarrhea is the cornmonest symptom among patients experiencing invasive disease, although many infected individuals remain asymptomatic (244)-

The diarrhea itseIfcan vary fiom a few bose stools to febriie, bloody dysentery. Antibody responses to amebic proteins have been found more commckLy with those experïencing symptomatic infection (235, 103, 114, 117). This observation is critical since it ailows investigators to assess invasive disease (such as colitis or abscess) compared to merely infection

(stool positivity)).A higher relative risk of Section bas been observed in men in their 3rd - 4" decade, and in pregnant or postpartum women (2, 11)- in Europe and North Amerka, where access to safe drinking water is almost guaranteed, amebiasis seems to occur in defined groups, including institutionalized persons, recent traveliers to or immigrants fkom endemic areas, and sexually active male homosexuals (244).

Walsh performed a metaanalysis of recent articles reporting epidemiological swvey results and projected estimates of the number of infections, number of cases of disease, and the number of deaths by region (Table 1.1.1). These resuIts suggest that 480 million people were infected with "E.histolyticd' in 198 1 (assuming a single species, see 1.1.3). However, only 34-50 million individuals experienced symptomatic disease, with 40-1 10,000 resulting fatalities- This estimate makes amebiasis the third Iargest killer of humans by parasitic disease, behind malaria and schÏstosomiasis. Interestingly, oniy a fiaction (10%) of those infected with Enîamoeba actuaily experïence invasive disease, consistent with the possibüity of two organisms, one pathogenic and the other non-pathogenic, both of which are morphologicaily identical.

1.1.4 Zymodeme analysis: moleciilar differentiation into two species

For aimost 50 years, parasitologists who were driven by morphological identification, refused to accept the Brumpt two-species hypothesis: E.histo2ytica was the potential pathogen and E-dispar was a gut commensal, in spite of morphological identity- In 1968, Reeves and

Bischoff were the fim to classify Entamoeba species based on the electrophoretic mobiiity of amoebal enzymes (186). Sargeaunt and workers extended this initial observation by orchestrathg a global effort to class* Enimnoebae based on isoenzyme mobility (207-209). These workers demonstrated that, in over 10,000 isolates, certain isoenzyme profiles (or c'non-pathogenic zymodemes") were never associated with invasive disease. Further immunological evidence was provided by John Ackers and colleagues who demonstrated that certain antigens were unique to the pathogenic isolates (227). More recently, several investigators have demonstrated that non- pathogenic isolates identified in a global survey are geneticaiiy distinct fiom pathogenic isolates

(reviewed in 233). This dowed for a formal redescription of the non-pathogenic qmodemes as

"Entamoeba dispar", in keeping with Bnunpt's prescient observation 73 years earlier, and the pathogenic zymodemes as "Entamoeba histolyria" (60).This nomenclature is now widely accepted in the liteninue. The implications of this redescription were signiscant since molecular differentiation of pathogen nom non-pathogen suggested that epidemiological questions had to be revisited and treatment protocols reconsidered. For example, it is uncertain whether the 90% of individuals, who according to Walsh, are infected with E-histolylica but are asymptomatic, Table 1.1.1 Global p~valenceand incidence of amebiasis in 198 1 (245).

Continent No- infections No. coiitis or abscess cases No-deaths

(miilïons) (miilions) (tbousands)

North & South America 90 10 10-30

Asia 290 15-3 O 20-50

Afi-ica 80 10 10-30

Europe 20 O None reported

Total 480 34-50 40-1 10 are actuaiiy Mected with E-dispar.Similarly, it remains unclear whether the 10% of infected and symptomatic individuals are infectecl with E-hiktolyticasensu stn'cto- Institutionaiïzed persons, active gay males, and chilciken in day cate centers are at risk for infection with E-hÏstoCytica or E-dispar- In an effort to describe the diagnostic approach to amebic infection, a case report of a 30 year-old male with multiple liver abscesses is presented-

This individuai denied risk factors for local transmission and had not traveIed to an endemlc area since emigrating fiom SnLanka seven years ago, Liver aspirate ELISA and serological analysis revealed infection with E-histolytica and not E-dispar. The presence of E-histolyfica in an individuai with no reported risk factors for local transmission is highly unlikeiy. One speculates that this individual contracted E. histolyrica in an endemic area years earlier and remained asymptomatic unt.corticosteroid immunosuppressivetherapy led to invasive disease,

1.2-1 Introduction

E. histoi'ytica is the causative agent of invasive disease resulting in amebic colitis or liver abscess. E-dispar infection redts in asymptomatic colonization of the large intestine. Both species are morphologicaliy identical but geneticaily distinct (60). Chical studies in both endemic and non-endemic settings have reixûiorced the new paradigm which posits that E-dispar is a gut commensal, while E-histolytica is responsible for both intestinal and extra-intestinat disease (92). However, asymptomatic infection with E-histolytica (pathogenic strain) is not uncornmon in individuals who have lived in an endemic area for many years (98). Definitive diagnosis requires demonstration of E.histolyrica cysts oc trophozoites in the stool or positivity for E. histoiytica coproantigen, and signifïcant (21 :5 12) serological anti-amebic titres (1 02).

However, both sporadic shedding of cysts and the low specificity of anti-amebic titres in residents of or immigrants hmthe tropics confou1ids labotatory dlagnosïs. Non-endemlc setting transmission of Ehfstoiytrca is rare, wheres Edrspw transmission is evident among institutionaIized individuais, active gay males, and day-care infants (1 15). Pathogenesis of

E. histolyrica infection leading to invasive disease Iikely involves adherence to colonk mucin, cytolysis of intestinal epitheiiai and immune effector celIs, and modulation of host hune firnctions (37).

1.2.2 Case Report

A 30 year old male, who had emigrated 7 years earlier fiom Sri Lanka, was admitted to a hospital in Toronto, Canada, in early May 1998 with a diagnosis of dergic bronchitis and asthma- His respiratory symptoms resolved foIlowing treatment with a quinolone antibiotic and prednisone (1 Smg, daily). He was well until -3 weeks later when he developed right upper quadrant pain, fever, and fatigue and was re-admitted to hospital. BLood work performed at that time reveaied an elevated white blood count (175 10'~).Neutrophils (13.5~109/L), monocytes

(2.2 x IO'L), and eosinophils (0.8~10~~)were aiso elevated. Abdominai ultrasound perfomed on June 24, 1998, revealed three hypoechoic lesions in the liver, with the largest measuring 7cm x 8.5cm x 6.3cm. One week later computed tomography (Figure 1.2-1) of the abdomen confïrmed the presence of cysts in the right lobe of the liver. Daughter cysts, mural nodules, and calcifications were absent. Remarkably, during the first few days ofhis hospitalization, the right upper quadrant pain and fever spontaneously resolved despite the absence of any specific therapy. The patient remained well with no fever, abdominal pain, anorexia, nausea, vomiting or diarrhea. Blood work perfonned in early August showed no irregularitïes with leukocytes and granulocytes retuming to baseline levels. Figure 1-2-1 : Coniputed tomography scan with contrast dye indicoting amebic liver abscess in nght lobe of liver of patient infected with E-histolytica. 00August 4,1998, he was referred to the Tropical Disease Unit in order to determine the etiology of his liver lesions. He was essentiaiiy asymptomatic aithough he did admit to some fatigue and mild shortness of breath on exertion- Amebic and hydatid serology was performed-

An ultrasound-guided percutaneous nedeaspiration of the liver cysts was performed whiie awaiting the results of the serologic investigations The cyst aspirate was subjected to enzyme- linked immunosorbent asçays (ELISA) in order to determine if E- histoZj&a antigen was present in the aspirate. Both Techlab's Entamoeba test and Alexon's ProSpecT gave a positive readùig for Entamoeba surface antigens. Techlab's E. histolyiica test (specinc for E. histolytica antigen) was positive (Table 1-2.1)-

Amebic serology (by indirect hernagglutiaation m])was subsequentiy shown to be positive at a titer of 1:4096, Hydatid serology as assessed by enzyme immunoassay and immunoblot at the Centers for Disease Control and Prevention (CDC) in Atlanta was negative-

The patient was treated with 2g of metronidazole (a tissue-active amebicide) for five days, followed by the lumenal agent r0d0quinol(650mg tid. for 20 days). When seen in foilow up at 6 months, the patient was asymptomatic and the iiver lesions had largely resolved as determined by ultrasound- Table 12.1 Redts of ELISA performed on the amebic Liver abscess aspirate

ELISA OD 1nterpretation3

~echlab' Entamoeba test 0.566 positive (EdIEh)

Alexon-a rend^ ProSpecT O. 161 positive @d/Eh)

Techlab E. histol'ytica test 0.904 positive (Eh ody)

1. ELISA kit manufactured by Techlab (BIacksburg, VA). 2. ELISA kit manufactured by Alexon-Trend (Ramsey, MN). 3. c%d/Eh" implies test is positivefnegative for either E-di- or E-histolylica; "Eh only" Unplies test is positivefnegative for the pathogen E-histolylica alone 4. Cut-off vaiues for al1 kits are OD=0,050. 1.23 Discussion

Prolonged treatment with oral steroids in the management ofthis patient's bronchitis and asthma likely had a significant immunosuppressive effect. Prior infection with EhisrolyhCa presumAbly occurred years earlier in Sn Lanlca with no symptoms until the theofprednisone dnig therapy. Prednisone has kenpreviously reported to provoke rapid progression of fulmuiaat amebic disease following systemic dmg therapy in patients who have spent the or are currently living in the tropics (68,70,86,104,105,129, 154,159). One patient had last been in an endemic area as long as 25 years before corticosteroid-induced Invasive amebiasis (228). This suggests that pathogenic E-IzrStolytica can remain latent for years.

Recent developments at the bench ailow us to revisit out understanding of how glucocorticoid-mediated immunosuppression may predispose an individuai to invasive disease in spite of long-standing amebic colonization in the large intestine. Small animal studies

CO- that corticosteroid therapy coupled with thymectomy exacerbates invasive amebiasis to a greater extent than thymectomy or sham-operation aione (80)- Moreover, investigations performed in viîro have shown that amebic antigens specificaiiy induce pro-infiammatory cytokines IL-8, IL-la, gro-a, GM-CSF, and IL-6 from colonic epithelial celis (67).

Dexamethasone (a glucocorticoid) specincaliy inhibits amebic induction of IL-8 from huma. colonic epithelial cells (256). Pro-idammatory cytokines recmit neutrophils and other leukocytes to the site of amebic mucosal innltration (202). Since glucocorticoids are potent suppressors of pro-innamatory cytokine secretion, it is possible that prednisone therapy resulted in inhibition of immune effector cell recruilment and defense at the intestinal epithelium.

This rnay alter the balance in favor of the pathogen to invade and metastasize, most commonly to the liver. E. histolytica infection bas the potentiai to be sukllnical for many years and it therefore remaius crucial to nile out E-hrStolyhCa Infectioa FoUowing a course of immunosuppressive therapy, amebic infection shouid be incfuded in the differential of fever and bloody diarrhea or right upper quadrant pain in any individuai who has Llved in an endemlc are% To this end, we have recentiy shown that serological tests Iike indirect hemaggiutination test (EW) anaor stool antigen ELISA are specific for E. histolyticu infection (95).

1.2.4 Conclusions

In sumrnary, corticosteroid-induced immunosuppression iikely Ied to the rapid onset of invasive disease, including metastasis of amebic trophozoites to the nght lobe of the iiver. Multiple abscess formation ensued Initial symptoms comprishg right upper quadrant pain and fever had resolved following cessation of prednisone treatment. Only mild shortness of breath persisted, however, until the time of definitive diagnosis. The patient responded weii to 2g of metronidazole (a tissue-active amebicide) for five days, followed by the Lumenai agent

Iodoquinol(650mg ti.d. for 20 daYs).The liver lesions were almost entirely resolved on foilow- up ultrasound 6 months later- The treatment of amebiasis has two fiindamental objectives: (0 to cure invasive disease at both bintestinal and extra-intestinal sites and (fi) to eluninate the passage of cysts fiom the bowel lumen. Metronîdazole has emerged as the dmg of choice for the treatment of invasive disease.

Other 5-nitroimidazole denvatives such as tinidazole, oniidazole, and secnidazole, available outside of Noah Amenca, may be equally efficacious. Several lumenal agents are available to eradicate cyst passage. DDixanide furoate is the preferred agent but is not readily available.

Other effective agents include paromomycin and iodoquinol. One complicating factor in the use of these agents has been the overwhelming evidence that there exists a morphologicaliy identical but genetically distinct non-pathogenic species, Entamoeba dispar. This organism does not cause disease and therefore requires no chemotherapy. Controlied trials with antigen detection kits and PCR-based assays that cmdistinguish Entamoeba hi~olyticafiom E. dispar are in progress (Chapter 6 & 7). However, until these techniques are Myevduated and widely available, treatment of asymptomatic cyst passers (iikely E-dispur infection) in non-endemic areas is suggested. Treatment of asymptomatic cyst passers in areas where Entamoeba dispar/ histoZytica is endemic remains a futile public health policy due to high rates of re-infection with

E. dispar. In addition to chemotherapy, drainage of amebic liver abscesses using fine needle aspiration, guided by ultrasound or computed tomography, is a useful therapeutic intervention when (i) abscesses are extremely large and there is danger of imminent rupture, (ü) medical therapy has failed, (ii) a left lobe abscess threatens to rupture into the pericardium, or (iv) immediate symtomatic relief is required. Surgical intervention is reserved for toxic megacolon compfica~gamebic coiitis or in cases where hepatic abscesses have ruptured into theperitoneal cavity.

1.3.1 Non-invasive colonization (Asymptomatic CysüTrophozoite Passers)

Most asymptomatic individuals diagnosed with ccamebiasis",based on stoof microscopy, wiU in fact be infected with the harmless commensal E- dispar. Preliminary data indicate that

ELISA and PCR-based tests can distinguish E. dispar fiom E. histolylica but these assays remain research tools at present (4,30,93,237, Chapter 3). Amebic serology may be an alternative strategy for the diagnosis of E. histolytica in non-endemic areas, sinçe several studies have shown that antibodies develop only duruig infection with E. hisfolylica, even in those who are asymptomatic (99, 176,243). In these studies, negative serology virtually dedout E.. histolylica infection. However, a recent study of German traivelers showed a disturbingiy low sensitivity for amebic serology among those uifected with E. histoi'ytica (243). One possible factor for the poor results may be the fact that antibodies are often not detectable during the first week of infection.

Furthemore, serology is of limited use in endemic areas since antibodies may persist for years derinfection. Therefore, unless serology is readily avdable, or until new antigen or PCR-based assays are criticalIy evaluated and available, aii aqmptomatic cyst passers and excreters of trophozoites should be treated with a lumen-active agent alone (see Table 1.3.1).

Diloxanide furoate (Furatlll-de) is the preferred agent for eradicating intralumenal cyst passage

(146). A 10-day course has an 85% success rate in elïminating the organism. Less than 10% of the oral dose is excreted in the feces, but sufficient amounts of the active fonn accumulate in the colonic Lumen. The remainder is hydrolyzed within the intestinal mucosa, absorbed, and ultimately excreted in the urine. Treatment is regarded as successfül if the stools remain fiee of cysts and trophozoites for one month. Due to the irreguIar excretion pattern of the organism in Table 1-3.1 Treatment of the Asymptomatic Cyst Passer

Dmg Aduit Dosage Pediatnc Dosage

(Po) @O)

Diloxanide 500 mgtïdx 10 20 mg/kg/daya x 10

Furoatev days days (maximum

(Furamide) 1500 mg/day)

Paromomycin 30 mg/kg/day 25 mg/kg/daya x 7

(batin) x7 days days

(maxÏmum 2 g/day)

Iodoquino 1 650 mg tid x 20

(Yodoxin) days 20 days

(maximum 2 g/day)

In U.S.A. available only through the Centers for Disease Control : telephone (404) 639-3670 or (404) 639-3356 days and (404) 639-2888 nights and emergencies. In Canada: Bureau of Pharmaceutical Assessment, Health Protection Branch, Ottawa (6 13) 941 -21 08. Administered in three divided doses. stool, several specirnens should be evaluated to document cure. Diioxanide is relatively nontoxic; occasional mild gastrointestinal symptoms and increased flatuience are the most fkequent side effects (253). This drug is available only through the Centers for Disease Control and Prevention

(CDC) in the USA and Bureau of Pharmaceuticai Assessment, HeaIth Protection Branch (ZfPB) in Canada.

An alternative and equaüy efficacious lumenal agent is paromomycin (Humath)), an aminoglycoside which is not absorbed fkom the gastrointestinal tract and is die in pregnancy

(23,39, 146,229,254). Diarrhea and abdominal cramps are the only signincant side effects

(1 12). Iodoquhol (Yodoxin) is also marketed in North America and is equally effective to diioxanide fiuoate but requires a longer treatment course (20 days compared to 10 days for diloxanide and 7 days for paromomycin). In addition to gastrointestinai side effects, iodoquinol may interfere with thyroid fünction tests due to its hi& iodine content. Rare dose-related neurotoxicity, such as optic newitis, has been reported with prolonged dmg use or inappropriately high doses (76). As with diloxanide, follow up stool examinations are essentiai to monitor the success of treatment since faiiure rates occur in 1045% of individuais treated with any of the lumenai agents. In vitro studies with bacitracin zinc suggest it may also be usehl as an alternative non-toxic lumenal agent (9). Notably, ~izetmnidazoleVlagyl) or its counterpart 5-nitroimidazole denvatives should not be used to treat the asymptomatic cyst or trophozoite passer since these dmgs are active in tissue Section but ofien fail to eradicate organisms fÎom the lumen. 13.2 Invasive Dlrcwe: Intestinal Sites (Amebic Coütis)

Metronidazole (Flagyl) is dedmg of choice for patients with acute and chronic fonns of invasive intestinal disease including ameôoma Table 2 summarizes alternative treatment strategies. Metronidazole is a 5-nitroimidazole derivative with anti-microbiai activity against some anaerobic bacteria as well as protozoa nich as E. hislolyrica and Gidalamblia.

Metronidazole is almost completely absorbed foliowing oral adminidministration.It is highly efficacious with 90% cure rates in many weli-controlled studies (17,204). Despite widespread use, E. histolyhca resistance to rnetronidazole has not yet becorne a clinical problem. It is important to note however, that the complement of genes associated with multi-drug resistance and the drug-efflux mechankm do exist in this organism (98). Although most authorities recommend that amebic colitis be treated for 7 to 10 days, studies indïcate that 2.4 g of metronidazole given p.o. once daiiy for 3 days is equally effective (29). The use of metronidazole during pregnancy has been controversial with one study suggesting teratogenic eEects when used in the fist trimester and others suggestuig safety in pregnancy (34,168). A cornparison of one cohort of 1000 patients who received the dmg with those who did not showed no adverse effects to the fetus (1 19). Indeed, therapy during pregnancy is warranted for invasive amebic disease because of the increased risk of £idminant amebiasis and severe complications,

Common side effects of metronidazole are primarily gastrointestllia.1

(nausea, headache, metallic taste, abdominal discornfort) and central newous system

(headache, dizziness, and drowsiness) in ongin (192). Other side effects include dark coloured urine, urticaria, and transient neutropenia (193). Patients should avoid aicoholic beverages due to metronidazole's disulfhm-like properties (169). The dmg is also Table 1.3 2 Treatment of Invasive Disease - Amebic Colitis* Dmg Adult Dosage Pedlatnc Dosage (Po) (Do) MetroaZdazolew 750 mg tid x 5- 30-50 mg/kg/daY@x (Fla~l) 10 days IO days or (maximum 2250 2-5 g once daily mg/da~) x 3 days

2 g once daiiy x 50 mg/kg/daymx 3 3 days days

Alternative: ~ehydroernetine~ 1-1.5 1-1 -5mglkgldaym x 20 (Mebadin) mfidkdday days x 5 days (maximum 90 mg/day) (maximum 90 mdday) * Treatment of invasive disease is followed by a complete course of a Luminal agent as descnbed in Table 1.3.1 for the asymptomatic cyst passer In USA: available only through the Centers for Disease Control: telephone (404) 639-3670 or (404) 639-3356 days and (404) 639-2888 nights and emergencies. In Canada: Bureau of Pharmaceutical Assessment, Health Protection Branch, Ottawa (613) 941-2108. Administered in three divided doses. "arketed outside of USA- and Canada May be admuiistered intravenously if patient unable to take by mouth (500 mg IV every 6 h x 5-10 days) avdable for intravenous administration in patients unable to take oral meciications. Shce metronidazole usually fails to eIiminate intesha1 camiage, a lumen-active agent should dways be given in conjuction for any form of invasive disease (amebic colitis or amebic liver abscess

- see next section) whether or not parasites are detected on stool examination (29). To reduce the likeiihood of additive side effects, these drugs may be given sequentially (see Tables 2 and

3). Outside of the United States and Canada, Midazole (Fasigyn), a 5-nitroim*dazole denvative, is the preferred drug for treatment of invasive disease (17). It is generally associated with fewer gastrointestinal side effects. As with metronidazole, a lumen-active agent must aLso be administered. Tetracyciine or erythromycin foLlowed by a lumenal agent are effective alternatives for patients with mild amebic colitis who are unable to tolerate metronidazole.

However, these regimens do not eradicate parasites in the liver and care must be taken to assess the patient for hepatic involvement. An alternative dmg is dehydroemetine (Mebadin) which is rapidly amebicidal in tissues but not the lumen (203). This drug is associated with significant gastrointestinal, cardiac, and neuromuscular side effects. Neuromuscular reactions include muscle weakness and pain, stiffhess of the hbsand neck, while cardiotoxicity results in ECG changes (T-wave flattening or inversion), fdin blood pressure, and tachycardia (203) (230).

There is no clear evidence that cornbinhg dehydroemethe with metronidazole is more efficacious than metronidazole alone in the treatment of invasive amebiasis. Dehydroemethe is only avaiiable in North America with the authorkation of the CDC or HPB. As in asymptomatic cyst passers, follow-up stool examinations should be perfomed 2-4 weeks after the completion of therapy to detect treatment failures. Intestinal perforation during amebic colitis is best managed with therapy consisting of antiarnebic and antibacterial agents. Surgicd resection (usually in the form of total colectomy) is generally reserved for patients wÏtii toxic

megacolon,

1.3.3 Invasive Diserise: Extra-Intestinal Sites

Outside of the gastrointestind tract, Entmoeba hiktofytica may metastasize to the ber, hg, brain, and skh The most common site of tissue invasion is the Iiver. Patients with amebic Liver abscess (ALA) generdy present with fever, weight Loss, abdominal pain, and hepatic tendemess.

The diagnosis is established by demonstrating a space-occupyïng Lesion in the liver through either uitrasonography, CT or MRI scan (71, lS6,2 16,242). Amebic serology is positive in 99% of patients, but may occasionally be negative withui the fïrst week of symptom omet (21 6).

ALA and invasion of other sites can usually be managed by medical therapy alone (see

Table 3). The preferred drug is metronidazole foilowed by an agent to eradicate intralumenal infection. In uncomplicated mild to moderate disease, single dose therapy with 2.5 g of metronidazole has been effective, and provides an alternative to those patients who cannot tolerate the standard 5 to 10 day regimen (29). Chloroquine is known to be a potent arnebicide in hepatic tissue alone and some clinicians have recommended the addition of chloroquine to metronidazole when treating critically iiL patients. However, there are no well-controlled trials to support the use of combination therapy.

Percutaneous drainage of iiver abscess should be reserved for patients who are not responding to medical therapy or who have large abscesses that threaten to rupture (173), (75). In addition, left lobe abscesses can potentialiy rupture into the pericardium and therefore require surgical drainage ifpercutaneous access to the abscess is not possible. Aspiration may also provide prompt symptomatic relief to those with severe right upper quadrant abdominal ph. Table 1.3.3 Treatment of invasive disase (amebic Ilver abscess)*

Dmg AduIt Dosage Pediatric Dosage (Po) (Po) Metronidazolew@a&) 750 mg tid x 5-1 0 days 30-50 mg/kg/dayo x 5- 10 days

2.5 g oraiiy once

Alternative: ~e hydroemetineG 1- 1-5 mg/kg/day IM x 5 1-1 -5 mg/kg/day* x 20 days (Mebadin) da~s (maximum 90 mgfday) (maximum 90 mg/day) *~reatmentof invasive disease is foUowed by a complete course of a Iuminal agent as described in Table 1.3.1 for the asymptomatic cyst passer In U.SA. available only through the Centers for Disease Control: telephone (404) 639-3670 or (404) 639-3356 days and (404) 639-2888 nights and emergencies. In Canada-- Bureau of Pharmaceutical Assessment, Health Protection Branch, Ottawa (613) 941-2 108. O Administered in three divided doses. May be adrninistered ïntravenously if patient unable to take by mouth (500 mg N every 6 h x 5-10 days) There is some recent evidence that dtrasound- and cornputer tomography-guided drainage in combination with intraiesional chemotherapy ïmproves outcomes in AL& compared to medical therapy alone, or medical therapy with surgical drainage (171). However, in general, drainage is recommended only for the complicated case ofALA (see Table 4). The abscess cavity ofien resolves slowly over severd months as determined by ultrasonography and up to 20% of patients wiU have a permanent hepatic cystic cave after the completion of therapy (7). Amebic antibodies as dete-ed by indirect hemagglutination and ELISA may take up 2 to 10 years to disappear after successfûi treatment- Table 1.3.4 Indications for Aspiration ofAmebic Liver Abscess

1. Failure of medicd therapy ( no response within 3 days to metronldazole)

2. Left lobe abscesses with danger of rupture into the pericardium

3. Large abscesses where rupture is believed to be imminent

4. To nile out pyogenic abscess when there are multiple lesions

5. For prompt symptomatic relief of severe pain The enterïc protozoan E-histolytica k the etiologiç agent of human amebiasisS The pathogenesis of this organism has vexed researchers for many years since Koch's postdate is not always met following infection with this organism- That is to Say, only 10% of individuals who are idected actually develop symptomatic disease, while the remaInder are asymptomatic cyst passers. The dual speciation of E-histoi'ytica into pathogen (E-histolytica sensu stricto) and non- pathogen (E-dispar) has forced investigators to revisit epidemiologicd questions since the vast majority are likely infected with E-dispar-However, even ifwe restrict our discussion to

E. histolytica, it is stiU not entirely clear why only a fiaction succumb while many will tolerate infection indefinitely.

Ingested cysts release motile trophozoites in the smdintestine. The factors that contribute to E.histolytica cxcystation and encystation are not well undertood, and have not been replicated in ce11 culture systems (Le- ody the trophozoite form is reiiably cultured in vitro),

However, some evidence has been presented that amebic cysteine proteases rnay in part regulate the encystation (215). Trophozoite attachent to mucin and intestinal epitheliurn cmresult in a local Uiflafllffilatoryresponse, which may be foilowed by mucosal ulceration (144). It is thought that penetration of the intestinal epithelium and lamina propna are required for symptomatic disease. Systemic infection results when trophozoites penetrate the bowel wali and enter the portal circulation where hematogenous dissemination, most commonly to the liver, occurs. In an effort to provide a mechanistic expianation for this invasive scenario, several laboratories have focused on putative virulence factors that facilitate these events. The primary candidate adhesion molecule, which likely mediates adhesion to intestinal mucin, is the galactose/N- acetylgalactosamuie lech (44). This protein is discussedat length in section 1.7. However, 0th- candidate vinilence factors have enpositedr (i) Degeneration of epithebi cells, dissolution of the basement membrane, and degradation of host immune effector molecules (IgG and IgA) are iikely mediated by extracellular cysteine proteases, several of which have been cloned, characterized, and implicated in invasive disease- (ïi) A family of pore-fomiing amphipathic peptides (amoebapores) have ken identified in E-histolytia and Unplicated in its potent cytolytic capacity.

1.4.1 Extracellular cysteine proteases

Multiple potential effects of extracellular cysteine proteases in E.histo&tica pathogenesis have been suggested- Prior to traversing the bowel wd, it is likely that these enzymes degrade host extracellular matrix (ECM) as weU as the basement membrane. Evidence for this has been presented by correlating the expression levels of the cysteine proteases with the cytopathic effect axnoebae have on mamrnahn target cell monolayers (106,107,13 8,179,183). Mutant

E. histolylica strains deficient in cysteine protease activity were diminished in their capacity to disrupt target monolayers. Specifïc cysteine but not serine protease inhibitors attenuated the cytopathic effect of trophozoites on the same monolayers. Furthemore, purifïed amebic cysteine proteases were able to mimic the amoeba-driven cytopathic effect on monolayers. While it is speculated that these proteases also degrade laminin and fibronectin, major ECM proteins found in association with the intestinal epithelium, no direct in vitro evidence has been provided.

Several reports suggest that amoeba-denved cysteine proteases may exacerbate inflammation and tissue damage associated with granuloma and abscess formation in animal models. Protease inhibitors were able to decrease tissue damage in rats that had been subcutaneously inoculated with amebic trophozoites (18)- h another study, severe combined immunodeflclency (SCID) mice, which lack both £3 and T celis due to a genetic defam hematopoietïc development, were challenged intrahepaticaiiy with trophozoites. These workers noted that preincubation of amoebae with the highiy specinc cysteine protease inhibitor, E-64 CL-~m-epoxysuccinyl- leucylarnid0-(4-~dino)butanel, was able to reduce abscess size in the liver when assessed

48hrs post-challenge (223)- Additionallyi a MAb directed against amebic cysteine protease demonstrated that this enzyme is expressed within the abscess of infected anïmals. Co-injection of amoebae and laminin also reduced abscess size, suggesting Iaminin served as a cornpetitive substrate for protease activity (132). The exiracellular location of these proteases was Mer supported by evidence that 83% of patients succumbing to invasive disease mount an antibody reponse specinc to the amebic cysteine protease (1 83). Lysis of invasive amoebae may also result in such a response and was not distinguished in this study.

Since symptomatic infection requires penetration of the intestinal mucus layer and adherence to underlying mucosa, amoebae must overcome locdued secretory IgA (sIgA) reponses in the gut. Experiments were performed where culture medium fiom which growing amoeba had been depleted was exposed to sIgA (109). In this way, degradation of sIgA was blocked specincally by

E-64 and shown to be optimal at neutral pH, suggesting the action of extracellular amebic cysteine proteases. Simila-activity has been reported against IgG, which may be relevant in Iight of recent reports that IgG can be translocated across polarized intestinal epitheiia (61). Reed and colieagues have also shown that amebic cysteine proteases have the capacity to both exacerbate and reduce the action of cornplement (1 80-1 82, 184). Cysteine proteases can specincally cleave component C3, thereby activating the altemate pathway of complement-mediated lysis. On the other hand, cysteine proteases degrade the anaphylatoxins C3a and CSa, which have been associated with pro-iaflammatory ecects, including neutrophil chemotaxk, macrophage activation, enhanced vascular pemieability, smooth muscle contraction, and release of IL-1,IL-6, and IL-8 (1 8 1). While these results are interesting, a strict correlation of complement-resistance with the action of cysteine proteases has not been shown for E.histoIytica in any animal mode1 to date,

Attempts to geneticaüy isolate and sequence the neutral extracellular cysteine proteases of

E- histoiytica have resulted in the idenfication of at least six conserved homologues (tenned

"acp"). Gene cloning was achieved by polymerase chah reaction (PCR) ampüncation using primers based on the conserved active site of known eukaryotic cysteine proteases (3 1, 179).

E. histoi'ytica cysteine proteases share sequence identity for cntical active site residues with the papain sub-family of cysteine proteases which ail contain the GLU-Arg-Phe-Asn-Iso-Asn amino acid consensus motif. Aithough cathepsin B does not contain this motif, the amebic cysteine proteases appear to have similai substrate specifkity to this verterbrate hctional homologue rather than cathepsin L which does contain the Glu-Arg-Phe-Asn-Iso-Asn motif(106). Lnitial studies had suggested that E. histowca retained cysteine proteases which were absent in

E-dispar, thereby contributing to its increased virulence, but this result has been refuted by two separate groups (32, 153, 179). It is possible, however, that E-hisrolytfca expresses these proteolytic enzymes to a greater degree tban E-dispar. This is difficult to show ni viiro because

E. histolytica and E. dispar culture conditions are usually different, with E-dispar ofîen relying on

CO-cultivationwith bacteria However, the recent developement of axenic (no associated bacteria) culture conditions for E-histolytia and E-diqar should make it possible to examine this question

(48, 113)- 1.4.2 Amoebapores: a famiïy of pore4orming peptides

Cytolysis of mammalian cells by E-histolyticatrophozoites is a rapid event initiated by intimate contact of effector and target cell. Within a matter of minutes, sweliing and widespread surface blebbing occurs in the target. WhiIe initial adherence is likely mediated by the

E. histolytica galactoselN-acetylgalactosaminelectin (discussed in section 1.7), proteinaceous factors are released which compromise the mget membrane. Failure to exclude dyes such as propidium iodide clearly demonstrated target celi de&, Moreover, the killing capacity of amoebae is not restricted to a single celi type,

The first mechanistic evidence for the cytolytic event was the discovery that amoeba are capable of inducing a pore in the lipid bilayer of the target, which resulted in depolarization of the ce11 (1 39,255). Subsequent protein purification and genetic cloning revealed that this protein was a pore-forming peptide, comprised of at least three family members (126, 128). Termed

"Amoebapore" A, B and C, these genes encode a 77 amino acid mature peptide with 6 cysteine residues and 1 histidine residue in conserved positions. The three isoforms retain 3547% oved identity based on arnino acid sequenchg and mas spectroscopy (125). A single trophozoite was shown to express al1 three ameoebapores in a ratio of 35(A): IO@): l(C) (128). Amoebapores share significant sequence identity (25-30% based on exact amino acid identity and 4540% based on functional amino acid identity) with NK-lysin fkom mammalian natural killer (NK) and

T cells (122). Structural analysis using circular dicroism spectroscopy indicated that amoebapore

A has a compact a-helical structure. Secondary structure prediction has demonstrated that the polypeptide chain folds into four, principaily amphipathic, helicai elements connected by 3 amino acid loops (128). The amoebapore is also rehctory to proteolytk cleavage (except for the loops) and denaturation, and retains no £ketbiol groups, These data suggested that thiï amphipathic pore-fonning peptide was highly compact and sTable, consistent wîth its 4u-heloc bundle predicted stucture- A similar predicted structure was obtained for NK-lysin using the same experimental approach, and is unlike other cytolytic peptides (54)-

In order to meet the task of being soluble upon release hmthe amoeba and then buqhg itseLf in the target celI membrane, investigators have suggested that ameobapore conforms to the

"barrel-stave" model (123). The model proposes that pore-forming peptide monomers oligomerize upon membrane insertion, thereby formïng a -water-filled channel through which smalI molecules cmpass. Chernical cross-linhg studies demonstrated that amoebapore peptides oligomerize to form larger complexes through specinc peptide-peptide interactions (126). This iikely results in chamel formation in the target celi membrane that uItïmately disrupts the osmotic potential of the celi. Evidence to this effect was presented in the form of synthetic peptides derived from amoebapore. Biological activity as measured by permeathg of target cells showed that putative helk 3 (amino acids 50-63) had the highest membmnolytic activity (124).

Modification of specific cationic residues within amoebapore aected binding to phospholipid vesicles (lysine) or oligomerization of the peptide (histidine) when membrane-associated. Target specificity studies of amoebapore have shown that these peptides have a wide cellular range, with activity against granulocytes @do), T lymphocytes (Jurkat), and monocytes (U-93 7) (1 23).

Resistance of E.histo[yrica to amoebapore activity may be due to the presence of ceramide in the trophozoite plasma membrane(8). Although phagocytosis of bacteria have not been directly related to pathogenesis of E-histoZytica, amoeba actively phagocytose and lyse bactena at a high rate (1000 bacteria per hour) (152). Bacterial lysis within phagocytic vacuoles is likely the pnmary role of amoebapores. This was supported by the observation that non-pathogenic but bacteria-consuming E-dispar also retains and expresses near identicai homobgues to

E. histolytica (123). Contact with host cells probabIy tnggers granuiar exocytosis that results in cellular lysis, rather than spontaneous release as observed for cysteine proteases (127). Whether host ce11 killiag is a chance event is not chsince there îs no seIective advantage for this to occur, although there might be a selective advantage to periodic cytoiysis ofhost immune effector ceiis in order to pemst in mucosa Recent data suggest that E-histolyrcu trophozoites initiate caspase-dependent apoptosis in target neutropids upon intimate contact (C. Huston, personal communication). This iraplies that either pore-forming peptides and/or a yet-to-be- identified receptor on the trophozoite surface may trigger the apoptotic pathway on target ho* cells. Considerable debate exists in the fiterature concedg the possibility of acqeg

irnmunity to E. hisroZyrrCa colonization and subsequent invasive disease- On the one hand, studies have reported the recurrence of intestinal amebiasis, and an elevated iikeIihood of infection with

increasing age, both suggesting the absence of acquired Tmmuni~(129,236). On the other han& a large (aibeit uncontrolled) study pedormed in Mexico City demonstrated a recurrence rate of

0-29%among patients (1,021) who had prior amebic liver abscess, a much Iower rate than in the general population (55). Furthemore, immunosuppressive states, including unrelated corticosteroid therapy, pregnancy, inhented immunodeficiency, and malnutrition, have been associated with exacerbated invasive amebiasis (3,19 1,228). Invasive amebic infection provokes both humoral and ceil-mediated immune responses that are subject to manipulation by the parasite in order to prolong colonization.

As mentioned in section 1.1 ,3,8 1-1 00% of patients with invasive intestinal or hepatic disease develop systemic IgG antibodies specific to E. histolytica (103, 1 1 1, 114, 116, 160,235).

These circulating dbodyresponses can persist fiom 2-1 1 years. Similady, secretory IgA responses have been transiently observed with invasive disease in stool and saliva (1 43, 108,

2 14). Animais irnrnunized with various amebic hctions can be protected against hepatic

challenge with virulent trophozoites, but protection does not correlate with anti-E-histoljtica

titres (83). However, a role for anti-amebic antibodies cannot be completely dedout. Seydel and

colleagues passively transferred human anti-amebic antisera 24 hours pior to intrahepatic

challenge of SCID mice with vident trophozoites, and found that abscesses were up to 24.5%

smaller than in controi animais (2 13). Amebic evasion of humoral responses has been demonstrated in vitro, where trophozoites are actively able to aggregate and shed antibodies that attach to their surface (12,36). As mentioned earlier, it has been shown that cysteine proteases secreted by the trophozoites have the capacity to degrade both anti-amebic sIgA and IgG (109).

Taken together, anti-amebk antibodies appear to be a sensitive marker for exÏsting or previous invasive disease rather thau a sign of immunoprotection in human disease,

Complement-mediated lysis of E- histolytica trophozoites has been demonstrated in infected patients (97,157). In the hamster model, complement depletion pnor to intrahepatic challenge of trophozoites results in greater susceptibility to abscess than in undepleted controls

(79). While there is some controversy over whether pathogenic isolates are always resistant to complement, it is clear that E-histolytica has the capacity to modulate the effects of both altemate and classic complement activation (1 80-1 82, 184, L85).

As early as 1964, the observation was made that E-histolytica has the potential to phagocytose huma.polymorphonuclear neutrophils (PMNs), and not the reverse (10 1)- More recent evidence has supported the toxic effects of virulent strains of E-histolyrica on neutrophils in well-defined medium (199). Killing of neutrophils was contact-dependent and inhibited with

N-acetyl-D-galactosamine,suggesting a role for the amebic le& (177) (discussed in section

1-7). Suice neutrophils are most likely the first cells recruited to the site of trophozoite ~lîration at the intestinal mucosa or liver, degradation of neutrophils may contribute to the tissue damage seen with invasive amebiasis (43,238). Release of pro-inflammatory cytokines by intestinal epithelium might trigger neutrophil recruitment upon contact with amebic trophozoites (67). Co- incubation of hepatocytes and trophozoites results in greater damage of liver ceil monolayers in the presence of neutrophils, possibly due to release of toxk vacuolar enzymes fiom neutrophils

(200). The neutrophils by themselves had no effects on the hepatocytes in these experiments. The role of neutrophils in host defense has remained puzzIing since these immune ceils appear to contribute to disease in most studies-

However, several experiments pefiomed by Ghadikïmin the early 1980s (at the hstitute for

Parasitology at McGiU University) have demonstrated a generalized role for cellular immun@.

Relying on the hamster model of hepatic amebiasis, these workers showed that corticosteroid treatment, neonatal thymectomy, splenectomy, y -atÏon, silica therapy, anti'body-mediated macrophage or Lymphocyte attenuation resdted in exacerbated formation of Iiver abscess following trophozoite challenge (79-84). These data suggested that celi-mediated immunity plays a role in host resistance. This conclusion is supported by hamster immunization studies where amebicidal peripherd blood Lymphocytes (PBLs), cytotoxic T lymphocytes (CTLs), and peritoneal macrophages were identined in immunized animals alone. Using a different model, certain investigators found that macrophages limit the extent of metastasis in a nude

(thymectomized) mouse model of amebiasis (226). Antibody dependent cellular cytotoxicity

(ADCC) by human monocytes of erythrocytes coated with amebic antigens in the presence of immune serum has also been demonstrated (225). Sharma demonstrated that monocyte-derived macrophages, activated with concanavalin A (Con.)- or phytohemaggiutinin (PHA)-eiicited cytokines such as interferon-gamma ([FN-y), were able to kill vident amoebae at effector to target ratios of 10: 1 (198). Killïng was not elevated with exposure to immune serum, suggesting that ADCC was not the mechanisrn of cytotoxicity. Phase contrast microscopy relying on

Nomarsky optics revealed that killing was contact-dependent and inhibited when cells were resuspended in 10% dextran. Enzymatic depletion of hydrogen peroxide eliminated amebicidal activity of macrophages, suggesting that killing was oxidative in nature. CTL activity could be elicited fiom purifîed lymphocytes agaEl histolytica following stimulation with PHA over 18 hours (196). Guerrero demonstrated that patients who had been cured of amebic liver abscess by anti'biotics possessed lymphocytes that were spontaneously cytotoxic to amebic trophozoites (reviewed in 202). However, other workers have suggested that in viim stimulation of Lymphocytes with soluble amebic antigen (SAA) is required for cytotoxicity (197). In addition to prllning lymphocytes for killing, they found that SAA can induce proliferation of T lymphocytes in healthy uninfected individuals in the absence of LPS

(20 1). Maximal lymphocyte proliferation occurred in response to 100 pg of amebic proteid mL for 5-7 days in the presence of monocytes. However, proliferation was significantly Iess when compared to ConA induction. More recentiy, the mitogenic properties of SAA were apparently locaüzed to the amebic galactose/N-acetyl-D-galactosamine(WGaiNAc) dace1ecti.n.

Proliferation was inhibited by asialofetuin (terminal GaNAc-containhg glycoprotein) but not fetuin (terminal sialic acid). However, this result diBiers fiom previous studies suggesthg that

SAA did not elicit blastogenesis of naïve Lymphocytes (78,94,210).

In order to survive potentiaily amebicidal encounters with macrophages and T cells,

E- histolytica has adapted evasive manoeuvra. Ortiz-Ortiz had suggested early on that survival of amebic trophozoites was likely favoured by a transient state of immunosuppression associated with hepatic infection (1 58). This state can be induced by corticosteroid therapy as aiready mentioned, but there is some evidence that the amoeba themselves downregulate both macrophage and T ce11 effector fimctions. Chadee and colleagues have recently described the role of activated macrophages in amebicidal activity (36). In addition to IFN-y, -or necrosis factor- alpha (TNF-a)and colony stimdating factor4 (CSF-1) are capable of activating macrophages for amebicidal activity (58, 133). Killing was shown to have oxidative and non-oxidative propertïes- Nitrïc oxide (NO)was the principal effector molecuie against E.hr'sroCytica, followlng upregulation of inducible NO synthase (NOS)gene expression, The requirernent for macrophage activation suggests that a Thl-type T ceU response wouid be prefemed for protection.

Several investigators have shown that macrophages located adjacent to the abscess are suppressed in cornparison to peritoneal and splenic macrophages, suggesting localized suppression (59). It appears that E.histolyhCa is responsible for this down-regdation by several means. Macrophage antigen presentation is suppressed by downreguiating EN-y-induced major histocompatibility complex II (MHC II) expression (247). Pre-treatment of naïve murine macrophages with SAA (soluble amebic antigens) reduces IFN-y-induced Ia expression on their ce11 surface by dllninishïng Iap transcription. This effect appears to be mediated by SAA-induced production of arachidonic acid metabolites prostaglandin Et (PGE2) observed both in vitro and in animal models (246). There is also some evidence that amoebae retain a homologue to monocyte locomotion-inhibitory factor (MLIF) which upregulates CAMP in human macrophages, possibly leading to reduced MKC II expression (187).

