Journal of Medical Virology 72:424–435 (2004)

Mucosal Arenavirus Infection of Primates CanProtect Them From Lethal Hemorrhagic Fever

Juan D. Rodas, Igor S. Lukashevich, Juan C. Zapata, Cristiana Cairo, Ilia Tikhonov,Mahmoud Djavani, C. David Pauza, and Maria S. Salvato*

Institute of Human Virology, University of Maryland Biotechnology Institute, Baltimore, Maryland

Arenaviruses are transmitted from rodents tohuman beings by blood or mucosal exposure.The most devastating arenavirus in terms ofhuman disease is Lassa fever virus, causing upto 300,000 annual infections in West Africa. Weused a model for Lassa fever in which Rhesusmacaques were infected with a related virus,lymphocytic choriomeningitis virus (LCMV). Ourgoals were to determine the outcome of infectionafter mucosal inoculation and later lethal chal-lenge, to characterize protective immune re-sponses, and to test cross-protection between avirulent (LCMV-WE) and an avirulent (LCMV-ARM) strain of virus. Although intravenous in-fections in the monkey model were uniformlylethal, intragastric infections recapitulated thespectrum of clinical outcomes seen in humanexposure to Lassa fever virus: death, recoveryfrom disease, and most often, subclinical infec-tion. Plaque neutralization, ELISA, lymphocyteproliferation, and chromium-release assayswereused to monitor humoral and cellular immuneresponses. Cross protection between the twostrains was observed. The three out of sevenmonkeys that experienced protection were alsothe three with the strongest cell-mediated immu-nity. J. Med. Virol. 72:424–435, 2004.� 2004 Wiley-Liss, Inc.

KEY WORDS: arenavirus; mucosal infection;primates; hemorrhagic fever;immunity

INTRODUCTION

Arenaviruses are rodent-borne pathogens that cancause lethal disease in guinea pigs, primates, and hu-man beings [Oldstone, 2002]. Numerousmurine studiesof viral persistence and cell-mediated immunity featurethe prototype arenavirus, lymphocytic choriomeningitisvirus (LCMV) [Zinkernagel and Doherty, 1974; Ahmedet al., 1984; Salvato et al., 1991; Oldstone, 2002]. Incontrast to murine LCMV infections, LCMV infection ofRhesus macaques can resemble Lassa hemorrhagic

fever in human beings [Jahrling et al., 1980; Peterset al., 1987; Lukashevich et al., 2002, 2003]. This studydescribes the consequences of intragastric inoculation ofRhesus macaques with LCMV.

Mucosal exposure accounts for a significant propor-tion of natural Lassa fever virus infections. Every yearLassa fever virus (LAS) infects approximately 300,000human beings with 30% morbidity and 16% mortality[Jahrling et al., 1985; McCormick et al., 1986a, 1987;McCormick, 1990; McCormick and Fisher-Hoch, 2002].The natural reservoir for LAS is the rat Mastomysnatalensis, and the acquisition, preparation, and con-sumption of rodentmeat is amajor risk factor for rodent-to-human transmission [ter Meulen et al., 1996].Another indication that ingestion of arenaviruses canlead to disease is that some hepatitis outbreaks in zoo-kept tamarinds and marmosets were traced to feedingwith LCMV-infected mice [Montali et al., 1993, 1995].Our experimental studies showed thatmucosal inocula-tion leads to attenuated LCMV infection in mice [Raiet al., 1996, 1997] and in macaques [Lukashevich et al.,2002, 2003] in comparison with same-dose inoculationsby the intravenous route.

The WE strain of LCMV used in the present study ishepatotropic inmice, guinea pigs, and primates [Riviereet al., 1985; Zinkernagel et al., 1986; Lukashevich et al.,2002, 2003]. When inoculated intravenously, LCMV-WE, induces a lethal hemorrhagic fever in macaquesthat resembles human hemorrhagic fever [Jahrlinget al., 1980; Peters et al., 1987; Lukashevich et al.,2002, 2003]. TheArmstrong (ARM) strain of LCMVdoes

Abbreviations: moi, multiplicity of infection; pfu, plaque-forming unit; IU/L, international units per liter (referring toaminotransferase activity).

Grant sponsor: NIH; Grant numbers: RR138980, AI5252367,AI53620, AI53619.