Low levels of TNFa favour grandoma formation -- a hospitable environment for trophozoites - while high TNF-a promotes macrophage killing of E-histol'ytica (135). The growth of abscesses in the gerbil mode1 appears to be consistent with low levels of TNF-a production by abscess-derived macrophages (AMO) (249). Amoebae may antagonize (by upregulating PGE2) TNF-aproduction by abscess macrophages (212,248,249). This may attenuate upregulation of reactive oxygen intermediates (ROI), which would otherwise serve as

CO-factorsin NO-mediated killing of amoebae by these macrophages (134). In addition, AMOS are hyporesponsive to PMA-triggered oxidative bursts when comparai to distal macrophages, suggesthg localized cihinuition of AMû activity by E. histolyrica (59). Since passive transfer of immunity to hepatic abscess formation in hamsters is abolished

when T ceils are depleted, It is thought that dgen-specXc T ce11 activatioa, cytokine production, and effector ceii activity may play a key role Ïn immunity (84)- T ceiis hyporesponsive to in w-PO stimulation with mitogens @hytohemaggfutinÏnand Con A) or amebic antigen are found in patients and experirnentai anmials wîtb hepatic amebiasis as compared to uninfected controls (217,241)- Kyporesponsiveness can be induced in naïve splenocytes in vi'o by exposure to SAA (42). In addition, overail T ceus nmbers are reduced in patients infected with E. histolytca, with a skew in population due to reduced CD~+and elevated CD~+cells (77,

197). Reduction in the CD~+population may be the cause for reduced antigen-specific T ceil responsiveness and reduced macrophage activation. GlycosyIation of proteins is ubiquitous: se- extracellularmatrix @CM), membrane- spanning and intraceMar proteins are known to exist as glycoproteins, The biosynthetic ceiluiar apparatus responsibk for assembhg carbohydrate structures relies on glycosyItransferases and glycosidases that have specinc substrate requirernents. Many sugar-rnod@hg enzymes exïst allowing for a diversity of carbohydrate structures, It is thought that celiular messages are encoded within the various linkages of sugar groups. Proteins that specifïcally decipher messages encoded by these sugar groups have been well described, Lectins are considered to be non- enqmatic proteins which selectively bind to specific carbohydrate structures. Structural studies allow us to organize known animal lectins into two broad categones, although others exist that do not conform to this paradigm. The fitst group (I)of lectins is comprised of proteins containing a buried ligand-binding site that redts in engulfhent of the sugar. The second group 0 - including animal C-type and S-type Iectins, bacterial toxins, and virai proteins - has a shaiiow sugar-binding site on the protein surface. This overview will focus on group II lectins since these proteins are more closely aligned with amebic lectin experiments presented in Iater Chapters.

1.6.1 Classification and structural information

Carbohydrates intetact with lectins through hydrogen bonds, divalent metal ion coordination, and van der Waals and hydrophobie interactions (69). The abundance of hydroxyl groups on monosaccharides make sugars ideai for hydrogen-bonding with critical residues in lectins which Table 1.6.1 :Cornparison of features which distinguish C-type and S-type lectins (65)

Pro~ert~ C-Type Lectins S-Type Lectins (Galectins)

ca2+Requirement YeS No

Cysteine residues ciidfides fiee SUlfbydryIs

Solubiiity variable bsersoluble

Location extracellular Intra- and extracellular

Carbohydrate various Mostly f3-galactosides

Specificity appear to be conserved- Hydroxyl goups serve as both hydrogen bond donors and acceptors with side chain atoms such as the carbony1 oxygen in the presence and absence of water- DLvalent cations can alter the shape of the binding site or directly coordinate with the sugar (250).

Hydrophobie interactions InvoLve aromatic (non-polar) side chaias and the B face of the hexose ring which has charged ali'phatic (polar) protons. The C-type (ca2%iependent) animal lectin can be readily distinguished fiom the S-type (thiol-dependent) animai Iectins, The canonicai C-type lectin is the mammalian asialogiycoprotein@epatic lectin) receptor, responsible for clearance of desialylated glycoproteins fkom the circulation. Receptor-bound asialoglycoprotein is endocytosed and transported to intracellular lysosomes for rapid degradation of ligand. Structural studies of hepatic lectins from rat, human, chicken as well as other species have been instructive in identwgcommon motifs (21,63,9 1)- Each hepatic lectin contains a transmembrane domain composed of 25 amino acids and an N-terminal cytoplasmic domain. The remaining polypeptide is extracellular and may oligomerize with adjacent identicai subunits. Partial proteokysis that generates a C-terminal fragment, foilowed by selective binding to carbohydrate affinity columns, was used to ascribe the carbohydmte recognition domain (CRD of -15-kDa) (46). Comparison of homologous hepatic lectins suggested conservation of 18 invariant residues required for carbohydrate binding, but divergence of selected residues responsible for ensuring sugar specificity (65). Several key cysteine residues are conserved within the CRD which Likely form disulfide bonds. CRD binding of Ligand requires the presence of ca2* and a pH greater than 6.5.

Homologous C-type lectin CRDs have also ken identified in invertebrates such as the sea urchin and fkom the hemolymph of Sarcophaga peregrina (85,23 1). Other C-type lectins include the selectins, macrophage mannose receptor, and the soluble collectins (66, 118). A flurry of recent structural analyses on C-type lectins have dernonstrated both convergent and divergent evolution (64). Convergent evolution of folding topology fiom a wide-variety of proteins such as endostath, CD44, and bacterial intimin has been observe& none of which share ligand specificity with the classical C-type Iectin mannose-binding protein. Carbohydrate recognition domains (CRDs) are therefore a sub-family of moddes withïn a larger superfâmily of C-type lectin-like domains (CTLDs).

The second broad group of shallow-pocket lectins is termed S-type animai lectins or galectins. The fiuidamental merence £tom C-type lectins is the Iack of requirement for divalent cations in sugar-binding (15)- In addition, S-type lectins usually retain thiol groups in reduced form and the lectin itselfis usually soluble, with both intracellular and extracellular IocalIzation.

Primary sugar specScity is commoniy for fLgaIactosides. Cornparison of dl homologous S-type lectins has demonstrated sequence conservation within the CRD (also - 15-kDa) (87). Absolute conservation of His45, Asn47, and Arg49 has kendemonstrated. These residues are thought to hydrogen bond with the C4 hydroxyl group of galactose which is axial in position. The C6 hydroxyl group of galactose also participates in hydrogen-bonding, whereas the C3 group is not

(and is stencally available for substitution) (16). Ia addition, the C3, C4, CS, and C6 carbon atoms of gdactose form a plane with a weIi-defhed spatial relationship with an arornatic side chain of the lectin (190). In contrast, the C4 -OH of mannose and glucose is equatonal in position and is thought to be a key discnminator fiom glucose/mannose-binding lectins such as

Lathyms ochrus lectin (24). SurprisingIy, the primary specificity of lectins does not always detemiine overall selectivity. For example, the Gnxonia simpZzf?coZia lectin requires galactose to bind the Lewis b blood group oligosaccharide but does not show specificity for aqparticdar monosaccharide (190). Oxidation of the S-type lectin renders them unable to bind sugar likely due to the loss of fiee s&ydryl groups adjacent to the CRD. The spatiai arrangement of the lectin domains Iargely detemiines the bction of this protein- in the fist casey Iectin subunits cmcontain several CRDs whiçh combme to provide hi@-afbïty binding of the Ligand, called "subsite rnuitivaiencf' (190). The macrophage mannose receptor

(MMR) recognizes carbohydrate ceil dacecomponents of pathogenic microbes, A single polypeptide of MMR contains 8 C-type lectin CRDs which in concert buid long chab polysaccharides present on the microorganism dace. h the second case, such as that observed for asialoglycoprotein receptor, each monomer contaios a nngle CRD that associates in the membrane to form oligomers, termed ccsubuoitmultivaIency". Experiments with branched oligosaccharïdes have demonstrated that two different CRDs of asialogiycoprotein receptor fom a specific spatial orgamanrzationthat binds to a sùigle arm of an N-linked oligosaccharide (121).

Another group ofweli-characterized mammalian lectins9the selectins, likely mediate low-afkïty interactions by which leukocytes roll on endothefim. Consistent with this, selectins apparently do not oligomerke, eliminating the possibility of multivalency conferring hi&-affhity interaction. Aside fiom multiple CRDs and lectin oligomenzation, it is dso possible that recruitment of Lectin to the ceii sdace by upregulation of lectin gene expression andor capping of lectin on the surface through cytoskeletai changes can provide high receptor density and appropriate spatial arrangement to ailow foc high-aflkity multivalency.

1.6.2 Lectins and microbial virulence

Bacterial toxins are among the bat-characterized rnicrobial lectins. Both Escherichia coli heat-labile enterotoxin (LT) and Vibrio cholera toxin (CT), which are composed of a cytotoxic A subunit coupled to 5 identicd B subunits each containing a CRD, have been stmcturally resolved with their respective carbohydrate ligands (1 5 1,218). Cornparison of the x-ray crystal structures of the LT-lactose and CT-pentasaccharïde complex show similarities with respect to termuial galactose-CRD interaction. CT has the added feature of a non-btïmate interaction with sÎaiÏc acid through 2 hydrogen bonds and weak van der Waals interaction. The dual specincity of CT for terminal galactose and sialic acid is a good example of how subsïte multivalency can also lead to overd specïfïcity. Pertussis toxin (PT) is heteropentamenc (with 4 distinct submits, one of which is repeated) but retains structurai homology to the CT fdymembers (224). lanuenza virus hemagglurinin (HA) is required for virai attachment to host ceils as weli as membrance fusion (25 1). HA is a trimer composed of triple-heiical coiled coils that Lie beneath globular CRD domains that bind sialic acid (252)-

Parasitic protozoa aIso rely on lectin-like interactions for host ceU attachment and colonization in the smali intestine (100). Little structural data exist for parasite lectins, although initial studies suggest some sunilarites to C-type and S-type animal lectins descnied earlier. The lumenal parasite Giardia lamblia has a membrane-associated lecth termed taglin (28/30-kDa isoforms), which has specificity for temuaal phosphomannosyl residues. As observed for C-type lectins, taglin requires pH > 6.5 and divalent cations for carbohydrate binding (74). Trichomonas vaginafisuses a N-afetyl glucosaminefmannose-specific lectin in mediating its cytopathic effect on target ceIl monolayers (1 94). Promastigotes of Lesmania brmiliensis possess a N- acetylglucosamine-specific lectin which may play a role in adherence to macrophage-like cells

(96). Sirnilarly, the merozoite stage of PZasmod~fakiparum expresses lectin-like surface molecules which mediate adherence to erythrocyte sugar groups, including N-acetylglucosamine,

N-acetylgala~tosamine~and N-acetylneuraminioc acid (102). Once iafected with P.faZc@arurn, parasitized red blood ceils form rosettes with uninfected ceils, relying on more promiscuous lectin-sugar interactions (38). This may explain, in part, the pathogenesis of cerebral malaria wbich occurs due to occfusion of microvascuIature as a result of erythrocyyte rosetting-

The amebic ceil surface lectin was nrst implicated in adherence and contact-dependent cytolysis ofhuman cells by Ravdïn and Guerrant In 1981- Centrifllgation bindïng assays performed at 4 OC have demonstrated that amoebae interact with human colonic mucin, human colonic epitheliurn, human neutrophils, and erythrocytes in a galactose- and lG acetylgalactosamine-inhibitable manner. Assays are performed at 4 OCin order to isolate the adherence properties of the amoeba since trophozoites wiil lyse most eukaryotic cells at physiologic temperatures. Chadee and workers demonstrated that purified native lectin bound human coIonic mucin with very high aff?nity (8.2 x 10'~'MI). Mucin fikely represents the pnmary interface of amoeba and human colon during colonization following uifection.

1.7.1 Sugar specificity of the E.kbtoIytica Iectin

Elegant studies determinuig the sugar-specificity of the amoeba and target cens relied on

Chinese Hamster Ovary (CHO) ce11 glycosyIation pathway mutants (1 30, 13 1, 178). The striking obsewation was made that a particuiar CHO cell line (ldlD.Lecl), which lacks N- and O-linked galactose and N-acetylgalactosamine sugars, adhered to amoeba 95% less than wild type CHO cells. Data fiom several CHO ce11 Iines lacking various sugars is summarized in Table 1.7.1.

Interestingly, cytotoxicity was also diminished by 90%, suggestuig that lectin-sugar interactions Table 1 -7.1 Chinese Hamster Ovary ceil mutants exhibit WGaiNAc-specinc adherence and cytotoxicity to arnoeba in centrifugation binding assays (1 3 l), (1 3 0) (1 78).

CHO ceU line Terminal Sugar Adherence (4OC) Cytotoxkity (3 7OC)

N-linked O-Iinked

Parent Gat-2 Sialic Sialic 65 +,2

acid acid

Lec I Mannose Sialic

acid

Mannose None

Lec 2 Galactose Galactose 92+ 1 58k3 are necessary for contact-dependent küiïng oftarget celis in this assay. Cytotoxicity =fers hem to the artificid apposition of aawebae and 51~r-~oadedtarget CHO cei&fofiowed by Lysis of the target ce11 which aiiows measurement of5'cr release Further evidence that cytotoxicity was contact-dependent came by way centrifugation assays pesonned in 10% dextran, which prevented appositon of amoebae and CHO ceïis, and redted in signincant inhibition of adherence and cytotoxicity (175).

Carbohydrate recognition by the lectin was merelucidated through the use neoglycoproteins engineered to express multiple sugar groups coupled to bovine serum albumin

(BSA). Membrane preparations of E.histoZjtr'ca were made and exposed to various radioactively- labefled mono- and oligosaccharides, as weil as neoglycoproteins in order to measure their affïnity for native lectin. Here the striking remit was that polyvalent neoglycoproteins

(G~NAC~~BSA)were >100,000-fold more potent than simple sugars (GalNAc), strongly suggesting that the lectin relies on either subsite or subunit multivalency in order to aord hi&- affinity binding. Gel filtration analysis of the native lectin using a Sephacryl S-300 columns resuited in elution of the protein as a broad peak of approximately 660-Da, suggesting the lech oligomerizes to possibly allow subunit multivalency (162). Additionally, data presented in

Chapter 2 lend greater weight to the possibility that several CRDs reside within a single lectin molecule. N-acetylgalactosamine is favoured over galactose by a factor of 7 when presented as a monosacchat-ide, and not mannose or glucose. Lactose (galactose-P 1-4-glucose) and N- acety llactosamine (wacetylgalactose-P 1-4-glucose) were not of significantly higher affinity than the monosaccharides galactose or N-acetylgalactosamlne atone, indicating that the amebic lectin likely binds to a single (terminal) sugar. Apparently, lectin binding to galactose or N- Table 1 .%îCarbohydrate specificty ofthe E-h'sfo&ticaGal/GmAc Iectin (6)

Saccharide E.?zisrulj&5ëa GaVWActectln Asialoglycoprotein

(~CSOYP) @c~~~P)

Monosaccharide

Ga1 2,500 10,000

GaiNAc 710 275

p-NP-PGaINAc 37 LOO

Man >250,000 >100,000

Glc >250,000 >200,000

Disaccharide

Lactose 710 6,000 N-acetyiiactosaniine 480 -

Synthetic Conjugate

NAc-YD(G-GalNAce >1,600 0.003

Tris-aminohexyl-lactoside 110 4

Tyr-Asp(GLy3aminohexyt-lac)l 130 11

NeoGlycoprotein

GaoBSA 0.15 0.007

G~NAc~~BSA 0.005 0.0006

GlcslBSA >11 0.0053

FucZ8BSA 2.7 >8.9 acetylgaiactosamine is improved in the presence of an aromatic group (such as p-nïtrophenyl),

reminiscent of C-type lectins. A useW compatison is made with the C-type gaIa~tose/~

acetylgaiactosamine lectin asidogiycoprotein in Table 1.7.2. While both lectins share many

features in common, the amebic Iectin has - based on its selective binding to srnail synthetic conjugates -- different stereospecincity to asialoglycoprotein-

1.7.2 Structure of the E.hiStoCjîka GaVGalNAc lectin:

Despite commonalities with the C-type Eectin, asialogiycoprotein, the amebic lectin does not contain prirnary sequence compatible with well-described CRDs- The amebic Iectin is a 260-kDa heterodimer, composed of a 170-kDa heavy subunit and a 3 1/3S-kDa light subunit which is glycosylphosphatidyiinositol (GPQ-anchored into the dacemembrane bilayer (Figure 1-7.1).

The native lectin was purifïed by both galactose-affhity chromatogmphy and demonstrated to be the lectin by: (i) galactose-specific CHO ce11 binding; (ii) identification by MAb antibodies subsequently generated against the recombinant heavy subunit; (iii) bctional studies showhg that anti-native lectin antisera, but not pre-immune sera, blocked CHO ceIl binding. At least five genes encode the heavy subunit, three of which have been Wysequenced. The predicted amino acid sequences of hgZ1 and hg13 are 95.2% identical, and share 89.2% and 89.4% identity to hgZ2, respectively. The heavy subunit has severai interesthg features: (î) The amino terminal (residues

1- 1 87) has an abundance of cysteuie (3 -2%)and tryptophan (2.1%) residues. (ii) Amino acids Figure 1.7.1 Subunit structure of the E-lrzktolyticaGaVGaiNAc lectin

pseudorepeat cysteine-rie h

SH2 binding motif

CD59 homology

Cytoplasm 187-378 are characterized by the absence of cysteines. (K) A repeating region of -30 amino

acids was identified between residues 379-654. (iv) Residues 379-1209 have an extnrordinariiy

high number of cysteines (97 or 10.8%) and a domain which bars significant identity to the

mammalian CDS9 protein associated with cornpiement resistance. (v) Hydrophobicity plots

suggest residues 1210-1235 likely constitute the single membrane-spanning region. (ii)The

remaining carboxy-terminal residues (1 236-1 277) contain an epidermal growth factor receptor

(EGFR)-like tyrosine phosphorylation motif (Am-Glu-Am-Ma-Glu-Tyr). (vii) Potential N-

Iinked glycosylation sites (16 in hgll and hgZ3,9 in hgZ2) are present throughout the protek

The small subunit has a calculated molecuiar weight of 32-kDa with hydrophobie amino- and

carboxy-terminal signal sequences (147, 148,234). Both IgZI and Zgf2 (80.1% identical) contain

putative N-glycosylation sites and possess no simiifTcant homologies with other known proteins.

However, a putative GPI anchor cleavage or addition site was identined. The Light subunit exists

as either a 3 1-kDa GPI-anchored or 35-kDa non-anchored isofonn on the native lectin of

amoeba Nitrous acid deamination of the [3~palmitate-labelled3 1-kDa, but not the 35-kDa,

isomer resulted in acyl-phosphatidyl inositol release, supporting the presence of a GPI anchor

(147). The disparïty in molecuiar weight for the native light subunits likely arises fiom variation

in lipidation and glycosylation. It is unusud that the 3 1-kDa isomer contains both a putative transmembrane domain and a GPI anchor, and has not been reported in the Iiterature previously.

As alluded to earlier, none of the genes encoding either heavy or light subunit in E-histolytica

bears homology to C-type or S-type lectins.

Severai monoclonal antibodies (MAbs) have been generated against epitopes resident on the heavy subunit. Interestingly, while some MAbs block adherence of amoeba to CHO cells, others enhance adherence. Some but not aU anti-heavy sub-t MAbs have the potentid to also block complernent resistance and cytolysis of target cells (listed in TabIe 1.8.3). These data suggest, albeit indirectly, that the heavy subunit passesses the hctional CRD because anti-heavy subunit antibodies can block adherence, cytotoxinty, as well as complement resistance. MAbs directed against the Light subunit have no effect on adherence, cytotoxkity, or complement resistance mediated the native lectin (149). Two mAbs (3F4 and 8A3) generated against the170-kDa subunit enhance adherence of trophozoites to target host cens- 3F4 and 8A3 were only 2 of 7 mAbs (see Table 1.7.3) that detected a putative E-dispar lectin homologue by radioimmunoassay-

These authors concluded that E-dispar infection may therefore elicit IgG responses that predispose an individual to more facile colonkation by E-histoiytica. This hypothesis has wt been corroborated in clinical epidemiological surveys. Moreover, it is unlikely that fiilminant amebic infection (leading to liver abscess) confers any selective advantage to E-histolylica since cyst passage by the host is the only requirement for propagation of this organism.

1-73Complement resistance conferred by the E.h&o&tica lectin:

The mechanism of cornpiement resistance (conferred by the E. histolytica lectin) was demonstrated by Petri and colleagues (26). Im.munoafThity-purified iectin bound components C8 and C9 of the complement system and conferred resistance to C5bC9 lysis of complement sensitized-amoebae. hterestingly, pre-incubation of the lectin with Ga1 enhanced complement- mediated lysis of amoeba. These authors argued that lectin CRD occupancy prevented the CD59- like domain binding CS and C9.

The two complement pathways, classical and alternative, contain 25 separate activation and regdatory molecules and at least 7 complement receptors. A well-characterized sequence of enzymatic cleavage*prote&pmtein interaction, and conforrnatlonal change occurs in complement activation, the details of which are not discussed here. However, perturbation of compIement activation by parasites can occur at diffexent stages dong the pathway. Cl bindïng and activation, C3 enzymatic regulation, control of alternative pathway activation, C3 receptos,

CSb-C9 and poly-C9 formation (as seen for E-histo&tica Lech), and complement regulation by molecules such as decay accelerating factors PAF) and homologous restriction factor (HRF) are all targets for parasite evasion. h has been demonstrated that Schistosoma MQ~ISO~and

Trypanosoma bnrcei cm bind CLq, which may facilitate host cell attachent and invasion, resulting in parasite avoidance of a host-protective role for C 1 Înthe classical pathway (1 55,

206). T~anosomamui trypomastigotes are capable of binding to Clq, as weii as whole C 1, and subsequently degrade these proteins in the presence or absence of C 1 inhibitor (188,189).

Shedding of complement components cmrender the parasite inert to the lytic cascade. For exarnple, S mansoni schistosomulae bind C3 on their surface but subsequently lose these molecules fiom theu outer membrane (205). Molecular mimicry of proteins that regdate complement activation bas also been demonstrated in T.cmzi. The C3b fragment is deposited inefficiently on trypomastigotes and when deposited is cleaved to iC3b, which prevents CS convertase and CSb-C9 membrane attack complex formation (188)- While the inefficient deposition is due to shedding, the cleavage to iC3b has been amibuted to a parasite protein that has fiuictional homology to the negative regulator of the complement, DAF. In addition, unrelated to the E, histolytica Iectin-mediated complement evasion, Reed and colleagues have demonstrated that an amebic cysteine protease is able to cleave C3, rendering it inactive (184).

Certain parasites are also capable of inhibithg the membrane attack complex (MAC) derthe lytic components are deposited on the surface. This is seen with Leishmania donovrmi promastigotes where rapid cleavage of C3b to iC3b occurs on the parasite dacedue to

Leihania-derived proteolytic cleavage (72)- Fiiy, stable insertion of a non-lytic C9 on the microbial dacemay resuit from parasite proteins which bind other components of the MAC such as C8, as seen for the E-hr'stol'yficuIectin (26). This was demonstrated dlrectly by showing that '2S~-labelIedlectin bound C8 and C9,but not CSb6 and C7,transfeued on to PVDF membranes.

1.7.4 E-hiktolytica lectin: immunodominance, antigenic conservation, and vaccinology

Seroepidemiological surveys have demonstrated that the arnebic lectin is an immunodominant molecule, Sera obtained fkom patients with amebic berabscess in disparate regions such as South mca, Egypt, and Mexico aii recopize the lectin pdedfiom the standard laboratory strain HM1:IMSS (1, 164). This is not surprising in üght of the remarkable lectin gene conservation across E-histolyfica isoiates (1 72). The conservation of the lectin and its imrnunodominmce in human infection have sparked two potential applications: (i) the use of anti-Iectin MAbs in stool antigen ELISA detection kits in order to distinguish E-histolytica fiom

E-dispar (93) and (iï) the use of serologicd ELISA to determine ifa patient is infected with an invasive strain of E-histolyrica (as opposed to non-invasive E-dispar)(1).

The amebic lectin is also useful vaccine candidate for several reasons: (i) The lech appears to play a crucial role in the pathogenesis of invasive amebiasis since it likely mediates Table 1-7-3Features ascriid to anti-16monoclonal antiiesgenerated against the heavy subunit of the GaL/GaNAc lectin of E-tristolyfica

Epitope Isotype Effect on Effect on Effect on ~eactivity~ LocationC

Ad.erencea cytotoxici$ C5b-C9 P NP WCg

3F4 IgGl Increases Decreases No effect + + 482-998 (C)

8A3 IgG 1 hcreases No effect Decreases + + 482-998 (L)

7F4 IgG2b No effect No effect Decreases + - 482-1 138 (C)

8C12 IgGl Inhibits Inhibits Demases + - 895-998 (L)

1G7 IgG2b Inhibits Inhibits Decreases + - 482-8 18 (L) H85 IgG2b Inhibits Inhibits Bloch + - 482-1082 (L)

3D12 IgG 1 No effect Not tested Bloch + ND' 482-998 (C)

a Adherence was measured by centrifügation binding of E-histolyrica trophozoites to CHO cells and by binduig to ~'~-labelledpurifÏed colonic mucin (167). b. Cytotoxicity was measwed by 5'~r-releasenom CHO ceUs foltowing addition of E. histoiytica trophozoites (1 95). c, CSb-C9 resistance was assayed by addition of purified complement to E-histolytica trophozoites and Lysis measured microscopically (27). d. N refers to non-pathogenic (E-dispar)and P refers to pathogenic (E-histolyrica) amoebae determined by radioimmunoassay (RIA) (163). e- Mapping of epitopes was carried out by GST-lectin fksion protein expression in Escherichia coli followed by western immunoblot (141). f. ND, not determined g, L (linear) and C (conformational) nature of epitopes are based on western immunoblot performed under reduced/alkyIated and non-reducedhative conditions, respectively (142). adherence to colonic mucosa and cytolysis ofhost intestinal epitheliai and immune celis (1 74,

175). (ii) Homology with CD59 confprs to the Iectin the capacity to mediate evasion of the host cornplement pathway at the stage of the membrane attack complex (C5b-Cg) by binding to components CS and C9 (27). (iii) The lectin Is immunodominant in human infionand Îs recognized by the systemic humoral immune system (161). (iv) The lectin stimulates proliferation of human lymphocytes that have been primed during infection to kili amoebae and enhances the production of protectke cytohes nom immune ceils (T Lymphocytes and macrophages) (37,202). (v) The lectin is a highly conserved antigen in al1 isolates of

E. histolytica determined to date (172). (vi) In a cohort of 102 1 patients who had been sucessfdly treated for amebic liver abscess, a lower rate of recurrence (0.29% over 5 years) occurred than in a non-immune control group (24% over 5 years), suggesting the possibility of acquired protective immunity (55).

Immunization studies in gerbils with crude amebic lysates have induced partial or cornplete immety to experimentally induced infection in several animal models (222)-

Immunization with affinity purified native lectin in complete Freund's adjuvant (CFA) has demonstrated 43-86% protective efficacy in preventing experimentdly induced amebic Iiver abscess (165). A key observation made in this study was that immunized animals that succumbed to hepatic inoculation had iiver abscesses greater in mean sue and weight thm in sham- immunized controls. The presence of lectin mAb epitopes (3F4,8A3)that result in increased adherence in vitro rnay explain this phenornenon (Table 1.7.3). Owing to the possibility that disease-exacerbating epitopes exist within the native Lech, recombinant lectin fragments have been produced and purified in prokaryotic systems for immunization purposes (219,257). These studies have demonstrated that protective efficacies of 83% and 71% cm be achieved by the recombinant cysteine-rich region (6494202 and 758-1 134) done in gerbils, without the exacerbating effects seen with the native lectin- Greater than 90% of sham-immunized conîrol animals develop abscesses in the gerbü model foiiowing direct hepatic inoculation of 106

E-Aisrolytica trophozoites (43)-

In this model, experïmental infection completely bypasses the role of mucosal immunity in preventing invasive disease and remW a simiificant criticism. Therefore efforts have been made to generate an animai model of intestinal amebiasis. Staniey and colleagues have recentiy shown that inoculation of E-hr'srofytica trophozoites into human intesthal xenografb in severe combined irnmunodeficiency (SCID-hu-int) induces intestinal colitis (47). However, the absence of both T and B cells in the SCID mouse does not ailow for protective acquired immunity to be evaluated. Another shortcoming with intramuscuiar or intraperitoneal immuni7ation schemes is the failure to induce mucosaI immunity that tikely plays a key role in preventing the infiltration of trophozoites across the intestinal epithelium. Secretory immunogtobulin A (sIgA) directed against the lectin have been demonstrated in patients infected with E- histolytica (1 1O), (20).

Moreover, anti-lectin sIgA responses can block the adherence of amoeba to mammaiian ce1Is in vitro, suggesting that mucosal immuaity in the form of sIgA might be immunoprotective, Use of the amebic Iectin as an oral immunogen recombinantly expressed in Salmonella did not elicit IgG antibodies but did result in significant reduction in abscess size following intrahepatic challenge

(140). These data suggest a non-specific cellular response as sficient for protection against abscess formation. Clearly, when taken together, a role for both cellular (macrophages NO- mediated killing) and humoral (mucosai IgA and systemic IgG) fünctions is supported in the 1-7-5 Signalling via the Eh&o@tîka GaUGaNAc lectin

The first demonstration of lectln-rnedïated sigrdikg was made by CO-incubationof target ceus and amoeba, which resulted in amebic actin polymerization. Observed cytoskeletai changes could be inhibited by pre-incubation of the trophozoites with either galactose or IV-acetyl-D- galactosamine, suggesting involvement of the amebic lectin (13, 14)- Cytoskeletal changes appear to be vitai to cytoIysis of intestinal epithelial cells since over-expression of amebic myosin in stable-transfected amoebae redted in fdure to kill target cells in vitro (10)- These mutant amoebae exbibited abnonnal movement and faiied to cluster bound Ligands (a normal feature of arnoebic cccapping").Capping also occurs in response to exposure of trophozoites

ECM proteins iike fibronectin and is contemporaneous with nlamentous actin CO-localization with fibronectin receptor, myosin II, a-acthin, and vinculin (240). Modulation of the actin cytoskeleton is iikely mediated by amebic Rac, a smali GTP-binding protein, which has been implicated in phagocytosis and capping (89, 136). While the lectin retains a.SH2-binding domain in the cytoplasmic tail of the heavy subunit, there has been no clear role demonstrated for

SHî-binding proteins in signalling mediated by the lectine

1.7.6 GaVGaïNAc lectin gene regulation

Association of trophozoites with bactena is thought to affect amebic virulence, suggesting dynamic interaction of E-hisrolytica with its immediate environment (25). Several investigators have looked into transcriptional regulation of the lectin. Analysis of the lectin heavy subunit promoter, identified 5 major upstream regdatory elements -1 -5) and three core promoter elements (CREI -3) (35,170). The core cis-acting elements contain the initiator sequence, TATA-like sequence, and GAAC element, ali of which are necessary for optimal lech transcription- However, DNA-bindlng protek responsiile for regdation of these core sites have

not been demonstrated,

The early work of Ravdia and Guerrant clearly demonstrated a role for carbohydrate-

dependent adherence of amoebae to severai relevant mammaliau ceU lines (1 74,175)-

Subsequent to this, Petri and colIeagues were able to aff?nify puri@the native 260-kDa

GaVGalNAc hololectin fiom amebic Lysates (166). Monoclonal antibodies were generated against the heavy (170-ma) and light (3 5-kDa) subunits (l4L, 149)- Centrifugation binding assays demonstrated that MAbs directed at the heavy, but not the light, subunit were able to inhibit GaVGalNAc-dependent adherence to mammaLianceUs (141). This suggested indirectly that the heavy subunit possessed sugar-binding activity, despite no apparent primary sequence homology with welI-characterized C- and S-type lectins. Adherence-inhibithg MAbs were mapped to the cysteine-rich (356-1 143) region of the 170-kDa subunit, suggesting that these residues may comprise one or more carbohydrate recognition domains (CRDs). We hypothesized that fragments of the heavy subunit, which span the entire extracellular portion, could be synthesized in vitro and added to the neoglycoconjugate, GalNAc19BSA,previously shown to bind hololectin resident on amebic membranes with hi& aff?nity (6).In Chapter 2, experiments that attempt to directly demonstrate the presence or absence of CRD(s) in the 170-kDa cysteine- nch region are described. The dichotomy in pathogenesis between Ehiktolytica and E-dispmprovided a nattuaiiy- occrrrring avident cornterpart for molecular andysis. Ediqm7sapparent n~~pathogenicity also begged the question: does tbîs orgauism harbour one or more homoIogous genes to the

GaVGaiNAc lectin of E-hiscolylica- Previous work had demomtrated that MAbs directed against

E-histolyrica 170-kDa subunit cross-reacted with an antigen on the surface of E.dispw (163).

However, the E-dispar homologue had not kengenetically isolated. We hypothesized that a homologous PCR approach couid be used to identifil the E-di- Lectin heavy subimit gene.

Chapter 3 describes experiments that led to the identification of "dhgl2) a gene homologous to hgl2 of E-histolyiica, îhat was both present and expressed in Edispar- A novel PCR-RFLP approach to distinguish E-dispar fiom E. histoZytica, based on a genetic polymorphism between the two genes, is identified. We Merhypothesized that rnolecular differences between the

E-histolytca and E-dispar Iectin genes would be informative in understanding the merential vinilence of these two related organisms. Chapter 4 details the molecdar characterization of the

E-dispar GaVGaNAc lectin gene and protein, ident-g dinerentid gene and conformational regulation in pathogen versus non-pathogen. Evidence that both distinct and common epitopes are present on the E-dispar and E-histolylica lectin (170-kDa) homologues is presented.

While GaVGaNAc-dependent binding to mammalian cells appears to be important, other adhesion molecules may be expressed on the trophozoite surface. Marnmaiian lymphocytes rely on both lectin-iike (for rolling) and integrin-Ilke (for sticking) interactions with endotheiial cells.

For instance, during hematogenous spread, both lectin- and integrin-mediated interactions are required for lymphocytes to extravasate in respoase to the appropriate biochernical cues fiom a site of inflammation. Since parasites often usurp host responses for their own &val and propogation, experiments were performed to idente a novel integrin-like molecule for E. histo&h'ca. Previous work suggested that a MAb (MHM23) directed against human 02- integrin (CD 18) cross-reacted with an undetermineci Ehistolytica sudace aatigen (5). We hypothesized that thÏs putative B2 integrin was involved in amoeba-host cell adherence- Data presented in Chapter 5 provides evidence that a novel CD 184ike molecule, is present on the trophozoite surface, and participates in adherence to activated endotheliai cens, Moreover, the putative amebic integrin rp4T') migrates at approximately 47-kDa following

ïmmunoprecipitation and shows apparent specincity for hman ICAM-1 Ïnamoeba?iost ceU adherence assays.

The emergence of stool antigen ELISA kits that are able to distinguish E-histolyria nom

E-dtspnr ushg MAbs directed against the GaVGaiNAc Lectin permitted not only evaluation of the efficacy of such disgnostic kits compared to traditiond methods, but aiso a molecular epidemiological survey in a non-endemic settïng (93). We hypothesïzed that these new assays would aiiow us to revisit epidemiological questions concerning the relative preiavence of

E.hisîolyrica and E-dispar in a non-endemic setting. We aimed to identify what proportion of

Entmebae infectioris, previously called "E.histoiyricd' by traditional rnicroscopy, were actually

E-histobtka and not E-dispar.These results are presented in Chapter 6 and have important clinical implications since E- histolyrica misdiagnosis leads to unnecessary dnig therapy with potentiai side-effects and the possibility of fùehgdrug-cesistance. In addition, as an ancillary objective, a new immunochromatographic stnp that is capable of simuitaneously detecting several dysentery-associated gut pathogens was evaluated for the ihttime (Chapter 7). The impact of these findings are placed in context and future directions are discussed in Chapter 8. Abd-- M. D., T. F. Jackson, V' Gathiram, A, nk el-Hawey, and 8.1. RI1vdin. 1993. DEerentiation of pathogenic Entamoeba histolytÏca infections fkom nonpathogenic infections by detection of galactose-inhibitable adherence protein antigen in sera and feces. J Clin Microbiol. 31~2845-50. Abioye, A. A, 1971. Establishment and propagation of Ibadan strains of Entamoeba histolytica in vitro. Trans R Soc Trop Med Hyg. 65369-75. Abioye, A. A. 1973. Fatal amoebic colitis in pregnancy and puerperium: a new clinico- pathological entity. J Trop Med Hyg- 76:97-100. Acuna-Soto, R, Samuelson, P-De Gùolami, L. Zarate, F. MiIlan-Velasco, G. Schoolnick, and D. Wirth. 1993. Application of the polymerase chain reaction to the epidemiology of pathogenic and nonpathogenic Entamoeba histolytica Am I Trop Med Hyg. 48~58-70. Adams, S. A., S. C. Robsoii, V. Gathiram, T. F. Jacksoa, T. S. Pillay, R E. Kirsch, and M. W. Makgoba. 1993. Immunological similanty between the 170 kD amoebic adherence gtycoprotein and human beta 2 integrins, Lancet 341: 17-9. Adler, P., S. J. Wood, Y. C-Lee, R TVLee, WvA. Petri, Jr., and R L. Schnaar. 1995. High affTnity binding of the Entamoeba histolytica lectin to polyvalent N- acetylgalactosaminides~J Bi01 Chem, 270:s 164-7 1. Ahmed, L., 2. A. Salama, A. El Rooby, and G. T. Strickland 1989. Ultrasonographic resohtion time for arnebic iiver abscess. Am J Trop Med Hyg. 4:406-10. Aley, S. B., W. A-Scott, and Z. A*Cohn. 1980- PIasma membrane of Entamoeba histolytica. J Exp Med. l52:39 1-404- Andrews, B. J, and Bo Bjorvatn. 1994. Chemotherapy of Entamoeba histolytica: studies in vitro with bacitracin and its zinc sait Tram R Soc Trop Med Hyg. 88%- 100. Arhets, P., J. C. Olivo, P. Gounon, P. Sansonetti, and N, GuilIen. 1998. Virulence and fiinctions of myosin II are inhibited by overexpression of Light meromyosin in Entamoeba histolytica Mol Bi01 Cell. 9: 1537-47. Atakpa, T-O. U., Me Damisah, He C. Whittle, and B. M. Greenwood. 1978. Amoebiask in northem Nigeria: an evaluation of the gel diffusion test AM Trop Med Parasitoi. 72:123-6, Aust-Kettis, A., and K. G. Smdqvist, 1978- Dymmks of the interaction between Entamoeba histolytica and components of the immune response. 1. Capping and endocytosis; influence of inhibithg and acceierating factors; variation of the expression of daceantigens. Scand J Immunol- E35-44. Bdey, G. B., DmB. Day, and J. W. Gasque. 1985. Rapid polymerization ofEntamoeba histolytica actui induced by interaction with target cells. J Exp Med. 162:546-588 Baüey, GeB., J. R Gilmour, and N. E. McCoorner. 1990. Roles of target ceii membrane carbohydrate and lipid in Entamoeba histolytica interaction with mammalian cells. Infect Immun. 58:2389-9 1, Barondes, S. H. 198 1. Lectins: theii multiple endogenous cellular fiinctions. Annu Rev Biochem. 50:207-3 1. Baroïides, SeHm, D. N. Cooper, M. A. Gitt, and H. Leffier. 1994. Galectins. Structure and f'unction of a large family of animal lectins. J Bi01 Chem. 269:20807-10. Bassily, S., Z. Farid, N. A. el-Masry, and E. M. Mikhail. 1987. Treatment of intestinal E. histolytica and G. lamblia with metronidazole, tinidazole and ornidazole: a comparative study. J Trop Med Hyg. 90:9-12. Becker, I., R Perez-Tamayo, 1. Montfort, A. M. Aivizouri, and R Perez-Montfort. 1988. Entamoeba histolytica: role of amebic proteinases and polymorphonuclear leukocytes in acute experiinental amebiasis in the rat. Exp Parasitol. 67:268-80. Berlin, E. A., and E. K. Markeu- 1977. An assessrnent of the nutritional and health status of a.Aguaruna Jivaro community. Ecology and Food Nutrition. 6:69-8 1. Beving, D. E., C. J. Soong, and J. 1. Ravdin. 1996. Oral immunkation with a recombinant cysteine-rich section of the Entamoeba histolytica galactose-inhibitable lectin elicits an intestinal secretory immunoglobulin A response that has in vitro adherence inhibition activity. Infect Immun. 64: 1473-6. Bischoff, J., and H. FmLodish. 1987. Two asialoglycoprotein receptor polypeptides in human hepatoma cells. J Biol Chem. 262: 11825-32. Bos, E. J., W. J-Schouten, H, Noordpooi, M. Makbh, and B. F. Oostburgi 1980. A seroepidemiological study of amebiasis in Surinam by the enzyme- iinked immunosorbent assay (ELISA), Am J Trop Med EZyg. 29:3 58-63 - Botero, D. R L964. Treatment ofacute and chronic intestinal amebiasis with entamide furoate. Trans R Soc Trop Med Hyg, 58:419-21. Bourne, Y., P. Rouge, and C. Cambüiau- 1992- X-ray structure of a biantennary octasaccharide-lectin complex eedat 2-3-A resolution. J Bi01 Chem. 267: 197-203. Bracha, R, and D. Mirelman. 1984. Vinilence of Entamoeba histolytica trophozoites. Effects of bacteria, microaerobic conditions, and metronldazole. J Exp Med. 160:353-68. Braga, L. L., H. Ninomiya, J. J. McCoy, K. Adal, T. Wiedmer, C. Pham, P. J. Sims, and W. A. Petri, Jr. 1992. Inhibition of the complement membrane attack complex by the galactose- specific adhesin of Entamoeba histolytica Arch Med Res- 23: 133. Braga, L. L., H. Ninomiya, J. J. McCoy, S. Eacker, T. Wiedmer, C. Pham, S. Wood, P. J. Shs, and W. A Petri, Jr. 1992. Inhibition of the complement membrane attack complex by the galactose- specinc adhesion of Entamoeba histolytica. J Clin Invest. 90: 113 1-7. Bray, R S., and W. G. Harris. 1977. The epidemiology of infection with Entamoeba histolytica in the Gambia, West Afnca. Trans R Soc Trop Med Hyg. 71:401-7-

Briggs, P., R K. Freeman, and S. Jo Yaffk 1994- Drugs in pregnancy and lactation. A iference guide to fetal and neonatal risk (4th ed-). Baltiimore:WilIams & Wilkins585-7. Britten, D., S. M. Wilson, R McNerney, A. H. Moody, P. L. Chiodini, and J. P. Ackers. 1997. An improved coIorimetric PCR-based method for detection and differentiation of Entamoeba histoiytica and Entamoeba dispar in feces. J Clin Microbiol. 35: 1108-1 1. Bruchhaus, I., T. Jacobs, M. Leippe, and E. Tamich. 1996. Entamoeba histolytica and Entamoeba dispar. differences in numbers and expression of cysteine proteinase genes. Mol Microbiol. 22:255-63. Bruchhaus, I., and E. Tannich. 1996. A gene highly homologous to ACPl encoding cysteine proteinase 3 in Entamoeba histolytica is present and expressed in E. dispar. Parasitol Res. 82: 189-92. Brumpt, E. 1925. Etude sommaÏre de I'Entamoeba dispar ILsp- Ami& a kyste quadruiucIees, parasite de l'homme Bulletin de l'AcademÎe de Mediçine (Paris). 94:943- 952. Burtin, P., A. Taddio, O. Aribumu, T. R Einanan, and G. Koren. 1995, Safety of metronidazole in pregnancy: a meta-analysis, Am J Obst Gyn- 172 (2 Pt 1):525-9- Buss, Ho,C. Lioutas, S. Dobinsky, R Nickel, and E, Tannich. 1995. Analysis of the 170-kDa lectin gene promoter of Entamoeba histolytica, Mol Bioçhem Parasitol, 72: 1-10, Calderon, J., M. de Lourdes Munoq and 8. M. Acosta. 1980, Surface redistribution and release of antibody-induced caps in entamoebae, J Exp Med, 151: 184-93. Campbell, D., and K. Chadea 1997. Survivd strategies of Entamoeba histolytica: Modulation of ceil-mediated immune responses. Parasitol Today. 13: 184-9. CarIson, J., and M. WahIgren. 1992. Plasmodium fdciparum erytbrocyte rosetting is mediated by promiscuous lech-like interactions. J Exp Med. 176: 13 11-7. Carter, C. H., A. Bayfes, and P. E. Thompson. 1968. Effects of paromomycin sulfate in man against Entamoeba histolytica and other intestinal protozoa. Am J Trop Med Hyg. ll:448-S 1. Chacin BoniMa, L., E. Chacin-Martin- E. Espinoza, and B. Cardenas. 1982. A seroepidemiological study of amebiasis in children of low socioeconomic Level in Maracaibo, Venezuela. Am J Trop Med Hyg. 31: 1 2 O3 -6. Chacin-Bonilla, L., and D. Bonpart. 198 1. A seroepidemiological study of amebiasis in ad& in Maracaibo, Venezuela. Am J Trop Med Hyg. 30: 1201-5. Chadee, K, M. Denis, and K. KeUer. 1991. Down-regdation of murine lymphocyte responsiveness to mitogens after treatment with antigens of Entamoeba histolytica Parasitol Res, 77572-6, Chadee, K., and E. Meerovitch. 1984. The MongoLian gerbil (Meriones unguicuiatus) as an experimental host for Entamoeba histolytica Am J Trop Med Hyg. 33:47-54. Chadee, K., W. A. Petri, Jr., D. JeInnes, and J. 1, Ravdin. 1987. Rat and human colonic mucins bind to and inhibit adhetence Iectùi of Entamoeba histolytica, J Clin Invest. 80: 1245-54. Cheng, X. J., H. Tachibana, S, Kabayashi, Y. ffineda, and M, Y, Huang, 1993. Pathogenicity of Entamoeba histolytica isolates from Shanghai, China ParasitoL Res- 79~608-10. Chiacchia, K. B., and IL Drickamer. 1984, Direct evidence for the transmembrane orientation of the hepatic giycoprotein receptors, J Biol Chem, 259: 15440-6- Cieslak, P. R, H. W. ti Virgin, and S. L. Stanley, Jr. 1992. A severe combined imrnunodeficient (SCID) mouse mode1 for infection with Entamoeba histolytica J Exp Med. 176: 1605-9. Clark, C. G. 1995. Axenic cdtivation of Entamoeba dispar Brumpt 1925, Entamoeba insolita Geiman and Wichtennan 1937 and Entamoeba ranarum Grassi 1879- J Eukaryot Microbioi- 42:590-3. Collins, R F., and L. O. Edwards. 198 1. Prevalence of intestinai helminths and protozoans in a rural population segment of the Dominican Repubiic. Tram R Soc Trop Med Hyg, 75549-5 1. Councilman, W. T., and H. A. Lafleur. 1891. Amoebic dysentery, Johns Hopkins Hosp Bull, 2:395-548. Crevenna, P. B. 1977. Epidemologia de la amibiasis. Salud Publica de Mexico. 19:411- 420- Crevenna, P. B. 1971. Pruebas immunologicas de amibiasis en un grupo de poblacion de la ciudad de Mexico, Rev Gastroenterol Mex. 36: 100- 105, Cuadrado, R R, and 1. G. Kagan. 1967. The prevalence of antibodies to parasitic diseases in sera of young Army recruits fiom the United States and Brazil. Am J Epidemiol. 86:330-40. Dandekar, T., and M. Leippe. 1997. Molecular rnodeling of amoebapore and NK-lysin: a four-alpha-helix bundle motif of cytolytic peptides fiom distantly related organisms. Fold Des- 2~47-52. De Leon, A. 1970. Delayed progaosis in amebic hepatic abscess. Archiv hvest Med. 1: S205-6. De Muyk, A., and M. Silva de Lagrava. 1977. Survey of intestinal parasites in school children fiom Yapacani, Bolivia. Boletin Chileno de Parasitologia. 32:71-2. Delgado y Garnicri, R,and E, Lopex A- 1971. Frecuencia de la ami'biasis y de otras parasitosis intestinales en una zona de Ia ciudad de Mexico. Rev Invest Saiud Publica- 31:I 14-127, Denis, M., and K. Chadee. 1989, Cytokine activation of mwine macrophages for in vitro kihgof Entamoeba histolytica trophozoites. Mect Tmmun- 57: 17506, De-, M., and K Chadee. 1988- In vitro and in vivo studies of macrophage fiinctions in amebiasis. Infect Immun- S6:3 126-3 1. Diamond, L, S., and C. G. Clark 1993-A redescription of Entamoeba histolytica Schaudinn, L903 Fmended Waker, 1911) separating it fiom Entamoeba dispar Bnimpt, 1925. J Eukaryot Microbiol. 40~340-4. Dickinson, B. L., K. Badizadegan, 2. Wu, and e. al. 1999. Bidirectional FcRn- dependent IgG transport in a polatized human intestinai epithelial ceii he. I Clin Inv. lO4:9O3-9 11, Dobell, C. 1919. The amoebae in living man, A zoological monograph. .London: John Bales & Sons, Danielsson.:. Drickamer, K. 198 1. Complete amino acid sequence of a membrane receptor for glycoproteins. Sequence of the chicken hepatic lectin. J Bi01 Chem, 256:5827-39- DRckamer, K. 1999. C-type lectin-like domains. Curr Opin Stmct Biol, 9:585-90. Drickamer, K. 1988. Two distinct classes of carbohydrate-recognition domains in animal lectins. J Bi01 Chem. 263:9557-60. Drickamer, K., and M. E. Taylor. 1993. Biology of animal lectins. AMU Rev Ce11 Biol. 9:23 7-64. Eckmann, L., S. L. Reed, J. R Smith, and M. F. Kagnon. 1995. Entamoeba histolytica trophozoites induce an infiammatory cytokine response by cultured human ceiis through the paracrine action of cytolytically released interleukin-l alpha.. J Clin Invest. 96: 1269- 79. Eisert, J., J. E. Hannibal Jr, and S. L. Sanders. 1959. Fatal amebiasis complicating corticosteroid management of pemphigus vulgaris. N EngI J Med. 261:843-5. Elgavish, S., and B. Shaanan. 1997. Lectin-carhhydrate interactions: different folds, common recognition princples. Trends Biochem Sci. 22:462-7, El-Bennawy, M., and H. Abd-Rabbo. 1978. Hazards ofcortisone therapy in hepatic amoebiasis, 5 Trop Med Hyg, 81:71-3, Elizondo, G., R Weissleder, D. D. Stark, L, E. Todd, C. Compton, J. Wittenberg, and J. T. Femtcei. 1987. Amebic Liver abscess: Diagnosis and treatment evaiuation with MR imaging. Radiology. 165:795-800. Etges, R, J. Bouvier, and C. Bordier. 1986, The major surface protein of Leishmani promastigotes is a protease. J Bi01 Chem- ZO:9098-lO 1, Ewald, P. W. 1995. The evolution of vinilence: A umgLink between parasitology and ecology. J ParasitoI- 81:659-669, Farthing, M. J., M. E. Pereira, and G. T. Keasch. 1986. Description and characterization of a dacelectin fiom Giardia lamblia- Infect Immun- 51:661-7- Filice, C., G. Di Pem, M. Strosselli, E. Brunetti, S. Dughetti, D. H. Van Thiel, and C. Scotti-Fogkni. 1992. Outcome of hepatic amebic abscesses managed with three dierent therapeutic strategies. Dig Dis Sci. 37:240-7. Fisher, A. IL, F. G. Walter, and S. Szabo. 1993. Iodoquhol associated seinires and radiopacity. J Toxicol-Clin Toxicol. 31: 113-20. Gandhi, B. M., H. Gupta, M. Irshad, T. C. Chawala, and B. N. Tandon. 1986. CelI- mediated imrnunity in amoebic liver abscess. J Trop Med Hyg. 89: 163-6- Ganguly, N. K., R C. Mahajan, N. J. GüI, and A. Koshy. 1981- Kinetics of lymphocyte subpopdations and their hctions in cases of amoebic liver abscess- Tram R Soc Trop Med Hyg, 75:807-20. Ghadirian, E., and E. Meerovitch. 1982. Effect of complement depletion on hepatic amoebiasis in hamsters. Clin Immun01 Immunopathol. 24:3 15-9. Ghadirian, E., and E. Meerovitch. 1981. Effect of immunosuppression on the size and metastasis of amoebic Liver abscesses in hamsters. Parasite IÏnmunol- 3:329-38. Ghadirian, E., and E. Meerovitch. 1981. Effect of splenectomy on the size of amoebic iiver abscesses and metastatic foci in hamsters. Infect immun. 31:571-3. Ghadirian, E., and E. Meerovitch, 1982. Macrophage requirement for host defence against experimental hepatic amoebiasis in the hamster- Parasite hmunol. 4:219-25. Ghadinan, E., E. Meemvitch, and D, P. Hartmann. 1980, Protection against amebic liver abscess in hamsters by means of immunizatïon with amebic antigen and some of its fractions. Am J Trop Med Hyg. 29:779-84- Ghadirian, E., E. Meerovitch, and P. A. Kongshan. 1983, Role of macrophages in host defense against hepatic amoebiasis in hamsters- Infect Immunc 42: 1017-9. Giga, Y., A. hi,and K. Takahashi. 1987, The complete amino acid sequence of echoidin, a Iectin fkom the coelomic fluid of the sea urchin Anthocidaris crassispina Homologies with mammajian and insect lectins. J Bi01 Chem, 26236197-203.

Gijsbers, A, J., J. A. Millar, K, O'SulIivan, N. M. Thomson, J. P. Dowhg, and D. Scott. 1978. Activation of occult amebiasis by corticosteroids [letter]. Br Med J. 2: 1372- 3- Gitt, M. A., and S. H. Barondes. 1986. Evidence that a human soluble beta-gdactoside- binding lectin is encoded by a fdyof genes. Proc Nat1 Acad Sci USA- 83:7603-7- Group., W. S. W. 1980. Parasite-related diarrhoeas. Bull WHO. 58: 8 19-830- Guillen, N., P. Boquet, and P. Sansonetti. 1998. The small GTP-binding protein RacG reguiates woid formation in the protozoan parasite Entamoeba histolytka. J Ceil Sci. 111:1729-39. Gutierrez, G., J. C. Margain, R Castaneda, O. Enriquez, M. Aubanel, and B. Sepulveda. 1974. vational serologic survey. 1. Study of antibodies against E. histolytica in a serniniral locality on the Gulfof Mexico Coast]. Arch Invest Med (Mex). 5475-80. Halberg, D. F., R E. Wager, D. C. Farrell, J. Hildreth 4th, M. S. Quesenbeny, J. A. Loeb, E. C. Holland, and K. Drickamer. 1987. Major and minor fonns of the rat Liver asialoglycoprotein receptor are independent galactose-binding proteins: Primary structure and glycosylation heterogeneity of minor receptor foms. J Bi01 Chem. 262:9828-38, Haque, R, A. S. Faruque, P. Hahn, D. M. Lyeriy, and W. A. Petri, Jr. 1997. Entamoeba histolytica and Entamoeba dispar infection in children in Bangladesh. J Infïect Dis- 175734-6- Haque, R, K. Kress, S. Wood, T. F. Jackson, D. Lyerly, T. Wilkins, and W. A. Petri, Jr. 1993. Diagnosis of pathogenic Entamoeba histolytica infection using a stool ELISA based on monoclonal ann'bodies to the galactose-specinc adhesin [see commentsl- J Infect Dis, 167t247-9, Hams, W. G., and R S. Bray. 1976, Cellular sensitivSty in amoebiasis-preliminary resuits of lymphocytic transformaon in response to specsc antigen and to mitogen in carrier and disease States, TmR Soc Trop Med Hyg, 70:340-3, Healy, G. R 1986- Immunologie tools in the diagnosis of amebiasis: epidemiology in the United States- Rev Infect Dis- 8~239-46- Hernandez, A. G., N. Rodriguez, D. Stojanovic, and D. Candek 1986. The localization of a lectin-Like component on the LeishmatÜaceil dace. Mol Bi01 Rep. 11: 149-53. Huldt, G., P. Davies, A, C. Allison, and H. U. Schorlemmer. 1979- Interactions between Entamoeba histolytica and complement. Nature-277:214-6. Irusen, E. M., T. P. Jackson, and A. E, Simjee. 1992. Asymptomatic intestinal colonization by pathogenic Entamoeba histolytica in amebic liver abscess: prevalence, response to therapy, and pathogenic potentid [see comments]. Clin Infect Dis. 14:889-93. Jackson, T. K, V. Gathiram, and A. E. Simjee. 1985. Seroepidemiological study of antibody responses to the zymodemes of Entamoeba histolytica. Lancet 1: 716-9. Jacobson, R L., and R J. Doyle. 1996. Lectin-Parasite interactions. Parasitol Today. 12: 55-60, Jarumilinta, R, and F. Kradolfer. 1964. The toxic effect of Entamoeba histoLytica on leukocytes. Ann Trop Med ParasitoL 58:375-8 1. 102. Jungery, M. 1985. Studies on the biochemical basis of the interaction of the merozoites of Plasmodium fdcipanun and the human red ceii. Tram R Soc Trop Med Hyg. 79591- 7. 103. Juniper, IL, Jr., C. L. Worreil, M. C. Minshew, L. S. Roth, H. Cypert, and R E. Lloyd. 1972. Serologic diagnosis of amebiasis. Am J Trop Med Hyg. 21: 1 57-68. 104. Kanani, S. R, and R Knight. 1969. Amoebic dysentery precipitated by corticosteroids. Br Med J. 3: 114. 105. Kanani, S. R., and R Knight. 1969. Relapsing amoebic colitis of 12 year's standing exacerbated by corticosteroids. Br Med J- 2:613-4. Keene, W. E., M. E, Hidalgo, E, Orozco, and J, H, McKemwC1990, Entamoeba histolytica: correlation of the cytopathic effect of vident trophozoites wïth secretion of a cysteine proteinase. Exp Parasitol- 71:199-206- Keene, W. E., M. G. Petitt, S. Aiien, and J, 8, McKerrow. 1986. The major neutml protelliase of Entamoeba histolytica J Exp Med, 163~5364% Kelsaii, B. L., T. G. Jackson, V. Gathh, S. B, Saïïg, M. Vaithilingum, R D. Pearson, and J. 1. Ravdin. 1994- Secretory imrnunoglobului A anti'bodies to the galactose-inhibitable adherence protein in the saliva of patients with arnebic iiver disease. Am J Trop Med Hyg, 51:454-9. Kelsaii, B. L., and J. 1 Ravdin. 1993. Degradation of human IgA by Entamoeba histoLytica. J Mect Dis, 168: 13 19-22- Kelsall, B. L., and J. 1, Ravdin. 1995- Immunkation of rats with the 260-kilodalton Entamoeba histolytica galactose-inhibitabIe lectüi elicits an intestinal secretory immunoglobulin A response that has in vitro adherence-inhibitory activity Infect Immun. 63~686-9. Kessel, Je F., W. P. Lewis, C. M. Pasquel, and Jœ A. Turner. 1965. Indirect hemagglutination and complement fixation tests in amebiasiw Am J Trop Med Hyg. 14: 540-50, Keusch, G. T., F. J. Troncale, and R D. Buchanan. 1970. Maiabsorption due to paromomycin. Arch Intern Med. 125273-6. Kobayashi, S., E. imai, H. Tachibana, T. Fujiwara, and T. Takeuchi. 1998. Entamoeba dispar: cultivation with sterilized Crithidia fasciculata, J Eukaryot Microbiol. 453s-8s- Kotcher, E., M. Miranda, and V. Garcia de Sdgado. 1970. Correlation of dinical, parasitological, and serological data of individuals ùifected with Entamoeba histolytica. Gastroenterology. S8:3 88-9 1. Krogstad, D. Je,H. C. Spencer, Jr., and G. R Heaiy. 1978, Current concepts in parasitology. Amebiasis. N Engl J Med. 298:262-5. Krupp, 1. M. 1970. Antibody response in intestinal and extraintestinal amebiasis. Am J Trop Med Hyg. 1957-62. Krupp, I, M., and S, 8. PoweU. IWf. Comparative study of the anti'body response in amebiasis. Persistence after successflll. treatment, Am J Trop Med Hyg. 20:421-4. Lasky, L. A. 1992, Selectins: interpreters of ceii-specinc carbohydrate Information during inflammation. Science. 258:964-9. Lau, A, H., N. P, Lam, and S. S, Pisciteiii. 1992, CIinicaI pharmacokinetics of metronidazole and other nitrounidazole anti-infectives, Ciin Pharmacolchet, 23~328-64 Lawrence, D. N., J. V. Neel, S. H, Abadie, L. L. Moore, L. J. Adams, G. R Eealy, and 1. G. Kagan. 1980- Epidemiologic studies among Amerindian populations of Amazonia m. Intestinal parasitoses in newly contacted and acculturating viUages. Am J Trop Med Hyg. 29530-7, Lee, Y. C. 1992. Biochemïstry of carbohydrate-protein interaction. Faseb J. 6:3 193-200. Leippe, M. 1995. Mentweapons: NK-lysin, is a mammalian homolog to pore-fmiing peptides of a protozoan parasite petter]. Cell. 83: 17-8. Leippe, M. 1999. Antimicrobial and cytolytic polypeptides ofamoeboid protozoa-- effector molecules ofprimitive phagocytes. Dev Comp Immunol. 23:267-79. Leippe, M., J. Andra, and H. J. Muller-Eberhard. 1994. Cytolytic and antibacterial activity of synthetic peptides derived fiom amoebapore, the pore-forming peptide of Entamoeba histolytica. Proc Natl Acad Sci U S A. 91:2602-6, Leippe, M., J. Andra, R Nickel, E. Tannich, and H. 3. Mulier-Eberhard. 1994. Amoebapores, a family of membranolytic peptides fiom cytoplasmic granules of Entamoeba histolytica: isolation, primary structure, and pore formation in bacterial cytoplasmic membranes. Mol Microbiol. 14:895-904. Leippe, M., S, Ebel, O. L. Schoenberger, R D. Horstmann, and H. J. Muller- Eberhard. 1991. Pore-forrning peptide of pathogenic Entamoeba histolytica. Proc Nat1 Acad Sci U S A- 88:7659-63. Leippe, M., H. J. Sievertsen, E. Tannich, and R D. Horstmann. 1995. Spontaneous release of cysteine proteinases but not of pore-forming peptides by viable Entamoeba histolytica. Parasitology. 111569-74. Leippe, M., E. Tannicb, R. Nickel, G. van der Goot, F. Pattris, R. D. Horstmann, and H. J. Muller-Eberhard. 1992, Primary and secondary structure ofthe pore-forming peptide of pathogenic Entamoeba histolytica. Embo J, ll:XO 1-6. Lewis, E. A., and A. U. Antia. 1969. Amoebic colitis: Review of 295 cases. T'ans R Soc Trop Med Hyg- 63:633-8- Li, E., A. Becker, and S. L. Stanley, Jr. 1989, Chinese hamster ovary cells deficient in N- acetylglucosaminyltransferase I activity are resistant to Entamoeba histolytica- mediated cytotoxicity. Infect Tmmun, 57: 8- 12. Li, E., A Becker, and S. L. Stanley, .k. 1988. Use of Chinese hamster ovary cells with altered glycosylation patterns to define the carbohydrate specificity of Entamoeba histolytica adhesion. J Exp Med- 167: 1725-30. Li, E., W. G. Yang, T. Zhang, and S. L. Stanley, Jr. 1995- Interaction of laminin with Entamoeba histolytica cysteine proteinases and its effect on amebic pathogenesis. Ihfiect Immun. 63:4150-3. Lin, J. Y., and K. Chadee. 1992. Macrophage cytotoxicity against Entamoeba histolytica trophozoites is mediated by nitric oxide fiom L-arginine. J Immunol. 148:3999-4005. Lin, J. Y., K KeiIer, and K. Chadee. 1993. Entamoeba histolytica proteins modulate the respiratory burst potential by murine macrophages. Immunology. 78:29 1-7. Lin, J. Y., R Seguin, K. Keiier, and K Chadee. 1994. Tumor necrosis factor alpha augments nitric oxide-dependent macrophage cytotoxicity against Entamoeba histolytica by enhanced expression of the nitric oxide synthase gene. Infect Immun. 62: 1534-41. Lohia, A., and J. Samuelson. 1996. Heterogeneity of Entamoeba histolytica rac genes encoding p2 lrac homotogues. Gene. 173~205-8. Losch, F. 1875. Massenhafte entwickelung von amoben im dikdann, Anat Physiol. 65: 196-21 1. Luaces, A. L., and A. J. Barrett. 1988. AEtypurification and biochemical characterization of histolysin, the major cysteine proteinase of Entamoeba histolytica. Biochem J- 250:903-9, Lynch, E. C., 1. M. Rosenberg, and C. Gitkr. 1982. An ion-channel forming protein produced by Entamoeba histolytica- Embo J. 1:801-4- Mann, B. J., B. V. Butkholder, ond L. A, Lockhacti 1997, Protection in a gerbii mode1 of amebiasis by oral hunization wïth Salmonella expressing the gdactose/N-acew D- gdactosamine inhibitable lechof Entamoeba histolytica Vaccine. 15:65963- Mann, B. J., C. Y. Chung, 3. M. Dodson, L. S. Ashley, L. L. Braga, and T. Lw Snodgrass. 1993. Neutralizing monoclonal antibody epitopes of the Entamoeba histolytica galactose adhesin map to the cysteine-nch extracellular domain of the 170- kilodalton subu~lx*t.Infect Immun, 61: 1772-8- Mann, B. J, and L. A. Lockhact. 1998. Molecular anaiysis of the GaVGalNAc adhesin of Entamoeba histolytica 5 Eukaryot Microbiol. 45: L3 S-16s- Martiaez-Cairo, S., A Goirb, O. mu no^ and M. Reyes. 1979. [Coproantibodies in intestinal amebiasis]. Arch Invest Med- 10: 121-6. Martinez-Palorno, A., V. Tsutsumi, F. Anaya-Velazquez, and A Gonzalez-Robks. 1989. Utrastructute of experimental intestinal invasive amebiasis. Am J Trop Med Hyg- 41:273-9- Mathur, T. N, and J. Kaur. 1973. The fiequency of excretion of cysts of Entamoeba histolytica in known cases of non-dysenterie amoebic colitis based on 2 1 stool examinations. Indian J Med Res, 6l:3 30-4. McAuley, J. B., B. L. Herwaldt, S. L. Stokes, J. A. Becher, J. M. Roberts, M. K. Michelson, and D. D. Juranek 1992. Diloxanide furoate for treating asymptomatic Entamoeba histolytica cyst passers: 14 years' expenence in the United States [see comments] .Clin infect Dis. 15:464-8. McCoy, JOJ., B. J. Mann, T. S. Vedviek, Y. Pak, D. B. Heimark, and W. A Petri, Jr. 1993. Structural analysis of the iight subunit of the Entamoeba histolytica galactose- specinc adherence lech J Bi01 Chem. 268:24223-3 1. McCoy, J. J., B. J. Mann, T. S. Vedvick, and W. A. Petri, Jr. 1993. Sequence analysis of genes encodiig the light subunit of the Entamoeba histolytica galactose-specific adhesin. Mol Biochem Parasitol. 61~325-8- McCoy, J. J., A. M. Weaver, and W. A. Petri, Jr. 1994. Use of monoclonal anti-light subunit antibodies to study the structure and fiuiction of the Entamoeba histolytica GaUGaINAc adherence lectin. Glycoconj J. 1l:43 2-6. Meivin, D. M., and L. J, Mata. 1971. Intestinai parasites in a Mayan-Indian village of Guatemala- Rev Latlnoamericana Mimbiol, 13: 15-19, Merritt, E. A., S. Sarfaty, F. van den Akker, CaL'Hoir, J. A. Martial, and W. G. Hoi. 1994. Crystai structure of the cholera toxh B-pentamer bound to receptor GMl pentasaccharide. Protein Sci, 3: 166-75, Mirelman, Do 1987. Ameba-bacteh relationship in amebiasis, Microbiol Rev. 51:272- 84. Mirelman, D., Y. Nuchamowitz, B. Bohm-Gloning, and B. Walderich. 1996. A homologue of the cysteine proteinase gene (ACP 1or Eh-CPp3) ofpathogenic Entamoeba histolytica is present in non-pathogenic E. dispar strains. Mol Biochem Parasitol. 78:47- 54. Mody, V. R 1959. Corticosteroids in latent amoebiasis Detter]. Br Med J. 2: 1399. Musoke, A. JO, and A. F. Barbet. 1977, Activation of complement by variant specific antigen of Tsrpanosoma brucei. Nature. 27O:M 8-40. Nordestgaarci, A. G., L. Stapleford, N. Worthen, F. S. Bongad, and S. R Klein. 1992. Contemporary management of amebic liver abscess. Am Surg. 58:3 15-20. Orth-Ortiz, L., R Capin, N. R. Capin, B. SepuMa, and G. Zamacona. 1978. Activation of the alternative pathway of complement by Entamoeba histolytica. Clin Exp Immunol, 34: 10-8. Ortiz-Ortu, L., G. Zamacona, B. Sepuiveda, and N. R Capin. 1975. Ceii-mediated immunity in patients with amebic abscess of the Iiver. Clin Immun01 Immunopathol. 4: 127-34. Palau, L. A., and S. A. Kemmeriy. 1997. Fust report of invasive amebiasis in an organ transplant recipient. Transplantation. 64:93 6-7. Pattemon, M., G. R Heaiy, and 8. M. Shabot. 1980. Serologic testing for amoebiasis. Gastroenterology. 78:. Petri, W. A., Jr., J. Broman, G. Heaiy, T. Quinn, and J. 1. Ravdin. 1989. Antigenic stability and immunodorninance of the GaVGalNAc adhaence lectin of Entamoeba histolytica. Am J Med Sci. 297: 163-5. 162. Petri, W. A., Jr., M. D. Chapman, TeSnodgrass, B. J. Mann, 3. Broman, and 3. 1. Ravdin, 1989. Subunit structure of the galactose and N-acetyl-D-gaIactosamine- inhibitable adherence lectin of Entamoeba histoIytica J Bi01 Chem. 264=3007-12. Petri, W. A., Jr., T. FeJackson, V, Gathirant, K. Kress, L-D, Saer, TT,L. Snodgrass, M. D. Chapman, Z. Kecen, and D, MireIrnaa. 1990. Pathogenic and nonpathogenic sttains of Entamoeba histolytica can be differentiated by monoclonal antibodies to the gaIactose-specific adherence lectin- Infect Immun. 58: L802-6. Petri, W. A., Jr, M. P. Joyce, J. Broman, R D. Smith, C. F. Murphy, and J. 1. Ravdin. 1987. Recognition of the galactose- or N-acetylgdactosamIlie-bhding lectui of Entamoeba histolytica by human hunesera, Infect immun- 55:2327-3 1- Petri, W. A., Jr., and J. 1. Ravdin. 1991. Protection of gerbas nom amebic liver abscess by immunhtion with the galactose-specinc adherence lectin of Entamoeba histolytica Infect immun, S9:97-lO 1. Petri, W. A., Jr., R D. Smith, P. H. Schiesiager, C. F. Murphy, and J. 1. Rrivdin. 1987. Isolation of the galactose-binding lectin that mediates the in vitro adherence of Entamoeba histolytica, J Clin Invest- 80: 1238-44. Petri, W. A., Jr., T. L. Snodgrass, T. F. Jackson, V. Gathiram, A. E. Simjee, K. Chadee, and M. D. Chapman. 1990. Monoclonal antibodies directed against the galactose-binding lectin of Entamoeba histolytica enhance adherence. I Tmniunol. 144~4803-9. Piper, J. M., E. F. Mitchell, and W. A. Ray. 1993-Prenatai use of metronidazole and birth defects: no association. Obst Gyn, 82:348-52. Powell, S. W., A. J. Wümost, and 1. N. e. a. McLeod. 1965. Dehydroemetine in the treatment of amebic liver abscess- An.Trop Med Parasitol. 59:208-9. Purdy, J. E, L. T. Pho, B. J. Mann, and W. A. Petrî, Jr. 1996. Upstream regdatory elements controllhg expression of the Entamoeba histolytica Lectin. Mol Biochem Parasitol. 78:91-103. Rab, P. W., M. F. Quinn, W. D. BosweU Jr, and e. ai. 1983- Pattern of resolution in successfidly treated hepatic amebic abscess: sonographic evaiuation. Radiology. l49:54l- 3. d RamakrWhnan, G., B. D. RigLnd, J. E, Rudy, and B. J, Mam. 1996. Physical mapping and expression of gene familes encoding the N-acetyi D-gaiactosamhe adherence lectin of Entamoeba histolytica- Mol MicrobioL W:91-100, Ramani, A., R Ramani, M. S, Kumar, B. N, Lakhar, and G. N. Kundaje. 1993- Ultrasound-guided £5~needle aspiration of amebic berabscess. Postgrad Med J. 69~381-3. Ravdin, J. I., B. Y, Croft, and R L. Guerrank 1980. Cytopathogenic mechanisms of Entamoeba histolytica. J Exp Med. l52:3 77-90.

Ravdin, JwI., and R L. Guerranti 1981, Rob of adhexence in cytopathogenic mechanisms of Entamoeba histolytica. Study with mammalian tissue culture cens and human erythrocytes. J Clin Invest- 68: l3OS-l3- Ravdin, J. I., Te F. Jackson, W. A Pem*,Jt., C. F. Miupby, B. L. Ungar, V. Gathiram, Js Skiïogiannis, and A. EwSimjw. 1990. Association of semm antribodies to adherence Iectin with invasive amebiasis and asymptomatic infection with pathogenic Entamoeba histolytica, J Mect Dis. 162:768-72. Ravdin, J. I., C. F. Murphy, R A. Salata, R L. Guerrant, and E. L. Hewlett. 1985. N-Acetyl-D-galactosamine-inhibitableadherence lectin of Entamoeba histolytica- 1- Partial purification and reIation to amoebic virulence in vitro, J Mect Dis- 151:804-15. Ravdin, J. I., P. Stanley, C. FeMurphy, and W. A. Petri, Jr. 1989. Charactenzation of ce11 surface carbohydrate receptors for Entamoeba histolytica adherence lectin. Infect Immun. 57:2179-86. Reed, S., 3. Bouvier, A. S. Pollack, J. CeEngel, M. Brown, K. Hirata, X Que, A. Eakin, P. Eagblom, F. Gillin, and et al. 1993. Cloning of a virulence factor of Entamoeba histolytica. Pathogenic strains possess a unique cysteine proteinase gene. I ChInvest. 91: 1532-40, Reed, S. L., JI G. Curd, 1. Gigli, F. DwCillin, and A. 1, Braude. 1986. Activation of complement by pathogenic and nonpathogenic Entamoeba histolytica. J Irnrnunol. - 136:2265-70. Reed, S. L., J. A. Ember, D. S, Herdman, R G. DiScipio, TI E. Hu@, and 1. Gigi, 1995. The extracellular neutral cysteine proteinase of Entamoeba histolytica degrades anaphylatoxhs C3a and C5a. J Unmunol. 155:266-74, Reed, S. L., and 1. Gigli. 1990. Lys% ofcomplement-sensitive Entamoeba histolytica by activated terminal complement components. hitiation of complement activation by an extracellular neutral cysteine pro teinase. J Ciin Invest, 86: 18 15-22. Reed, S. L., W. E, Keene, and J. HoMeKerrow. 1989. Thiol proteinase expression and pathogenicity of Entamoeba histolytica J Cün Microbiol. 27:2772-7. Reed, S. L., W, E. Keene, J. H* McKerrow, and IIGigli. 1989. CIeavage ofC3 by a neutral cysteine proteinase of Entamoeba histolytica. J Tmmunol. 143: 189-95. Reed, S. L., P. G. Sargeaunt, and A. 1. Braude. 1983. Resistance to lysis by human senun of pathogenic Entamoeba histolytica. Tram R Soc Trop Med Hyg- 77:248-53. Reeves, R E., and J. M. Bischoff. 1968- Classincation of Entamoeba species by means of electrophoretic properties of amebal enzymes. J Parasitol. 54594-600. Rico, G., O. Diaz-Guerra, and R R Kretschrner. 1995- Cyclic nucleotide changes induced in human Ieukocytes by a product of axenicaily grown Entamoeba histolytica that inhibits human monocyte locomotion, Parasitol Res. 81: 158-62. Rimoldi, M. T., A. Sher, S. Hehy, C. EL Hanmer, and IL Joiner. 1988. Developrnentally regulated expression by Trypanosoma cnizi of molecules which accelerate the decay of complement C3 convertases. Proc Nat1 Acad Sci USA. 85: 193-7. Rimoldi, M. T., A. Tenner, D. Bobak, and K. A-Joiner. 1987. Clq enhances Trypanosoma cnizi internalization by human mononuclear cells. Complement. 4:2l7. Rini, J. M. 1995. Lectin structure. Annu Rev Biophys Biomol Struct. 24:SS 1-77. Rode, H., M. R. Q. Davies, and S. Cywes. 1978. Amoebic liver abscesses in infancy and childhood. S Afi J Surg. 16: 131-8. Roe, F. J. C. 1985. Safety of nitroirnidazoles. Scand J Infect Dis. 46:72-8 1. Rothsiein, E., and D. D. CIancy. 1969. Toxicity of disulfiram combined with metronidazole. N Eng J Med. 280: 1006-7. Rousell, F., and o. al. 199 1. A cytopathic effect of Trichomonas vaginaiis probably mediated by a mamose/N-acetyl-glucosamine binding lectin. ht J Parasitol. 21:9414. SaHiet, L. D., and W. J. Petri. 1991- Role of the galactose lectin ofEntamoeba histolytica in adherence- dependent killing of mammafim celis, Infect Tmmm, 59:468 1-3. Salata, R A., J. G, Cox, and J. 1. Ravdin. 1984. The killing of virulent Entamoeba histolytica trophozoites by phytohernaggfutt'nin-elicitedcytotoxic T-lymphocytes. Clin Res. 32:365A. Salata, R A., A. Martinez-Palomo, H. W. Murray, L. Conales, N. Tcevino, E. Segovia, C. F. Murphy, and J, 1, Ravdin. 1986. Patients freated for amebic her abscess deveIop cell-mediated immune responses effective in vitro against Entamoeba histolytica J Tmmunol, l36:263 3-9. Salata, R A., EL W. Murray, B. Y. Rubin, and J. 1. Ravdin. L987. The role of gamma interferon in the generation of human macrophages cytotoxic for Entamoeba histolytica trophozoites. Am J Trop Med Hyg. 3'1:72-8. Salata, R A, R D. Pearson, C. F. Murphy, and J. 1. Ravdin. 1982. The interaction of Entamoeba histolytica with human polymorphonuclear leukocytes and monocytes. Clin Res. 30:853A. Salata, R A., and JI1. Ravdin. 1986. The interaction of human neutrophils and Entamoeba histolytica increases cytopathogenicity for liver cell monoIayers. J Mect Dis. 154: 19-26, Salata, R A., and J. 1. Ravdin. 1985. N-Acetyl-D-gdactosamine-inhibitableadherence iectin of Entamoeba histolytica. II. Mitogenic activity for human lymphocytes. J Iafect Dis. lSl:8 16-22. SaIata, R A., and J, 1. Ravdin. L986, Review of the human immune mechanisms directed against Entamoeba histolytica. Rev Infect Dis. 8:26 1-72. Salem, H. H., and H. Abd-Rabbo. 1964. CLinical trials with dehydroemetine dihydrochloride in the treatment of acute arnebiasis. J Trop Med Hyg. 67: 137-4 1. Samuelson, J. C., A. Burke, and J. M. Courval. 1992. Susceptibility of an emetine- resistant mutant of Entamoeba histolytica to multiple dmgs and to channel blockers. Antirnicrob Agents Chemother. 36:2392-7. Samuelson, JeC., A. Sher, ruid J* PDCaultieId, 1980. Newly transformeci Schîstosomula spontaneously lose surface antigens and C3 acceptor sites during culture. J ~unoic124:2055-7.

Santoro, Fe,P. J, Lachmam, A. Capron, and M. Capron. 1979- ActÏvation of complement by Schlstosoma manson.schistosomuIar killing of parasites by the alternative pathway and requirement of IgG for classical pathway activation-3 TmmunoI. f23:lSSl-7, Sargeaunt, P. G., T. Fe Jackson, and Am Simjee. 1982. Biochemical homogeneity of Entamoeba histolytica isolates, especially those fiom Liver abscess. Lancet. 1: 1386-8- Sargeaunt, P. G., and JmE. Wiiams. 1979. Electrophoretic isoenzyme patterns of the pathogenic and non-pathogenic intestinai amoebae of man. Trans R Soc Trop Med Hyg. 73: 225-7. Sargeaunt, P. Ge,J. E. Wiitiams, and Je DI Grene. 1978, The differentiation of invasive and non-invasive Entamoeba histolytica by isoenzyme electrophoresis- Trans R Soc Trop Med Hyg. 72519-21. Savanat, T., P. Viinond, and N. Niiongkol. 1973. Blast transformation of lymphocytes in amebiasis. Am J Trop Med Hyg. 22:705-10. Schaudinn, F. 1903. Untersuchungen uber die Fortpflanzung einiger Rhizopoden (Vorflaufige Mittheilung ), Arbeiten der Kaiserlichen Gesundheitsamte. 19547-576. Seguin, R., and K. Chadee. I992. Entamoeba histolytica modulates TNF-alpha, IL-1 alphaheta and c-fos gene expression in macrophages. Arch Med Res. 23: 147-9- Seydel, K. B., IL L. Braun, T. Zhang, TgF. Jackson, and S. L. Stanley, Jr. 1996. Protection against amebic liver abscess formation in the severe combined irnmmodeficient mouse by human anti-arnebic antibodies. Am J Trop Med Hyg. 55:330- 2. Shaalan, M., and R P. Baker. 1970. Detection of coproantibodies in amebiasis of the colon: a prehnhry report. Am J Clin Pathol. S4:6lW. Sharma, M., K. Hirata, S. Herdman, and S. Reed. 1996, Entamoeba invadens: characterization of cysteine proteinases. Exp Parasitol. 84:84-91. Sharma, MmP., R R Rai, S. K. Acharya, J. C. S. Ray, and B. N. Tindon. 1989. Needie aspiration of amebic liver abscess. Br Med J, 299:1308-9. Simjee, A. E., V. Gathicam, H, M. Coovadia, T. FaJackson, P. Kiepiela, DeJe Pudifin, and M. Strettoa 1985. Celi-mediated immunity in hepatic amoebiasis. Tram R Soc Trop Med Hyg- 79: 165-8. SUma, T. K., S. E. Pronk, Ki HeKak, B. A. van Zanten, A. M. Berghuîs, and W. G. HoL 1992. Lactose binding to heat-labile enterotoxin revealed by X-ray crystaiiography. Nature. 355256 1-4, Soong, C. .JO, B, E. Tocian, M, D, Abd-AUa, T, F, Jackson, V, Gatharim, and J. 1. Ravdiii. 1995. Protection ofgerbils fkom amebic iiver abscess by immunkition with recombinant Entamoeba histo1ytïca 29-kilodaiton antigen- Mect Immun- 63:472-7. Soto, J., H. Sagua, A. Fuentes, and M. Raby. 1979. Tasas de prevaiencia de enteroparasitos en preecolares de la ciudad de Antofagasta, Rev Chil Peàiatr- 50:75-9. Spencer, H. C., JeJ. SuIlBran, H. M. Mathews, M. Sauerbrey, M. Bloch, W. Chin, and G. R Healy. L98 1. Serologic and parasitoIogic studies of Entamoeba histolytica in El Salvador, 19741978, Am J Trop Med Hyg. 30:63-8. Stanley, S. L., Jr. 1997. Progress towacds development of a vaccine for amebiasis. Clin Microbiol Rev- lO:63 749. Stanley, S. L., Jr., T. Zhang, D. Rubin, and E. Li 1995. Role of the Entamoeba histolytica cysteine proteinase in amebic liver abscess formation in severe combined immunodeficient mice. Mect Immun,63: 1587-90. Stein, P. E., A. Boodhoo, G. D. Armstrong, L. D. -ne, S. A. Cockïe, M. H. Klein, and R J. Read. 1994. Structure of a pertussis toxin-sugar complex as a mode1 for receptor binding. Nat Struct Biol. 1591-6. Stemberger, & 1978. [Cytolytic immtmreactions in vitro agaiast trophozoites of E. histolytica (author's transi)] ,Tmmun Lafekt- 6:7 1-8. Stern, J. J., J. R Graybill, and D. J. Dmtz 1984. Murine amebiasis: the role of the macrophage in host defense. Am J Trop Med Hyg. 33:372-80. Strachan, W* D., P. L. Cbiodhi, W. M. Spice, k H, Moody, and 5, P. Ackers. 1988- Tmmunologicai differentiation of pathogenic and non-pathogenic isdates of Entamoeba histolytica Lancet- l:S61-3- Stuiver, P. C., and T. J. L. M. Goud. 1978, Corticosteroids and beramoebiasis. Br Med J, 2:394-5, Suchak, N. G., R. S. Satoskar, and U, EL Sheth. 1962- Entamide fiitoate in the treatrnent of intestinal amebiasis, Am J Trop Med Hyg. 11:330-2. Sukov, R J., L. J. Cohen, and W. F. Sample. 1980. Sonography of hepatic amebic abscesses, Am J Roentgenol. lM:9 11-5, Tabhashi, H., H. Komano, N. Knwaguchi, NoKitamura, S. Nakanishi, and S. Naton 1985. Clonhg and sequencing of cDNA of Sarcopbaga peregrina humoral lectin induced on injury of the body wall. J Bi01 Chem. 260:12228-33. ~akemoto,T., T. Sonilci, E. Kashirvazaki, S. Mon, F. -ta, O. Taji, and E. Vexina. 1981. The human impact ofcolonization and parasite infestation in subtropical lowlands of Bolivia. Soc Sci Med- 1SD= 133-9. Tannich, E. 1998. Royal Society of Tropical Medicine and Hygiene Meeting at Manson House, London, 19 February 1998. Amoebic disease. Entamoeba histolytica and E. dispar: comparison of molecules considered important for host tissue destruction. Tram R Soc Trop Med Hyg. 92593-6. Tannich, E., F. Ebert, and R D. Horstmann. 1992. Molecular cloning of cDNA and genomic sequences coding for the 35- kilodalton subunit of the galactose-inhibitable lectin of pathogenic Entamoeba histolytica. Mol Biochem Parasitol. 55:225-7. Thompson, P. E., S. K. Graedel, C. R Schneider, W. B. Stucki, and R M. Gordon. 1968. Preparation and evaluation of standardized amoeba antigen fiom axenic cultures of Entamoeba histolytica. Bull World Health Organ. 39:349-65. Trissl, D. 1982. fmmunology of Entamoeba histolytica in human and animal hosts. Rev Infect Dis. 4: 1154-84. Trou, Ho,H. Marti, and No Weiss. 1997. Simple differential detection of Entamoeba histolytica and Entamoeba dispar in fiesh stool specimens by sodium acetate-acetic acid- formalin concentration and PCR. J Clin Microbiol. 35: 170 1-5- Tsutsumi, V., R Mena-Lopez, E Anaya-Vehzquez, and A. Bhrtïnez-Palomo. 1984- Cellular bases of expeimental amebic liver abscess formation. Am J Pathol. llR81-9 1, Urdaneta, A. B., ADV. MoriiIo, A. P. Nunez, and H. B. Albarmm 1975, Ami'biasis intestinal- Gen Orgauo de Ia Sociedad VenezoIana de Gastroenterologk 29:125-lSl. Vazquez, J., E. Franco, G. Reyes, and 1. Meza. 1995, Characterization of adhesion plates induced by the interaction of Entamoeba histoIytica trophozoites with fibronectin. Cell Motil Cytoskeleton. 32~3745. Vinayak, V. K., and P. Sharma. 1989. Kinetics of the immune responses during the course of hepatic amoebic infection-an experimental study. Trans R Soc Trop Med Hyg. 83:349-53 . Wagner-Mmslau, C., M. Reiser, and P. Lukas. 1985- Computed tomography fhdings in amebidsS Radiologe. 25597-8. Walderich, B., A. Weber, and J. KnobIoch. 1997. Differentiation of Entamoeba histolytica and Entamoeba dispar fiom German travelers and residents of endemic areas. Am J Trop Med Hyg. 57:704. Waiker, E. L., and A. W. SeIlards. 19 13, Experimental entamoebic dysentery. Philipp J Sci. 8B:253-33 1. Walsh, J. A. 1986. Problems in recognition and diagnosis of amebiasis: estimation of the global magnitude of morbidity and mortality, Rev Infect Dis. 8:228-38. Wang, W., and K. Cbadee. 1992, Entamoeba histolytica alters arachidonic acid metabolism in macrophages in vitro and in vivo. Immunology. 76:242-50. Wang, W., and K. Chadee. 1995. Entamoeba histolytica suppresses gamma interferon- induced macrophage class II major histocompatibility complex Ia molecule and LA beta mRNA expression by a prostaglandin ES-dependent mechanism. Infect Immun. 63: 1089- 94. Wang, W., K. Kelier, and K. Chadee. 1994- Entamoeba histolytica modulates the nitric oxide synthase gene and nitric oxide production by macrophages for cytotoxicity against amoebae and tumour cells. Immmology. 83:6O 1-10. Wang, W, K Keller, and K. Chadee. 1992, Modulation of tumor necrosis façtor production by macrophages in Entamoeba hÏstoLytica infection, Mect Inimun- 60:3 169- 74. Weis, W. l., and Ki Drickamer. 1996. Strucnual basis offectin-carbohydrate recognition. AMU Rev Biochem- 65:44l-73. Wiiey, D. C., and 5.5. Skehel. 1987. The structure and hction of the hemag@utinfn membrane glycoprotein of influenza vim. Ann Rev Biochem. 56:365-94. Wilson, 1. A., J*.Je Skehel, and D. C. Wïiey.1981. Structure of the haemagglirtioin membrane glycoprotein of influenza virus at 3-A resolutiofl- Nature- 289:366-73. WoKe, M. S. 1973. Nondysenteric intestinal amebiasis. Treatment with diloxanide fbroate. Jama 224: 160 1-4. Woodnifi, A.. W., and S. BeU. 1967. The evduation of amoebicides. Trans R Soc Trop Med Hyg. 61:435-9. Young, J. De,T. M. Young, L. P. Lu, JeC. Unkekss, and 2. A Cohn. 1982. Characterization of a membrane pore-forming protein fiom Entamoeba histolytica. J Exp Med. 156: 1677-90. Yu, Y., and K. Cbadee. 1997. Entamoeba histolytica stunuiates interIeukiri 8 fkom human colonic epithelial cells without parasite-enterocyte contact Gastroenterology. 112:1536-47. Zhang, T., P. R Cieslak, and S. L. Stanley, Jr. 1994. Protection of gerbils fkom amebic liver abscess by immunization with a recombinant Entamoeba histolytica antigen. Infect IMm~fl.62: 11 06-70.

Adherence of Entmoeba hÏsto&tica trophozoites to colonic mucin, epithelium and other

target ceils is mediated by the amebic Gal/GaINAc lectin. We constructed in vitro

expression vectors containing fùii-length (residues 1-1 BO), cysteine-poor (1-3 53, 1-48O),

and cysteine-rich (3 56-1 143,480-900) hgments of the gene encoding the heavy subimit

of the adherence lectin, hgl2, In vitro transcription foiiowed by translation using a

nuclease-treated rabbit reticuiocyte Iysate system was carried out- Immunoreactivity of riz

vitro translated Hg12 was confïrmed by immunoprecipitation with lectin-specific

monoclonal antibodies 1G7 and 8A3 which recogaize liaear epitopes. Protein dinilnde

isomerase-refolding of Hg12 enhanced immunoreactivity (P c 0.05) with the

conformationally-dependent mAb 3F4- Binding of PDI-refolded tùll-le~~.~Oth(P < 0.001)

and cysteine-rich (P = 0,005) Hg12 to Chinese Hamster Ovary celis was galactose-

dependent and competitively hhibited by native hololectin (ICso= 39.6ng/mL). The

cysteine-poor region (1 -353) did not bind CHO ceus, Both Ml-length (1-1280) and

cysteine-rich (356-1 143) Hg12 bound the glyconeoconjugate GalNAcci9BSAin a GalNAc-

specific manner. The smalier cysteine-rich hgment (480-900) also exhibited GalNAc-

specific binding but to a lesser extent (p<0.05) than residues 1-1280 and 356-1 143.