*Correspondence to: Maria S. Salvato, Institute of HumanVirology, University of Maryland Biotechnology Institute, 725West Lombard St., Baltimore, MD 21201.E-mail: salvato@umbi.umd.edu

Accepted 13 October 2003

DOI 10.1002/jmv.20000

Published online in Wiley InterScience(www.interscience.wiley.com)

� 2004 WILEY-LISS, INC.

not cause overt disease, even after intravenous inocula-tion [Danes et al., 1963; Peter et al., 1987; Lukashevichet al., 2002].

Our goals in this study were to establish a model formucosal infection of non-human primates using theLCMV-WE strain, and to determine whether mucosalinoculation protected against a subsequent lethal viruschallenge. Mucosal inoculation of Rhesus macaqueswith LCMV-WE induced the range of outcomes seen inhumans naturally infected with Lassa fever virus, fromsubclinical to fatal, and varied in its ability to elicitprotective immunity. Cell-mediated immunity was re-corded in four of five mucosally inoculated monkeys;the fifth animal died before immune responses could beassessed. Two of five mucosally inoculated monkeyswere protected from lethal challenge: one protectedanimal was inoculated with LCMV-WE, and one pro-tected animal was inoculated with LCMV-ARM, that isknown to confer cross protection after intravenousinfection [Peters et al., 1987].

MATERIALS AND METHODS

Virus Stocks and Cell Culture

Large stocks of serum-free LCMV were produced inVero E6 cells and stored at 107 to 108 plaque-formingunits (pfu)/ml for use in all monkey inoculations. TheLCMV-WE and LCMV-ARM strains have been wellcharacterized [Salvato and Shimomaye, 1989; Djavaniet al., 1998], and their use in monkeys has beendescribed [Lukashevich et al., 2002]. Viral stocks usedfor proliferation assays were heat-inactivated at 568Cfor 30min. Stocks of vaccinia viruswild type (VVwt) andvaccinia viruses recombinant for LCMV glycoprotein(GP) and nucleocapsid protein (NP) genes, VV-GPLCMV

and VV-NPLCMV, were described elsewhere [Whittonet al., 1988]. Vaccinia viruses were titered as described

[Mackett et al., 1984]. All cell cultureswere incubated inhumidified chambers at 378C with 5% CO2.

Rhesus Macaque Inoculations and Challenges

Seven healthy rhesusmonkeyswere housed in aBSL-2/3 facility. They were all 2–5 year old females withnormal weights and activity levels. Four monkeys wereinoculated by gavage (i.g.) with the LCMV-WE strain:two with 107 pfu (Rh-ig7a; Rh-ig7b) and two others with108 pfu (Rh-ig8a; Rh-ig8b). One monkey (Rh-iv3b)received a standard lethal challenge of 103 pfu LCMV-WE intravenously (i.v.) as a positive control for thevirulence of the virus stock. Two more Rhesus maca-ques, Rh-ig8 (ARM) and Rh-iv3 (ARM), were used toassess the cross-protective ability of LCMV-ARM afteri.g. (108 pfu) or i.v. (103 pfu) inoculation. Animals werefirst anesthetized (ketamine, 20 mg/kg) and then eithergiven 1 ml of virus in PBS by intragastric gavage, orgiven 0.5 ml of virus in PBS via the left saphenous vein.Animalswere observeddaily. Body temperatures (rectalthermometer) and weights (digital balance) were takenbefore every bleeding. For the first lethal challenge, allmonkeys that survived the primary inoculation wereanesthetized as previously described, and inoculatedintravenously with 103 pfu of LCMV-WE. A secondlethal challenge was administered to the three animalsthat survived the first inoculation, and all three sur-vived. Table I hasa comprehensive summary of animals,routes of inoculation, virus doses, and outcomes.