Neither the cysteine-poor fragment (1-480), luciferase (protein control), nor control

translation reactions (without hg12 lectin mRNA) bound GalNAc19BSA.Binding to

GalNAc~~BSAwas shown to be dependent on the concentration of GalNAcigBSAcoated

in each well or 35~-lectinadded (KD= 0.85 t O.37pM). Binding was competitively

inhibiteci by the terminal GaiNAc-containhg glycoprotein, asiaiofetuin (P c 0.005). Taken together, these data provide direct evidence that the entire cysteiae-rich region of the GaVGalNAc lectin heavy subunit is necessary and suflncient for complete carbohydrate binding-

Entmoeba hhtolyrica is the causative agent responsÏbIe for amebiasis. Infection leads to an estimated 40-50 miilion cases of amebic colitis or Iiver abscess annuaiiy- Arnebiasis is surpassed ody by malaria and schistosomiasis as a leading cause of death by parasitic disease (32). The pathogenesis of E.hisf02ytica infection involves adherence to colonic mucin (9,cytolysis of host epithelial and immune effector celis (8,9,22), and modulation of host immune hctions including: proteolysis of secretory IgA (24,3 L), complement evasion (26), and inhibition of macrophage defense rnechmisms (3).

Adherence to several celi types is mediated by the WGaINAc-specific lectin which is composed of a singie membrane-spanning 170-kDa heavy subunit (13,30) linked by disdide bonds to either a 3 1 or 35-kDa llght subunit (14,29). The 3 1-kDa isomer is thought to be GPI-anchored, the significance of which is unclear (1 5). Both the heavy and light subunits are encoded by multiple genes (15,21). InterestingIy, a homologous

Gal/GalNAc lectin is also present and expressed in the morphologicaily identical but geneticaily distinct non-pathogenic ameba, Entmoeba dispm (20).

The heavy subunit is an immunodominant amebic surface protein and is recognized by antisera fiom patients with invasive disease (16). Monoclonal antibodies generated against the heavy subunit have been reported to both inhibit and enhance adherence, possÏi1y owing to conformatiod regdation ofiigand interaction (19,28).

Epitopes recognized by adherence-Znhi'bitory mAbs map to the cysteine-rich segment

(residues 596-1 082) of the heqsubufiif, suggesting induectly that the carbohydrate bInding domain(s) LieCs) within thÏs region (1 1). Others have suggested that a sugar- binding domai.Lies within the pseudorepeat region (436-624) (IO). BÏnding studies using amebic membranes with glyconeoconjugates show that the WGaiNAc lectin probably relies on subsite and subunit multivalency in order to achieve avid adherence (1)-

In the present study, we used the cDNA encoding hg12 in order to construct in vitro expression vectors. Full-length (FL; 1-1280) and cysteine-rich (CR; 356-1 143 and

480-900) Hgi2 were translated in a ceU-fiee system, shown to be immunoreactive with lectin heavy subunit-specinc monoclonal antibodies @Ab), and PDErefolded ïnto a more

"native" conformation. Using this approach, we directly demonstrate GaVGaiNAc- inhibitable binding by the cysteine-rich region of the heavy subunit.

2.3.1 Strains and culture condition

Axenic E. histolytica strain HM1 :IMSS (ATCC 30459; Arnerican Type Culture

Collection, Rockville, MD), were gmwn in TYI-S-33 medium supplemented with

100U/mL penicillin and 100mg/mL streptomycùi sulfate (Life Technologies,

Gaithersburg, MD) as defined by Diamond (6)- E-histolyticaGaUGalNAc lectin and anti- lectin monoclonal antibodies were obtained as previously described ((1 8,25)- CHO cells were grown in Dulbecco's MEM supplemented with 10% FCS and 100mg/mL gentamicin (aü obtauied fkom Life Technologies, Gaithersburg, MD) in 75 cm2 plastic tissue culture flasks (Cornkg Costar, Cambridge, MA), CeiIs were harvested with 0.25% trypsin in Dulbecco's PBS (DPBS) without ca2+and MC(LX~ Technologies,

Gaithersburg, MD).

23.2 In vitro expression vectors and constnicts

Olïgonucleotides KK79 (5'-ACGT TCT AGA TTA AAT ATC TTA TTA TTA TGT-3') and KK80 (5'-ACGT GTC GAC TTA TCC ATT GTA AGT AGC TGC-3') were synthesized correspondhg to the 5' and 3' ends of hg12 (Genbank L00636) in order to generate the Mi-length (1-1280) heavy subunit which contaias a portion of the signal peptide. The sense and antisense oligos were designed with restriction sites for XaaI and

Sa& respectively. A 3-8kb fhgment was amplined fkoom HM1 :IMSS genomic DNA ushg high-fidelity TaqpLus DNA polyrnerase (Stratagene, La loua, CA). The arnpiîfïed product was digested with XbaI and Sa4 cloned into modified pGEM-42 (Promega Corp.,

Madison, WI) containhg the AMV UTL and SP6 promoter upstream of the koaI site

(pGEM-42-FL). In similar fahion, the following constructs were generated: (i) Plasmid construct pGEM-42-CP (1-353) using PCR sense primer KK79 and anti-sense KK114

(5'-ATG TGG TCC AAG TGA AAC AAG-3'). (ii) Plasmid construct pDISP-CP (1 -480) using sense primer DP50 @'-CGC CCG GGA TGA AAT TAT TAT TAT TAA ATA

TG) and anti-sense primer DPSl(5'-GAC CTG CAG ATT AGC TTT TTG TTT ACA

TAC). (iii) Plasmid construct pDISP-CR2 (480-900) using sense primer DP48 (s'-CGC

CCG GGT GTA AAC AGA AAG CTA ATT G)and anti-sense primer DP49 (5'-GAC

CTG CAG ATA TGC TTC TTT GTA TGC CTC). (iv) Plasmid construct pcDNA3 CR1 (356-1 143) was generated by restriction digestion of in-fhme EcoRI sites at 1070 and

3428 in hg12- Fùefly luciferase mRNA (Promega Corp.) was used as a control for translation experhents. PIasmid DNA was purilied fiom WormedE- coli @H5a straïn) using QlAEilter columns (Qiagen, Chatsworth, CA). PIasmid constructs were

Iinearized withBaI CpcDNA3) or San (pGEM-4Z, pDISP) for in vitro transcription.

233 In vitro transcription and translation

Sufi- or XbaI-linearized fidi-length and cysteiae-rich coLlStNcts were phenoVchloroform extracted and ethanol precipitated, foiiowed by resuspension in DEPC-treated water.

Linearized plasmid DNA (100-800 ng) was added to in vitro transcription reagents as per manufacturer's recommendations (Promega Corp., Madison, WI). T7 or SP6 RNA polyrnerase transcription was carried out at 37OC for 1 hour. RNA concentration was approximated by 1% agarose gel electcophoresis. RNA transcrïpt was added directiy to the rabbit reticulocyte lysate system reagents as outhed by the manufacturer (Protocols

& Applications Guide, 2nd ed., Promega Corp., Madison, WI). Translation was carried out in the presence of ~-[~'~]~~steineor c5~]~ethionine ([CN Pham., Costa Mesa, CA) for 1 -5 hours at 30°C. Translation products were analyzed by 7.5% SDS-polyacrylamide gel electrophoresis. Gels were fked for 30min in 20% methanol/lO% acetic acid followed by the organic scintillant EN~HANCE(NEN Research Products, Boston, MA) for 45 min, rinsed for 1 min. in distilled water, dried and exposed to HypefilmTM

(HyperfihTM,Amersham Life Science)- 23.4 TrichIomacetic acid (TA)precipitation of in vihtranslated proteins

En order to detemilne the efficiency of translation for each transcript, in vitro synthesized proteins (2&) were added to GF/A giass fiber filter circles (Fisherbrand, Pittsburgh, PA), dried, placed in 5mL scintillation fluid (CytoscintTM) and counted (LS 1701, Beckman

Instruments hc., Mississauga, ON) in order to determine the Total RadioactiMty per

Reaction- Simiiarly, 2pL of translated protein was added to LN NaOH/2% m2(248p.L) for 10 min. at 37'C in order to hydroIyze amino acyf tRNAs present in the lysate. At the end of 1O&. incubation, 1mL of ice-cold 25% TCA/2% casamino acids were added

(30min, on ice) to precipitate the synthesized protein. The sample was then added to a

GF/A glass fiber fXter circle @re-wetted with ice-cold S%TCA), washed with 3 mL of ice-cold S%TCA (3x) and rinsed wïth ImL acetone (lx). Filter circles were dried at 7S°C for 1 Ornin, and added to scintiliation fluid vials and counted (Total hcorporated

Radioacitvity pet Reaction). Translation efficiency for each protein was deterxnïned with the foliowing formula: % INCORPORATION = (Total hcorporated Radioactvity per

Reaction /Total Radioactivity per Reaction) x 100%. These data were used to standardize equivalent counts of incorporated radioactivity for the various synthesized proteins used in the binding experiments. The amount of 3S~-lectinsynthesized was determined based on the specific activity of the ~-[~*~]~~steine(1075 Ci/ mmol) or MethiOnine (1 175

Cil mmol) incorporated during in viîro synthesis. The number of cysteine or methionine residues present in 1-1270 FL (98 cysteines, 19 methionines), 356-1 143 CR1 (84 cysteines, 10 methionines), 480-900 CR2 (46 cysteines, 6 methionines), 1-480 CP (25 cysteines, IO methionines) and firefly luciferase (4 cysteines, 14 methionines) were used to determine amounts of protein translated for experiments perfonned with either L- [35~l~ysteineor fSs]~ethionine labelling. The procedure outhed here is descriid in the Protocols and Applications Guide, 2nd ed, (Promega Corp., Madison, w-

2.3.5 Treatment with Protein Disuffide Isomerase (PD9

PD1 was purchased fiom Panvera Corp. (Madison, WI). Typicaily, a 40 pL aliquot of the in vitro translated protein was treated with 20 pL of 0.2mM oxidized giutathione ,20 p.L of 2 mM reduced glutathione ,and 20 pL of 20 mg/mL of PD1 in 200mM sodium phosphate [email protected]) at room temperature for 15 min. For a negative control, additional afiquots of translated protein were treated with reduced glutathione, oxidized giutathione, and buffer (no PDI).

2.3.6 Immunoprecipitation of in vitro translation products

A panel of mAb directed agdthe lectin heavy subunit (25) was used to irnmunoprecipitate both M-length and cysteine-rîch translation products. Recombinant

Protein G Sepharose beads (Zymed, South San Francisco, CA) were washed 3x in l5OmM NaCl, ImM EDTA, 0.5% Triton X-100,0.5% BSA, lOmM Tris buffer pH8.O

(NETTfBSA). The beads were then incubated with 5pg 1G7 (IgG2b), 3F4 (IgGl), SB8

(IgGl), 7F4 (TgG2b), 8A3 (IgGl), 8C 12 (IgGl), CLB (IgGl), or VM58 (IgGl) for 30 min. Equivalent counts of in vitro translated 35~-cys-labelledfull-length and cystehe-rich

Hg12 were added to the various Protein G Sepharose-coupled mAbs and incubated for 4-6 hours at room temperature and rotated end over end. The beads were centrifitged (16000 x g, lmin), the supernatant discarded, and then washed in the following sequence: 1 x

NETT + BSA (OS%), lx NEIT, 1 x NETT + NaC1 (OS%), and 1 x NETT. The pelleted beads were added to an equal volume of lx sodium dodecyl sulfate (SDS) Laemmli sample buffier, boiled for 10min with vortex mixïng at 5m.bFoiiowlng centrifbgation

(16000 x g, le),the supernatant was loaded on to a pre-poured 7.5% polyacryIamide gel (SDS-PAGE) and analyzed by fhorography as above.

2.3.7 Immunoprecipitation of native Iectin heavy subunit

10' trophozoites at stationary phsae were harvested following 3 days growth in Y[-S medium. Tropozoites were washed 3x with PBS and resuspended in 500 pg/mL of Sulfo-

NHS-Biotin (PIERCE, Rockford, IL) for 30 min. at room temperature in order to biotinylate ali surface antigens with primary amines. Ceiis were then washed 3x with PBS to remove unbound biotin and cmde plasma membranes prepared according to the method of Aley et al (2). Extracted membranes were solubilized in trophozoite solubilization buBer as described by Petri and Schnaar, with the exception that 2mM

PMSF (Sigma Chemical Co., St.Louis, MO) and 50pg/mL ttaas-epoxysuccinyl-L-

IeucyIamide(4-puanidino)-butane (E-64)(Sigma Chemical Co.) were used for protease inhibitors in order to reduce degradation of the 170-kDa heavy subunit (1 7). Solubüized membranes (100 pg) were added to 20 pg of monoclonal Ab (1G7,3F4or CLB) coupled to Protein G Sepharose beads as described earlier. hunopreciptation buffer used was

NETT + 0.5% BSA containing 50 pg/mL E-64and 0.4m.M PMSF. Antibody-coupled beads and solubilized membrane fraction were rotated end-over-end overnight at 4OC in immunoprecipitation buEer TT + 0.5% BSA containiag 50 pg/mL E-64 and 0.4mM

PMSF). Beads were centrifbged (16000 x g, 1min), the supernatant discarded, and then washed as outüned above. Pelieted beads were added to an equai volume of 1 x SDS

Laemmli sample buffer, boiled for 10mÎn with vortex mixing at 5 min. Following centrifugation (16000 x g, 1 min), the supernatant was rm on a pre-poured 7.5% SDS-

PAGE and transferred onto nitrocellulose (Life Technologies). Blots were blocked in 5% milk proteïn/Tris-buffered Saline containhg 0.05% Tween-20 (ïTBS) for 2 hrs at room tempera-, incubated with 15000 dilution of avidin-HRP in 2% milk proteinlTIBS for

1hr (BIORAD), washed 3x with TTBS for 5 min- and 1x with TBS for 7 min, and developed using ECL reagents (Amersham Life Science).

2.3.8 Flow cytometry

HM1 trophozoites (1 x 10' ) at stationary phase were harvested following 3 days of growth in YI-S medium. Trophozoites were peileted and washed 3x with PBS and resuspended in 0.5% BSA/PBS contaùung 1:100 dilution of monoclonal antibody (1G7,

3F4,7F4, or isotype control CLB) fox 1 hr on ice. Cells were washed 3x with PBS and incubated with a 1: 100 dilution of goat-anti-mouse-fluoroscein isothiocyanate in 0.5%

BSMBS (Biorad Iab., Hercules, CA) for 30 min, on ice. Cells were washed 3x with PBS and finally resuspended in 2% padormddehyde/PBS until analysis by flow cytometg-

(Coulter EpicsQ Elite@ ESP) for fluorescence intensity.

2.3.9 CHO binding studies

CHO cells were trypsinized and resuspended in MEM, centrfiged (1000 x g for 3 min) and washed 2x in DPBS. The ceii pellet was resuspended in either DPBS, DPBS + 55 mM glucose, DPBS + 55 mM mannose, DPBS + 55 rnM D-galactos, or DPBS + pWed native lectin (I,10, or 50 nglmL). AU sugars were obtauied fiom Sigma-AIdrich

(Oakville, ON). The final suspension of celis was counted on a hemocytometer and concentrations were adjusted to 1o6 celis/mL. 5 x 1O' cells were aliquoted into 12x7Smm polystyrene culture tubes. PDI-treated in vitro translated protein (100,000 cpm; dete-ned as descriid above) was added and incubated with the ceus at 4OC for 60 min.

After incubation, the ceils were washed 2x in DPBS, DPBS + sugar, or DPBS + native lectin. The washed ceiis were layered onto a 4: 1 mixture of siLicon oil (Accumetnc,

Elizabeth, KY) and minerai oil (Sigma-Aidrich) in a 1.5 rnL microfige tube. The cells were pelleted at 9000 x g for 1 min. The tips of the microfuge tubes containhg the ce11 pellets were cut off and dropped into lOmL of scintillation fluid (CytoScintTM,ICN

Pharm) and counted. Absolute counts were at least 30x higher than background. AU bhding experiments were carried out in duplicate or trïplicate and presented as mean + standard error (n=3). AU CHO celi experiments were performed by Dr. Peter Wan

(University of Toronto).

2.3.10 GaiNAc-BSA binding studies

The glyconeoconjugate GaINAci9BSA (Sigma-Aldrich,) was resuspended in PBS at

lmg/mL. 50 per weii of Ga[NAci9BSAwas added to Immuion@ 2 Removawell strips

(Dynatech Laboratories, hc, Chantilly, VA) and incubated at room temperature for 2 hours. The unbound glyconeoconjugate was then pipetted and discarded. Control wells were coated with 50 pg of BSA (Sigma-Aldrich, Oakviiie, ON) in the same way. Equal counts (1 00,000 cpm) of 35~-labelledin vitro translated and PDI-refolded Lectin heavy subunit were resupended in 0-1% BSA-PBS brrffer, then added to GaNAclgBSA and

BSA (controf weii) in paraiiel and lncubated for 2hrs at room temperature- The unbound lectin was then discarded and the weiïs washed 3x with PBS-The wells were snapped off each of the RemovaweU strips and added to scintillation vials (5 mL, CytoScintTM) and counted (cpm with a% error per reading), GaINAc-specific binding (cpm) = Lectin bound to GaINAcisBSA(cpm) - Lectin bound to BSA (cpm)-

For inhibition experiments, equai amormts of 35~-~ectuiwere pre-incubated with asiaiofetuin (25 or 50 CIM) or the irreievant glycoprotein holotransferrin (50 p.M) and incubated for Ihr at room temperature with end over end rotation. Lectin plus inhibitor were then added to glyconeoconjugate- or BSA-coated well and analyzed as above.

Radiolabelled protein added in each bindhg expriment was standardized by TCA precipitation. SDS-PAGE of the same proteins was carried out in order to finther ensure specific translation. AU MAC-BSAbinding experiments were perforxned at Least 3 times in duplicate or triplicate-

2.3.11 Densitometric Analysis

Ail films were photographed (Gel Documentation System, U-V.P. Inc., San Gabriel, CA) and analyzed using software by Scion Image (Scion Corp., Fredenck, MD). 2.4.1 Immunoprecipitation of fiill-Iength and cysteine-rich HgU

Both full-length oand cysteine-rich (CR) Hg12 were translated in viîro using a rabbit reticulocyte Lysate system. FL Hg12 (1- 1280) migrated at -1 700kDa, CRI Hg12 (3 56-

1143) at -1 10-kDa, and CR2 Hgl2 (480-900) at 4-kDaby SDS-PAGE under denaturing and reducing conditions. ui order to ensure that Iectin translation products were immunoreactive, FL, CRI, and CR2 Hg12 '*s-labelled proteins were immunoprecipitated with a panel of anti-lectin mAbs. Equai amounts of radiolabelied protein were added to Protein G sepharose-coupled mAb, CO-incubated,washed, and bound protein run on a 7.5%polyacrylamide gel under both reducing and denaturing conditions. FL, CRI, and CR2 were immunoprecipitated with 1G7 and 8A3, and to a lesser extent with 3F4 and 7F4 (Fig. 2.1 shows data for CR1 Hgl2). h agreement with previous findings (12), the strong immunoreactivity of 1G7 and 8A3 with the translated proteins suggests that they recognize linear epitopes, whiie mAb 3F4 and 7F4 are Eely conformationally4ependent. In order to support this conclusion, unfixed HM1 trophozoites expressing the GaVGaINAc lectin were shown to be immunoreactive with

3F4 but not an isotype control CLB (Fig, 2.2). Moreover, the heavy subunit of the 1ecti.n was specifically immunoprecipitated by 3F4 under native conditions, suggesting that 3F4 recognizes the heavy subunit when conformationally in tact on the trophozoite (Fig. 2.2).

The observed 170-Baband was confirmed to be the Iectin heavy subunit Western immunoblot (data not shown). Figure 2-1 Tmmrnoprecipitation of in vitro synthesized cysteine-rich (CRI) region (3 56-

1 143) of 35~-~abelledHgE CR1 Hg12 was immunoprecipitated with a panel ofanti-lectin mAbs lG7 (IgGZb), 3F4 (IgGl), 7F4 @Wb), 8A3 (IgG1) and irrelevant isotype VM58

(lgG1). In order to improve detection of in vitro translated Hg12 by putative confonnationaiIy-dependentmAb 3F4, transIation products were treated with the folding catalyst protein diSulfide isomerase @DI) in the presence of giutathione redox buffer- The thioredoxin PD1 has been shown to promote native painng of cysteïnes required to re-fold several proteins (5,23). Consistent with proper refoldhg, PDI-treated FL and CR Hg12 demonstrated reproducibly improved immunoreactiviity with 3F4 @ c 0.05 by ONE

WAY ANOVA; n=3; based on densitometrïc analysis) as compared to no PD1 treatment

(Fig- 2.3 shows data for CR1 HG). The same experiment performed with an isotype control mAb CLB showed no merence in immunoreactivity (Fig. 2.3)- Furthemore, mAbs 1G7 and 8A3, reported to recognize Linear epitopes (12), showed no Herence in immunoreactivity with and without PD1 treatment (data not show)-

2.4.2 Binding of full-length and cysteine-rich Hglt to CHO cells

CHO binding studies were penormed with PDI-refolded '*s-labelled Hgl2. TCA precipitation was used to quantitate radiolabel incorporation and standardize each protein synthesis in order to ensure that ciifferences in binding were not due to discrepancies in translation efficiency. Consistent with the established carbohydrate specincity of the amebic lectin (23)' FL and CR1 Hg12 bound tu CHO cells with binding specifically inhibited (60 t 4.5%; P 5 0-005; Student's t-Test) by 55mM galactose but not by 55mM glucose or 55mM mannose. In con- radiolabelled protein outside the cysteine-rich domain (1 -3 53) did not exhibit galactose-inhibitable binding- As Merevidence for specificity, native purified hololectin (260-kDa) significantiy competed binding of FL

Hg12 with an ICso of 39.6 nghL (Fig 2.4). 2.43 Binding of hii-length udcysteincrich Hgi2 to GaiNAc-BSA

To confirm and extend these ITnduigs, 3S~-labeLledFL (1-1280), CRL (356-1 I43), CR2

(480-900), and CP (1-480) Hg12 were PDkefolded and added to microtiter plate welis coated with the glyconeoconjugate GaINAci9BSAorwith BSA alone. As with CHO cell experiments, TCA precipitation was used to quantitate radiolabel incorporation for the various constructs and standardlze each radiolabelled protein for binding studies-

GaiNAc-specinc binciing was observed for FL, CRI, and CR2 Hgi2 but no sigdicant binding was observed for CP HglZ the irrelevant protein luciferase, or control translation reactions pe~omedin the absence of exogenous mRNA. (Fig. 2.6). Similar redts (to those in Fig. 2.5) were obtained for experiments performed with either ~-[~*~]~~steineor

35~-Methionine, suggesting that TCA precipitation was an effective method of standadizing radiolabel incorporation for the difFerent proteins (Fig. 2.5). CR2 (480-900) exhibited lower maximal binding @<0.05, n=3) compared to EX (1-1280) and CR1 (356-

1 143) which bound equally well to GaiNAc-BSA. Binding of CRI Hgl2 to

GalNAcc19BSAwas signincantly inhibited by 25 ph4 and 50 pM asialofetuin (53.1% and

64.8% inhibition, respectively) but not by equimolar amounts of an irrelevant glycoprotein (holotransferrin at 50 pM) (Fig. 2.6). Figure 22 Immunoreactivity of3F4 wÏth the native lechheavy subunit- (a) Indirect

immunofluorescence of HM1:IMSS trophozoites by flow cytometric anaiysis. PrUnary

ad-lectin monoclonal antibodies (mAb) 1G7,3F4,and 8A3 (CLB was the isotype

control used to mesure background) were followed by secondary goat-anti-mouse-

flouroscein isothiocyanate (FITC), Data rIepict fluorescence intensity on a logarithmic

scale for 104 trophoPtes hma cepresentative experiment (b) ,)unoprecipitation of

biotinylated native lectin heavy subunit (170-kDa) corn solubilized amebic membranes

using anti-lectin mAb 3F4 (IgG1) and inelevant isotype control CLB (IgG1).

Tmmunoprecipitates were nui on 7.5% SDS-PAGE, transferred to nitrocellulose, and probed with avidii-HRP.

3F4

1G7

CLB

b 3F4 CLB 170 kDa Figure 2.3 Tmmunoprecipitation of PDI-treated 35~-~abe~edcysteine-rich Hg12 (356-

1 143)- CRI Hgi2 (1 IO-kDa) was immunopreclpitated m-th and without PDI-treatment by the anti-Lech mAb 3F4 OgG1) and the irrelevant isotype CLB (igG1). Figure 2.4. ~uiduigof 35~-~ecthCR1 Hg12 to CHO ceils Cornpetitive inhi'bition of35~-

IabeUed CR1 Hg12 binding to CHO cells by native hololectul @Cso= 39.6ng/mL ). Data are presented as mean f SE for duplicate determinations (data kindly provided by Dr.

Peter Wan, University of Toronto). Figure 2.5 Bïnding of3'$-~etand 35~-~yslabeiied CR1 Hg12 to GaiNAclcisBSA.In order to demonstrate that observed differences in bhdlng were GalNAc-specIfic and not due to aitered efficiency in radiolabel incorporation or translational efficiency, CR1 (356-1 143;

84 Cys, 10 Met) and CP (1-480; 25 Cys, 10 Met) were synthesited in the presence of 3S~-

Met or 3S~-~ys.NO signincant Werence PO.05, n=2, ONE WAY ANOVA) in binding was observed for each 3S~-~et-labelledor 35~-~ys-~abelled protein when equivaient counts were added. Figure 2.6 Binduig of 3S~-labelledHg12 (FL, 1-1280;CP, 1-480;CRI, 356-1 143; CE.

480-900) to WAC~~BSA,(a) GaiNAc-specific bindlng of FL, CP, CRI, and CR2 KgIî, lucifierase, or samples fiom control translation reaction without tectln mRNA added-

GaNAc-specinc binding (cpm) = 35~protein bound to GaiNAcL9-BSA(cpm) - 35~ protein bound to BSA (cpm). Equivalent counts of 3S~-labelledprotein used in each corresponding binding experiment were anaiyzed by SDS-PAGE and fluorography in order to ensure specific in Mh.0 synthesis (shown below eachcorresponding bar). CR2

Hg12 binds to Lower maximal activity compared to FL or CR1 Hg12 (*p<0.05) (b)

Inhibition of GaiNAc-specific binding of cysteine-nch Hg12 to GdNAci9BSA with asialofetuin (ASF) and holotransfemin (hT) at 25 or 50p.M- Data are mean t SE fiom 3 separate experiments performed in duplicate. Statisticai signifïcance shown ushg a ONE

WAY ANOVA test (*P

35~-labe~edprotein (100,000 cpm) as a fiuiction of GalNAcL9BSAconcentration for CR1

Hg12 (ckcles) or Luciferase (triangles). Non-speclfic binding of CR HglZ to BSA is ais0 shown (squares). (b) Specinc binding of 35~-~~~gL2(90.54mCilnmoL) to 50pg of

GdNAcigBSA as as a fùnction of the amount of radiolabelled lectin added to each weii.

GaINAc-specinc binding (cpm) = '*s protein bound to GalNAccis-BSA(cpm) - "S protein bound to BSA (cpm).The herepresents the least squares fit of the data to a rectanguiar hyperbola with the K' (0.85 + 0.37pbQ and B,, (18 125 + 2103 cpm). Data presented in (a) and (b) are mean f SE for duplicate determinations Binding of '*s-labened CR ~gl2was dependent on the concentration of GaINAclgBSA

coated in each weIl and was sahirabIe Pig. 2.6). Background binding to BSA was not

dependent on the concentration of BSA coated Furthenn~re~the irrelevant control

protein luciferase fded to bind GalNAci9BSAin concentration-dependent manner (Fig.

2.6). In order to estimate the affiniîy of bindingy"s-CR Hg12 (9054rnCi/nmol) was

added in a dose-dependent manner to GaINAc-BSA and BSA (background control).

Specifïc bhding was detemiined and the least squares fit of the data to a rectanguiar

hyperbola was used to calculate the KD (0.85 & 0.37pM) and B,, (1 8 125 f 2102cpm)

(Fig*2.6).

An understanding of the molecular basis for amoebic adherence may provide

insight into the pathogenesis of invasive disease and a possible basis for the design of

interventions that would disrupt amoebic attachment and cytolysis. For iostance, delineating the carbohydrate binding capacity of the E-histolyrca GaVGaiNAc lectin may

facilitate the rational design of recombinant protein or DNA vaccines (33). To date, there has been no direct evidence suppocting the assumption that the heavy subunit of the

GaVGalNAc lectin contains the carbohydrate binding domain(s). in this study, we have

localized the carbohydrate binding domains of the E- histoZyt&caGaVGaNAc lectin to the cysteine-rich region (356-1 143) of the heavy subunit This is the fist direct evidence that the cysteine-rich region of the 170-kDa subunit is sufncient for high carbohydrate binding. The in vitro synthesized GaVGaïNAc Iectin heavy subuoit was speCpfr.caLiy recognized by ad-lectin mAbs that bind hear epitopes (1G7 and SM). This is in agreement with previous observations that 1G7 and SA3 recognize bacteridy-produced recombinant hgrnents of the 170-kDa subunit when reduced and denatured (12). PD1 treatment of the transiated heavy subunit impmves its irnmmoreactivity with the putative conformationaily-depeiadent antÏ-lectin mAb, 3F4, suggestïng that it has been folded into a more "native" conformation. The folduig cataiyst PD1 6as been previously used to achieve bctional "natiive"proteins following in vitro synthesis (7,27). Aithough not tested here, it is possible that PD1 treatment makes no ciifference in binding activity since others have recently shown -Ac-specifk binding of glyconeoconjugates by lectin transferred to PVDF membranes following SDS-PAGE @r. W. Petri Jr, personal communication). Binding studies ushg 35~-labelledFL and CR Hg12 indicated that millimolar concentrations of the monosaccharide galactose was able to specificaiiy inhibit binding to CHO cells. Moreover, purined native hololectin competed biriding of 35~- lectin to CHO cells in a dose-dependent fashion. The hgment containùig residues 1-353 failed to bind CHO cells in a galactose-dependent manner. These results suggested that

PDI-refolded FL and CR Hg12 are fiinctionai.

In order to confirm and extend these observations, a more potent multivalent ligand was evaluated. Previous studies using glyconeoconjugates have shown that they bind E. histolylica membranes with high aff?nity (Ko=lO+/-3nM) and this is likeiy due to the clustered arrangement of the multiple GalNAc residues present on the glyconeoconjugate (1). G~NAc~~BSAwas able to mhiiit binding ofE.hiktolytica membranes to the same radiofigaud at 200,000-fold lower concentrations tfian GaiNAc alone (1)- Here we show that în *tro syntheslzed FL (1-L280), CR1 (356-1 L43) and CR2

(480-900) HglS bind to G~WAC~~BSAin a -Ac-specific manner. Equivalent counts of radiolabeiIed fiirefly luciferase or sarnples fiom control translation reactions (no added lectin rnRNA) synthesized in identical fashion did not bind GaiNAci9BSABIndlng of FL and CRI Hg12 to the glyconeoconjugate occurred with the similar maxhaI activityr suggesting that CR2 (356-1 143) contaîns aii carbohydrate binding activity. Kowever, the smaller CR2 (480-900) fhgment bound with lower maximal acitivity, suggesting that residues between 356-480 anaor 900-1 143 contribute to high aEtycarbohydrate binding. Binding of CR1 was concentration-dependent, appeared to be saturable, and was specifically inhibited by asialofetuin, The terminai GaiNAc-containing glycoprotein asialofetuin @Cso-3 0p.M) binds to E. histolytica membrane-bound lectin but with lower afEnity than glyconeoconjugates like G~INAC~~BSA(ICs0=0.005pM) (1)- Additionally, commercial preparations of asialofetuin are often incompletely desialylated and may therefore contribute to incomplete inhibition by this compound- The bindiig aflkity seen for CRI Hg12 (KD= 0.85 f 0.37pM) was of similar order of magnitude to that observed for the native lectin resident on amebic membranes when bond to mucin (& = 8.2~10-

11M) -- a naturally-occurring glyconeoconjugate found in the human colon(4)- These data are consistent with discontinuous residues contained within CR1 Hg12 (356-1 143) contributing to high aff?nity sugar-binding conferred by the GaVGaiNAc lectin or multiple CRDs in the heavy subunit (356-1 143). Lotter and colleagues had shown that antisera against the pseudorepeat region (379-654) blocked adherence to a greater extent than other portions ofthe cysteïue-rich region (10 ). Hence, sugar bindhg activity present withthe pseudorepeat (3 7964) and by extension CRI-Hgi2 (356-1 143) ma.provide the additionai residues for maxima1 activity, not seen in CR2-FI@ (480-900). The experimentd strategy presented here provides a usefid alternative approach to Mer delineating the carbohydrate binding capacity of the 170-kDa subunit. Adler, P., S. J. Wood, Y. C. Lee, R T. Lee, W. A. Petri, Jr., and R L.

Sehnaar. 1995. High dhity bÏndZag ofthe Entamoeba histoiytica iecth to

polyvalent N-acetylgalactosaminides. J Biol Chem. 27O(lO):S 164-71.

Aley, S. B., W. A. Scott, and 2. A. Cohn. 1980- Plasma membrane of

Entamoeba histolytica. J Exp Med. lS2(2):39 1-404.

CampbeU, D., and K, Chadee. L997. SUCViVal strategies of Entamoeba

histolytica: Modulation of cebmediated immune responses. Parasitol.Today.

13:184-189.

Chadee, K., M. L. Johnson, E. Orozco, W. A. Petri, Jr., and J. 1. Ravdin.

1988. Binding and internaikation of rat colonic mucuis by the galactoseAG acetyl-D-galactosamine adherence lectin of Entamoeba histolytica J Infect Dis.

158(2):398-406.

Chadee, K., W. A. Petri, Jr., D. J. Innes, and J. 1. Ravdin, 1987. Rat and human colonic mucins bind to and inhibit adherence lectin of Entamoeba histolytica. J Clin Invest. 80(5): 1245-54.

Diamond, L. S., D. R Harlow, and C. C. Cunnick. 1978. A new medium for the axenic cultivation of Entamoeba histolytica and other Entamoeba. Tram R Soc

Trop Med Hyg. 72(4):43 1-2.

Gilbert, H. F. 1997. Protein disulnde isomerase and assisted protein foldïng.

J-BioKhem. 272:29399-29402. 8. Guerrant, R L., J. Bmsh, J. 1. Ravdin, JeA. Suihan, and G. L. Mandell.

1981, Interaction between Entamoeba histol~ticaand human polymorphonuclear

neutrophüs. J Wect Dis, l43(1):83-93,

9. Leippe, M., J. Andra, and H. J. Muiier-Eberhard. 1994- Cytolytic and

antibacterial activity of synthetic peptides derived fiom amoebapore, the pore-

fonning peptide of Entamoeba histolytica- Proc Natl Acad Sci U S A. 91(7):2602-

6,

10. Lotter, H., T. Zhang, IL B. SeydeI, S. L. Stanley, Jr., and E. Tannich. 1997.

Identification of an epitope on the Entamoeba histolytica 170-kD lectin conferring

antibody-mediated protection against invasive amebiasis. I Exp Med.

l8S(lO): 1793-80 1,

12. - Mann, B. J., C. Y. Chang, Je M. Dodson, L. S. Ashley, L. LeBraga, and T. L.

Snodgrass. 1993, Neutralizing monoclonal antibody epitopes of the Entamoeba

histolytica galactose adhesin map to the cysteine-rich extracellular domain of the

170-kilodalton subunit, Infect Immun, 6l(5):1772-8.

12. Mann, B. J., and L. A. Lockhart. 1998. Molecular anaiysis of the GaVGalNAc

adhesin of Entamoeba histolytica- J Eukaryot Microbiol. 4S(2): 13 S- 16s.

13. Mann, B. J., B. E. Torian, T. S. Vedviek, and W. A. Petri, Jr. 1991. Sequence

of a cysteine-rich galactose-specinc lecth of Entamoeba histolytica Proc Natl

Acad Sci U S A- 88(8):3248-52.

14. McCoy, J. J., B. JeMann, T. S. Vedvick, Y. Pak, D. B. Heimark, and W. A.

Petri, Jr. 1993. Structural analysis of the light subunit ofthe Entamoeba

histolytica galactose-specifk adherence lectineJ Bi01 Chem. 268(32):24223-3 1. 15. McCoy, JeJ., B. J. Mann, T. S, Vedvick and W. A. Petri, Jr. 1993. Sequence

adysk of genes encoding the light subunit of the Entamoeba histolytica

galactose-specific adhesin. Mol Biochem Parasitol. 61(2):325-8.

16. Petri, W. A., JE, M. P. Joyce, J. Bmman, R D. Smith, C. F, Murphy, and J.

1. Ravdin. 198%Recognition of the gaiactose- or N-ace~lgalactosamine-binding

lectin of Entamoeba histolytïca by human immune sera, Mit Immun-

55(10):232?-3 1,

17. Petri, W. A., Jr., and R L. Schnaar. 1995 Punfrcation and characterization of

galactose- and N- acetylgalactosamine-specilïc adhesin lectin of Entamoeba

histolytica. Methods Enzyrnol. 253:98-104.

18. Petri, W. A., Ji=, R D. Smith, P. H. Schlesinger, C. F. Murphy, and J. 1.

Ravdin. 1987. Isolation of the galactose-binding lectin that mediates the in vitro

adherence of Entamoeba histolytica. J Clin Invest, 8O(S): 1238-44-

19. Petri, WeA., Jr., T. L. Snodgrass, T, F. Jackson, V. Gathiram, A. E. Simjee,

K. Chadee, and M. D. Cbapman. 1990- Monoclonal antibodies directed against

the galactose-binding lectin of Entamoeba histoIytica enhance adherence. J

Immunol, l44(12):4803-9.

20. Pillai, D. R, D. Britten, J. P, Ackers, J. 1. Ravdin, and IL CmKain. 1997- A

gene homologous to hg12 of Entamoeba histolytica is present and expressed in

Entamoeba dispar. Mol Biochem Parasitol. 87(l): 10 1-5.

21. Purdy, J. E., B. J. Mann, E. C. Shugart, and W. A. Petri, Jr. 1993. Analysis of

the gene family encoding the Entamoeba histolytica galactose-specific adhesin

170-kDa subunit. Mo1 Biochem Parasitol. 62(1):53-9. Ravdin, J. I., B. Y. Croit, and R L-Guemt* 1980, Cytopathogenk mechanisms of Entamoeba hïstoIytlca- J Exp Med, 152(2):377-90-

Ravdin, J, Io,and R L. Guerranti 1982 - Role of adherence in cytopathogenic mechanisms of Entamoeba histolytica, Study with mammallan tissue culture ceUs and human erythrocytes- J CIin Invest. 68(5):1305-13-

Ravdin, J. I., and B. L, Kelsaii. 1994. Role of mucosai secretory irnmunity in the development of an amebiasïs vaccine- Am J Trop Med Hyg. 50(!5):3641.

Ravdin, J. Io,W. A. Petri, C. F. Murphy, and R D. Smith. 1986. Production of mouse monoclonal antibodies which inhibit in vitro adherence of Entamoeba histolytica trophozoites. lnfect Immun. 53(1):1-5.

Reed, S. L., JOG. Curd, 1. Gigli, F. D-Gillin, and A. 1. Braude. 1986.

Activation of complement by pathogenk and nonpathogenic Entamoeba histolytica. J Imniunol. 136(6)2265-70.

Ruddon, R W., and E. Bedows. 1997. Assisted protein folding. J.Biol.Chem.

272:3 125-3 128.

Saffer, L. D., and W. 1 Petri. 1991. Role of the galactose lectin of Entamoeba histolytica in adherence- dependent killing of marnrnalian cens. Infect Immun.

59(12):468 1-3.

Tannich, E., F. Ebert, and R D. Horstmann. 1992. Molecular cloning of cDNA and genomic sequences coding for the 35- kilodaiton subunit of the galactose- inhibitable lectin of pathogenic Entamoeba histolytica. Mol Biochem Parasitol.

55(1-2) :225-7. 30- Tannich, E., F. Ebe* and R D. Horstmann. 199 1- Primary structure of the

170-kDa surface iectin of pathogenk Entamoeba histolytica- Proc Nat1 Acad Sci

U S A- 88(5): 1849-53-

3 1. Tannich, E., H. Scholze, R Nickel, amtd R D. Horstmann, 1991, Homologous

cysteine proteinases of pathogenic and nonpathogenic Entamoeba histolytica

Differences in structure and expression. J Bi01 Chem- 266(8):4798-803.

32- Walsh, J, A, 1986, Problems in recognition and diagnosis of amebiasis:

estimation of the globai magnitude of morbidity and mortality. Rev Infect Dis-

8(2):228-3 8.

33. Zhang, T., and S. LmStanley, Sr. 1994. Protection of gerbas fiom amebic Iiver

abscess by immunkation with a recombinant protein derived fiom the 170-

kilodalton surface adhesin of Entamoeba histolytica. Wect immun, 62(6):2605-8.

Entmoeba dispar coIonizes the large intestine whereas E.hfsfoZyticanot ody colonizes but also has the potential to penetrate acmss the bowel wd. The GaüGINAc tectin has been irnplicated in attachment to and cytolysis of target ceiis by E- hÏstolyticaaSeveral genes have been shown to encode the Iectin heavy subunit, but no homologue(s) have been identined in E-dispur. A compaxison of the prhary sequence and expression of the

Iectin heavy subunit gene may prove informative in understanding the d8erentia-i vinilence between these two organisms, Here we ident* a homologue ot hgZZ in

E-dispar and demonstrate that it retains -87% amino acid sequence identity, that it is distinct to E-dispar, and that the gene is actively transcribed in the non-pathogen- The enteric protozoan Entamoeba hisfolytica is associated with invasive amebiasis and is the thkd leading cause of death fiom parasitic disease (1)- Entamoeba dispar is a rnorphologically identicai but geneticdy distinct protozoan that is considered non- pathogenic and therefore a commensal of the large intestine (2). In an effort to delineate the pathogenesis of amebiasis and as an approach to iden* novel therapeutic interventions, severai investigators have focused on characterizhg merences at the molecular level between E- histdytica and E. dispur (3).

Adherence to colonic mucins and epithelial ceus is postulated as the f%st step in colonization of the large intesthe by E. histo&tica (4). Adherence is mediated by a

Galactose/N-Acetyl Gaiactosamine-inhibitablelech (5-8). The gene felywhich encodes the heavy subunit in E. histolytica is composed of at least five genes (hgll-5) (9,

1O, 1 1, 12). hgll, hg23, hg14 and hg25 share greater identity tban with hg12 indicating that hgll. 3, 4, 5 are likely derived fiom a more recent gene duplication event (9-12)- hg24 and hg25 are, however, paaiai sequences representing approximately 15% of the estirnated full length (12). We show here that E. dipar also possesses a gene encoding the fiinctional subunit of the GdGalNAc-inhibitable lectin which is distinct fiom the genes identified to date in E. histolytica and that it is expressed in the non-pathogenic species. 33.1 Gene cloning and DNA sequencing.

PCR primers (DP8: 5'-TGGGAATTCTïCTGA GACTGTACA; DP9: 5'-ACTG

GGATCCGTTAAATATCTTATTATTATGTT)were designed based on the known sequence of the E_ histolytica gene hg12 (1 1)- These primers were used to amplean expected 3-4kb product fkom E- dispar genomic DNA (SAW760) with the high-fidelity thermostable enzyme Taqplus DNA poIymerase (16)- The 3-4 kb product was cloned and manually sequenced- PCR products were anaiyzed by electrophoresis on 1%agarose gels.

PCR products were digested with shrimp alkaline phosphatase and exonuclease prior to sequencing with Sequenase version 2.0 (Sequenase PCR Product Sequencing Kit, USB,

Cleveland, OH) (1 7). One-step cloaing of PCR produc& was achieved using p~~?I vector (TA Cloning Kit, Invitrogen, San Diego, CA). Sequence analysis revealed that the nucleotides 0-3425 contained an open reading frame (ORF). PCR @P 10: 5'-

AGAAGAATTCCCAATAATT ACACCA; KK8Or 5'-ACGTGTCGACT

TATCCATTGTAAGTCAGCTGC) was used to amplift, clone, and sequence the remaining ORF encoded by nucleotides 3425-3768. Overlapping PCR products were arnplified at the 5' and 3' ends of the gene using lectin gene-specific primers in combination with hZAP T3 and T7 primers (Stratagene) on a EcoRI-digested SAW760 genomic DNA library which served as template DNA. Sequence analysis of these products confinned the 5' and 3' sequence of the gene as weii as flanking nucleotides. 3.3.2 Southern Blot-

Southern blot of E. hisroij&' (clone 96-27) and E. dispar (clone 96-53) genomic DNA.

The E- hisrolyfica strain HM4 5MSS was cultured In axenic medium YI-S using flat- sided 16 x 125mm culture tubes (13)- E. dispur dones and sttains were cdtured xenicaüy in Robinson's medium supplemented with sterile rice starch (14). Trophozoites were harvested in Iate logarïthmic phase by chilling on ice for 5 min, centrifugation at

500 x g, and washing the pellet in PBS. DNA was subsequently extracted usuig previous methods (15). RNA was extracted using a phenoVguanidinium thiocyanate-based total

RNA extraction protocoI (RNA STAT-5om LS, Tel-Test Inc., Friendswood, TX); carrier tRNA was used to CO-precipitatetotal RNA.

3.3.3 Reverse transcription (RT) - poIymerase chah reaction (PCR).

Total RNA was extracted fiom log-phase E. histolytica (strain HM-1) and E. dispar

(clone 96-53) using a phenollguanidinium thiocyanate-based method Figure 3-4 shows that RT-PCR amplincation @P 11 : 5'-CTTGATAAACAAGATlTTTGT; DP 12: 5'-

CAGnGCATATGAATCTTCTT) of total RNA digested with DNase 1yields the expected 400bp product for both E. histolytica and E. dispar

3.3.4 PCR- restriction hgment length polymorphism (RFLP). nie presence of the Hindm site at nucleotide position 417 of dhgl2 permitted the design of a PCR strategy using consensus primers -78: 5'-TGGCACTGTTAA

ATATCTTATTATTATGTTGTC;KK1145'- ATGTGGTCCAAGTGAAACAAG) followed by restriction enzyme digestion with Hindm as a method to distinguish clones Figure 3.1. The complete pfedIfedIctedamino acid sequence of the GaVGaNAc-inhiiitabfe iectin gene dhgZ2 of E-dispar. Sequence aIignment of rlhg2 and hgl2 is shown using

DNASIS v3 -2software (Hitachi Amenca Ltd., Brisbane, CA), The cysteine-nch region runs fiom residue 3 79-12 10- DHGL2 ,AMI HGL2. AMI

DEGL2 .AMI HGLS .AMI 1260 1270 1280 1290 1300 1251 ,...... ,...... -..O 1251 and sîrains of E. dispar and E, hl'sto&tica. Al1 E_ dispar and E, hisrolytfcaclones and shains subjected to PCR amplincation remit in a 1.0 kb expected product However, digestion with HincUiI cuts the dispur PCR product into two smaüer hgments of expected size wMe the E- histolytica PCR product remains uncut-

The open reading &me comprises a gene (&gC2, &par heavy chain of the @I/GaiNAc- inhibitable lectin most homologous to hgl' which, ai the predicted amino acid level, is

86.7% identical to hgl2,84.3% identical to hgZ3,83.4% identical to hgll, 79.4% identical to hgl4, and 78 -9% to hgZ5 (Fig. 3.1). Further analysis of the predicted amino acid sequence reveals that 97 cysteines contained in the putative carbohydrate binding domain of the E. histolytica heavy chain are conserved in Dhgl2- Parentheticaiiy, the cysteine- rich region runs fiom 379-1210 in Dhgl2 versus 361-1 182 in Hgi2. However, a search for the N-X4/T motif identified only 5 cooserved Nlinked glycosylation sites of the 9 potential sites identified in Hg12. Dhgl2 aiso maintains sequence homology to CDS9 between residue 744-779 and 801-834. Of the 54 residues of Hg12 implicated in CD59 homology, 49 are conserved in Dhgiî (13). The cytoplasmic tail of this adhesin contains an epidennal growth factor rececptor tyrosine phosphorylation motif_

A Southem blot was performed with 10 pg of E. hisfolytica genomic DNA (clone

96-27) and 10 pg of E. dispar genomic DNA (clone 96-53) digested with HindIII. The digests were electrophoresed on an agamse gel and blotted onto a nyIon membrane- The Figure 32Southern bbt ofE- hrStoljtica and E- dispar genomlc DNA digested to

completion, electrophoresed, blotted and hybridized with a n~-dA~~-Iab~edprobe

(nucleotides 1-3425 of &gZ2), Lane 1: E, histoi"yricagenomic DNA digested to completion with HrndQ Iane 2: E- dispur genomk DNA digested to completion with

HincUU. Size markers are indicated in nucleotide base pairs- membrane was probed with the 3-4 kb &gL? random-prhed PCR product @rimers

DPWDP9) labelied with o-32~-d~TP(Hg. 32). The &lg2 probe clearIy cross- hybridizes with members of the hgl gene famiIy in E. histo&tica. The observed restriction fkagments for hg12 were as expected based on the known restriction map for hgll, 2, and 3 (for andysis see Purdy et al (9)). Bnefly, there are no ffindm sites in h@ly

2, and 3 (hg14 and hg5are not fûüy sequenced) and therefore ali hgments should be

Iarger than 3.8kb (the nucleotide length of hgl) as was observed (Fig. 3.2, Ime 1)-

However, dhgZ2 contains a restriction site at nucleotide position 41 7. The anticipated differences in the RFLP pattern for Hindm-digested E. dwar genomic DNA compared to E. histoiyticu are shown in Figure 3.2. Similar to E. histo~tica,the presence of at least

2 bands in agarose gel electrophoresis suggests that the GaVGalNAc-inhibitable Iectin of

E. dispur is also encoded by a gene family. As shown elsewhere, the use of CHEF gel analysis may be better able to resolve additional members of the lectin gene fdy(12).

Total RNA was extracted fiom log-phase E, histolytica (strain HM-1) and E. dispar (clone 96-53) using a phenol/guanidinium thiocyanate-based method. Figure 3 -3 shows that RT-PCR amplification @Pl 1: 5'-CTTGATAAACAAGAmGT; DP12:

5 '-CAGTTGCATATGAATCTTCTT) of total RNA digested with DNaseI yields the expected 400bp prodiict for both E. histo2ytica and E. dispo. Digestion with HindIlI confirmed that the RT-PCR product amplified fiom E. dispar total RNA was the cihg2 transcript: two smaller fragments of appropriate nucleotide length were seen for the RT-

PCR product associated with amplification fiom E. dispar total RNA and not fiom E. histolytica total RNA (data not shown). Figure 3 -3. RT-PCR of JEhistolytica and E,dispm- total RNA-Total RNA extracted ftom

1O' E. histoiytica (strain HM-1) trophozoites subjected to RT-PCR amplincation before

(lane 1) and after (iane 2) treatment with RQ1 DNase. Total RNA extracted fkom 10' E dispar (clone 96-53) trophozoites with identicai RT-PCR conditions before (lane 3) and

&er (Iane 4) RQI DNase treatment, Lane 5: negative control; Iane 6: 1kb ladder (GIBCO

BRL, Gaithersburg, MD), Figure 3.4. PCR-RFLP analysis differentiates E- hfstoC~"cufmm E. disparsparE- histoiytica and E. dispar genomk DNA was subjected to PCR with consensus primers (lanes 1-6) and foliowed by RFLP anaiysis with HindIII (ianes 8-13). Lane 1,8: clone 94-45 (Ed), lane 2,9: clone 96-53 (Ed), lane 3,l O: strain SAW760 (Ed), lane 4,11: clone 96-27 (Eh), lane 5,12: clone 94-80 (Eh), lane 6,13: strain HM-13MSS (Eh); lane 7: Ikb ladder The presence ofthe HindIII site at nucleotide position 417 ofmigl2 permitted the design of a PCR strategy usïng consensus primers (KK78:5'-TGGCACTGTTAA

ATATCTTATTATTATGTC; KR1 145'- ATGTGGTCCAAGTGAAACAAG) foiiowed by restriction enzyme digestion with Hr'ndm as a method to dlstinguish clones and strains of E. dispar and E. histolytica- Aii E. di- and E. histolylica clones and strains subjected to PCR amplincation redt in a 1.0 kb expected product However, digestion with Hurdm cuts the E. dispar PCR prodnct into two smaüer bgments of expected size while the E. histolytica PCR product rem& uncut (Fig. 3 -4).

In conclusion, dhg12 represents a single, transcribed gene encoding the heavy subunit of the GaVGaiNAc-inhibitable lectin of E. dispar. Data presented here are consistent with the presence of more than one gene encoding the heavy subunit in E. dispar. This resembles the gene family known to exist in E. histolytia. The putative amino acid sequence suggests the retention of carbohydrate-binding kdcomplement evasion activity conferred by this adhesin. Sequence divergence fiom hg12 of E. histolytica has provided a unique Hindm site in dhgl2 absent from the hg1 gene family: this site has been used to develop a PCR-RFLP strategy to distinguish E. dispar and E. histolylica that may have diagnostic appIications- 1- Braga, L.L., Niaomiyrr, H., MCCoy, JJ., Eacker, S., Wiedmec, 'î., Pham, C.,

Wood, S., Sims, P.J., and Petri, W.A. (1992) Inhibition of the complement membrane

attack complex by the galactose-specific adhesïn of Entamoeba hktolytica- J- Clin.

Invest. 90,1131-1137.

2. Chadee, K., Petri, W.A., Innes, D.J., and Ravdin, J.I. (1987) Rat and human colonic mucms bind to and inhibit the adherence lectin of Entmoeba hr'stolj#icu. J. Clin.

Invest- 80,1245-1254,

3. Chadee, EL, Petri, W.A., Johnson, M., Orozco, M.E., and Ravdin, J.I. (1988)

Binding and intemaiization of purifieci rat colonic mucin by the GaVGaiNAc adherence lectin of Entamoeba histo[ytica- J. Id Dis. 158,398406.

4. Clark, C.G., and Diamond, L.S. (1991) Ribosomai RNA genes of 'pathogenic' and

'non-pathogenic' Entamoeba histolytica are distinct. Mol. Biochem. Parasitol. 49,297-

3 02.

5. Diamond, L.S., and Clark, C.G. (1 993) A redescription of Entamoeba histolytica

Schauninn, 1903 (Emended Walker, 191 1) separating it fiom Entamoeba dispar Brumpt,

1925. J. Euk. Microbiol. 40,340-34.4,

6. Diamond, L.S., Clark, C.G. and Cunnick, C.C. (1995) YI-S, a casein-fiee medium for axenic cultivation of Entamoeba histolytica, related Enrmoeba,

Giardia intestinalis and Trichomonas vagzhaIis. J. Euk Microbiol. 42,

277-278.

7. Horstmann, RD., Leippe, M., and Tannich, E. (1992) Recent progress in the molecdar biology of Entmoeba hisfolytica. Trop. Med. Parasitol. 43,2 13-2 18. 8. Mann, B.J., Torian, B.E., Vedvick, TS*,and Petri, WA(1991) Sequence of a cysteine-rich galactose-specific lectin of Entmoeba fikto&tica- Proc- Natl, Acad, Sci-

USA 88,3248-3252.

9. Nielson, K, Scott, B., Biuer, J-C,, and Knb, K. (1994) TCI~IUS~~DNA polymerase for more robust PCR. Strategies 7,6465.

10. Petri, W.A., Smith, WD., Schlesinger, P-H., and Ravdh, JJ. (1987) Isolation of the galactose-binding Lectin which mediates in vÎîro adherence to Entamoeba hitolytica-

J. Clin. Inv- 80, 1238-1244,

11. Purdy, J.E.9 Mann, B.J., Shugart, E.C., and Petri, W.A. (1993) Analysis of the gene family encoding îhe Enfamoeba histolytica galactose-specific adhesin 170-kDa subunit. Mol. Biochem, ParasitoL. 62,53-60,

12. Ramaknshnan G., Ragland, B.D., Purdy, J.E., and Mann, B.J. (1996) PhysicaI rnapping and expression of gene families encoding the N-acetyl D-Galactosamine adherence lectin of Entamoeba hisruiytica. Mol, Microbiol. 19,91-100.

13. Ravdin, J.I. & Guerrant, RC. (198 1) Role of adherence in cytopathogenic mechanisrns of Entamoeba hÏstoLytÏcaaJ. Clin. Invest. 68, 1305- 13 13.

14. Ravdin, J.I., John, J.E., Johnson, L.I., Innes, D.J., and Guerrant, RL. (1985)

Adherence of Entamoeba histolytica trophozoites to rat and human colonic mucosa.

Infect. Immun- 48,292-297.

15. Robinson, GoLa(1968) Laboratory cultivation of some human parasitic amoebae- J-

Gen. Microbiol. 53,6949-

16. Sanger, F., Niciclen, S., and Coulson, AR (1 977) DNA sequencing with chah terminating inhibitors. Proc. Natl. Acad. Sci. USA 74, 546305467. 17. Tannich, E., Ebert, Fm,and Horstmnnn, RD. (1991) Primary stnicture ofthe 170-

kDa surface lectin of pathogenic Entmoeba histoCyrica. Proc- NatL Acad, Sci- USA 88,

1849-1 853,

18. Walsh, J.A. (1986) Problems in the recognition and diagnosis of amebiasis: estimation of the gIobai magnitude of motbidity and mortaIity Rev- Infect- Dis. 8,228-

238.

Amebiasis is a major cause of dysentery in the deveIoping world and contnaues to sigoincant morbidity in approximately 50 miIlion people worldwide. Invasion begb with adherence of Entamoeba histoCytica trophozoites to colonic mucin and intestinal epithelium, a key step mediated by the amebic GaUGaNAc lectin. Cornparison of the lectin fiom E histol'ytica (pathogenic in humans) and Entanoeba dispur (non-pathogenic in humans) was undertaken in order to elucidate the differentiai role of this molecde Ïn colonization versus invasion. Surface lectin was less abundant on axenic Bdispm when comp~lredto axenic E. histolytica, cornmensurate with Herences in lectin (heavy and light subunits) RNA by serni-quantitative RT-PCR.Sequence alignment of homologous lectin genes indicated -90% conservation at the amino acid level. In vitro translation studies showed that both Hg12 (E.hisro[ylica) and Dhgl2 (E-dispm) were Micient for high-etybinding to multivalent GaNAcci9BSA.The 1G7 epitope (mapped to residues

482-8 18) was constitutively present for native lectin on the surface ofunfixed, axenicdy- cultured E. histolytica (HM1 :MSS) trophozoites but absent on axenic E-dispm

(SAW 1734cIAR). However, fixation with methanoVacetone restored the 1G7 epitope on

E-dispar trophozoites. Furthemore, 1G7 also recognized E-dispar lectui (residues 480-

900) when transiently transfected into COS7 cells. Anti-sera nom patients with amebic liver abscess (ALA) indicated that prior ALA results in a systemic IgG response which recognized E. dispar lectin and correlated with clinical cross-protection. E- dispar infection alone is capable of elicithg a humoral response specific to E-dispar lectin (480-

900), suggesting that both shared as well as discrete epitopes are present for homologous lectins. These data, when taken together, provide evidence for molecular ciifferences between the E.hisro&tica and E-dispar GaI/GaINAc Iecth that may in part explain the avirulence of E-disparwhen compared to pathogenlc E-hrSto&ticaCT

Amebiasis is responsLbIe for 50-100,000 deaths each year (38). E-oeba histotytica is the causative agent of invasive disease redtiag in amebic colitïs or liver abscess-

Entamoeba dispar uifection redts in asymptomatic coIonization of the large intestine.

Both species are morphoIogicdy identicai but geneticaIiy distinct (l0,37). Clinical studies in both endemic and non-endemic settings have reinforced the new paradigm which proposes that E-dispar is a gut commemal, while E-hl'srolytica is responsible for both intestinal and extra-intestinal disease (14,27). Stool antigen ELISA targetting the

GaVGaiNAc lectin and which distinguishes pathogen fiom non-pathogen has indicated that E-dispar accounts for >95% of infections previously cailed E-hisloiytica in a setting of non-endemicity (26,27). Pathogenesis of E-histolytica infection leading to invasive disease requires adherence to colonic mucin, cytolysis of uitestind epitheiial and immune effector cens, and modulation of host immune f'unctions (6,8,29). In an effort to explain the virulence of E. hisrobtica, several laboratories have doned genes that encode putative vinilence factors, including lectin-like adhesion molecules, cysteine proteases, and cytolytic peptides cailed "amoebapores" (18,21,32,36).

The galactosdN-Acetyl-galactosamhe (GaüGaINAc) lectin (a 170-kDa and 3 1 or

3 5-kDa heterodimer) is expressed on the surface of E.hisro&tica trophozoites. Inhibition of adherence with anti-heavy subunit antisera or monoclonal antibodies, by denudation of terminai GaiNAc on target ce& or by cornpetition with excess GaINAc suggests that the

lectin is necessary for amichment to CEIO ceiis (19,20,30). Sero-epidemiologicai surveys

demonstrate that the GaVGaiNAc lectin is an hmu11odominant moIecule (23). AU

patients experiencing invasive disease mount a systemic humoral response against the

heavy subunit (170-ma) but not the Light subunit isoforms (3 1135-ma) of the lectin.

Various monoclonai antibodies generated against the lectin heavy subunit either ùihibit or

enhance adherence of trophozoites to target celis, suggesting confonnationai regdation of

carbohydrate afkity (20,24). BindÏng studies with fragments of the heavy subunit

suggest that one or more carbohydrate-binding domains reside within the cysteine-nch region (1 3,28). Moreover, binding of Eshistolytica membranes to glyconeoconjugates

(G~~NAC~~BSA)show that the lectin likely relies on subsite and subunit multivalency to achieve high-aff?nity adherence (2).

Several studies have identified homologous genes encoding putative virulence factors in Enrarnoeba dispar (5, 17,ZS). Since both species retain a high degree of amino acid sequence identity (40%) for homologous vinilence genes, it is possible that gene regdation or modulation of protein bction may account for dinerentid invasiveness.

We reported previously that the lectin gene (dhgl2) is present and expressed in E.dispm

(25). Whole cell studies have only recentiy been made possible since both E-histolylica and E. dispar can now be cultwed axenicaiiy (9, 16). Axenic E-dispar appears to attach and destroy target cells to a lesser extent (4x) than its invasive counterpart in GaVGaiNAc lectin-dependent centrifugation binding studies using Chuiese Hamster Ovary (CHO) cells (12). Furthermore, evidence that Edhpm is less Wulent in an experllwntai hepatic amebiasis mode1 has also been presented (M. Espinosa-Cantellano, Seminario in Amebiasis, Mexico City, 1997). However, a fiiller cornparison of E.disparfEEhistoCytica lectin gene regdation, carbohydrate aflinity, epitope structure7coaformational modulation, and E-dispar lectïn seropositivity in human disease has not been undertaken.

Here we show that homologous genes encoding the GaVGaINAc lectin share both sequence and epitope identity- However, both Lectin RNA and Wace protein expression

Ievels were Iess in E-dispar than E. histolytica when cuitured under axenic conditions.

Additionally, the cysteine-rich region (480-900), previously reported to be immunodominant in human disease, immunoprotective in animal models, and sufEcient for high afEnity carbohydrate-binding, is conformationally regulated such that E-dispar conceals portions of this domain. Patients with previous amebic liver abscess appear to be protected from subsequent Esdispar infection. We show that, following amebic liver abscess in human subjects, systemic IgG responses were able to recopize both

E- histolytica and E-dispm lectin. These data cowlated with observed cross-protection to

E-dispar infection subsequent to ALA caused by Ehistolytica infection. Interestingly, analysis of patient sera from South Aûica and Egypt indicate that, in some cases (12%),

E-dispar infection alone is able to elicit an E-dispur lectin-specific humoral response.

This fmding runs counter to current thinking and suggests that E-dispar cmbe minimally or sub-clinically invasive in a subset of individuals.

--

4.3.1 Strains and culture condition

Axenic E. histoiytica strain HM1 :IMSS (ATCC 30459; American Type Culture

Collection, Rockville, MD) was grown in YI-S medium containhg 100 mg/mL gentamicin &Xe Technologies, Gaithersburg, MD) (1 1)- Anti-lectin heavy subunit

monoclonal antibodies were obtained as previously described (3 1)- Axenic E-dispar straIn

SAW 1734cIAR and CYNO 16 were grown in YIGADHA-S mebas described

previously CM), and axe& E-dispar strain SAW760 was grown in YI-S medium

supplemented with much (9)- COS7 ceiis (ATCC CRL 1651) were grown in RPMI

supplemented with 10% FCS and 100 mg/mL gentamicin (aii obtained fiom Life

Technologies, Gaithersburg, MD) in 75 cm2 plastic tissue culture flash (Coming hc.,

Coming, NY). Ceils were harvested with 0.25% trypsin in Dulbecco's phosphate buffered

saline (DPBS) without ca2+and ~g+(LifeTechnologies, Gaithersburg, MD) foiiowed

by washes in PBS.

4.3.2 Flow cytometry

HMl:IMSS, SAW1734cIAR, SAW760, and Cm16trophozoites at stationary phase were harvested following 3 days of growth in axenic medium. For the Iectin surface expression experiments, trophozoites (1 x 103 were îïrst fixed in 2% paraformaldehyde

/PBS,washed 3x with PBS, and then stained with primary mAb 8A3 (1 :100 dilution in

0.5% BSNPBS). Ceiis were washed 3x with PBS and incubated with goat anti-mouse- fluorescein isothiocyanate (1 A00 dilution in OS%BSA/PBS) (BIORAD, Hercules, CA) for 30 min. on ice. Cells were washed 3x with PBS and fïnally resuspended in 2% parafiormaldehyde/PBS until evaluation by flow cytometry (Codter Epics Elite ESP) for

FITC staining. Mean fluorescence was quantifred by Couiter Epics Elite software. For epitope identification experiments without prïor fixation, trophozoites (1 x 103 were pelleted and washed 3x with PBS and resuspended in 0.5% BSAIPBS contamg 1: 100 dilution of monoclonal antiidy (1G7,3F4,8A3, or CLB) for L hr on ice- Secondary antibody staining and aow cytometry were perfonned as above*

4.3.3 PIasmid constructs and COS7 transféction

Plasmi-d constmct pDLSP-hgl2CR encodiag the cysteine-rich region (480-900) of the

E. histotytica lectin gene hg12 (Genbank Accession #L00636) was generated by hi&- fidelity Taqplus polymerase chah reaction (TaqplusPCR, Stratagene, La Joua, CA) using sense primer DP48h (s'-CGC CCG GGT GTA AAC AGA AAG CTA ATT G) and an& sense primer DP49h (5'-GAC CTG CAG ATA TGC TTC TTT GTA TGC CTC)- PCR products were digested with BgiIi and PstI mew EngIand Biolabs, Beverly, MA) and

Iigated into vector pDISPLAY (Invitrogen, Carlsbad, CA) for expression in COS7 ceils

(ATCC CRI, 1651). Surface expression is facilitated by heterologous signal

(immu11oglobuii.n G-kappa chain) and transmembrane (platelet derhed growth factor receptor) sequence in the pDISPLAY vector, The homologous region fiom the E-dispar lectin gene dhgl2 (Genbank accession #U737 10) was ampiified using sense primer

DP48d (s'-CGC CCG GGT GTA AAC AAA AGC CTA ACT G)and DP49d (5'-GAC

CTG CAG ATA TGC TTC TTT GTA TGC CTC) and cloned into pDISPLAY in the same way. COS7 cells were transfected using chloroquinef DEAE-dextrau as described previously (4).

4.3.4 RNA isolation and reverse transcription-poiymerase chah reaction (RT-PCR)

Total RNA was extracted fiom 10' axenicaiiy cultured E.histoZyficaand E.dispm trophozoites usïng a phenoVguanidinium thiocyanate-based total RNA extraction protocol (RNA STAT-SO~~LS, Tel-Test, Friendswood, ïTQ- Total RNA was digested with RQl

DNase (Promega Corp.) for 30 min at 37Cand qmtitated using standard

spectrophotometric anaiysis- Reverse transcription wïth alfia mosaic virus (AMV)

reverse transcriptase (Lifie techno Iogies, Gaithersburg, MD) was carried out as follows:

2.5 mM dNTPs (2.5 w),O-LM DTT (2 pL), 40,000 UfmL Rnase inhibitor (0.5 pL), 25

Ufdrandom hexamer nucleotides (2 pL), and 2 pg RNA (1 1.8 pL) were incubated

together at 70°C for 5 mh, chilied on ice, and centrifbged to coilect any condensate.

Reverse transcription 5x baer(5 pL) and AMV reverse transcriptase (1.2 pL) were then

added, the mix incubated at 37OC for 60 min., heated to 90°C for 5 min, and finally on ice

for 5 min. Sarnples were centrifuged briefly and stored at -20°C until PCR amplincation.

Consensus primers were used for PCR amplification of lectin heavy subunit and actin

(Genbank #M16339) genes. Lectin heavy subunit was amplified using sense primer DP 11

(5'-CTT GAT AAA CAA GAT TTT TGT) and antisense primer DP12 (5'-CAGTTG

CAT ATG AAT CTT C'Ti). Consensus prime15 were not feasible for lïght subunits due to sequence divergence, therefore primers were optunized for each species. E. hisfolyria lectin light subunit (Genbank #L20898) was amplified usiag sense primer DP6 1 (s'-TTC

T'TA AAC GAT AGA GAC ATC) and antisense primer DP62 (5'-CCT GTA TTG GTT

CAG TTG C). E-dispur lectïn light subunit (Genbank #U85823) was amplified using sense primer DP33 (S-ACT GGA TCC ATG CAT CTA AAT TAT TAT C) and antisense primer DP34 (s'-ACT GAA TTC TTA GGC AAA AAC TGG AAT AAA TAT

C). Amebic actin was amplined using consensus sense primer DP28 (5'-GGA TCA GGA

ATG TGT AAA) and antisense primer DP29 (5'-TCC ATC ACC TGA ATC CAT).

Amplification mix was as follows: Sterile water (34.5 pL), dNTPs (8 pL), 5x EZ baer (5 a),Taq DNA polymerase (0.5 pL), 25 pmol DPL 1or DP28 primer (0.5 fi),25pmol

DPIS or DP29 primer (0.5 pi,), and RNA (1 pL), TempIate concentration-dependent amptification was demonstrated by adding OSpt,, 1 pi,, and 2 pL of cDNA to PCR amplification reactions. Amplification was performed for 40 cycles (94OC / 1 min denaturation step, 54OC / 1 min, 72OC / 1 min extension step), foliowed by a hai extension at 72OC for 7 min. ALI temperature cycling was performed ushg Robocycler

Gradient 40 (Stratagene).

4.3.5 In vifro binding studies

Binding studies were perfonned as described previously (28). Briefly, homologous genes encoding the full-length lectin heavy subunit fiom B histolyrca (hgl2) and E. dispw

(dhgl2) were PCR-amplified ushg high4delity Tuqplus DNA poiymerase (Stratagene) and cloned into pcDNAl (InMtrogen). PCR amplification was canied out using consensus sense primer DP9 (5'-ACT GGG ATC CGT TAA ATA TCT TAT TAT TAT

GTT) and consensus antisense primer DP8 (S-TTG GGA ATT CTT CTG AGA CTG

TAC A)- Bal-linearized Ml-length constructs were in vitro transcribed and translated using a nuclease-treated rabbit reticulocyte system accordhg to the manufacturer's recommendations. Equimolar amounts ~-[~*~]~~steinelabelled in vitro translated and protein disulfide isomerase-refolded (Panvera Corp., Madison, WI) lectin heavy subunit were resupended in O. 1% BSA-PBS bufTer, then added to GalNAc i9BSA- and BSA- coated wells in parailel and incubated for 2hrs at room temperature. The unbound lectin was then discarded and the wells washed 3x with PBS. The welis were snapped off each of the Removawell strips (Dynatech Laboratories, Inc., ChantillyyVA) and added to scintillation vids (SmL, CytoScht-, ICN Ph-, Costa Mesa, CA) and counted (cpm with 4%error per reading) using a scintiuation counter (LS 170 1, Beckman Instruments

Inc., Mississauga, ON). GaiNAc-specific binding (cpm) = Lech bound to GaiNAcr9BSA

(cpm) - Lech bound to BSA (cpm)-

4.3.6 Indirect Immunotluorescence

At 24 hours post-aa~sfection,COS7 cells were passaged on to 8-chamber weil slides

(Nunc) and fixed in 50% methanof/ 50% acetone (vfv) 48 hours later. CeNs were then incubated with monoclonai anhidy 1G7 (1:IOO dilution in PBSf 0.5% BSA) for 1 hour at room temperature- Fixed cells were washed 5x with PBS, then incubated with goat anti-mouse IgG antibody conjugated to FITC (1 :100 dilution in PBS/0.5% BSA).

Following 5 washes with PBS, mounting medium (Vector Labs) was added, and a 24 x

50mm cover slip placed over the slide. Fluorescence detection was carried out using a

Nikon fluorescence microscope.

4.3.7 Immunoprecipitation

In vih-O translation followed by immunoprecipitation of radidabeled lectin was performed as described (28). Bnefiy, recombinant Protein G Sepharose beads (Zymed,

South San Francisco, CA) were washed 3x in 150 mM NaCl, 1 mM EDTA, 0.5% Triton

X- 100,0,5% BSA, 10 mM Tris buffier pH8.0 (NETT+BSA). The beads were then incubated with 5 pg 1G7 (IgG2b) or isotype control for 30 min. Equivalent arnounts

(molar) of in virro translated ~-~~~]~~ee-laedHgl2, Hg12 (480-900), Dhgl2,

Dhgl2 (480-900) were added to Protein G Sepharose-coupled mAb and incubated for 4-6 hours at room temperature and rotated end-ovr-end The beads were centrfiged (16000 x g, 1 min), the supernatant discardeci, and then washed in the following sequencer 1 x

NETT + BSA (OS%), lx NETT, 1 x NETT +NaCL (OS%), and 1 x NETT, The pelleted beads were added to an equal volume of lx SDS Laemmli sample buffer, boiled for 10 min with vortex mixïng at 5 min- Foliowing centrifùgation (16000 x g, lmui), the supernatant was loaded on to a pre-poured 7.5% polyacrylamide gel and anaiyzed by fluorography-

4.3.8 ELISA to detect anti-lectin antibodies in human antisera

COS7 cells were transiently transfected with plasmid pDISPhgl2 (480-900), pDISP.dhgl2 (480-900), and vector alone. Transfected cells were harvested at 72 hours post-transfection (checked by indirect immunofluorescence for efficiency) by washuig 3x in PBS, followed by resuspension in solubilization bufTer (50 mM Tris, 150 mM NaCI,

0.5% Nonidet P-40 (Sigma), 5 mM EDTA, 50 pg/d E-64 (Sigma), and 2 mM PMSF izife Technologies)). Mock-, Dhgl2 (480-900)-, and Hg12 (480-900) -transfected whole ce]! proteins were resuspended in 1% BSA/ 0.1% Tween-201 PBS and coated onto Costar

(Coming Inc., Corning, NY) microtitre 96-weU plates under non-reducing and non- denaturing conditions. Mole ce11 proteins (10 pg / well) were incubated at room temperature for 2 hrs. Humau antisera from E- histo&tr'ca-infecte and E-dispar-Sected patients, as well as fiom uninfected control patients were senally diluted and incubated for 1 hr at room temp (1). Wells were washed 4x with PBS then incubated with goat anti- human-HRP conjugate (1 :50,000, Sigma, Mississauga, ON). Wells were washed 4x with PBS, and incubated with 100 pL TMB (tetramethylbenzÏdineySigma) for ttp to 30 min at room temp. Reactions were stopped using 0.5 N Eiydrochloric acid and plates read at

450nm ushg an ELISA microplate reader (ThetmornmTM,MoIecular Devices Corp.,

Smyvale, CA). OD readings nom mock-traasfected weiis were subtracted fiom dhgl2- and hgi2- transfected weils, to get a specific OD reading for each titre of human antisera added. Specinc OD = OD (lectui transfectants) - OD (mock transfectants)- Samples with specific OD 2 0.1 AND 2 mean + 3 standard deviations for uninfected control anti-sera were considered positive-

4.3.9 Densitometric Anabsis

Al1 films were photographed (Gel Documentation System, U-VP. hc., San Gabriel, CA) and anaiyzed using software by Scion Image (Scion Corp., Frederick, MD).

4.4.1 Relative expression of lectin message is less in axenic E.dispar compared to

E. histoiytica

In order to quant* the relative amount of lectin (1 70-kDa and 35-KDa) RNA between axenically cultured E-dispm (SAW760, SAW 1734, CYNO 16) and E.histolytiu

(HM1 :IMSS), total RNA was extracted and subjected to RT-PCR.Consensus primers were designed to homologous lectin (170-kDa) and actin genes so that ciiffierences in amplification were due to discrepant levels of messenger RNA and not difïerences in Figure 4.1. Relative expression of lectin (170-kDa and 35-kDa) genes is greater h

E- histolytica than E-dÏspar under axenic cuiture conditions, Total RNA was extracted fkom stationary phase axenic HM1 :IMSS and CYNO L6:TPC trophozoites, quantifieci, and subject to AMV reverse transcription and amplFfTcation with Taq DNA polymerase.

Template concentration-dependent amplification was demonstrated by adding 0.5 jL

(lanes 1,4), L pL (lanes 2, S), and 2 p.L (ianes 3,Q of cDNA to PCR ampfication reactions. (A) RT-PCR amplicons are indicated for Iectin heavy subunit (1 70-kDa), lectin light subunit (35-kDa), and amebic actin. (B) Resuits of scanning densitometric andysis of representative RT-PCR gels are presented as a histogram with the corresponding Iane designation- Light subunit amplicons were seen for CYNO 16 at higher concentrations of input cDNA- Densitometric values are normallzed to the level of actin RT-PCR amplicons perfïormed under identical conditions (OS, 1, and 2 pL template cDNA).

Results are representative of 4 independent experiments. primer compati'biIity- These data were normalized to the levels of endogenous acti. message detected using the same cDNA preparation, thereby serving as an internal control(39). E. hisfolyrica Iectin heavy subunit (hg[2)message was significantly higher than E-dispar (dhgl2) using densitometrk andysis (Fig- 4.1, ONE WAY ANOVA, p4.05). Similarly, lectin Iight subunit levels of RNA were reproducibly higher in

E-hisfolytica(ZgZI) than in E.di' (CilgII).Light subunit amplicons were present at higher concentrations of input cDNA (eg. 5pL) and comparable data were obtained for

RNA extracted fiom 2 different axenicaily cultured E-dispar strains (SAW 760 or

CYNO 16) (data not shown). AI1 strains conform to the zymodeme classification developed previously (34).

4.4.2 Expression of surface hololectin is less in E-dispar compared to E.hrStolyrica under axenic conditions

In order to confirm that differences in RNA levels resulted in aitered protein expression, we compared levels of surface hololectin fiom axenically cultured E-dispar and

E. histolytica Trophozoites were collected and immunostained with anti-lectin monoclonal antibody 8A3 which recognizes both E. histolyiica and E-dispar lectin.

Monoclonal antibody 8A3 was chosen since it was able to Mmunoprecipitate the lectin heavy subunit nom E.histoijtica (Hg12) and E-di- (Dhgl2) with equal efficiency (Fig.

4.3A). Evaluation of mean fluorescence intensity by flow cytometry demonstrated that the lectin was -2.5-fold less abundant on axenic E-dispar (SAW1734) than E-hisrolyrica

(HM1 :IMSS) (Fig. 4.2). ki contrast, immunostaining ushg a separate monoclonal antibody MHM.23 that detects a P2 integrin-like antigen on the surface of Figure 42. Relative expression of native lech(260kDa) on amebic daceis pater on

E. histolytica (HM1:IMSS) than E-dispar (SAW 1734) under axeaic culture conditions.

Mean fluorescence ïntenslty + standard deviation was detemiined for 10,000 trophozoites harvested at stationary phase by flow cytometric analysis foliowing indirect immunofluorescence using anti-lectin MAb 8A3 (igG1) or isotype control8C12 (TgG1) followed by goat anti-mouse-fluoroscein isothiocyanate (FITC) fiom a single representative experiment. E-hfstolyticaand E-dispar trophozoites showed that this motecule was expressed to higher levels in E-dispar (see chapter 5)-

4.4.3 E.dWpar and E.fiiSioi&ricrr lectïn (170-Ba) are sufficient for high-affhity binding to GaiNAcWBSA

To test whether the heavy subunit of the GaVGaiNAc lectin Iiom E.di'ur was dcient to bind carbohydrate, we pedormed a microplate assay using in vih.0 translated radiolabeled lectin developed in our Iaboratory (28). We have shown previously that, using this assay, binding of the E-histolytica heavy subunit to the glyconeoconjugate

GalNAc19BSAwas GaNAc-specific (KD= 0.85 +: 0.3 7pM), saturable, and inhibitable by the terminal GalNAc-containing protein asialofetuin (28). Here we extend this observation to EAispar lectin heavy subunit. Immunoreactivity with the Iectin-specific

MAb (8A3) was confïrrned by immunoprecipitation of the in vitro translated 170-kDa subunit encoded by dhgl2 and hgZ2 (Fig 4.3A). Equimolar amounts of lectin @hg12 and

Hgl2,2.3 x 10' cpm/£inol) were added to GaNAccisBSAand shown to bind in a dose- dependent manner (Fig.4.3B). The irrelevant protein luciferase was added to

GalNAc19BSAin the same way to indicate levels of non-specific binding-

4.4.4 Epitope conservation in E.di;spar and E.krStolytica cultured under axenic conditions

Several monoclonal antibodies (MAbs) have been previously generated against the heavy Figure 4.3. Lech heavy subunÏt (170-kDa) fiom E.itisto&tica (He) and Edispar

(Dhgl2) are sufncient for high affinity GaiNAc-specSc bindlng in vitro- (A)

Immunoprecipitation of in vitro synthesized 35~-~abe~edheavy subunit (1 70-kDa) fiom

E-dispar (DhgI2). Dhgl2 was immunoprecipitated wÏth ad-lectin mAb BA3 (IgGl, lane

1) and irrelevant isotype VM58 (IgGl, lane 2). (B) Binding o~"s-H&~(2.3 x los cpdhol) and 3S~-~hgL2(2.3 x lo5 cpm/hol) to 50pg of GalNAcci9BSAas a bction of the amount of radiolabelled lecth added to each weU. GalNAc-specinc binding ofHgl2,

Dhgl2, and Luciferase, are plotted fiom a representatve experiment. GaINAc-specinc

3 5 binding (cpm) = 35~protein bound to GaiNAcis-BSA (cpm) - S protein bound to BSA

(cpm). Equivalent mounts ~f'~~-labeIIedprotein used in each corresponding bkding experiment were analyzed by SDS-PAGE to ensure specific in vitro synthesis and trichloroacetic acid (TCA) precipitation to ensure equd radiolabel incorporation (28). Figure 4.4- Immunoreactivlty of axenk E-hisfolytica@Ml -IMSS) and axenic E-dispar

(SAW 1734) trophozoites without £kation by fIow cytometric adysis. Primary anti-latin monoclonal antiies(mAb) 1G?,3F4, and 8A3 (CU3 was the isotype control used to mesure background) were foiiowed by secondary goat anti-mouse-fluoroscein isothiocyanate (FITC). Data depict fluorescence intensity on a logarïthmic sale for 1o4 trophozoites fkom a representative expriment

1G7

3F4

8A3

CLB subunit of E histolytica Iectin (3 1)- With the tecent development of axenic culture

conditions for E-dispar, we undertook to evaiuate theu cross-reactivity to Edispm lectin

(16). Tmmunostaining of u&ed trophozoites demonstrated that mAbs 3F4 and 8A3

detected both E-histoiytica (HM1:IMSS) and E-dispar (SAWl734) native lectin, whereas

1G7 oniy detected E-histdyticanative Iectb (Fig. 4-4)- Trophozoites were then fkced in

methanoVacetone and immunostained as before. Foiiowing fixation all three mAbs cross-

reacted between E.hisrolytica (HML:IMSS) and E.dispm (SAWi734), suggesting the

retention of al1 three epitopes (Fig. 4.5). These data suggested that the cysteine-nch

region, characterked by the 1G7 epitope which had been mapped previously to residues

482-8 18, was conformationally concealed on the amebic surface under native conditions.

4.4.5 E.di;spar lectin (170-kDa) cysteine-Rch region retains both sequence and

epitope identity to its E.hiitoiytica homologue

The cysteine-rich region of the E. hist02ytica Lectin is sufncient for carbohydrate binding,

immunodominant in human disease, and partidy protective in animal immuflzation

studies. In light of these fidings, we carried out an alignment of Dhgl2 (residues 480-

900) and Hg12 (480-900). Cornparison of primary sequence indicated that Ehistolytica

and E-dispar lectin homologues retained considerable overall identity (90%), with 41 of

44 cysteines conserved, Mersupporthg retention of the 1G7 hear epitope (Fig. 3.1).

To test whether the 1G7 epitope was indeed retahed in E-dispar, Wection of COS7 cells with the cysteine-rich region of Dhgl2 (480-900) and Hg12 (480-900) was carried out using a plasmid engineered to expose the cysteine-nch region at the membrane Figure 4.5. Immunoreactivity of axenic (A) E- histobtica (HM1 rIMSS) and axenic @)

Edispar (SAW 1734) trophozoites with fixation. Primary aati-Lech monoclonal antiiodies (mAbs) (a, e) 1G7, (b, f) 3F4, and (c, g) SA3 ,and (d, h) CLB (isotype control) were foiiowed by secondary goat aoti-mouse-FJTC. Trophomites were fked with methanoVacetone prior to antibody staining. Panels depict field ofview under 400x

(a-d) and 1000~(e-h) magnincation; bar = 15 jm (a-@ and 6 pm (e-h). Panels d and h show increased time ofexposure (automatic seaùig) in order to detect Iow Ievel fluorescence due to non-specinc staining. A B dace-Immunofluorescence detectïon confirmed the presence of the 1G7 epitope on

Dhglî-transfected cells but not mock-transfected ceiIs (Fig. 4.7A). Furthemore, 1G7 was used to successfuly immunoptecipitate Dhgl2 (480-900) when translateci in vitro (Fig.

4-73).

4.4.6 Seroposltivity to E.dkpar and E.histoiytica lectin following invasive and non- invasive disease

In order to evaiuate whether E-dispar infection done is able to elicit a systernic immunoglobulin G (IgG) response directeci at E-disparbut not E-hisroi'ytica lectin, semm was collected fiom individuals diagnosed by stool antigen ELISA (Egypt) or moderne analysis (South Afnca). Microtitre wells were coated with recombinant lectin (480-900 of

Dhgl2 and Hglî) and incubated with sera from E-dispar-infected individuais as well as sera nom uninfected controls. Ten of 83 (12.0%) individuais with confïrmed E-dispar infection mounted a specinc IgG response to recombinant Dhgl2 (480-900) but not Hg12

(480-900) at a 1A000 titre (Fig. 4-7B). One of 83 (1.2%) sera in this group recognized both E-disparand E. histoiyficalectins, The remaining 72 (86.8%) sera fiom E-dispar infected persons and 22 American control sera failed to react with either Lectin. An identical ELISA strategy was used to determine the lectin seroreactivity of patients succumbing to amebic iiver abscess (ALA). AU 24 patients with documented ALA following E. histoIytica infection were seroreactive to E. hisfolyticalectin (480-900) (Fig.

4.7A). Using the same critena for positivity, 18 of 24 (75%) ALA sera cross-reacted with

E.dispar lectin (480-900). The mean ELISA OD values for ALA patients were significantly higher @

Indirect ~unofioutescenceofCos7 cektransfected with (a) Hg12 (480-900) (b)

Dhgl2 (480-900) or (c) vector alone- Primary anti-lectin monoclonal anti'bodies (mAb)

1G7 were followed by secondary goat anti-mouse-fluoroscein isothiocyanate (FITC).

Panels depict field of vîew under Lûûûx magnification; bar = 6 p-(B)

~tmoprecïpitationof in vitro synthesized '*s-labelled Dhglî (480-900) with LG7 (lm

1) or isotype control (lane 2)- Figure 4.7- Human antï-sera fiom patients Infected with non-invasive E.dispur compared to invasive E- histol'ytica react wÏth lectln, Human and-sera fkom (A) amebic Lver abscess patients infected with E.histolytica and (B) asymptomatic Edivinfècted indivi'dds were incubated with residues 480-900 of the Iectin heavy sub-t fiom E-histulytica &- axis) and E-dispar(x-axis)- Horse-radish peroxidase (HRP)-conjugated mouse ad- human IgG was used as secondary antibody. ELISA plates were read at 450 nm and absorbance values are pIotted. Data are representatne of 3 independent experiments. - --

We have mercharacterized rnolecular ciifferences between the pathogenic and non- pathogenic lectin homologues in au effort undestand the differential role of each in colonization (E.dispar) versus invasion (E.histolytica)). E-dispar, the avident counterpart to E. histolytica, bas hasen assocîated with asymptomatic colonization of the large intestine in several cluiicai diagnostic -dies (15). Previously, we showed that oniy 3/72 (4.2%) prospectively enroiied patients - who were determined Entamoeba coproantigen-positive by more than one stool ELISA - actually harboured the pathogen (26). These three individuals had the highest anti-amebic titres when tested by [HA, and two of three had signs and symptoms suggestive of invasive disease. None of the remairhg 95.8% infected with E-dispm were symptomatic-

In spite of this, E-di- possesses several homologues of putative E-histolylica vinilence factors (5,17,25). For example, we and others have demonstrated that E-di- retains and expresses a near identical homologue (Dhg12) to the E-histolytica GaUGaINAc lectin (Hg12) heavy subunit (12,25). Severai studies have shown that the GaVGalNAc

Iectin is necessary for adherence of E. hisrolytica to host ceils and colonic minin vitro

(7,30). However, attachment to CHO cells (-2.5-fold less) and GaiNAccBSA (-2-fold

Iess) as weii as cytolysis of CHO cells (4fold Iess) was signincantly duninished in

E. dispar when compared to E. histolytia (12). Ih addition, axenicdy-cuiîured Edispur failed to form abscesses in the hamster mode1 foliowing intra-hepatic inoculation, whereas E-histolyticafonned abscesses with typical centrai necrosis (Dr. Martha

Espinosa-Canteiiano, Semùiano in Amebiasis, Mexico City, 1997). Data presented here provide a molecdar basis for the differences observed in Edispar trophozoite-CHO ceil adherence and cytotoxicity assays, as well as amebic Wulence in experïmentai amebiasis.

Expression of the GaVGaMAc iecth in axeaicaliy cultured E-dispm was compared to E- hisfoljtl'ca-RT-PCR studies demonstrated that RNA Ievels ofthe heavy subunit (dhgi2) and light subunit (mgll) in E-dispar was Iess than in E.histo&tka-

Consistent with this finding, surface expression of lectin was aiso lower in E-dispar when assessed by flow cytometry under axenic culture conditions. It was shown previously that antisense inhibition of lectin light subunit diminishes lectin-dependent cytotoxicity, raising the possibility that iight subunit RNA levels determine the level of hololectin on the amebic dace(3). In vitro binding studies showed that Dhgl2 and Hg12 bound

GalNAcigBSA with high affinity when compared to protein controls. This fïnding was not surprising in Light of the 87% sequence idenitity aod near complete conservation of 97 cysteines present in the functioaal heavy subunit between species. Taken together, these data suggest that the trophozoite expression of the E-dispar GaIlGaiNAc Lectin gene

(dhg22) was relatively lower than its homologue in Eliistolytr'ca, whereas carbohydrate affinity of homologous lectin was not significantly different foilowing in vitro synthesis lectins.

Since several studies have suggested that the cysteine-rich region of the lech is

ïmmunodorninant in human serological studies, immunogenic in animai models, and sufficient to bind carbohydrate in vitro, we focused on region 480-900 of DhgE and

Hglî. The lectin homologues share 90% identity at the predicted amino acid level in this region. T-ecction Ïnto COS7 ceUs ofDhgl2 (480-900) and HglZ (480-900) coupled to

heteroigous signal and transrnembrane sequences foiIowed by indirect immmo-

flourescenceas weii as irnmunoprecipitation studies demonstrated that residues that

comprise the lG7 eptiope (482-8 18) are highiy conserved, However, ffow cytometrïc

studies of unfixed trophozoites cultureci under axenic conditions showed that the LG7

lectin epitope was concealed on the daceof EAi'sprr but ~o~tutivelypresent for

E- histolytia (22). These data rnay suggest that the cysteine-rich region may be either (0

conformationaUy regulated so as to conceal this domain on the native E-dispar lectin or

(ii) the presence of media additives (mucin for SAW 760 and formalin-ked Crithidia fascicuIata components for SAW 1734cIAR or CYNO 16 ) in axenic E-dispar cuftures

may block LG7 recognition. Much may, for exampie, bind E-dikpar lectin (which is

fiinctional in vitro) and occupy residues that recognize the 1G7 Fab. However, both 3F4 and 8A3, which map to adjacent regions of the heavy subunit, did identi@ the E-dispar

1ecti.n in the presence of high molecdar weight (Ho6daltons) mucin. Furthemore, media additives may modulate lecth expression levels in E-dispar such that lectin message and surface protein are diminished. Notwithstanding these differences, we propose that Lower lectin gene expression, less abundant surface hololectin, and concealment of the cysteine- rich region of the lech (in particular, the IG7 epitope), rather than insuisic differences in carbohydrate aff?nity, may account for reduced GaiNAc-dependent adherence and cytotoxicity observed in vïîro and reduced vinilence in experimental amebiasis for axenic

E-dispar (12).

Since the lectin is a purported vinilence factor but is present, expressed and hctional for the non-pathogen, we sought to ascertain if E-dis-infection alone is sficient to eiïcit a specifk humoral response (35)- We assessed the sera of 87 patients infected with E.dispar. Diagnois was conhned by both stooL antigen ELISA and zyrnodeme analysis. We found that lof83 (12%) individuais had mounted a systemic QG response specinc to E-dispar Iectin (480-900) but not E-hisfolyticalectin (480-900). This finhg is surprising in Iight of cEcal diagnostic studies that, mostiy, suggest that

E-dispar causes no cLinicaL symptoms (33)- The possibility that E-dispar may be capable of rninimaiiy invasive disease without causing symptoms is therefore raised. Aithough it is possible that immune and nutritional deficiencies, as weii as other comorbid States, may permit E-dispur invasiveness, most immunocompromised AIDS patients with

E. dispar infections remain asymptomatic. Since E-dispar and E. histolytica lectin are

-90% identical in region 480-900, elicited antibodies specific to E-dispar lectin probably result fiom either (i) unique conformations absent in E.hisfoïytica Iectin or (ii) diffierences in exposure of certain epitopes. Finaiiy, we demonstrate that patients succumbing to amebic liver abscess following E. histolytica infection mount a strong humoral response that recognizes both Hg12 (480-900) and Dhgl2 (480-900). These data are consistent with a prospective study performed in South Mcawhere prior ALA protected individuais fiom subsequent E-dispm Section (LI. Ravdin, University of Minnesota, personal communication). Abd-Ma, M. D., T. F. Jackson, V, Grrthiram, A. MI el-Hawey, and J. 1.

Ravdin* 1993. Differentiation of pathogenic Entamoeba histolytica infections

fiom nonpathogenic infections by detection of gafactose-inhibitable adherence

protein antigen in sera and feces- J Clin h@crobiol. 31~2845-50.

Adler, P., S. J, Wood, Y. C. Lee, R T. Lee, W. A. PetrÏ, Jr., and R L.

Schnaar. 1995. High affinity binding of the Entamoeba histolytica lectin to polyvalent N-acetyigdactosaminidedes. J Bi01 Chem. 270:s 164-71,

Ankri, S., F. Padilla-Vaca, T. Stohrsky, L. Koole, U. Katz, and D. Mirelman.

1999. Antisense inhibition of expression of the lïght subunit (3 5 kDa) of the

GdGaNac Iectin complex inhiiits entamoeba histolytica virulence. Mol

Microbioi. 33:327-3 7.

Ausubel, F. M., R Brent, R E. Kingston, D. D. Moore, J, G. Seidman, J. A.

Smith, and K. Shuhl. 1992- Current Protocols in Molecular Biology. Wiley htescience. Volume 2: Chapter 16.

Bruchhaus, Io,T. Jacobs, M. Leippe, and E. Tannich. 1996. Entamoeba histolytica and Entamoeba dispar: clifferences in nurnbers and expression of cysteine proteinase genes. Mol Microbiol. 22~255-63.

CampbeU, D., and K Chadee. 1997. SUTVival strategies of Entamoeba histolytica: Modulation of cell-mediated immune i-esponses. Parasitol.Today.

13: 184- 189.

Chadee, K., M. L. Johnson, E. Orozco, W. A. Petri, Jr., and J. 1. Ravdin. 1988. Binding and intesnaiization ofraî coIonic muclns by the galactose/N-

acetyl-D-gdactosamine adherence lectin ofEntamoeba hktolytica J Infect Dis.

158~398-406-

8. Chadee, K,W. A. Petri, Jr., DeJ. limes, and C Ravdin. 1987- Rat and

human colonic mucins bind to and *'bit adherence lectin of Entamoeba

histolytica- J Clin Invest. 80: 1245-54.

9. Ciark, C. G. 1995. Axenic cultivation of Entamoeba dispar Brumpt 1925,

Entamoeba insolita Geiman and Wichterman 1937 and Entamoeba rananim

Grassi 1879. J Eukaryot Microbiol. 42590-3.

10. Diamond, L. S., and C. G. Clark 1993. A redescription ofEntamoeba

histolytica Schaudim, 1903 (Emended Walker, 1911) separating it fiom

Entamoeba dispar Brurnpt, 192S.J Eukqot Microbiol. 40:340-4.

1 1. Diamond, L. S., C. G. Clark, and C. C. Cunnick. 1995. YI-S, a casein-fiee

medium for axenic cultivation of Entamoeba histolytica, related Entamoeba,

Giardia intestinalis and Trichomonas vaginalis. J Eukaryot Microbiol. 42:277-8.

12. Dodson, J. M., C. G. Clark, L. A. Lockhart, B. M. Leo, J. W. Schroeder, and

B. J. Mann. 1997. Comparison of adherence, cytotoxicity, and GaVGalNAc 1ecti.n

gene structure in Entamoeba histolytica and Entamoeba dispar. Parasitol

International. 46:225-23 5.

13. Dodson, J. M., P. W. Lenkowski, Jr., A. C. Eubanks, TeF. Jackson, J.

Napodano, D. M. Lyerly, L. A. Lockhart, B. J. Mann, and W. A. Petri, Jr.

1999. Infection and immunity mediated by the carbohydrate recognition domain

of the Entamoeba histolytica GaVGaiNAc lectin. J Mect Dis. 179:460-6. 14. Haque, R,A. S. Faruque, P*Hahn, D. M, Lyerfy, and W. A Petri, Jr. 1997-

Entamoeba histolytica and Entamoeba dispar infection in children in Bangladesh.

J Infect Dis. 175:734-6,

15- Jackson, TIE'. 1998, Entamoeba histolytïca and Entamoeba dispar are distinct

species; chical, epidemiological and serological evidence. Int J Pa.to1.

28:181-6.

16. Kobayashi, S., E. bai, HeTachibana, T*Fujiwlvr, and T. Takeucbi. 1998.

Entamoeba dispar: cultivation with sterilized Crithidia fasciculata J Eukaryot

Microbiol. 45:3S-8s.

17. Leippe, M., E. Bahr, E. Tannich, and R D. Horstmaan. 1993. Cornparison of

pore-forrnîng peptides fiom pathogenic and nonpathogenic Entamoeba histolytica.

Mol Biochem Parasitol, 59: 101-9.

18. Leippe, M., S. Ebel, O. L. Sehoenberger, R D. Horstmann, and B. J. Muller-

Eberhard. 1991. Pore-forming peptide of pathogenic Entamoeba histolytica. Proc

Nat1 Acad Sci U S A- 88:7659-63.

19. Li, E., A. Becker, and S. L. Stanley, Jr. 1988. Use of Chinese hamster ovary

ceils with altered glycosylation patterns to define the carbohydrate specincity of

Entamoeba histolytica adhesion. J Exp Med. 167: 1725-30.

20. Mann, B. J., C. Y. Chung, J. M. Dodson, L. S. Ashley, L. L. Braga, and TmL.

Snodgrus. 1993. Neutralizing monoclonal antibody epitopes of the Entamoeba

histolytica galactose adhesin map to the cysteine-rich extracellular domain of the

170-kilodalton subunit. Infect Immun. 61: 1772-8 -

21. Mann, B. J., B. E. Torian, TmS. Vedviclc, and W. A. Petri, Jr. 1991. Sequence of a cysteine-rkh galactose-specifïc Iectïn of Entamoeba histolytica hoc Natl

Acad Sci U S A, 88r3248-52,

Petri, W. A., Jr., TcF. Jackson, V. Gathiram, K, Ktess, L. D. Saffer, T, L.

Snodgrass, M. D. Chapman, 2. Keren, and D, Mirelman. 1990-Pathogeaic and nonpathogenlc strains of Entamoeba histolytica can be differentiated by monoclonal antibodles to the galactose-specif?cadherence lectin. Infect Immun-

58: 1802-6.

Petri, W. A., Jr., M. P. Joyce, Broman, R D. Smith, C. F. Murphy, and J.

1. Ravdin. 1987. Recognition of the galactose- or N-acetyIgalactosamine-binding lectin of Entamoeba histolytica by human immune sera. Infect Tmmw1.55:2327-

31.

Petri, W,A., Jr., T. L. Snodgrass, T. F. Jackson, V. Gathimm, A. E. Simjee,

K. Chadee, and M. D. Chapman. 1990, MonocIonal antibodies directed against the galactose-binding lectin of Entamoeba histolytica enhance adherence, J

Immunol- 144~4803-9.

Pillai, D. R, D. Britten, J. P. Ackers, J. 1. Ravdin, and K. C. Kain. 1997. A gene homologous to hg12 of Entamoeba histolytica is present and expressed in

Entamoeba dispar. Mol Biochem Parasitol. 8210 1-5 (chapter 3).

Pillai, D. R., K. J.S., D. C. Sheppard, J. D. MacLean, D. W. Macpherson, and

K. C. Kain. 1999, Entamoeba histolytica and Entamoeba dispar: Epidemiology and cornparison of diagnostic methods in a seîting of non-endemicity. Clin. Infect.

Dis. 29:l3 15-1318 (chapter 6).

Pillai, D. R, and K. C. Kriin. 1999. Tmmun0~hromatographicstrip based detection of Entamoeba histolytical E-disparand Giardia Iambiïa, J. Ch.

Microbiol, 37: 30 I 7-9 (chapter 7).

28- Pillai, DeR, P. S. Wan, Y. CmYau, JO1, Ravdin, and K. C. Kain. 1999- The

cysteine-rich region of the Entamoeba histoIytïca adherence lectin (1 70-kilodaiton

subunit) is sufEïcient for hi@-affTnrity Gal/GalNAc- specific bhdïng in vitro.

Mect Immun. 6R3836-41(chapter 2).

29. Ravdin, J. I., B. Y. Croft, and R L. Guemnt. 1980. Cytopathogenic

mechanisms of Entamoeba histolytica J Exp Med. 152:377-90.

30. Ravdin, J. 1, and R L. Guemt- 198 1. Role of adherence in cytopathogenic

mechanisms of Entamoeba histolytica Study with mammalian tissue culture celis

and human erythrocytes. J Clin Invest. 68: 1305-13.

3 1. Ravdin, J. I., W. A Petri, C. L Murphy, and R D. Smith. 1986. Production of

mouse monoclonal antibodies which inhibit in viîro adherence of Entamoeba

histolytica trophozoites. Infect Immun. 53: 1-5.

32. Reed, S., J. Bouvier, A. S. PoUack, J. C. Engel, M. Brown, K. Hirata, X Que,

A. Eakin, P. Hagblom, F. Gillin, and et al. 1993. Clonhg of a virulence factor

of Entamoeba histolytica Pathogenic straïns possess a unique cysteine proteinase

gene. J Clin Invest. 91: 153 2-40,

3 3. Sargeaunt, P. G. 1994. Entamoeba histolytica and E. dispar Detter; comment].

Tram R Soc Trop Med Hyg. 88:254-5.

34. Sargeaunt, P. Ga,J. E. Wiams, and JeD, Grene. 1978. The differentiation of

invasive and wn-Invasive Entamoeba histolytica by isoenyme electrophoresis.

Tram R Soc Trop Med Hyg. 72:s 19-21. 35. Tannich, E. 1998. Royal Society of Tropical Medicine and Hygiene Meeting at

Marison House, London, 19 February 1998. Amoebic disease- Entamoeba

histolytica and E. dispar: cornparison ofmolecules considered important for host

tissue destruction. Tram R Soc Trop Med Hyg. 92593-6,

36. Tannich, E., F. Ebert, and R D. Horstmann. 1991. Primary structure of the

170-kDasurface lectin ofpathogenic Entamoeba histolytica. Proc Natl Acad Sci

U S A. 88~1849-53..

37. Tannich, Ew, R D. Horstmann, J. Knobloch, and Hw 8. Arnold. 1989.

Genomic DNA merences befween pathogenic and nonpathogenic Entamoeba

histolytica. Proc Natl Acad Sci U S A- 86:s 118-22.

3 8. Walsh, J. A. 1986. Problems in recognition and diagnosis ofamebiasis:

estimation of the global magnitude of morbidity and mortalitytyRev Wect Dis.

8:228-38.

39. White, Bw A. 1993, PCR Protocols: Current Methods and Applications. Humana

Press. Volume 15: 177-188.

Entamoeba histdytica infection causes dysentery, intestinal colitis, and hepatic abscess in

an estimated 50 million people worldwide, Attachent of EXstolytica trophozoites to

intestinal epithelium and vascular endothefiun during iiver metastasis resdts in an

inflammatory process. We report the identification of a distinct amebic P2 integrin

(CD 18)-Like molecule that affords adherence to TNF-a-activated endotheliai ceiis, Data

fiom flow cytometry and indirect immunofluorescence assays suggest the amebic P2

integrin was localized to focaI adhesion plates and was present in both E-histolytica and

E-dispar.The amebic P2 integrin appeared to be distinct fiom the amebic GaVGaINAc

lectin based on recombinant expression, amebic colocalizattion and ELISA studies.

Trophozoite adherence to endothelial ceIls expressing IC AM- I (CDS4) following

activation with TNF-a or ICAM-1-transfected CHO cells was specifïcdiy inhibited with

anti-CD 18 or anti-CDS4 MAbs. Tmmunoprecipitation of trophozoite lysates with anti-

CD 18 polyclonal antisera followed by SDS-PAGE identified a candidate 47-kDa ce11

surface protein. In summary, evidence in support of a novel B2 integrin-iike moIecuie, p47, participahg in amebic adherence to TNF-a-activted endothelial ceiis expressing

ICAM-1 is presented. Integrin-dependent binding may dowtrophozoites to opportunistically adhere to activated intestinal epithelium or vascular endothelium expressing ICAM- 1 during amebic colitis or hepatic abscess. Entamoeba histolyrica is the intestinal protozod parasite that is responsi'ble for an

estimated 50 million cases of invasive amebiasis anndy(37). Amebic virulence is

initiated when trophozoites adhere to intestinal epitheiium and mucin, thought to be mediated by the amebic galactoseN-acetyl-D-galactosamine (GaVGalNAc) lectin (6,27).

Intimate contact of ameba with host celi results in rapid cytolysis effected by a potent combination of degradative cysteine proteases and cytolytic pore-fonning peptides (3,4,

29). Amebic contact with and diwption of the epitheiial layer results in secretion of chemoattractant and pro-infiammatory C-X-C chemokines (IL-8, GROa) and cytokines

(GM-CSF, IL-la,IL-1P, and IL6) both in vitro and in viim (13,30,3 1).

Recruitment of neutrophils during an acute intestinal idlammatory response begins with extravasation- Neutrophil transendothelial migration relies on lectin-like (E-,

P-, and L-selectins), integrin @2), and ce11 adhesion (ICAM-1, PECAM, CD47) molecules (22). Transepithelia1 migration of neutrophils dong the basolateral surface, across the tight junctions, and on to the Lumenal surface of intestinal epithelia is stimdated by IL8 secretion during pathogenic gastrointestinal infection (1 8)-

Interestingly, this process can be incited by gastrointestinai infections such as Salmonella fyphirnwium and Helicobacterpylori (8, 19). Migration is facztated by the temporo- spatial expression of adhesion molecules found on both neutrophil (P2 integrin) and epithelial ceil interceMar adhesion molecule-1 (ICAM-1) on the apical surface, and

CD47 on the basolateral surface. Localization of ICAM-1 indicates that this adhesin is up-regulated on the apical surface of the colonic lumen during inflammatory processes, probably serving as a tether for neutrophil f32 integrin (23,24), Severe combined immunodeficient (SCID) mice engrafted with fetd human uitestuie (HU-INT) have demonstrated that neutrophils are possÏily the fïrst line of defense against human amebic infection (3 1,32)- Recruitment cm be specifÏcaUy inhibited with anti-idammatory agents leading to increased epithelial permeabiiity to ameba (3 1)-

Mucosai infiammation and epitheliai destruction can result in movement of

Wulent E- histol'jrica trophozoites in the opposite direction to recruited neutrophiIs. A plausible mode1 wouId predict that trophozoites disnipt the epithelial barrier, traverse extracellular matrix and penetrate vascular endotheid celis, thereby accessing the portal circulation. In support of this, trophozoite penetration into the mucosa is associated with local infiltration with neutrophils in experimental modeIs (7,I7,34). Cysteine proteases are probably responsible for degradation of IgA, collagen and other extracellular matrix

@CM) proteins (3). Epitheliai ce11 cytolysis is effected by amphipathic pore-forrniing peptides (amoebapore) inserted into the target ceii membrane after intimate contact (16).

Adherence to the mucus layer prior to intestinal epithelial cytolysis is thought to be mediated by the GaVGalNAc Iectin suice this molecule affords hi&-aflkity binding to colonic mucin (6). It is Likely that amebic attachment to and migration across epithelium,

ECM, and endothelium prior to portal entry requires severai adhesion molecules durhg the acute inflammatory response induced by E-histolyfica. In this regard, it has been show that a 140-kDa B 1 integrin-like molecule is expressed on the amebic surface and binds fibronectin, a putative ECM ligand during migration (35).

We report that E. hisrofyticaexpresses a distinct P2 integrin-like, capped dace molecule with a homologous epitope to that of neutrophil P2 integrin. Amebic adherence to ICAM-1-expressing cens - either collstitutlvelyor activated by pro-lnfIammatoq cytokine TNF-a -was spedicaily inhibited by anti-p2 integrin, anti-ICAM-1 antibodies, and ~g+depletion. Tmmunoprecipitationofamebic proteins identifïed a unique 47-kDa molecule corresponding to the amebic 82 integrin. These data suggest that

E. histo2ytica may usurp innammatory pathways during invasive disease by expressing adhesion molecules able to specincaüy recognize up-reguIated counter-receptors on host celIs.

5.3.1 Strains, ce11 lines, and cultnre conditions.

Axenic E.histuZyticu, strain HM1 :MSS (ATCC 30459; American Type Culture

Collection, Rockville, MD), were grown in YI4 medium supplemented with 25 pg/mL gentamich (Life Technologies, Gaithersburg, MD) as defhed by Diarnmd (1 1). Chinese hamster ovary (CHO) celis (mock-transfected, ICAM-1, and CD36 stable transfectants) were grown in RPMI supplemented with 10% fetd calfserum (FCS) and 40 pg/mL G418 sulfate (LXe Technologies) (14). Bovine aortic endothelial cells (BAECs) were grown in in RPMI medium supplemented with 15% FCS and 100 Ufdpenicillin and 100 mg/mL streptomycin sulfate. BAECs were activated with medium containhg 1O0 ng/mL recombinant human TNF-a for 8 hours (33). Partidly adherent Spdopterafigiperda

(Sf2 1) cells expressing the GaVGalNAc lectin heavy chah (Hgiî) (kindly provided by

Dr. Yvonne Yau, University of Toronto) were grown at 27*C in Grace's Insect Medium

(Life Technologies) supplemented with 10% FCS and 25 pg/mL gentamich. Al1 cells were grown on standard 75 cm2 plastic tissue cuIture flasks (coated with 0.2% gelath for

BAECs), MammaIian ceiis were harvested with 0.25% trypsin m Dulbecco's PBS (DPBS) without CCand MC(&Se Technologies, Gaithersburg, MD). Trophozoites and ce11 lines were washed 3x with DPBS prior to adherence assays*

53.2 Fïow cytometry.

HM1 :IMSS trophozoites at stationary phase were harvested foilowing 3 days of growth in axenic medium. Trophozhoites (1 x 10') were washed 3 x with PBS and then stained with primary monoclonal antibody (MAb) ad-P2-integrin MHM.23 (IgGl), aati-amebic lectin MAbs 1G7 (IgG2b) and &A3 (IgGl), and an isotype control (IgG1) for lhour (hr) at room temperature (R) (20,28). PharyMAbs were diluted 1:lOO in 0.5% BSA/PBS.

Ceils were washed 3x with PBS and incubated with goat anti-mouse-fluorescein isothiocyanate (FITC) (1 :100 dilution in 0.5% BSA/PBS; Bio-Rad, Hercules, CA) for 30 min on ice. Cells were washed 3X with PBS and finally resuspended in 2% parafiormaldehyde/ PBS und evaluation by flow cytometry (Coulter Epics Elite ESP) for

FITC stainuig. An identical protocol was used for SfZ1-Hg12 and Sn1 cells. Mean fluorescence was quantified by Coulter Epics EKte software.

5.33 Enzyme-linked immunosorbent assay (ELISA).

Immunoaff?nity purifed native GaVGalNAc hololectin (1 pg) was resuspended in 50 pL distilled water and added to each well of 96-microweii ELISA plate (Coniing Costar,

Cambridge, MA) for 1 hr at RT (26). Blocking buffer (200 pL+1% BSA/Tris buffered saline (TBS)) was added for 1 hr at RT. Blocking buffer was fIïcked out and wells washed 3x wïth TBS. MAbs 1G7 (IgG2b), MHM.23 (IgGl), and CLB (IgGI) were serially diluted in l%%SA/PBSand added to each weU for Ihr at RT. Primary MAb was flicked out and wells washed 3%with TBS, Goat anti-mouse aikaline phosphatase secondary mtibody (1250 dilution in 1% BSA/TBS, Bio-Rad) was added to each well for 30 min at RT. Secondary antibody was flicked out and weils were washed 3X with

TBS. Plates were tapped mtiI dry on paper towels and incubated with 100 pL abstrate mix (Lmg/mL p-nitrophenyl phosphate in 10 mM Tris/ 10 mMNaCV 0.5 mM MgC12) for

30 min at RT. Colorimetric reaction was quanfied by measuring absorbance at 450nm using an ELISA microplate reader (ThetmomPcTM,Molecular Devices Corp., Siinnyde,

CA).

5.3.4 Immunoprecipitation and Wes tem blot.

10' trophozoites at stationary phase were harvested following 3 days growth in YI-S medium. Trophozoites were washed 3x with PBS and resuspended in 500 &mL of

Sulfo-NHS-Biotùi (PIERCE, Rockford, IL) for 30 min. at room temperature in ordec to biotinylate surface antigem. Amebae were washed 3x with PBS to remove unbound biotin. Trophomites were solubilued in trophozoite solubilization butTer (TSB) (50 mM

Tris, 150 mM NaCl, 0.5% Nonidet P-40 (Sigma), SmM EDTA, 50 pg/mL E-64 (Sigma), and 2mM PMSF (Life Technologies)) as described by Petri and Schnaar (25). Lysates were protein-quantified using a Micro-Lowrie assay @io-Rad) and resuspended in immunoprecipitation buffer (1 50 mMNaCl, 1 mM EDTA, 0.5% Triton X-100,0.5%

BSA, 10 mM Tris bufEer pH8.0, NETTtBSA). Immunoprecipitation buffet was supplemented with 2 mM PMSF (Sigma Chemical Co) and 50 pg/mL Pans- epoxysucchyl-L-Ieucylamide(4guanidino)-butane (E-64) (Sigma Chem-cal Co,) used as

addiional protease inhi'bitors in order to reduce protein degradation. Lysates (500 pg)

were pre-cleared with 15 pL packed volume of recombinant protein G sephatose beads

for 1 hr at 4OC, spun (16000 x g, 1 min), and supematants added to 10 pg rabbit

polyclonal antisera (anti-p2 integrin (human) or control antisera) or 20 pg of MAb (3F4

or CLB) and rotated end-over-end ovemitat 4OC (1 5). Protein G sepharose beads (50

pL packed volume) were added to lysate plus antibody for 2 hrs at 4OC. Beads were

centrifiiged (16000 x g, 10 min), the supernatant discarded, and then washed as outlined

above. Pelleted beads were added to an equal volume of lx SDS Laemmli sample buffer, boiled for 10 min with vortex mïxing at 5 min. FolIowing centrifbgation (16000 x g,

1 min), the supernatant was run on a pre-poured 7.5% SDS-PAGE and transferred onto

nitroceiIulose (Life Technologies). BIots were blocked in 5% milk proteMris-buffered

Saline containing 0.05% Tween-20 (m~)for 2 hrs at room temperature, incubated with

15000 dilution of avidin-HRP in 2% mdk protein/TTBS for 1 hr (Bio-Rad), washed 3x with TTBS for 5 min. and lx with TBS for 7 min, and deveIoped using ECL reagents

(Amersham Life Science).

5.3.5 Adherence assays.

Centrifugation binding assays were pebormed as described by Ravdin and Guemint (28).

Briefly, 1 x lo4trophomites and 2 x 10* mammalian cells (CHO,CHO-ICAM-1, CHO-

CD36, BAEC) were centrifbged at 150 x g and incubated at 4 OC for 2 hrs. Following disaggregation of ce11 pellet, the number of trophozoites with 23 adherent target ceiis (considered a positive rosette) were counted-. At Ieast IO0 trophozoites were counted per assay and expressed as a percentage, For ïnhi'bition experlments, either trophozoites

(MHM23 GgG1) or VM58 ('IgG1)) or target ceus (152 OgGl) or FA6 (TgG1)) were pre- incubated with IO pg MAb prior to centnfrigation. Ameba and target ceiis were washed

3 x with and resuspended m Hank's balanced Salt soIution (HBSS) without ca2+and MC for adherence assays performed in the absence of divalent cations. Activated BAECs were washed 3x with PBS prior to adherence assay in order to remove residual RJF-a following 8hr stimdation.

5.4.1 A distinct $2-htegrin epitope is present on the surface of E. hktoC&tica and

E. dbpar trophozoites.

Considerable evidence exists that leukocytes and neutrophils use 82 integrins to stick to vascular endotheiium and intestinal epitheiium (22). Previous data suggested that a MAb

(MHM.23) generated against human P2 integrin (CD 18) recognized a homologous epitope on the surface of E-histolytca (HM1 :IMSS) trophozoites (1). MHM.23 has been used extensively in immu~1cytochemistryand P2 integrin hction-blocking studies in other systems (20). To confkm and extend this observation, E. histolytica and E-dispar trophozoites were assessed for the expression of the P2-integrin-like antigen. The

MHM.23 epitope was identified by flow cytometric analysis on both E. histolytica and

E-dispar trophozoites weIi above background levels (Fig. 5.1). Since others have reported that the Iectin heavy subunit bears some homology to known P integrins, we tested if

MHM.23 detected either native or recombinant,titln (170-kDa subunit) (1). Weanti-

Iectin MAb LG7 detected immunoaff?nitv pui.lfied native GaVGaINAc lectïn in a dose-

dependent fashion, anti-f32 integrin MAb (MHM.23) did not Fïg. 5.2). Further, anti-

lectin MAb 1G7 identined SfLl ceus expresshg recombinant lectln, whereas -23

failed to so (Fig 5.3). Western immunoblot of -1 Iectin-positîve ceil lysates with

MHM.23 also faüed to ident* the 170-kDa Iectin heavy subunit (data not shown),

5.4.2 Immunoprecipitation of the amebic $2 integrin.

In an effort to identify the putative 82 integrin of E-histolytica,trophomite lysates were

immunoprecipitated with rabbit polyclonal anti-sera directed against human P2 integrin.

A unique band correspondhg to a molecular weight of 47,000 daitons (p47) was demonstrated absent with rabbit control antisera (Fig. 5.4A). Anti-Iectin MAb (3F4), or controi MAb did not puU down the same 47-kDa protein (Fig. 5.4B)-

5.4.3 Adherence of E.lirSlolytica trophozoites to TNF-a-activated ceils can be specificaiiy inhibited by anti-p2 integrin antibody.

To determine the fùnctional relevance of the amebic p2-integrin, we examined the ability of trophozoites to adhere to mammaiian cells stimuiated with pro-infiammatory cytokine

TNF-a.It has been demonstrated previously that BAEC stimdated with TNF-a upregulated ICAM-1 expression but down-regulated PECAM expression (3 3). Adherence assays using BAEC showed that amebae bound with greater than 2-fold higher avidity Figure 5.1, Amebic P2 integrin expression on E_fiso&t&caand E-disparstrains by flow cytometry. Mean fluorescence inteety @@i)ir standard deviation was determined for

10,000 trophozoites (HM1 3MSS) harvested at stationary phase by flow cytometrïc analysis. Ame6ae were immunostained with prhary anti+2 integrin MAb MHM.23

(IgG1) and isotype control CLB (IgGl) foliowed by secondary goat-anti-mouse- fluoroscein isothiocyanate (FITC) fiom a single representative experiment,

6-

ci- = 4-

2 - Figure 5.2- Detection ofamebic GaVGalEJAc hololectin by ELJSA, Anti-Iectia mAb 1G7

(IgG2b) recognizes the amebic lectin in a dosedependent mamer whiie MHM23 (TgG1) and CLB (IgGl) do not- Goat-anti-moue-AP secondary antibody was used withp- nitrophenyl substrate and absorbante values read at 450 nm, Figure 5-3- Flow cytometnc analysis of bacdovinis-transfectedSfl1 cells (kindly provided by DcYvome Yau, University of Toronto) expressing the heavy subunit (Ha) of the GaVGaNAc Iectin. Plots demonstrate HTC-staining for 5 x 103 SfZl cells foIIowing immunostaining antï-Iectin mAb lG7 (IgG2b), anti-CD 1%mAb MHM.23

(TgGl), and the istoype control CLB (IgGL), MHM.23 does not iden- recombinant fN- length heavy subunit in this bacdovirus expression system, Figure 5.4. Imm~no~pitatÏonofthe putative amebic f32 Uitegrin. (A) Rabbit anti+2

integrin (hwnan) polycfonal antisera (lane 1) was used to pulI down a 47,000 dalton band

@47, mw)by imrnunoprecipitation not seen with rabbit control antisera (iane 2) .Othet bands present are due to non-specinc recognition of biotinylated amebic protek. Whole amebae lysates (10' trophozoites) were added to mbbit antisera prior to incubation with protein G sepharose beads and nui on 10% SDS-PAGE and blotted with avidin-HRP. (B)

Immunoprecipites with anti-lectin mAb 3F4 (lane 1) or istoype control CLB (lane 2) were run on 7S%SDS-PAGE and blotted with avidin-HRP. These mAbs failed to immunoprecipitae p47, whereas the 170-kDa lectïn heavy subunit was puiied domwith

3F4 but not CLB (160-kDa non-specific band present following longer nIm exposure in lane 2). Figure 5.5, Adherence of activated endothelid celis to amebic trophozoites. BAECs (2 x

103 were StImuiated with TNF-a (100nglmL) for 8hrs prior to rossening with amebic trophozoites (1 X 10~).Activated BAECs (upreguiated ICAM-1 expression) adhere >2- fold greater to amebae. Increase in adherence is abrogated (* p<0.05 vs. VM58, a=2,

ONE WAY ANOVA) by p2 integrin blocking antiidy MHM.23 but not a control antibody (VM58). Figure 5.6. Adherence of CHO ceils expresshg human ICA.-1 with amebic trophomites. (a) CHO-ICA.-1, CHO-36, and CHO cells (2 X 103 were pre- incubated wïth antl-ICAM-1 bIochg anti'body (15.2), anti-CD36 blocking antibody

(FA6), and anti-ICAM- 1 (1 5.2), respectively, Anti-ICAM-l antibody significantiy reduced (* pcO.001 vs, control, n=8, ONE WAY ANOVA) adherence of CHO-ICAM-I cells to amebae (1 x 104) by rosette assay, Anti-CD36 blocking antibody (FA6) had no effect on amebaXHO-CD36 ce11 adherence. CHO ceiis blocked with anti-1CAM-f antibody (15.2) had no effect adherence. (b) Adherence assays were performed as in (a) with no mAb but in the presence (2.lmlM) or absence of MC.Signiscant inhibition

(p

ICAM-1 cells but not CHO-CD36 or CHO cells alone.

CHO-ICAM-1 CHO4036 CHO CHO-!CAM-¶ CHO-CD36 CHO upon RJF-a stimulation (Fig. 5.5). This Increase in adhetence was specificaiiy inhibited by blocking amebic B2 integrin with MHM.23 wO.05 vs. control, n=2). suggesting fhctionai involvement of this molecule. By blocking stable transfectants of human

ICAM-1 in CHO celIs with anti-1-1 mtibodies, we were able to si@cantly reduce

(-60%, p

However, CHO-CD36 blocked with anti-CD36 antibodies failed to reduce adherence to amebae, suggesting specifïc interaction with human ICAM-1. Additionally, amebic adherence to CHO-ICN-1 ceils was reduced (a%,p

M~~'',as predicted for integrîn-like binduig (Fig. 5.6B).

E. histolylica is capable of eliciting a specific repertoire of pro-inflammatory cytokines fiom intestinal epithelia (1 3, 30). These cytokines (IL-8, for example) are potent chemoattractants responsible for neutrophil recruïtment to sites of E.histolylica infiltration. This scenario is consistent with evidence fiom animai models that suggest neutrophils can be CO-locaiizedto sites of amebic coiitis (7, 17,32). In this model, neutrophil transepithelid migration and tethering at the Lumenal face of in testhal epithelia is mediated by its 82 integrin. These events are contemporaneous with up- regdation of the P2 integrin counter-receptor, ?CAM-1, by epithelial cells (22). Once metastasized to the liver, Merinfiammatory processes commence leading to granuloma formation, and up-regulation of ICAM-1 on vascular endothelid cells adjacent to liver parenchyma (36). We sought to investigate if: (i) E. histolytca parasites possess a distinct P2 integrin-like molecuie on theu surface and (5)Xso, whether thk

molecule is capable of recomii7Jng its cormter-receptor ICAM-1 when Ïnduced by pro-

idammatory cytokines on relevant host manmaLian celis.

Recent data had suggested that mAb MHM.23 identified a P2 Integrin epitope on

the surface of amebic trophozoites- We confjrmed and extended these obsewations using

flow cytometry in order to demonstrate that both E- histolyica (pathogen) and E-dispar

(non-pathogen) retain the MHM.23 epitope- However. in contrast to previous findhgs, we demonstrate that the 82 integrin homologous epitope does not lie on the amebic

GaVGaNAc lectin but is distinct based on flow cytometric anaiysis of Sf21 insect ceUs expressing recombinant iectin and ELISA studies with irnmunoafkity-pHed native hololectin. Polyclonai antiserum against human B2 integrin was able to immuno- precipitate a 47,000 dalton protein @47) not puiied domby pre-hune sera or anti- lecth mAbs. While mecpurification and peptide sequencing of p47 is needed in order to ensure specificity, these preliminary data suggest that p47 represents a novel B2 integrin-like molecule found on the protozoan parasite E. histdyticn.

We then investigated the bctional relevance of this molecuie. E.histolytica induces pro-inflarnmatory cytokines at the intestinal epithelium, leading to neutrophil recruitment, and the pathophysiology associated with amebic colitis. Wlethis has not been fomidy demonstrated with experirnental amebic infection, it is likely that trophozoite-induced IL-8 leads to neutrophil recruitment and ICAM-1 up-regdation.

Neutrophil transepithelid migration and tethering to the lumenal surface is a B2 integrin- dependent phenornenon. Tramendothelid migration is also a P2 integri.n/ICAM-1- dependent process. in fact, autopsies on human amebic liver abscesses followùig hematogenous spread via the portal circulation have shownup-regdation of ICAM-1 adjacent to amebic abscesses. This raises the possibiIïty that t6is protozod parasite may use 1CAM-l to both attach to intestinal epitheiial cells and migrate invade sub-epithelid layer and also to enter the portal circulation. Merprotozod parasites rely on similar opportunistlc adherence, most notably Piasmodiium faI&pnm-infected erythrocytes which adhere to human ICAM-1 expressed on host endotheiium (2,21)-

To test this, endothebal celis -Cs) that support adherence to human neutrophiis were stimuiated with the pro-innammatory cytokine TNF-a, which has been shown to increase ICAM-1 transcription and sudiace expression (9,33). Adherence of trophozoites to TNF-a-stimuiated BAECs Increased by >LOO% relative to control for eariy passage (6)endotheliai celIs. This rise in amebic adberence was specificaiiy inibited by anti-P2 integrin blocking antibody MHM.23, suggesting involvement of p47.

Moreover, attachent of amebae to CHO-ICAM-1 ceils was specifical1y inhibited (-

60%) by anti-ICAM-1 blocking antiiodies and was lower (40%) in the absence of divalent cation. Surprisingly, these amebae exhibited integrin-W

1CAM-l and are consistent with opportunistic adherence to host cells in a cytokine- induced pro-idammatory setting. In surnmaryywe present evidence here for a distinct and hctio~litllyrelevant P2 integrin-like molecde in E-histol'ytica~Future experïments will requïre Mer demonstmtion that p&ed p47 bears homology to known integrin proteins (especially in the region of the ligand-bùiding I domain) which are known to be present in several phyia

(5). For example, 2 dimensional gel eIectmphoresis coupled to MALDL-TOF mass spectroscopy may identify peptide sequence from minute quanaes of p47 present in irnmunoprecipitation experiments. Aitematively, homoIogous PCR based on conserved domains within P2 integrins or screening of an HM1 :IMSS expression h'brary with anti-

CD 18 mAbs can be used to geneticaiiy clone the amebic counterpart. Furthemore, since

B2 integrins usually occur as heteroduners composed of both a and P subunits, attempts to ident* an amebic a subunit homologue are undenvay. It has been demonstrated, however, that monomeric a and P subunits can be expressed and retain fiuiction in COS ceii transfection systems (12). The possibility that E-histoiytr'camay usurp host inflanmatory pathways in order to adhere and migrate may in part explain the tissue tropism of this parasite. The significance of 2-fold higher levels of P2 integrin on

E-dispar by flow cytometric analysis is not clear. However, sbce each putative virulence factor identined to date (amoebapores, cysteine proteases GaVGaiNAc lectin) is present and expressed in the non-pathogen, the presence of the B2 integrin on E-dispm is not surprising. Aithough not shown, it would be useful to test if the presence of other gut pathogens that up-regulate ICAM-1 on intestinal epithelium facilitate E. dispar colonization and invasiveness (opportunism). The amebic P2 integrin may be a usefid therapeutic target in vaccine strategies that airn to elicit mucosal immunity and thereby prevent invasive amebiasis. Adams, S. A., S. C. Robson, V. Gathiram, T. F*Jackson, T. S. Pillay, R Ec

Kirsch, and M. W- Makgoba. 1993-Tmmuno1oglça.i si-nilarity between the 170 kD amoebic adherence glycoprotein and human beta 2 intem. Lancet. 341: 17-

9-

Berendt, A. R, A McDowall, k G. Crnig, P. A Bat- M. J. Sternberg, K.

Marsh, C. 1. Newbold, and N. Hogg, 1992. The binding site on ICAM-1 for

PIasmodium falciparurn-uifected erythrocytes overlaps, but is distinct firom, the

LFA-1-binding site. Ceil. 68:7I-8 1.

Bruchhaus, L, T. Jacobs, M. Leippe, and E. Tannich- 1996. Entamoeba histolytica and Entamoeba dispar: ciifferences in numbers and expression of cysteine proteinase genes- Mol Microbiol. 22:2SS-63 -

Bmchhaus, I., and E. Tamich. 1996- A gene highiy homologous to ACP 1 encoding cysteine proteinase 3 in Entamoeba histolytica is present and expressed in E. dispar- Parasitol Res. 82: 189-92.

Burke, R D. 1999. Invertebrate integrins: structure, fûnction, and evolution. Int

Rev Cytol. 191:257-84.

Chadee, KI, M. LmJohnson, E. Orozco, W. A. Petri, Jr., and J. 1. Ravdin.

1988. Binding and internalization of rat colonic mucins by the galactose/N- acetyl-D-galactosamine adherence lectin of Entamoeba histolytica. J Mect Dis.

158:398-406. 7, Chadee, K., and E. Meerovitch. 1985- Entamoeba hktolytica: early progressive

pathorogy in the cecum of the gerbil monesunguiculahis). Am J Trop Med

Hyg- 34=283-91.

8- - Crowe, S. E., L. AhigM. Dyto~R HmHM~, MW Muller, P. Sherman, J.

Patel, Y. Jb, and P. B. Ernst. 1995. Expression of interleukin-8 and CD54 by

human gastrk epithelium after Helicobacter pylori infection in vitro.

Gastroenterology- 108:65-74.

9. Damtew, B., P. J. Spagnuolo, G. G. Goldsmith, and J. A. Marino. 1990.

Neutrophil adhesion in the elderly: inhibitory effects of plasma fiom elderly

patients. Clin Immun01 Immunopathol. 54:247-55.

10. Detmers, P. A., S. D. Wright, E. Oisen, B. KimbaU, and 2. A. Cohn. 1987.

Aggregation of complement receptors on human neutrophils in the absence of

ligand- J CeU Biol. 105: 1137-45.

11. Diamond, L. S., C. G. Clark, and C. C. Cunnick. 1995- YI-S, a casein-fiee

medium for axenic cultivation of Entamoeba histolytica, related Entamoeba,

Giardia intesthlis and Trichomonas vaginalis. J Eukaryot Microbiol, 42:277-8 -

12. Douglass, W. A., R. H. Hyland, C. D. Buckïey, A. Al-Sbamkhani, J. M. Shaw,

S. L. Scarth, D. L. Simmons, and S. K. A. Law. 1998. The role of the cysteine-

rich region of the B2 integrin subunit in the leukocyte fiinction-associated antigen-

1 CFA-1, aLP2, CD 1lalCD18) heterodimer formation and Ligand binding. FEBS

Letters. 440:414-8.

13. Eckmann, L., S. L. Reed, J. R Smith, and M. F. Kagnoff. 1995. Entamoeba

histolytica trophozoites induce an infiammatory cytokine response by cultured human cells through the paracrine action of cytolytically released interleukin-1

alpha J CIin Invest, 96: 1269-79-

14, Hasler, T., G. R Albrecht, MWR Van Schravendijk, JOC. Aguiar, K, E.

Morehead, B. L. Pasloske, C. Ma, J. W. BarnweU, B. Greenwood, and R J.

Howard. 1993. An improved microassay for Plasmodium falciparum

cytoadherence using stable transformants of Chinese hamster ovary ceUs

expressing CD36 or interceiIuiar adhesion molecule-1- Am J Trop Med Kyg.

48: 3 32-47.

15. Kantor, C., H. Suomdainen-Nevnnlinna, M. Patarroyo, K. Osterlund, T.

Bergman, H. Jornvall, J. Schroder, and C. Go Gahmberg. 1988- Purification

and large scale characterlzation of the human leukocyte adhesion glycoprotein

GP90 (CD 18). Eur J Biochem, 170:653-9.

16. Leippe, M., Eo Tannich, R Nickel, Gwvan der Goot, F. Pattus, Rw D.

Horstmann, and H. J. MaMer-Eberhard. 1992. Primary and secondary structure

of the pore-forming peptide of pathogenic Entamwba histolytica Embo J.

11:3501-6,

17. Martinez-Palomo, A., V. Tsubumi, F. Anaya-Velazquez, and A Gonzala-

Robles. 1989. Ultrastructure of experimentd intestinal invasive amebiasis. Am J

Trop Med Hyg. 41:273-9.

18. McCormick, B. A., P. MoHofinan, J. Kim, DmK Cames, S. 1. Miller, and Jw

L. Madara. 1995. Surface attachment of Salmonella typhimurium to intestinal

epithelia imprints the subepithelial ma& with gradient chernotactic for

neutrophils. J Celi Biol. 131: 1599-1608. 9 McCormick, B*A., C. A. Parkos, S, P, Coigan, D. Ki Cames, and J, L,

Madara. 1998. Apical secretion of a pathogen-eiicited epitheliai chemoattractant

activity in respome to dacecolonization of Intestinal epithelia by SaIrnonelia

typhimurium- J ~unol.f 60:455-66.

20- McMichael, A, J., F. M. Gotch, and A E. Eüdreth, 1982. Lysis of aiIogeneic

human lymphocytes by nonspecifIcaIiy activated T-like celis- Eur J ImmunoI-

12: 1002-5.

21. Ockenhouse, C, F., R Betageri, T. A. Springer, and D. E. Staunton. 1992.

Plasmodium falciparum-infected erythocytes bind ICAM-1 at a site distinct from

LFA- 1, Mac4 ,and human rbinovirus, Cell. 68~63-9.

22. Parkos, C. A. 1997. Moleculat events in neutrophil transepitheliai migration.

BioEssays- 19: 865-73,

23. Parkos, C. A., S. P. Colgan, M. S. Diamond, A. Nusrat, T. W. Liang, T. A.

Springer, and JOLm Madara. 1996- Expression and polarkation of intercellular

adhesion molecule-1 on human intestinal epithelia: consequences for

CD 11 b/CD 18-mediated interactions with neutrophils. Mol Med. 2:489-505.

24. Parkos, C. A., C. Delp, M. A. Arnaout, and J. L. Madara. 1991. Neutrophil

migration across a cultured intestinal epithelium: dependence on a CD 1I b/CD 18-

mediated event and enhanced efficency in the physiologie direction. J Clin

Invest. 88: 1605- 12.

25. Petri, W. A., Jr., and R L. Schnaar. 1995, Purification and characterizaion of

galactose- and N- acetylgalactosamine-specificadhesin Lectin of Entamoeba

histoiytica. Methods Enqmol. 253:98-104. 26. Petri, W. Am, Jr., R D. Smith, P. H. Schlesinger, C. F. Murphy, and J. 1.

Ravdin, 1987. Isolation of the galactose-binding Iectin that mediates the in vitro

adherence of Entamoeba histolytica J ChInvest. 80: 1238-44.

27. Ravdin, J. 1, and R L. Guemnt. 1981. Role of adherence in cytopathogenic

mechanisms of Entamoeba histolytica Study with mammalian tissue culture ceUs

and human erythrocytes. J Clin Invest. 68:1305-13.

28. Ravdin, J. I., W. A. Petri, C. F. Mruphy, and R DmSmith. 1986. Production of

mouse monoclonal antiidies which inhibit In vitro adherence of Entamoeba

histolytica trophozoites, Infect Immun. 53: 1-5.

29. Reed, S., J. Bouvier, A. S. Pollack, JO C. Engel, M. Brown, K. Hirata, X. Que,

A. Eakin, P. Hagblom, Fe Giiiin, and et al. 1993. Cloning of a vinilence factor

of Entamoeba histolytica Pathogenic strauis possess a unique cysteine proteinase

gene. J Clin Invest. 91: 1532-40.

30. Seydel, K. B., E. Li, P. E. Swanson, and S. L. Stanley, Jr. 1997. Human

intestinal epithelial cells produce proinflammatory cytokines in response to

infection in a SCID mouse-human intestinal xenograft mode1 of amebiasis. Infect

Immu11.65:163 1-9.

3 1. Seydel, K. B., E. Li, Z Zhang, and S. L. Stanley, Jr. 1998. Epithelial ceîi-

initiated inflammation plays a crucial role in early tissue damage in amebic

infection of human intestine. Gastroenterology. 115: 1446-53.

32. Seydel, K. B., T. Zhang, and S. L. Stanley, Jr. 1997. Neutrophils play s critical

role in early resistance to amebic liver abscesses in severe combined

immunodeficient mice. Mect Immun. 65395 1-3. 33 Stewart, R. J., T. S, Kas-, and P. A, Marsden. 1996- Vascdar endothelial

platelet endothefial adhesion molecule-1 (PECAM-1) expression is decreased by

RJF-alpha and IFN-gamma, Evidence for cytokine-induced destabillzattion of

messenger nimcIeic acid transcripts in bovine endothelid celis- J Tmmmol.

156: 1221-8.

34. Takeuchi, A., and B. P. PKüüps. 1975, Electron microscope studies of

experimental Entamoeba histolytica infection in the guinea pig. 1. Penetration of

the intestinal epithelium by trophozoites, Am .J Trop Med Hyg. 24:34-48.

35. Talamas-Robana, P., V. 1. Hernandez, and J. L. Rosaies-Encina. 1994. A beta

1 integrin-like molecde in Entamoeba histolytica Tram R Soc Trop Med Hyg.

88596-9.

36. Ventura-Juarez, J., R Campos-Rodrigua, HmA. Rodrigua-Martinez, A.

Rodriguez-Reyes, A. Martinez-Palomo, and V. Tsutsumi. 1997. Human

amebic liver abscess: expression of intercellular adhesion molecules 1 and 2 and

of von Willebrand factor in endothelial cells, Parasitol Res. 83:s 10-4.

37. Walsh, J. A. 1986. Problems in recognition and diagnosis of amebiasis:

estimation of the globai magnitude of morbidity and mortality. Rev Wect Dis.

8:228-38.

Recent studies suggest that stool antigen assays are more sensitive and speçinc than microscopy for the diagnosis ofEntmoe6a histo&zka Wectlon. Patients (n=l12) presenting at 3 North American centers with symptoms Mornsk factors suggestive of

E. histoiyticu infection were enrolled in this study to evaluate new diagnostic tests for E. histolyrica and E. dispar. Four ELISA-based stool antigen kits to detect

Ehistolyti~~disparwere blindly compared to stool microscopy, Patient sera were simdtaneously assessed for amebic antibodies by indirect hemagglutination (MA).

Detection of E-histolylcdE- dispm infection by the E~tarnoebatest@ (Techlab) and

ProSpecT@ (Alexon-Trend) showed good agreement (92.9%). Using stool antigen assays as the reference standard, stwl microscopy pe&ormed at referral centers was more specific (68.4% vs. 9.5%) but less sensitive (70.4% vs. 92.1 %) than microscopy performed in community laboratories. Diagnons by the E-histolyticatest@ (Techlab) and

Merlin Optimun S@ stool ELISA (Merlin Diagnostika) indicated that ody 3 of 72 (4.2%) coproantigen-positive stools were due to E.histolytica; the remainder were Edspar. IHA was a good predictor of histolyrica infection when amebic antibody titers were 21:s 12. Mection with Entamoeba histolfl2a has the potentiai to cause dysentery and extra- intestinal disease, notably hepatic abscess- Entamoeba dispar is not associated with disease and is considered a hardess commemai (27-29) .E-hisfolyttcu is responsible for approximately 100,000 deaths worldwide each year, making it second ody to malaria as a cause of mortality due to a protozoan parasite (39) .The World Health Organîzatïon has recommended that "E-histolyticashould be specincally idenaned and, if present, treated; if only E-dispar is identifie4 treatment is unnecessary" (41) .Traditionaily the diagnosis of E. histolytica infection has ken estabfished by the microscopic examination of either fiesh or fixed stool. However, microscopic diagnosis has several Limitations (10- 12,2 1) most importantly the inability to distinguhh E. hisfo&tica fiom E- dispar since they are morphologicaiiy identical (6) .

Arnebic culture and isoenzyme analysis is generdy considered a reference standard to differentiate E.histolyriu nom E-divar, but this method is not widely available nor practical for routine diagnostic laboratories (34) .Detection of&-amebic antibodies by indirect hemagglutination (HA) in patient sera has been reported to be indicative of E. histolytica Uifection. However, with serologic testing it may be dificuit to distinguish past nom present infection in individuals who emigrate fiorn, or currently reside in, an endemic area (12, 13, 16, 18).

Recent evidence has clearly demonstrated that E. histolytica and E-dispar are distinct species as shown by isoenzyme electrophoretic mobility and by the sequence and expression of certain genes (27-29) .The discovery of genetic sequence specificity and protein epitope specficity has allowed for the development of molecular strategies to distinguish the two organisms by PCR and mol antigen detectîon. PCR appears to be sensitive but is encumbered with laborious and inconsistent stm1DNA extraction protocois (2,4,26,3 7) .Stool antÏgen detection is relatively rapid and has been shown to be as sensitive and specifÎc as cuiture/isoenzyme differentiation and to outperform microscopy for the detection of E-trllstoiytica in endemic areas (1 1-13).

Four stool antigen ELISA kits are now commercially avaiIable. The Entmnoeba test@ (Techlab) detects the GallGalNAc Iectin of both E-histo&ticaand E-dispar whiie the E- hisrolytica te@ (Techlab) recognizes the lectin of the pathogen aione (1 1-1 3). The

ProSpecTB ELISA (Alexon-Trend) relies on polyclonal antibodies to E. histo&ticaï/E. dispur complex Maceantigens but canwt distinguÏsh pathogen fiom non-pathogen

(3 1). Meriin OptimunG9 S (Diagnostika) is reported to detect the serine-rich protein

(SREHP) of E-hr'stolytica alone (33). Compared to culture and isoenzyme analysis the

Techlab kits have reported high specificity and sensitivity in an endemic setting (1 1-13).

These assays provide a reference standard to evaluate new and old diagnostic assays as well as an opportwiity to revisit epidemiological questions conceming amebiasis in non- endemic settings.

We prospectively collected epidemiologic data, Sem, and fkesh stool samples, from patients presenting in a non-endemic area with suspected E. histoiyficainfection with the foilowing objectives: 1. To evaluate the performance of new diagnostic kits for

E. histoZytica(E- dispar infection; 2. To compare reference and community microscopie diagnosis of E- histoiytica[E. dispar complex; 3. To determine the prevalence of E. histolytica in a non-endemic area; 4. To examine the epidemiology of E- histoLytica/E. dispar in this setting; 5, To evaiuate the chicai utility ofserologic testing to dlfferentiate

E. histolytica from E. dipur in this popdation.

63.1 Study population

Patients presenting to the Tropical Disease Unit of The Toronto Hospital, McGill Centre for Tropical Disease at the Montreal General Hospital, and the Mectious Diseases and

Tropical Medicine Chic, Hamilton Health Sciences Corporation (McMaster Campus) with gastrointestinai symptoms (diarrhea, defïned as 23 loose (conforni to the container) bowel movemenWday, abdominal pain, nausea, weight loss, bloody stool, etc.) andlor risk factors for E. histohticdEEdispm uifection (recent travel (within 6 months), men who have sex with men, or recent immigrants fiom the tropicdsubtropics (< 2 years)) fiom 1993 to 1998 were eligible for inclusion in this study. Verbal inforrned consent was obtained fiom each patient and the study was approved by the Ethical Review Cornmittee of The Toronto Hospital and the Montreal General Hospital. Subjects were requested to provide fiesh stool samples (wimin 1 hour ofpassage) for microscopy and ELISA anaiysis, blood for serologic work-up, and background information in a directed questionnaire. Stool specimens were transported to the parasitology laboratory for routine ova and parasite evaluation by microscopy and aliquots of fksh unpreserved stool were fiozen for subsequent ELISA analysis. A questionnaire was administered to obtain the following infornation: age, gender, birthplace, immigration date, location of recent travel, duration of recent trtivel, sedorientation, and matment presctr'bed for the present ihess,

6.3.2 Microscopy and uidirect hemagglatination assay (IHA)

Stool sampies (min. of 2 stools) preserved Ïn Sodium Acetate Fornialin (SAF), were concentrated by formalin-ether or -ethy1 acetate sedimentation and exsmuied by microscopy for E. histolyticuE dispar cysts and trophozoites, as weil as other intestinal protozoa. Permanent stains were perfomied ushg Ïron hematoxyiin staining and organisms were microscopicaiiy examined with the aid of a micrometer. Serum was collected to assess the presence of amebic mtiidies by MA as described by the manufacturer (Celiognost@ Arnebiasis Kit, Behring Diagnostika, Marburg*Germany).

6.3.3 Enzyme linked immunosorbent assay (ELISA)

Frozen unpreserved stool samples were thawed and independently processed for stool antigen of E, histolyticdE- dispar complex (Entarnoeba test@; ProSpecB test) or E, historytica (Merlin Optimun@ S tesE. histolytica test@) accordhg to the manufacturer's instructions (AIexon-Trend, Ramsey, Merlin Diagnostika,

Bomheim-Hersel, Germany; Techlab, Blacksburg* VA). Discrepant ELISA results between the Entamoeba test@ (Techlab) and ProSpecTO (Alexon-Trend), which both detect E. histolyticdE. dLspm complex, were resolved by the use of the E. histoiytica specific kits and an additionai independent ELISA protocol for E. dispar as follows.

Bnefly, -0.2g of stool were renispended in buffered protein solution with 0.02% thimerosal and added to a weII containhg rabbit polyclonal antibody reactive against multipIe E-hïstoZytica/dispar daceantigens at room temp for 2 hts, Wells were washed

3x with wash solution VBS, detergent, 02%thherosal), încubated with mAb 3 t 8-28

(15000 dilution in 1%BSAIPBS, room temp, lhr.; mAb 3 18-28 kindly provided by D-

Mireiman, Weizmann Inaïtute of Science, Rehovot, bel)which detects E-dispur lysine-

rich sudace antigen (27) ,washed Sx, then incubated with goat anti-mouse conjugated to

horseradish peroxidase (1500 dilution in 1% BSAiPBS, room temp, 30 min). Weils

were washed 5x auddeveloped with 0.125% tetramethyIbenzidine (TMB) and the

reaction stopped with 16.6% hydrochioric acÏd- E-dispar surface antigen was used as a positive control and stool diluent as a negative contrai- A positive result was dehed as

an OD20.050 after subtracting the negative controi reading using a dud wavelength (b-

&j jO) ELISA microplate reader(Therm~~~~,Molecular Devices Corp., Sunnyvde, CA)-

Al1 ELISA assays were performed independently at least twice and blinded to the other

ELISA, rnicroscopic, and MA results.

During the study period, 112 patients were enroiled- Data pertaining to patient demographics, presenting symptoms, risk factors, location of recent travel, birthplace, other intestinal protozoan infections detected, and treatment drug for present iiiness were collected and are summarized in Figure 6.1. The study population was comprised of 72 males and 40 females (~112).Most patients were asymptomatic (33 -9%); chief cornplaints of symgtomatic individuals were diarrhea (30.4%) and Figure 6.1. Demographic and epidemiological characteristics of shidy population

(n=l12). (a) Age and gender of study population; @) symptoms at the time of presentation; (c) nskfactors; (d) other intestmal protozoan infections identlned by microscopy; (e) treatment presm'bed~

O 10 15 20 25 30 no- individuals O 5 10 15 20 25 30 35 40 no. ind~duais

0.fragiiis

O 20 40 60 80 no- inf-ons no- individuals Figure 6.2. Distribution of resuits nom ELISA kits performed on stool specimens

(n=112). OD values for Techlab's Eizfmnoeba test@ and Alexon-Trend's ProSpecT@ are plotted for each stool sample. OD values 10.050 after subtracting negative control are considered positive for each kit-

65 8 ! 8

1-. mm

â m m 0 8' a 0.1 ; rn E 2 I m- W : I - m C O p: 0.01 a ' + m

1E-3 0.0 1 0-1 1 Techfab ELISA (00) Figure 6.3. Prevaience ofE- histoljrtica versus E-dispmamong presenting patients based on consensus results of stooI antigen ELISA kits (n=112)

E-dispar/ E. histolytica (OD)

0.1 E. histolytica (00) abdominal pain (23 -2%). FoWthree percent of the study population were bom in an

areas where amebiasis is endemic (Scaand Asia accounting for 67-0%). Rkk factors

for North Amencan-born and Europem-bom Uid~dualsÎncluded travel to an endemic

area (64-1% of cases), men who have sex with men (12S%), or both (7.8%)- Travel was

primarily to Africa (29%), Mexico and Centrai America (28%), or Asia (23%). Of the 48

patients born in an endemic area, 60.4% had arrïved recently (withîn 2 years) and 3 5,4%

had retumed recently to visit their country of origin, Co-infections with B~àstocystis homink (58.9%), Endolimm nana (48.2%), Enîamoeba hartmanni (44.6%), and

Enramoeba coli (42.0%) were most often detected by stod microscopy- Sucty-five percent (7311 12) of participants received drug treatment and of these 76.7% (56/73) received iodoquinol, 21.9% (16/73) paromomycin, 5.5% (4173) diloxanïde furoate, and

9.6% (7/73) received metronidazole {some individuals received combinations of drugs)

(19).

Stool antigen detection by Techlab 's Entamoeba test@ and Alexon-Trend' s

ProSpecTB kits were used to identifl specimens positive for E. hisrolyficaLE-dispar cornplex. The MO kits agreed in 92.9% of the cases with 65 sarnples positive and 39 sampies negative for E. histolyticdE- dispur for both kits (Fig. 6.2). Eight discrepant redis were identified which were ail were positive by the ProSpecTB and negative by the Entamoeba [email protected] resolve this discrepancy, these 8 specimens were evaluated by the E. hisiolyica specific kits (1 of 8 was positive for E- hisrolytica (EH36) by both tests; see below) and by an independent ELISA utilipng mAb (3 18-28) directed against the 30- Table 6.1. Cornparison ofELISA redts with ProSpecTB, Entmnoebu test@, and mAb

3 18-28 for discrepant stool specimens shown in Figure 62.

Patient Isolate Aiexon-Trend Techlab mAb 318-28

ProSpecT@a Enfanoeba test

EH36 POS NEG POS

EH38 POS NEG POS

EH79 POS NEG NEG EH85 POS NEG POS

EH89 POS NEG POS

EH105 POS NEG POS EH1 IO POS NEG POS

EH1 11 POS NEG POS

" OD value 20.050 after subtracting negative control is regarded as positive with aii

ELISAs

b mAb 3 18-28 was a khd gift Dr, D. Mirehan, Weizmann Wtute of Science,

Rehovot, Israel(2 1) kDa lysine-nch SUZface antigen ofEl dispar (21). This strategy identified 7 of the 8

discrepants as positive for the Iysine-rich suffie antigen of E. dispm (Table 6.1),

suggesting that these were tme positnres, Furthemore, these resuits indicate that one

individual, EH36, was CO-infectedwith histolytica and E. dispar. Taken together, these observations are consistent with a previous report that the Entmoeba test@ is Less sensitive than the ProSpecTO at Iower parasite numbers (21).

The performance of antigen detection assays suggest that they may be considered as reference standards for the detection of E. histolyrica-and E. dispar (8,22-24). Based on these reports and the high correlation (>go%) between the Enfamuebutest@ and

ProSpecTQB, we used the consensus resdts obtained fkom these two ELISA kits as a reference standard (discrepants were excluded) to evaluate microscopy both in the community and referral setting. Compared to this standard, referrai microscopy was more specific (68.4% vs. 9.5%) but less sensitive (70.4% vs. 92.1%) than microscopy performed at community laboratories (Table 6.2). The positive predictive value (PPV) of a microscopie result was assessed for three principal hi&-risk groups based on the prevalence estimated previously by us and others for E. histoZytica/EEdispar infection

(1 7, 18) . While the PPV for referral centers was greater than that obtained for community laboratories in al1 three high-risk groups, it remained below 50% (Table 6.3).

Stool samples obtained fiom all patients were also blindly anaiyzed by the E. hirtoryrica test0 (Techlab) and Merlin Optimun SB ELISA (Diagnostika) which are reported to specifically detect only E- histolyh'ca. Both of these assays identified 3

(EH36,42,77)of 112 patients as positive for E. histolytca antigen (Table 6.4). Two of Table 6.2. Sensitivity and speci ficity for detection of E. hislolylica/E. dispar compiex by microscopy compared to the consensus result of the Entamoeba test@ (Techlab) and ProSpecTa ELISA (Alexon-Trend) kits as the reference standard.

Microscopy (O & P) Entamoeba@ and ProSpecT@ Specificity Sensitivity ELISA kitsC positiveb Negative TOTAL REFERRAL CENTERSa: Positive 39 8 47 Negat ive 18 19 37 TOTAL 57 27 84 68.4% 70.4% COMMUNITY LABORATORIESa: Positive 35 19 54 Negat ive 3. 2 5 TOTAL 38 21 59 9.5% 92,1%

'Referral O & P examination performed at a refenal hospital; Community Laboratories O & P examination performed by a vadety of privately-run laboratories. Note: Some patients were evaluated at both sites. 0,D. value 20.050 (after O.D. value of negative control is subtracted) Reference standard determined by consensus resdt of both Techlab (Entamoeba test@) and Alexon-Trend prospecT@ELISA kits, Table 6.3. Predictive value for a positive result by microscopy performed at

Referral or at Community Laboratorics for three high-risk groups in a non-endemic setting.

Risk Croup Referral Centers Communiîy

(Estima ted Prevalence) La boratories

Male Homosexuals (27%)(17)

Positive Predictive Value 45,2% 27,3%

Immigrants (2.4Oh) (12)

Positive Predictive Value

Travelers (4%)(40)

Positive Predictive Value 8.5% 4,1% Table 6.4. Results of ProSpecTO (Alexon-Trend), E,histolytica test0 (Techlab), and Merlin Optimun S@(Diagnostika)ELISA kits for patients with positive serology by IHA. Patient Isolate Alexon- Alexon-Trend Techlab Techlab Techlab ELISA Merlin Merlin ELISA (EndemicNon- Trend loterpretrtionb Entarnoeba E.liistolylica ~nter~retotionbptimunS@ lnterpretitionb endemic)' ProSpecTB ELISA ELISA ELISA EH13 (End) + Eh or Ed + - Ed - none EH36 (End) + Eh or Ed - + Eh -t- Eh EH40@nd) - uninfected - - uninfected - none EH42 (Non) + Eh or Ed + -1- Eh t Eh EH54 (End) + Eh or Ed + - Ed - none EH58(End) + Eh or Ed + - Ed - none EH66 (End) - uninfected - - uninfected - none EH70 (End) + Eh or Ed t - Ed - none EH74 (End) + Eh or Ed t - Ed - none EH77flon) + Eh or Ed -I- + Eh t Eh EH91 (End) 3. Eh or Ed + - Ed - none EH101 (End) + Eh or Ed t - Ed - none Note: Of the 52 patients with negative serology (4:64) and detecTable Entutnoeba coproantigen, al1 52 were interpreted as E,dispar by the E,histolytica test@ and Optimum SB ELISA kits. '"Endemic" is defined as 26 months residence in area of high E~nioebaincidence of infection. ~nter~retationaccording to manufacturer's instructions. Eh: Entamoeba histolytica; Ed: Entamoeba dispar; + : positive assay according to manufacturer's instructions; -: negative assay according to manufacturer's instructions. Table 6.5. Resuit of indirect hemaggiutination test (n=102) perfonned on patient sera

1'Titre No. Ind~duais these individuais had clinîcaily invasive disease (amebic liver abscess; dysentery). For isolate EH36, the Techlab's Entumoeba tesi@ and E. histoZytica test@ were intemaliy discordant suggesting that the respective monoclonal ann'bodies may have differential sensitivities-

Consenting patients (n=102) were aIso assessed serologically for amebic antibodies by MA. In 12 of 102 individuals (12.6%), antisera were positive (21:64) for antibodies (Table 6.5)- The 3 individuais who were positive for E, histolyrica antigen had the highest titres el :5 12) by MA. A krther 7 seropositive individuals were positive for

E. dispar done by antigen detection (Table 6.4). The remaining 3 seropositive patients had no detecTable Entamoeba coproantïgen by all ELISA kits- Ali homosexual males in this study (n=13) were seronegative and were infected with E. dispar done based on results of antigen detection-

Using E. hisfolyticuantigen detection as the reference standard, serologic titers recommended as positive by the manufacturer (>1:64) were 100% sensitive and 90.9% specific for E. histolytica infection; at a titer of 11512, serology was 100% sensitive and

100% specific.

The laboratory diagnosis of E. histoiytica has traditionally relied upon the microscopie examination of fecal samples. However, microscopy is iimited by mkidentification of E. histoiyrica/EEdispar complex (for leukocytes, E. harhnunni or other intestinal protozoa) and the inabiiiîy to distinguish pathogenlc E- histol~kahm îhe nonpathogenic E- diqar

(22)- Recent studies in endemic areas have suggested that antigen detection assays may

be an alternative strategy to idente E- hirstolytrca infection (1042)- This is the fïrst

study to blindly evaluate alL available stool antigen ELISA kits, microscopy, and serology

for the diagnosis E.histoiyticddispar infection in a non-endemic setting, We demonstrate that only a smali percentage of individuais in our setting identified with E- hisro~tic~- dispar complex by microscopy are acnially infected with E- hisfoljtica; that positive

serologic titers -4:5 12 are poor1y predictive of infection with E. histolytica; that the microscopic diagnosis ofamebiasis is problematic with respect to both sensitivity and specifkity; and that the majority ofindividuals diagnosed microscopicaiiy with E. histo2yticdE- dispar complex receive unnecessary treatment.

In this study, we used the consensus results of antigen detection assays as a reference standard to examine the epidemiology and microscopic diagnosis of E. histolyticdE- dispar in a North Amencan setting- The majority of subjects identified with E. histolyticuZ. dispar infection in this study had previously recognized risk factors

including travel to or recent emigration nom an area where amebiasis is endemic or were

men who have sex with men (22). Thc use ofthe Entamoeba test@ and ProSpecTQDkits to detect E. hLFloZytica/E.dispur infection in this population demonstrated excellent agreement (>go%). AU discrepancies between the kits were negative by the Entamoeba test0 and positive by ProSpecTB. When discrepancies were resolved by evaluation with a third independent ELISA, al1 but one discrepant were resolved as positive in agreement with ProSpecT@. Our results suggest that ProSpecT@ is more sensitive than the

Entamoeba test@ for the detection of E.hisrolyticddispar in stool and are in agreement with eariier reconstitution experiments with these kits (21). ProSpecTB may therefore be useful in screening for E-histo[yrca/disparcomplex infections but fails to distmguïsh pathogen fiom non-pathogen,

Techiab's E. histoljtica test@ and Merh Diagnostika's Optimun S@ have been reported to specificaiiy detect E hisrolifica antigen (10,21) . Both kits detected onIy 3 positive specimens out of 112 samples. Ofnote, these 3 individuais bad the highest titres by IHA (21512) and 2 had clinicai evidence of invasive disease. The thÏrd patient may represent an asymptomatic cyst passer of E. hisfolyfr'ca(1 3) .The remaining seropositive patients (ail bodresident for 26 months in endemic areas) either had E-dLFpm infections or were uninfected as determined by the ELISA kits. Previous studies indicate that amebic antibody titers may remain elevated for many years alter an invasive episode and confound attempts to distinguish current fiom past infections (see chapter 1.3). In this study the use of amebic serology to detect E. histolylica infection was non-specific at titers <1:5 12, especiaily in individds bom in endemic areas (1 5) .Consistent with previous studies, all homosexual males in this study harbored E. dispar infection as determined by stool ELISA and negative serology (40).

Agreement between ELISA kits and microscopy, as well as between referral centers and community laboratories for the detection of E. histolytica/E.dispar complex was problematic with respect to sensitivity, specificity, and interlaboratory agreement.

The low positive predictive value (<50%) for microscopy among aii three hi&-risk groups and the low prevdence of E. histolycica infection in this non-endemic setting, suggests that f5ture studies should examine the cost-effectiveness of replacing routine microscopy with ELISA kits for the detection of E- histolyrica, Co-infection with the non- pathogenic protozoaEndolinua nana and Enfamoebacoli was fiequent as detemiined by microscopy in study population. In contrast to a previous study, a high rate of

Entawoeba hartmanni co-infection was obsexved in our study population (2) ,

Concomitant protozoan infections are not unexpected since it is Wcely that a common mode of transmission occurs with these organisms.

The clinical implications of this study are signincant since oniy 3 of 65 (4.6%) unequivocaIIy coproantigen-positive stwl samples (positive by >2 assays) were positive for E. histolytica by ELISA; the remainder were identined as E. dispar infections. E. dispm is not associated with invasive disease and does not require drug treatment (22). If our results are generaiizable to other developed world settïngs, then the majority of patients identifïed with E. histoiyti~a~dispm complex by microscopy in non-endemic areas receive unnecessary therapy. The use of an E. histolyficatest (e.g. Techiab's

E-histolylica test@ or Merlin's Optimun S @) would allow for a specifïc diagnosis and obviate the need for unnecessary chemotherapy with its attendant costs, risk of side- effects, danger of dmg-resistance, and potential rnistreatment of another disease. It should be noted, however, that the true positive predictive value of stool antigen tests are difncult to measure in setting where oniy 3f 112 patients are tnie E.histoZjtica. Therefore the Merlin Optimun S 0 and E-hisrolyrica test @ shouid be tested in an endemic area as well where infections are more common. Since the majority of fecal specimens in the developed world are submitted in preservative, a major drawback to the current assays remains their requirement for a Eesh unpresewed fecal sarnpie. Aithough ours and previous studies suggest a LO-20% fdse-positive rate for serology (>l:64), higher seroiogic titers el :5 12) may help identifl E. histoiytica-infected patients. 1- Abuabara SF, Barrett JA, Hau T, Jonrsson O, Amebic Iïver abscess-Arch Surg

1982; 117-239-44,

2. Acana-Soto R, Samuebon 3, De Ghlami P, Zarate L, MilIan-Velasco F,

Schoolnick G, Wvth DmApplication of the poIymerase chain reaction to the epidemiology of pathogenic and nonpathogenîc Entamoeba histolytica Am J Trop Med

Hyg 1993; 48: 58-70,

3. Bracha R, NucharnoWitz Y, Mirelman D. MoIecuIar cloning of a 30-kilodalton lysine-rich daceantigen fiom a nonpathogenic Entamoeba histolytica strain and its expression in a pathogenic strain. Mect Immun 1995; 63: 9 17-25-

4. Britten D, Wilson SM, McNerney R, Moody AH, Chiodini PL, Ackers JP. An

ùnproved colorimetric PCR-based method for detection and differentiation of Entamoeba histolytica and Entamoeba dispar in feces. J Clin Microbiol 1997; 35: 1108-1 1.

5. Cabaiiero-Salcedo A, Viveros-Rogel M, Salvatierra B, Tapia-Conyer R,

Sepulveda-Amor J, Gutierrez G, Ortiz-Ortiz LmSeroepidemiology of amebiasis in

Mexico. Am J Trop Med Hyg 1994; 50: 412-9.

6. Diamond LS, Clark CG*A redescnption of Entamoeba histolytica Schaudinn, 1903

(Emended Waker, 19 11) separatirtg it fiom Entamoeba dispar Brumpt, 1925. J Eukaryot

MicrobioI 1993; 40: 340-4-

7. Gathiram V, Jackson TFG& Frequency distribution of Entamoeba histolytica zymodemes in a rural South Afiican population. Lancet 1985; 1:719-2 1. 8- Godez-Ru& A, Haque R, Aguirre A, Castanon G,HaU A, Guhl F, Ruiz-

Palacios G, MiIes MA, Warharst DC. Value of microscopy in the diagnosis of dysentery associated with invasive Entmueba hiirtoi'yticu- J Clin Pathol 1994; 47: 236-9-

9. Gutierrez G, Munoz O. Epidemiology of amebiasis Boca Raton, FL: CRC Press,

1990. h: Amebiasis - Infection and disease by Entmoeba histoTytica.

10. Haque R, Kress K, Wood S, Jackson TF, Lyeriy D, Wilkins T, Petri WA, Jr.

Diagnosis of pathogenic Entsmoeba histolytica infection using a stool ELISA based on monoclonal antibodies to the galactose-specXc adhesïn, J Infect Dis 1993 ;167: 247-9.

2 1. Haque R, Famque AS, Hahn P, Lyeriy DM, Petri WA, Jr. Entamoeba histolytica and Entamoeba dispar infection in children in Bangladesh. J Infect Dis 1997; 175: 734-6.

12. Healy GR hunologic tools in the diagnosis of amebiasis: epidemiology in the

United States, Rev Mect Dis 1986; 8: 239-46.

13. Insen EM, Jackson TE, Simjee AE. Asymptomatic intestinal colonization by pathogenic Entamoeba histolytica in amebic liver abscess: prevalence, response to therapy, and pathogenic potential. Clin Infect Dis 1992; 14: 889-93.

14. Jackson TF. Entamoeba histolytica and Entamoeba dispar are distinct species; clinical, epidemiologicai and serologicai evldence. Int J Parasitol 1998; 28: 18 1-6.

15. Juniper K, Sr., Worreii CL, Minshew MC, Roth LS, Cypert H, Lloyd RE.

Serologic diagnosis of amebiasis. Am J Trop Med Hyg 1972; 21: L57-68.

16. Kagan 1. Serologic diagnosis of parasitic diseases. N Engl J Med 1970; 282: 685-6.

17. Keystone JS, Keystone DL, Proctor EM. Intestinal parasitic infections in homosexual men: prevalence, symptoms and factors in transmission. CmMed Assoc J

1980; 123: 512-4. 18- Krogstad DJ, Spencer HC, JE, Heaiy GR, Gleason NN, Sexton DJ, Herron CA.

Amebiasis: epidemiologic studies in the United States, 19714974- Ann htem Med 1978;

88: 89-97-

29- Levine M, Lexchin J, Pellizzari R Drugs of choicer A formulary for general practice Ottawa, ON: Canadian Medical AssocÏation, 1997-

20. Mirelman D, Nuchamowitz Y, Stolanky T. Cornparison of use of enzyme-luiked immunosorbent assay-based kits and PCR ampiifkation of rRNA genes for simultaneous detection of Entamoeba histolytica and E, dispar. J Clin Microbiol 1997; 35: 2405-7.

2 1. Moshitch-Moshkovitch S, Petter R, Levitan A, Stolarsky T, Mirelman De

Regdation of expression of nisomal protein L-21 genes of Entamoeba histolytica and

E. dispar is at the pst-transcriptional level. Mol Microbiol 1998; 27: 677-85.

22. Petri WA, Jr. and U Singh. Diagnosis and management of amebiasis. Clin Infect

Dis. 1999; 29(5): 11 17-25

23- Pehrson, PO. Amoebiasis in a non-endemic country. Epidemiology, presenting syrnptoms and diagnostic methods. Scand J Infect Dis 1983; 15: 207-14.

24. Quinn TC, Stamm WE,GoodeU SE, Mkrtichian E, Benedetti J, Corey L,

Schufner MD, Holmes KKThe polymicrobial ongin of intestinal infections in homosexual men. N Engl I Med 1983; 309: 576-82.

25. Ravdin JI, Jackson TF, Petri WA, Sr., Murphy CF, Ungar BL, Gathhm V,

SkÏlogiannis J, Simjee AE. Association of serum antibodies to adherence lectin with invasive amebiasis and asymptomatic infection with pathogenic Entamoeba histolytica. J

Infect Dis 1990; 162: 768-72. 26-Sanuici J, Asai T, ObwaE, Kobayashf S, Takeuchi T. Identification of

Entamoeba histolytica ancl E. dispar cysts in stool by polymerase chain reaction. Parasitol

Res 1997; 83: 96-8,

27-Sargeaunt PG, Wiiams JE, Grene JD. The differentiation of Invasive and non-

invasive Entamoeba histoLytica by isoenzyme electrophoresis. Trans R Soc Trop Med

Hyg 1978; 72: 519-21.

28- Sargeaunt PG, Wjlliams JE, Electrophoretic isoenzyme patterns of the pathogenic and non-pathogenic intestinal amoebae of man, Tram R Soc Trop Med Hyg 1979; 73:

225-7,

29- Sargeaunt PG, Williams JE, Neal RA. A comparative study of Entamoeba histolytica (NIH :200, HK9, etc,), "E, histolytica-like" and other morphoIogically identical amoebae using isoenzyme electrophoresis. Tram R Soc Trop Med Hyg 1980;

74: 469-74.

30. Sehgal R, Abd-AiIa M, Moody AB,Chiodini PL, Ackers JP. Compatison of two media for the isolation and short-term culture of Entamoeba histolytica and E. dispar.

Trans R Soc Trop Med Hyg 1995; 89: 394.

3 1. Shion-Jeng O, Mei-Ying C, Kuo-Hui L, Chi-Byi He Use of the ProSpecT microplate enzyme immunoassay for the detection of pathogenic and non-pathogenic

Entamoeba histolytica in faecal specimens. Tram R Soc Trop Med Hyg 1996; 90: 248-9.

32. Smith PD, Lane EC, GU VJ, Manischewitz JF, Quinnan GV, Fauci AS, Masur

H. Intestinal infections in patients with the acquired immunodeficiency syndrome

(AIDS). Etiology and respome to therapy. Ann Intem Med 1988; 108: 328-33. 33, Stanley SL, Jr., Becker A, KunzJenkuls C, Foster L, Li E, Cloning and expression of a membrane antigen of Entamoeba histolytica possessing multiple tandem repeats, Proc Nat1 Acad Sci U S A 1990; 87: 4976-80,

34- Strachan WD,Chiodini PL, Spice WM, Moody AH, Ackers JP. Imniunological dserentiation of pathogenic and non-pathogenic isolates of Entamoeba histolytica

Lancet 1988; 1: 561-3.

35- Tannich E, Horstmann RD, Knobloch J, Amold HlL Genomic DNA Werences between pathogenic and nonpathogenic Entamoeba histolytica. Proc Natl Acad Sci U S A

1989; 86: 5 118-22,

36. Tannich E, Burchard GD. DBerentiation of pathogenic fiom nonpathogenic

Entamoeba histolytica by restriction fragment analysis of a single gene amplifïed in vitro.

J Clin Microbiol 1991; 29:250-5.

3 7. Troll H, Marti H, Weiss N. Simple differential detection of Entamoeba histolytica and Entamoeba dispar in fkesh stool specimens by sodium acetate-acetic acid-formalin concentration and PCR. J Ciin Microbiol 1997; 35: 170 1-5.

38. Upcroft J, Upcroft P. Future therapeutic alternatives for combating anaerobic protozoa- Parasitol, Today 1998; 14: 21 1-4,

39. Walsh JA. Problems in recognition and diagnosis of amebiasis: estimation of the global magnitude of morbidity and mortaiity- Rev uifect Dis 1986; 8: 228-38.

40. Weinke T, Friedrich-Janicke B, Hopp P, Janitschke K Prevdence and clinical importance of Entamoeba histolytica in NO high- risk groups: travelers retuming fiom the tropics and male homosexuals. J Wect Dis 1990; 161: 1029-3 1. 41 - WHO/PAHO/UNESCO report. A consultation with experts on amoebiasis, Mexico city, Mexico, 28-29 Jaauary, 1997. Epidemiol Bull 1997; 18: 134.

42- Xïmenez C, Leyva O, Moran P, Ranios F, Melendro EI, Rruniro M, Martinez

MC, Munoz O, Kretschmer R, Areliano J. Entamoeba histolytica: antibody response to recent and past invasiasiveevents. Ann Trop Med Parasitol 1993; 87: 3 1-9-

The efficacy of an immunochromatographîc strip (BIOSiTE Triage@) capable of detecting both Entamoebu histo&ticd E- dispar and Gimdia Zmblra coproantigen in fiesh-fiozen stool was evaluated (n=7L). BIOSITE Triage0 was 68.3% sensitive and

100%specific for E. histolyticd E. dispar detection compared to Alexon-Trend's

ProSpecTB test (reference standard). The ProSpecTB stool antigen ELISA was previously shown to outperform microscopy in both cornmunity laboratones and referral centers for the detection of E- histo&tica/ E- dispar complex. Neither test is able to distinguish E. histolytica fiom E. dispur. BIOSlTE Triage@was 83.3% sensitive and

100%specinc compared to microscopy (wet mounts and permanent stained smears) for the detection of CE lamblia. Entamoeba histolytica, Giardia lamblia, and Cryposporr'dimparvum are three of the major causes of protozoa-induced diarrheai disease (2,3,11). E. histolytica is responsible

for approximately 100,000 deaths worldwide each year, making it second oniy to malaria as a cause of mortaiity due to a protozoan parasite (1 1). Miction with G. lamblia is the most common parasitic infection reported in the United States and on a global scde giardiasis is responsible for approximateLy 100 million infections muaily (6). Infection with the coccidian parasite C.pmwm is usually ~e~limitingin immunocompetent individuais, but can be chronic and potentiaiiy Me-threatening in the immuw- compromised host (3). Contamination of municipal drinking water with C. pmm resulted in over 400,000 infections in Milwaukee during 1993 (7). Traditionally, the laboratory diagnosis of E. histolylica, lamblia, or C. pmvum infection is established by the microscopic examination of either fiesh or kedstool. However, microscopic diagnosis has several limitations and recent studies have reported that stool antigen imrnunoassays equal (G. lamblia} or surpass (E. histolyticd E dispar) microscopic detection of either of these pathogens (1,5,8,9).

Since multiple protomal infections may CO-existand in an effort reduce the cost per test, a stool enzyme immunoassay capable of simultaneously detecting E. histoZyticd

E. dispar cornplex, G. lamblia, and C .parvum has recently been developed (Triage0

Parasite Panel, Biosite Diagnositics, San Diego, CA). Triage@ is a single

immunochromatographic (IC) strip coated with monoclonai antibodies specific for 29- kDa surface antigen (E. histoljtica/ E .di'spm), alpha-1-giardin (G. lamblia), and protein disuffide isomerase (C pazyurn) (10)- In a previous study whkh assessed d available stool antigen ELISA kits (Techiab's E-oeba test@ and E- histolyitca test@, Merlin

Diagnostika's OptirnudB S, AIexon-Trend's ProSpecTB), we showed that ProSpecTO is the most sensitive and specific stool antigen ELISA currentiy available for the detection

E. htîrolyrcd E- dispar cornplex (sep chapter 6). In addition, ProSpecT@ outperfonns microscopy (wet mounts and permanent stained mnears) carried out in either co~rmunity laboratories or referral centers. In this study, we evaiuated the performance of'ïriagea compared blincüy to: 1. ProSpecTO for the detection of E hisrolyticd E- dr'ipar antigen in patient specirnens as well as in mock reconstitution experiments, 2, Microscopy (wet mounts and permanent stained smears) performed at a referral center for the detection of

G. lamblia in patient specimens. 7.3.1 Stndy popdation

Patients presenting to the Tropical Disease Unit of The Toronto Hospital, McGiil Centre

for Tropicai Disease at the Montreai General Hospital, and the Infectious Diseases and

Tropical Medicine CLinic, Hamilton Heaith Sciences Corporation (McMaster Campus) with gastrointestmal symptoms (diarrhea, defÏned as 13 loose [conform to the container] bowel movementdday, aabdominal pain, nausea, weight Ioss, bloody stooI, etc.) andor risk factors for E. histo&tica/EEdispar, G- Cablia, or C.pmvum infection (recent travel

[within 6 months], men who have sex with men, or recent immigrants fkom the tropics/subtropics [< 2 years]) fiom 1993 to 1998 were eligible for inclusion in this study.

Verbal ùIfocmed consent was obtained fiom each patient and the study was approved by the Ethical Review Cornmittee of The Toronto Hospital and the Montreal General

Hospital. Subjects were requested to provide fiesh stool sampies (within 1 hour of passage) for microscopy and ELISA analysis-

7.3.2 Stool antigen immunochromatographicstrip assay

Stool specimens (minimum of 2) were transported to the parasitology laboratov for routine ova and parasite evaluation by rnicroscopy (wet mounts, iron hematoxylin and modined acid-fast stains followhg formalin-ether concentration). Aliquots of fkesh unpreserved stool were fiozen (-20°C) for subsequent ELISA anaiysis.

Once thawed ,samples were resuspended in specimen dilution bu6er (bufféred protein solution/ 0.1% Na3)provided in each kit The assay procedure for both Triage@ and ProSpecT@ was camïed out according to the manufacturer's recommendations, A positive -on in the Triage kit is identifTai by a qualifative colorimetric reaction when the antibody conjugate (alkaline phophatase) reacts with the substrate (indoxyl phosphate) resulting in a dark blue/pucple heon the IC strip. The ProSpecT ELISA was read at bso using a microplate reader (Themom,-, MoIecdar Devices Corp., Sunnyvale, CA). The

ELISA plate format of the ProSpecTB allows for multiple tests (up to 96 per ELISA plate) which reduces the assay time per stoal. The Triage@ kit had the advantages of being completely self-contained and cmbe stored at room temperature. Triage@ dso contains an intemal positive and negative control for each test strip.

--

The results of stool antigen detection by Triage@ and ProSpecTB are summarized in

Table 7.1. Triage@ was 100% specinc and 68.3% sensitive compared to ProSpecTa

(used as the reference standard) for the detection of E.. hisrolyrcd E. dispm cornplex. The lower sensitivity of the Triage@ IC strip was explained in part by a 4-fold difference in the limit of detection for E. histolyticd E- dispm trophozoites. HM1 :IMSS trophozoites grown axenically in YI-S medium were seriaiiy diluted in specimen dilution buffer and subject to both Triage@ and ProSpecT@ assays (4). The results of these reconstitution experiments are indicated in Table 7.2. ProSpecTB was able to detect E. histolyticd E. dispm antigen at 250 trophozoites per mL, whereas Triage@ requked greater than 1000 trophozoites per mL for an unequivocal positive signal. Aithough Triage@ is a qualititative colorimetric test, we found that band intensity correlated well with Table 7-1: Results of Triage@ versus Aiexon-Trend ProSpecT@ for the detection of E-

- -- Alexon-Trend ProSpecT@ ELISA 1 Positive Negative TOTAL

Positive 28 13 41

Negative l O 1 30 1 30 Table 7.2: Limit of detection of E. histo&ticddispar by BIOSITE Triage@ IC stnp and

Aiexon-Trend ProSpecT@ ELISA.

Number of Alexon-Trend BIOSZTE Triage@ BIOSITE Triage@

Trophozoites/mL ProSpecT@ (oD)' IC Strip before IC Strip der

(HM1 :IMSS) f3tration2 fiItration2

105 ++f3

104

103

500

250 100 10 l.OD>O.100 was regarded positive(+) for ProSpecTB in this assay run.

2. Filtration of stool specimens resuspended in specirnen dilution baer(protein buffered solution containing 0.1% Na3)was carrîed out with the filter devices provided in the

Triage kit 0 accordhg to the manufacturer's recommendatiom (Biosite Diagnostics).

3. Band intensity using Triage0 measured visually by at least 2 readers: +++ very strong

; * strong ;+ medium; +/- weak; - none Table 7.3: Redts of Triage0 versus Standard O & P Mlcroscopy for the detection of E histo&ticd E- dispar

Standard O & P Microscopy

Positive Negative TOTAL

- - TOTAL trophozoite number independent ofthe stnp reader (Table 72)-At 500 trophozoites per mL (before filtration) or 1000 trophozoites per mL @£ter nitration), band intensity was weak and therefore prone to subjective reader error.

Cornparison of Triage@ rwïth microscopy for the detection of G, Imblia is sumrnarïzed in TabIe 7.3. A minimum of 2 stools per patient were used for both direct wet mounts and iron hematoxylidmodified acid fast staining- Aiiquoted fkesh-fiozen specimens were subsequentiy tested by IC strïp. Triage0 was 83.3% sensitive and 100% specific compared to microscopy perfonned blindly at our referral center (~71).

Triage@ was able to detect two mixed infections containing both E. histolyticd E. dispm and G. lamblia. No pantum infections were detected in these sampies by either microscopy or Triage0 IC strip.

In summary, the Triage@ IC strip is highiy specsc for the detection of E- histolytica/ E. di~purcornpiex. However, Triage0 is less sensitive (68 -3%) than an alternative ELISA diagnostic kit (ProSpecTB) for the detection of E. hisfuyticdE* dispm. The Lower sensitivity of Triage@ may be due to its inabiiity to detect E. histolyricd E- dispar antigen at or below 1000 trophozoites per d,especiaily following the required filtration of the sample (Table 7.2). ProSpecT@ remains positive at 250 trophozoites per mL. No association has ken made between parasite burden (number of cysts or trophozoites shed sporadically in stool) and clinical severity. Molecular epidemiological data suggest that E. histolyticu is associated with invasive disease, whereas the geneticdy distinct species E. dispar redts in asymptomatic infection (see chapter 6). Triage@has the advantage of being able to detect muitipIe protozod pathogens in a single test as evidenced by the 2 cases of mixed E. hÏstoCyhCd E_ dispar and GI lamblia infections in this study- Compared to reference microscopy, Triage@ was 83.3% sensitive and 100% specifïc for the detection of G. laddio. We propose that Triage@ is a useful alternative system for the detection of multiple pathogens in stool- Both of these tests are limited by an inability to distinguish pathogenic E. histoZytica fiom non-pathogenic E. dispar, and the requlrement for fiesh or fiesh-fiozen stool. Addiss, DeG., HmM. MathewspJ- Me Stewart, S. P. Walüquist, R M.

Williams, R J. Finton, H. C. Spenser, and J. D.D. 1991. Evaluation of a commercially available enzyme-linked immunosorbent assay for Giardia Iamblia antigen in stool. J.ClhMcrobiol,29: 1137-1 142.

Black, R E., A, C. Dykes, S. P. Sinclair, and J, G. WeUs. 1977. Giardiasis in day-care centers: evidence of person-to-person transmission_Pediatrics. 60:486-

491.

Chapman, P. A. 1988. Cryptosporidiosis: recent trends in epidemiology, diagnosis, and treatment. Serodiag, & Immunother. Infect, Dis- 2:3 1 1-3 17-

Diamond, L. S., C, G, Clark, and C. C. Cunnick 1995, M-S, a casein-fiee medium for axenic cultivation of Entamoeba histolytica, related Entamoeba,

Giardia intestindis and Trichomonas vagindis. J Eukaryot Microbiol. 42(3):277-

8.

Haque, R, 1. IL Ali, S. Akther, and W. A. Petri, Jr. 1998. Comparison of PCR isoenzyme analysis, and antigen detection for diagnosis of Entamoeba histolytica infection. J Clin Microbiol. 36(2):449-52.

Kappus, Id, and D. Juranek. 1988. Giardia in the well. JAMA. 259:1810.

MacKenzie, W. R, NoJ. Hoxie, M. E. Proctor, and et al. 1994. A massive outbreak in Milwaukee of cryptospondium infection transmitted through the public water supply. N Engl J Med, 33l(3): 161-7. 8, Mirriman, Dm,Y. Nuchamowitz, and T, Stolarsky. 1997- Cornparison of use of

enzyme-linked immmoso~ntassay-based kits and PCR amplification of rRNA

genes for simultaneous deteaion ofEntamoeba histolytica and E. dispar- I Clin

Microbiol- 35(9):2405-7,

9- Newman, R D., K L, Jaeger, T. Wuhib, A. A. M. Lima, R L. Guerrant, and

C. L. Sears. 1993. Evaiuation of an antigen capture enzyme-linked

immunosorbent assay for detection ofCryptosporidium oocytes. J-ClinMcro.

3 l:2O80-2084,

10. Reed, SeL., B. M. Flores, M. A. Batzer, M. A. Stein, VeL. Stroeher, J. E.

Carlton, D. L. Diedrich, and B. E. Torian. 1992. Molecular and cellular

characterization of the 29-kiiodalton periphecal membrane protein of Entamoeba

histolytica: differentiation between pathogenic and nonpathogenic isolates. Infect

Tmm~a.60(2):542-9.

11. Walsh, J. A. 1986. Problems in recognition and diagnosis of amebiasis:

estimation of the global magnitude of morbidity and mortality- Rev Infect Dis.

8(2):228-38.

Direct demonstration of WGaINAc lectin heavy subunit hctionality has been made possible through very high-aEhïty neogEycoconjugates and in viîm protein synthesis-

Recent data reported by the Schnaar group at Johns Hopkins University have shown that

E. histo2yticc1membranes bind to the neoglycoconjugate, GaNAcrnBSA, with over 10'- fold greater etythan GaiNAc alone (4)- This resuIt was intrïguing because it suggested that the amebic lecti.was capable of achieving very high affbïty association by either (i) the additive effect of multiple lechin-sugar interactions ador(ii) the additive effect of several CRD-sugar interactions within a single lectin molecule (30). However, these experiments relied on amebic membranes and therefore the qyestion of which of the two lecein subunits (1 70-kDa or 3 l/3 5-kDa) possessed CRDs remained unresolved.

Monoclonal antibodies (rnAbs) directed against the heavy subunit were adherence inhibitory in the standard amoeba-CHO ce11 adherence assay, suggesting the CRDs lay within the 170-kDa subunit. This dîd not provide sufEcient proof since antibodies are large molecules (-1 60kDa) and can exhibit steric hindrance (non-specific) of cellular adherence. The amoeba-CHO cell assay, developed by Ravdin and Guerrant and widely used in the literature, was shown to be solely lectin-dependent Since amoebae are cytotoxic to most mammalian ceiis at JPC, adherence assays are performed at the non- physiologic temperature of4OC (28).

We approached the CRD question in a different way: hgments of the lectin

(heavy and light subunit) were synthesized in viîro with 35~-aminoacid incorporation Figure 8-1- Schematic representation of in vitro binding assay, lir vitro traoscrïbed and translated 35~-labelledlecti. (heavy subunit) is added to microtiter weiIs coated with

GaINActsBSA, washed, and bound protein qumtifl:ed by ScintilIation coiniter-

scintillation counter using rabbit reticuiocyte Iysate. Since in viho tmmcription and translation yields minute

quantities of protein, a very hi&-affinity ligand was required in order to generate binding

isotherms and generate affinity constants- As showpreviously, neogIycoconjugates

bound to lectin in the nM range and were therefore s~ïtablecandidate ligands (4). We

obtained commerciaLLy avaiiable neoglycoconjugates, immobiüzed them to microwelIs,

and were able to directly demonstrate that region 356-1 143 (cysteine-rich region) of the

heavy subunit, but not other fragments (1-480) of the heavy subunit, was suflicient for

specinc, high-aff?nity binding (Chapter 2). A smaiier ment(480-900) within the cysteine-nch region did not bind to the same maximal activity as 356-1 143. The binding afflnity of the 3 56-1 143 was in the pM range, suggesting a ver-high afEnity association characteristic of multivalent interactions. It is unclear whether the high affinity achieved in this in viîro assay is due to the presence of several CRDs within the heavy subunit or the requirement for discontinuous residues to provide the correct conformation for the high-affinity state.

These data beg two Merquestions: (i) Do discontinuous residues between 3 56-

480 or 900-1 143 contribute to high-affuity carbohydrate binding by forming a conformation-sensitive binding pocket - "conformational model"? (ü) Do several CRDs, which Lie within the heavy subunit which then cluster to provide several low afnnity interactions that combine to provide an avid interaction with multivalent neoglyco- conjugates - " subsite and subunit multivalency modei"? Evidence both supporting and rehting each model exists in the literanire. On the one han& it was reported that two mAbs directed against the heavy subunit (3F4,8A3), enhanced adherence in the amoeba-

CHO ce11 adherence assay. Since this assay is perfonned at 4OC, it is likely that these antibodies induce a high-affinity binding conformation on the lectin rïther than rekthg

more receptor to the daceby means of increased gene expression. Such "ligand-

induced binding" has been observed in mammalian leukocyte adherence (39)-

Conformational regdation of a hi&-aflkiity state would be consistent with discontinuous

residues contributing to a hi@-aff?nity state as would be expected in a 'cconformattiond

modei". To test this, we compared anti-lectin (heavy subunit) mAb reactivity in its native

state and following fixation (Chapter 4). The lectin expressed on the daceof E..

histolytica under native conditions was detected by mAb 1G7 (linear epitope: 482-8 18)

(23), whereas the homologous epitope on the E. dispm lectin remained concealed Once

fixed, lectins nom both organisms were reactive with 1G7. Lectin homologues between

E. dispm and E. histolylica in region 482-8 18 are 90% identical. Whïle both lectins retain

the LG7 hear epitope, the homologous region of the E. dispar lech was concealed.

Differences in sugar etyhave been documented between the two organisms, and it is

tantalizing to suggest that conformational differences contribute to fiinctional merences

observed in adherence assays (discussed in section 8.2) (1 5).

Altematively, the mAbs 3F4 and 8A3 may (instead of conformational regdation)

cause clustering of the lectin on the amebic surface leadhg to spatial rearrangement of

lectin molecules that afTords high-afnnty interaction in whole cell adherence assays. This

assumes that lectin-mediated signalling to the amebic cytoskeleton is possible at 4'C, the

temperature at which the antibody effects were observed Mathematical modelling of

binding data obtained with neoglycoconjugates suggested that enhanced binding phenornena associated with polyvalent ligands, like rJaiNAc,,BSA, may be explained by the additive effects of simple multivalent interactions (20). This would suggest that subsite (ifthere are more than one CRDs per Iectin molecuie) and subunit multivdency

(when using whole membranes) are more significant than conformation of the leckThe role of the iight subunit remains elusive. Recent developments in antisense inhi'bition of

E. histolytica gene expression by David Mirelman's group (Weizmann Institutte) showed that suppression of Iight subunit gene expression resdts in diminished amebic cytotoxicity (5). These data wouid suggest the üght subunit either facilitates hololectin expression or has direct cytotoxic effects through an undehed mechanism.

Very recent work published by the Petri group at the University of Viginia has provided evidence for a CRD Iocated between residues 895-998 of the 170-kDasubunit

(1 6). These workers bound radiolabeiled GaiNAc I~BSAto recombinant lechproduced in a bacterid expression system. Recombinant fiagrnent 895-998 was imumunoreactive with mAb 8A3 and bound GdNAcci9BSAin a GaINAc-inhibitable rnanner, snggesting the presence of a CRD.Unfortunately, the binding affinity was not detemhed in this assay, which reiied on radiolabelled Ligand binding lectin blotted on to PVDF membranes followed by densitometric analysis of autoradiographs. Estimation of the association constants for this fragment would indicate whether the entire heavy subunit bound with greater or eqyai affinity. If -es were equai, it would suggest that 895-998 contaiwd the sole CRD, whereas if the aEnity of the smaller Fagment was lower, it would suggest the presence of several CRDs or absence of the appropriate ''fknework" of required non- contiguous residues, within the 170-kDa subunit, Region 895-998 is in a jwtamembrane position (TM=12 10-123 S), and seems counter-intuitive for a single CRD. These data also suggest binding is conformation-insensitive since recombinant hgments were run on

SDS-PAGE gels under reducing and denaturing conditions, and dl1remained fkctional. When taken together, the experimentai strategis presented by Dodson et d and those

presented in Chapter 2 allow us to systernaticaily generate even smaller fhgments

between 356-1 143, allowing for the fine-mapping of CRDs and estimation of their

individual carbohydrate affinity- A crystal solution structure of a bacteriaüy-expressed

CRD with Gai or GaiNAc wodd go a long way in understanding the binding mechanism

of the amebic lech.

The devance of high-aff?nity adherence to neoglycoconjugates is made apparent

when one considers the abundance ofmucin (a naturallyi)ccurrïng glycoconjugate) in the

large intestine. Chadee and colieagues demonstrated that E. histolyrica trophozoites

bound rat intestinal mucin and human colonic mucin with vexy high-aff?nity @M range) through engagement of the GaVGaNAc lectin (1 1). Chadee also demonstrated that

experimentd intestinal amebiasis was characterized by trophozoite induction of and

attachment to lumenal mucin. In certain instances, amebic mucosal erosion occurs accompanied by mucosai entry into the crypts, but not the lamina propria Epitheliai destruction ensues leading to neutrophil recruitment and activation of localized

in8ammation. Since mucin attachment appears to precede downstream cytolytic and

inflammatory events, it is conceivable that the generation of agents (antibodies or glycoconjugates) capable of specifically disrupting (and not enhaucing) the lectin-sugar

interaction may prove usefül in therapeutic interventions. To this end, a more detaïied understanding of the conformational regulation andor subsite multivalency of CRDs within the heavy subunit wiii permit a rationai approach. Perhaps the most profound development in Entamoeba research has been the near

universal acceptance of E- dispar as a distinct speciies, Early work by Sargeaunt @ondon

School of Hygiene and Tropical Medicine), in collaboration with researchers at the South

Afiican MedicaI Research Council, demonstrated that isoenzyme profiles (zymodemes)

correlated with virulence in the human host (33). Non-pathogenic zymodemes were coilectively renamed 6CEntmoebadispar", and pathogenic zymodemes, ccEntatnoeba histo&ticd', in keepùig with EdeBnunpt's prescient observations much earlier in the

1920s (1 4). Confounding matters, however, have been sporadic reports by investigatoa in endemic countries that so-cailed E. divar zymodemes have been observed in patients with symptomatic disease. In addition, data presented by Mirehan and colleagues have supported the conversion of E.. histoiyh'ca and E, dispar trophozoites cdtured in viîro

(25). These resuits rexnain controversial and have not been successfbiiy replicated by other laboratories,

The epidemiology of E. histolytica and E. diipm infection in a non-endemic setting was addressed by Weinke and coileagues in Berlin, Germany (46). These investigators followed over 2700 travelers and 320 active gay males (both high risk groups) and showed that 16.3% of the gay male cohort and 4.0% of the travelers had

E. hisçolytica infections. Amongst infected traveIers, 92.5% were E. dispar infections by zyrnodeme analysis, and ali 100% of infections in gay males (presumably due to local, direct traflsmission) were due to E. dispar. Of the 112 individuals with risk factors for Figure 8.2 Schematic representation of Chou & Fasmm plot of amebic heavy subunit homologues fiom E. dispm (Dhgi2) and E. histolytica mgi2) (L2). Secondary structure prediction demonstrates regions of a-helix, P-sheet, and B-tum conservation (N terminus) as weU as divergence (greater predicted a-helical secondary structure in Hgi2 in the extraceLiular, juxtamembrane region, see arrow)- Homology modehg and threading are not possible for the amebic lectùi since these genes bear no sequence similarity to crystallized Lectins, E- histoCytica infection in Toronto, 72 were actualiy infecteci, and of those 95.8% were

due to E. divar (Chapter 5). Therefore, while the inter-conversion of E- histoiytica and

E. di- in culture temains co~trove~al,it is becoming cIear that the vast majority of

Entamoeba human infections are due to an organism which appears to be a harmiess

commensal and non-invasive, while a very smail minority (-40%) are infected with E-

histolyria semstricto. Interestingly, infection with this organïsm does not guarantee an

invasive episode since individuais appear to be able to tolerate infection for years on end

(Chapter 1.3). These kuids of epidemiological and clinical scenarïos beg the question: are

structure-fünction merences between E. histo[ylica and E. dispar Wencefactors",

such as the Gal/GaiNAc lectin, informative in understanding the differential vinilence in

vnro?

For some tirne? a tecimical diniculty precluded a strict compare and con- study

between E. histolylica and E. dispar, namely that E. dispar appeared to require CO-culture

with bacteria (xenic culture) in order to survive, whereas E. histolyrica is able to survive

in the absence of other organisms (axenic culture). However, Seiki Kobayashi (Keio

University, Tokyo) and Graham Clark (London School of Hygiene & Tropical Medicine)

have recently devetoped axenic media for E. dispar culture (13), (21). Using a PCR

amplification approach with cDNA prepared fiom an axenic culture of strai. SAW 760, a

gene homologous to E. histolytca hg12 was identined in E. dispm (Chapter 3). Complete

sequence analysis, analysis by Southern blot hybridizatrion, and RT-PCR indicated that

this gene was not present in E. histolylica, was actively transcribed, and possessed a high degree of sequence identity (87% at the amino acid level) and predicted secondary

structure (Fig. 8.2). Cornparison of the CD59-Like domain on the heavy subunit impiicated in complement tesistance also showed conservation- A restriction fiagrnent

length polymorpbi'sm (.1P) in the lechhomologues (hg12 and dhgl2) pennitted a

straightforward PCR-RFLP approach to distinguishing E. dispar fiom Elhistolytia, and

may prove usefiil in diagnosis-

Immunoprecipitation of the lectin heavy subunits as weil as immunocyto- chemistry of E. histo&tica and E. dispar trophozoites demonstrated epitope conservation to a panel of mAbs (1G7,3F4,8A3) dïrected against the heavy subunit Whiie 3F4 and

8A3 recognized the native Iectin of E dispar, 1G7 did not detect this protein under native conditions on the ceii surface, only when fïxed, suggesting conformational regdation of the lectin .This may have important hctional impiications. For example, adhesion and c'de-adhesion" could be achieved by modulation of lectin conformation and cytoskeletal changes that allow trophozoîte movement The Mann group perfonned E. dispar trophozoite-CHO celi adherence and cytolysis assays, mdreported that the E. dispar adhered to and lysed target cells to a lesser extent than E. histolytica (15). Adâitiondy, axenicdy-cultured E-dispar failed to form abscesses in the hamster mode1 foIlowing intra-hepatic inoculation, whereas E- histoZyticu forrned abscesses with typical centra1 necrosis (Dr. Martha Espinosa-Canteiiano, Seminario in Amebiasis, Mexico City, 1997)-

The 1G7 epitope maps to 482-8 18 that falls within the cysteine-nch fiagrnent 480-900 shown earlier to bind GalNAci9BSAand therefore contains one or more CRDs (Chapter

2). The conceaiment or folding away of the 1G7 epitope on the native E. dispm lectin may explain the lower affinity for GaiNAc in whole ce11 adherence assays. In support of this, in viho bindiig assays of the homologous heavy subunits @hgU and Hgi2) demonstnited that both E. histolytica and E. dispar lectins when synthesized in vitro Figure 8.3 Workùig hypothesis ofE.hislojtka and E. diqm GaVGaiNAc lectin. Models of Lectin-dependent adherence for E. histolytica and E. d- are depicted. Transcnpt levels and &ce expression are higher ui E. histolytica- The heavy subunit CRD may be constitutively exposed on the surface of E. histolyrica but not E. dispar. intimate contact of target host cells leads to apoptosis either through receptor-mediated caspase activation or secondary to amoebapore-mediateci cytolysis and ca2+flux in target ceiis. bound GaiNAccisBSAwith equivaient affinity- In vilro synthesized Dhgl2 and Hg12 were immunoprecipitated by 1G7 suggesting that th3 epitope is exposed in this binding as-

A discrepancy in the relative amomts of lecth expressed on the amebic surface could also account for different adherence propeities of E. histolyrica and E. dispar-

Assessrnent of RNA transcript Ievels demonstrated that E. histu[ylica expresses more lectin @eavy and light subunits) traoscnpt than E. dispar, leading to more protein expressed on the E. histoiyticu surface- Greater Iectin expression may contribute to higher avidity bindùig of carbohydrate, especiaiiy since multivaiency bas been demonstrated to afford higher avidity adherence of amebic membanes to carbohydrate. However, media conditions for E. dispar straias, although axenic (no iive organisms present), are not exactly identical to E. histolyrica and it is possible that different nutrients (gluconic acid or dihydroxyacetone), mucin, or faricuIuta proteins may modulate Iectin gene expression and/or alter anti-lectin antibody binding.

From the parasite's eye view, water-borne passage only requires consistent release of cysts into the water, the cornmonest scenario in the developing world. Due to water filtration systems in non-endernic settings direct transmission, as seen in fecal-oral contact in gay males or institutionalized settings where unhygienic conditions may occur, is the only apparent mode of infection (18). Amoebae possess a potent cocktail of cysteine proteases and pore-forming peptides such that intimate contact with trophozoites usually result in lysis of the host cell, be it epitheiïal ceiis, neutrophiis, lymphocytes, or macrophages (8,Z). Retention ofcytolytk capammay be selected for evolutioaariiy

only in so far as theyaiiow amoebae to feed on enteric bacteria and defend against host

immune effector hctions- Since disruption of the intestinal epithelium, entry into the

portal circulation, and seeding of the liver confers no evolutionary advantageto the

parasite and often kills the host who cannot then pass on infective cysts' extra-intestinal

amebiasis iikely occurs "'by accident". Hence, many so-calied 'Wence factors"

descnbed in the Enrainoeba fiterature may in fact be housekeeping proteius gone awry.

Clinical epiderniologica1 data support the above view in the sense that only a smali

fraction (1 per 1000) of those infected with pathogenic E. histoijfïcaactuafly succumb to extra-intestinal symptoms, and of those many are imrnunocompromised during pregnancy, poor nutritional status, or immunosuppressive dnig therapy (glucocorticoid therapy) (1,4 1,45).

The lectin appears to mediate adherence to most host ce1 types (9,28).

However, it wodd be ditFcult to rationalize the abiiity of this motile parasite to traverse intestinal epithelium, extracellular matrix (-CM), as weil as endothehum with a single adhesion molecule. To this end, Talamas-Rohana's group at CINVESTAV in Mexico purified and characterized an amebic p 1 integrin that apparentiy bhds to fïbronectin, an

ECM constituent (42). Other investigators have identined a hyaluronic acid-binding protein expressed on the amebic surface that is antigenicdy similar to CD44 (29,3 1)-

Initial studies with anti-integrin antibodies had suggested the presence of 82 integrin

(CD 18)-like protein present in E. histolytica (3). Further investigation of the 82 integrin- like protein indicated that it was expressed on E. histolylica and E. di- and possibly Figure 8-4- A mode1 where trophozoites aâhere to mucus in the large intestine via clustered GaVGaINAc lectin is presented, Coionization of the bowel following initial adherence may be facfiitated by the degradative action of cysteine proteases on muci.,

IgA and other soluble factors. Localized infianunation and mucosal ulceration aiiows deeper penetration of trophozoites that may trigger cytokine release fiom and apoptotic death of intestinal epithelial tells, Interleukin-8 is induced by E. histofyticasecretory factors fiom colonic ceU hes zh vitro and Zn intestinal amebiasis moâels (see text), IL-8 recruïts neutrophils in part by up-regulating ICAM-1- This may dowtrophozoites

(expressing putative 82 integrin) to attach to intestinal epithelial ceils and penetrate into the submucosa- Here, trophozoites can traverse ECM proteins like flbronectin via amebic

B 1 integrin, and subsequentiy attach to endotheLiai cells (where 82 integrin-ICAM- 1 interaction may also occw) and enter the portal circulation.

-- Lumen required for adhesion at intestinai sites CE. diipar has never been reported to cause invasive disease),

The amebic B2 htegrin was distllict nom the OaVGaINAc Iectin and supported specific adherence of amoebae to target ceiis expressing ICAM-1 following stimulation with the pro-infiammatory cytokine TNFir (Chapter 5). These data are consistent with a mode1 where E, histolyticu trophozoites opportunisticaily adhere to up-reguiated ICAM-1 during parasite-induced inflammation that occurs at intestinai epithelium or vascuiar endotheiium. Interestingiy, immunohistochemicd studies have shown up-regdation of

[CAM-1 on endothelium adjacent to hepatic abscess sites (44)- Yi Yu in the Chadee lab

(McGiU University) showed that trophozoites stimuiated IL-8 in several colonic epitheliai ceiis -- which is known to up-reguiate ICAM-1- and thereby faciltate neutrophii transepitheliai migration (48). ICAM-1 may aüow trophozoites to tether at luminal surfaces of intestinal epitheiia or aord migration out of the lumen into subepithelial layers. The sigaificauce of increased levels of amebic PZ integrin in E- dispar (the converse of what was observed with the GaVGaINAc lectin) is not clear, but may suggest that integrins are aiso housekeeping proteins rather than c'viruiencefactors" per se. The action oflectin, proteases, pore-forming peptides, and integrins in concert, combined with the pro-infiammatory cytokine milieu and the modulatory effect of the normal gut flora, likely pennits the invasive behaviour of E. histolytica and not ''E. dispar". While no clear data exist, this information is testable using in vitro transepithelid migration assays (26). Insights into genetic and molecuiar differences between E_ histo[yrica and E- dispar have aiiowed for the deveiopment of rapid molec~detectïon assays (43)-

Genetic amplification by PCR of genes in which a poIymorphism exîsts between the two species permits differentiation of pathogen nom non-pathogen. The greatest Mcuity lies in the purification of suffiCient quantities of good quaiity genomic DNA. The source is invariably stool where contaminants and inhibitors of amplification persist-

Cornpounding this is the fact that a signincant portion of Entmoebu genomic DNA in stool is encapsulated in cyst fonn and therefore not easily purinable. Cyst rupture is not trivial although several labs have suggested apparently successful protocols (2,7).

Sequence anaiysis of homologous Lectin genes (hgl2 and dhgl2) established a simple

PCR-RFLP approach that distinguished E. dispar fkom E. histolytica (Chapter 3).

The Petri group has pursued the finding that certain mAbs generated against the

Iectin (7F4) are specifïc to the E. histolylia, whiIe others (3F4) recognize E. dispm and

E. hisralytica- Using an ELISA approach, patient stool can be rapidly (ehrs) screened for antigen (19). This strategy has proven more sensitive and specinc than traditionai microscopy and equal to isoenzyme analysis, which is impractical in a diagnostic setting

(Chapter 5). More important, the dissociation of pathogen from non-pathogen ailows for identification of patients who actually require treatment. Currently, diagnosis by microscopy does not aiiow for specinc identification of the pathogen and resdts in unnecessary chemotherapy that may lead to hg-resistance, adverse dnig reactions in situations of polyphanriacy, and attendant costs to the health care system. Inherent

problems with stool antigen diagnosis of Enfamoebuexist, Foremost among these are the

sporadic nature of cydtrophozoite shedding that rpquires multiple samphg in order to

deout infection and the requirement for fiesh or fieshaozen stool, Stool tïxatives such

as formalin and PVA appear to preclude both PCR and ELISA diagnosis. hterestingly,

antigen ELISA has been used suçcessfuly to detect amebic antigen in the amebic iiver

abscess aspintes (Chapter 1). Ultimately, the development of accurate, cheap, and fast

immunochromatographic stnps able to detect multiple entenc pathogens (which present

with similar signs and symptoms) in fixed stool sampies may revolutionize the diagnosis

of infectious diarrheal disease.

The separation of the human immune system into systemic @one maxrow, spleen, and lymph nodes) and mucosal (gastrointestinal (Go, genitourinary, and respiratory) compartments is well accepted (40). UWeboth compartments do interact, this separation allows to us to better understand what may constitute protective immunïty when considering a mucosal pathogen, iike E. histolytica. Whereas mucin, defensins, and flushing of lumenal pathogens through osmotic changes at the intestinal epitheiium compromise innate imrnunity to pathogens, the induced mucosal immune system has both humoral (IgA) and ceil-mediated (cD~+T helper lymphocytes and CD~+cytotoxic T lymphocytes (CTLs)) components (6). The major inductive sites in the GI tract are the gut-associated lyrnphoid tissue (Gan,such as the Peyer's patches (PP), appenduc,

mesenterk Lymph nodes, and intraepfieliai lymphocytes. The cellular architecture of the

PP is optunized for antigen uptake, presentation, and induction of antigen-specinc

afnnity maturation of 1gAf B cells i5om I@ B ce& under the influence of appropriate

cytokine milieu and T celi help (CD43 (32). CTLs that recognize microbiai antigeas

presented in the context of MHC class 1molecules are particuiarly effective aga&

intraceUular paîhogens. These antigen-specinc B and T ceiis migrate fiom inductive sites

to effector sites via a systemic route (mesenteric lymph nodes +thoracic duct -+ blood

-+ lamina propria). Mucosal defense mediated by IgA can occur in the lumen where

secretory IgA (SI@) can neutralize microbes, within epitheiiai celis where dimenc IgA cm bind iniracellular antigen, and within the Iamina propna where dimenc IgA complex with antigen and be transported into the lumen.

The role of mucosal immunity in combathg E. histolytcu infection of the gut is undoubtedly crucial in preventing invasive disease. In spite of this, the most comrnonly used animal model (the Mongolian gerbil) for vaccine efficacy studies relies on intra- hepatic challenge, which bypasses the role of mucosal immiinity (27,37,38). Lectin based immunization studies focusing on protection fiom hepatic disease were descnbed previously (section 1-7.4). A suitable model for vaccine studies of intestinal amebiasis has not been developed to date. This model would require oral challenge of infective cysts either in feed or nasogastric tube, or rectal challenge of infective cysts or trophozoites (since stomach acids would not be an impedement with this route)- A mouse model would be preferrd since reagents to assess immunological changes abound for this species when compared to gerbils. Stanley and colieagues (Washington University) have managed to engraft fetai human intestine into a SCID mouse host in order to evaluate human intestinal responses to trophozoite challenge (36). However, this model does not allow for immunization studies since SCID mice are not able to mount an immune response. A srnail animal model in which successful colonizatio~~of the gut with E. histoCyrica occurs iikely requires either transient immunosuppression, reduction of the enteric flora, or strain-specinc or knockout gene-specinc susceptibility.

The majority of mucosal vaccines employed currentiy in humans are live attenuated organisms, such as oral polio, Salmonella typhi TQla, and Vi6rio cholerae. h general, live attenuated pathogem can revert and pose a heaith risk, especidy to immunocompromised individuais. Antigens are poor mucosal immunogens in the absence of adjuvants such as cholera toxin (CT) or heat-labile enterotoxin (LT) likely due to antigen degradation prior to reaching inductive sites. Given World Health

Organization (WHO) critena for vaccines which indude affordability, heat stability (no requirement for cold-chah), single dose administration preferably mucosally, and multiple specificity, DNA immunization remains a desirable option. McCIuskie and

Davis recently reviewed the advantages of mucosaI DNA vaccines (24):

(i) no infiectious agents are used, hence no potential for reversion

(ii) applicable to immunocompromised hosts (eg. HIV-infected or mainourished

in developing counûies)

(iii) induction of both systemic and mucosai systems, with special utility in

generating CTLs agaïnst intraceliular pathogens

(iv) mucosal administration can eiicit responses at distal sites (eg. oral can elicit

GI response) (v) long-Iasting irnmunity possible due to persistent à, vivo antigen synthesis

(vi) ease of administration, heat stability, and low cost/ease of manufacture

(vii) multivalent vaccines possible against seved paihogens

DNA immunkation has proven successful in several models, especiaiiy where

cytotoxic CDS+T ceil (CTL) responses are requùed (17). Recent deveIopments in our

understanding of the imrnunobiology of E. histolytica infection suggest that a Thl-like

(cellular) immune responses, characterized by the requirement for elevated IFN-y and IL-

2, is crucial to host defense (10). T'hl-promting cytokines ([L-4) combined with mucosal

adjuvants (such as bacterial toxins or newer compounds iike the purifïed saponin, QS21)

can be CO-immunizedto potentlate the immune response. Other workers have shown that

CO-immunkationof DNA constructs encoding the immunogen of choice coupled to either

GM-CSF or B7.2 can vastly increase the potency of the cytotoxic response- Furtherrnore,

the route of administration (eg. intranasal, intraoral, or intrarectal) or method of

inoculation (eg. gene-gun immunkation of mucosal surfaces or Liposome delivery

systems) can boost the location of a specific huneresponse- DNA immunization

administered systemicaiiy is usefil in priming the immune system prior to boosting with viral protein immunization (34,35). For example, priming with DNA systernically

(intramuscular injection) foilowed by intranasai administration of replicationaeficient adenovirai recombinants expressing the antigen of choice has been shown to bolster vaginal IgA and IgG2a responses (47). The pnming effect was specinc to DNA immunization and couid be enhanceci by IL4 genetic adjuvants. These data suggest that

DNA can dso be coupled to other vaccination strategies to induce mucosal immunity at distai sites. These studies are now possible with the GaVGaiNAc Iectïn gene as DNA

immunogen since expression ofthis protein in COS7 cells has been achieved with high efficiency in viîro (Chapter 4). Proof of principle WU remdemonstration of anti-lectin humorai responses (systemic and mucosal) able to neutralize adherence of amoebae and host cells, lymphocyte proMeration in response to amebic lectui, induction of a Th1 -&e repertoire of cytokines, as weli as CTL responses to target celis expressing the amebic lech in the context of appropriate MHC.

1. Abioye, A. A. 1973. Fatal amoebic colitis in pregnancy and puerpenum: a new

clinico- pathological entity. J Trop Med Hyg. 76:97-100.

2. Acuna-Soto, R, J. Samuelson, P. De Girolami, L. Zarate, F. MiUan-Velasco,

G. Schoolnick, and D. Wirth. 1993. Application of the polymerase chah

reaction to the epidemiology of pathogenic and nonpathogenic Entamoeba

histolytica. Am J Trop Med Hyg. 4858-70.

3. Adams, S. A., S. C. Robson, V. Gathiram, T. F. Jackson, T. S. Pilïay, R E.

Kirsch, and M. W. Makgoba. 1993. Tmmunological similarity between the 170

kD amoebic adherence glycoprotein and humm beta 2 integrinS. Lmcet. 341: 17-

9. 4, Adler, P., SDJ. Wood, Y. C. Lee, R T. Lee, W- A. Petri, Jr,, and R L.

Schnaar. 1995, High affrnitv bhdlng of the Entamoeba histolytica lectin to

polyvalent N-acetylgdactosamlnides- J Bi01 Chem- 270:s 164-71 -

5. Ankri, S., F. Padilla-Vaca, T. Stolarsky, LmKoole, U. Kab, and DDMirelman.

1999. Antisense inhibition of expression of the light subunit (3 5 kDa) of the

WGaiNac lectin complex hhiiits entamoeba histolytica vinilence. Mol

Microbiol- 33:327-37.

6. Bevins, C. L., E. Martin-Porter, and T. Ganz. 1999. Defensins and innate host

defence of the gastrointestinal tract- Gut 45:9 11-5.

7. Britten, D., S. M. Wilson, R McNerney, A-H. Moody, P. L, Chiodini, and J.

P. Ackers. 1997. An ïmproved colorimetrk PCR-based method for detection and

differentiation of Entamoeba histolytica and Entamoeba dispar in feces. J Clin

Microbiol. 35: 1 108-1 1.

8. Bruchhaus, I., T. Jacobs, M. Leippe, and E. Tannich. 1996. Entamoeba

histolytica and Entamoeba dispar: merences in numbers and expression of

cysteine proteinase genes. Mol Microbiol. 22:255-63.

9. Burchard, G. D., and R Bilke. 1992, Adherence of pathogenic and non-

pathogenic Entamoeba histolytica strains to neutrophils. Parasitol Res. 78: 146-53.

10. Campbell, D., and K. Chadee. 1997, Survival strategies of Entamoeba

histolytka: Modulation of cell-mediated immune responses. Parasitol Today.

13: 184-9.

11. Chadee, K., M. L. Johnson, E. Orozco, W. A. Petri, Jr., and Je 1. Ravdin.

1988. Binding and internalization of rat colonic mucins by the galactose/N- acetyI-D-galactosamine adherence Iectin of Ehtamoeba histo1ytka J Infit Dis.

158:398-406-

12- Chou, F. Y., and G, D. Fasmm. 1974. Predîction of protein conformation-

Bioche~mktry.13222-45-

13. Clark, CoG. 1995, Axenic cdtivation of Entamoeba dispar Brumpt 1925,

Entamoeba insoIita Geiman and Wichterman 1937 and Entamoeba ranarum

Grassi 1879-J Eukaryot Microbiol- 42:590-3.

14. Diamond, L. S., and C. G. Clark. 1993. A redescrïption of Entamoeba

histolytica Schaudinn_ 1903 (Emended Walker, 1911) separating it fiom

Entamoeba dispar Brumpt, 1925.1 Eukaryot Microbiol. 40:3404.

15. Dodson, J. M., C. G. Clark, L. A. Lockhart, B. M. Leo, J. W. Schroeder, and

B. J. Mann. 1997. Cornparison of adherence, cytotoxicity, and GaVûaINAc lectin

gene structure in Entamoeba histoIytica and Entamoeba dispar. Parasitol

Internationai. 46:225-23 5.

16. Dodson, J. M., P. W. Lenkowski, Jr., A. CoEubanks, T. F. Jackson, J.

Napodano, D. M. Lyerly, L. A. Lockhart, B. JOMann, and W. A. Petri, Jr.

1999. Infection and immunïty mediated by the carbohydrate recognition domain

of the Entamoeba histolytica Gal/GaiNAc lectin. J Infect Dis. 179:460-6.

17. Donnelly, J. JO,JO B. Ulmer, J. W. Shiver, and M. A. Liu. 1997. DNA vaccines-

Ann Rev Immunol. lS:6 17-48-

18. Ewald, P. W. 1995.. The evolution of vuulence: A unmglink between

parasitology and ecology. J Parasitol, 81:659-669. Haque, R,K, Kress, S. Wood, T. F-Jackson, D. Lyerly, T. Wis,and W.

A. Petri, Jr. 1993. Diagnosis of pathogenic Entamoeba histolytica infection using a stoo1 ELISA based on monoclonal anti'bodles to the galactose-speclfic adhesin

[see comments]. J Infect Dis, 167:247-9.

Hubble, J. 1999- A mode1 of mdtivalent ligand-receptor equiliia which explains the effect of rnuitivalent binding inhibitors. MOL immunol. 36: 13-8.

Kobayashi, S., E. bai, HeTachibana, T. Fujiwara, and T, Takeuchi. 1998.

Entamoeba dispar: cultivation with sterilized Cnthidia fasciculata, J Eukaryot

Microbiol, 45:3S-83-

Leippe, M., E. Bahr, E. Tannich, and R DmHorstmano. 1993. Cornparison of pore-forming peptides nom pathogenic and nonpathogenic Entamoeba histolytica-

Mol Biochem Parasitol. 59: 101-9.

Mann, B. J., and L, A. LockharL 1998, Molecdar analysis ofthe GallGaNAc adhesh of Entamoeba histolytica. I Eukaryot Microbiol. 45: 13s-16s.

McCluskie, M, J., and HeL, Davis. 1999. Mucosal immrinization with DNA vaccines. Microbes and IBfection- 1:685-698-

Mirelman, Dm,R Bracha, A. Wexier, and A. Chayen. 1986, Changes in isoenzyme patterns of a cloned culture of nonpathogenic Entamoeba histolytica durhg axenization. Miect Immun. 54:827-32.

Parkos, C, A. 1997. Molecular events in neutrophil transepithelid migration.

BioEssays. 19: 865-73. Petri, W. A., Jr., and J. r. Ravdin, 1991- Protection of gerbils from arnebic liver abscess by immuni7ation with the galactose-specîfïc adherence lectïn of

Entamoeba histolytica- Infect humun, S9:97-IO 1,

Ravdin, J. I., and R L. Guerrant. 1981 - Role of adherence in cytopathogenic mechanisms of Entamoeba hlstolytka. Study with m~mmaliantissue culture cells and banerythrocytes, J Clin Invest. 68: 1305-1 3.

Renesto, P., P. J. Sansonetti, and N-Guiilen. 1997-Interaction between

Entamoeba histolytica and intestinal epitheliai cells involves a CD44 cross- reactive protein expressed on the parasite surface, Infect Tnimun, 6S:43 3 0-3.

Rini, J. M. 1995. Lectin structure. Annu Rev Biophys Biomol Stnict- 24:S 1-77,

Rosales-Encina, J. L., M. S. Campos-Salazar, and M. Rojkind Matluk 1992-

Entamoeba histolytica coiiagen binding proteins- Arch Med Res- 23: 109-13.

Sansonetti, P. J,, and A. Phalipon. 1999. M cells as ports of entry for enteroinvasive pathogens: mechanisms of interaction and consequences for the disease process. Semin ImmunoI. 11: 193-203-

Sargeaunt, P. G., J. E. Wiarns, and J. D. Greene. 1978. The differentiation of invasive and non-invasive Entamoeba histolytica by isoenzyme electrophoresis.

Tram R Soc Trop Med Hyg. 72519-21.

Schneider, J., S. C. Gilbert, TeJ. Blanchar, T. Hanke, K. J. Robson, C. M.

Hannan, M. Becker, R Sinden, G. L. Smith, and A. V. Hill. 1998. Enhanced irnmunogenicity for CDS+ T ce11 induction and complete protective efficacy of malaria DNA vaccination by boosting with modified vaccinia virus Ankara Nat

Med- 4:397-402. Sedegah, M., T. R Jones, M, ffiur, R Hdstrom, P. Hobart, J. A, Tine, and

S. L. Honinann. 1998, Boosting with recombinant vaccuiia increases immunogenicity and protective efficacy of malaria DNA vaccine. Proc Nat1 Acad

Sci USA. 957648-53-

Seydel, K. B., EoLi, P. Eo Swanson, and S. L. Stanley, JE 1997. Human intestinal epithelial ceUs produce proinflammatory cytokines in response to infection in a SCID mouse-human intestinal xenograft model of amebiasis- Mect

Immun-65: 1631-9.

Soong, C. J., K. C. Kain, M, Abd-Aiïa, T. F. Jackson, and J. 1. Ravdin. 1995.

A recombinant cysteine-rich section of the Entamoeba histolytica galactose- inhibitable lectin is efficacious as a subunit vaccine in the gerbil model of amebic liver abscess. J Mect Dis. 171:645-51.

Stanley, S. L., Jr. 1997. Progress towards development of a vaccine for amebiasis. Clin Microbiol Rev. 10~637-49,

Stewart, M., and NoHogg. 1996. Regulation of leukocyte integrin fûnction: afinity vs- avidity. J Cellular Biochem- 6155461.

Strober, W. 1998. Interactions between epithelial cells and immune cells in the intestine. Ann N Y Acad Sci. 859:37-45.

Stuiver, P. Co,and T. JOL. M. Gou~.1978. Corticosteroids and liver amoebiasis- Br Med J. 2: 394-5.

Talamas-Rohana, P., V. 1. Hernandez, and J. L. Rosales-Encina. 1994. A beta

1 integrin-like molecule in Entamoeba histolytica. Trans R Soc Trop Med Hyg,

88596-9- 43. Tannich, E., R. D. Horstmann, J, Knobloch, and H. H. Arnold* 1989-

Genomic DNA differences betweenpathogenic and nonpathogenic Entamoeba

histolytica, Proc Nat1 Acad Sci U S A. 865118-22,

44. Ventura-harez, J., R Campos-Rodriguez, H. A. Rodriguez-Martin- A.

Rodriguez-Reyes, A, Martin--Pdomo, and V. Tsutsumi. 1997- Human

arnebic liver abscess: expression of intercellular adhesion molecules 1 and 2 and

of von WUebrand factor in endothclid ceils, Parasitol Res, 83:s 10-4-

45. Walsh, J. A. 1986. Problems in recognition and diagnosis of amebiasis:

estimation of the global magnitude of morbidity and mortality- Rev Mect Dis.

8:228-38.

46. Weinke, Te,B. Friedrich-Janicke, P. Hopp, and K. Janifschke. 1990,

Prevalence and clinical importance of Entamoeba histolytica in two high- nsk

groups: travelers returning fiom the tropics and male homosexuals. J Infect Dis.

161: 1029-3 1.

47. Xiang, Q. X, S. Pasquini, and H. C. J. Erti. 1999. Induction of genitd

immunity by DNA priming and intranasal booster immunization with a

replication-defective adenoviral recombinant. J Immunol. l62:67 16-6723.

48. Yu, Y., and K. Chrdee. 1997. Entamoeba histolytica stimulates interleukin 8

ftom human colonic epithelial cells without parasite-enterocyte contact-

Gastroenterology. 112: 153 6-47.