Blood Collection

Blood samples were collected at biweekly intervalsand submitted to the clinical laboratory for completeblood counts and standard blood chemistry. Peripheralblood mononuclear cells (PBMC) were obtained fromanticoagulant-treated blood and centrifuged over Ficoll-

TABLE I. Outcome After Inoculation and Challenge of Rhesus macaques

MonkeyVirusa

(dose, route)Viremia(pfu/ml)

Outcome afterinfection

Challengeb

(days p.i.) Outcome

Antibodyresponse CMIe

ELISAc PRNd Prolif.CTL(LU)f

Rh-ig7a WE (107 pfu, i.g.) Neg. No disease 98 Died d14 <2 <1 � 2Rh-ig7b WE (107 pfu, i.g.) 8.8� 104 Disease/

grecovered112 Survived 8.7 2.5 þ 27

Rh-ig8a WE (108 pfu, i.g.) 8.0� 106 Died d10 N/A � <2 <1 ND �h

Rh-ig8b WE (108 pfu, i.g.) Neg. No disease 98 Died d13 <2 <1 � 1Rh-iv3 Arm (103 pfu, i.v.) Neg. No disease 56 Survived 5.9 1.9 þþþ NDRh-ig8 Arm (108 pfu, i.g.) Neg. No disease 56 Survived 5.8 <1 þ NDRh-iv3bi WE (103 pfu, i.v.) 2� 107 Died d14 N/A � NDj ND ND ND

aViremia was determined by plaque assay of plasma.bAll the animals that survived the primary inoculation were given a lethal challenge with 103 pfu LCMV WE i.v.cIgG levels are expressed as the inverse of the log10.dNeutralizing antibody titers are expressed as the inverse of the log10 dilution that blocked 50% of the plaques.eCell-mediated immunity (CMI) was measured by lymphocyte proliferation assays (prolif.) and cytotoxic T lymphocyte (CTL) assay.fLU is lytic units per 107 effector cells as described by Wunderlich and Schearer (64).gMonkey Rh-ig7b became sick 4 weeks after inoculation and recovered 8 weeks after inoculation. Virus was isolated from plasma and PBMC onlyduring the prodromal period.hOnly two time points were assessed: one before inoculation and one upon death. Both showed no CTL response above background.iThis monkey was only introduced into the experiment as a control for our viral stock during the primary inoculation.jND means not done.

Lethal Hemorrhagic Fever 425

Paque plus (Amersham Biosciences) as described else-where [Hinds et al., 1997]. Some remarkable changes inblood chemistry are noted in Table II.

Detection of Infectious Virus in Plasma

LCMV-WE and ARM were detected in plasma byplaque assay as described [Lukashevich et al., 2002]. Inbrief, ten-fold dilutions of plasma sampleswere added tomonolayers of Vero cells (about 2.5�105/well) in six-well plates, and incubated for 1 hr at 378C. After incuba-tion,wellswere overlayedwith1%agarose inMEMþ 2%FBS and incubated again at 378C. Five days later, thewells were treated with formaldehyde for 30 min. Thenthe agarose overlay was removed and monolayers werestained with crystal violet, washed with tap water,

dried, and visible plaques were counted to estimate thepfu/ml of virus in plasma.

RT-PCR for Detection of Viral RNA

RNA was extracted from PBMC or tissue samplesusing Trizol reagent (Invitrogen, Carlsbad, CA). Iso-latedRNAwas treatedwithDNAse (Promega,Madison,WI) 1 U/mg, at 378C for 30 min and purified again withthe RNAeasyMini Kit (Qiagen, Valencia, CA). RNAwasconverted to cDNA with the avian myeloblastosis virusreverse transcriptase (5 U, RT, Promega) using randomhexamers (Invitrogen) for 1 hr at 428C. The cDNA wassubsequently amplified using standard PCR conditionsas described [Lukashevich et al., 2002].

TABLE II. Blood Chemistries

Monkey Weeks after inoculation ASTa ALT GGT

Rh-ig7a 0 25 69 682 33 53 544 35 48 60

* (nothing abnormal for weeks)14 (challenge) 28 49 59

16 (morbid/death) 553 1,004 187Rh-ig7b 0 17 52 74

2 38 42 474 184 398 906 38 34 1258 40 36 94

10 16 30 65* (nothing abnormal for weeks)

16 (challenge) 30 36 56* (nothing abnormal for weeks)

26 41 93 5728 25 73 62

29 (healthy/sacrificed) 29 65 —Rh-ig8a 0 17 58 72

1.5 (morbid/death) 6,062 2,429 216Rh-ig8b 0 66 91 78

2 42 102 1244 26 54 85

* (nothing abnormal for weeks)14 (challenge) 34 92 79

16 (morbid/death) 1,302 1,312 217Rh-iv3 (ARM) 0 44 44 62

2 36 48 514 47 47 566 35 53 54

8 (challenge) 46 58 5010 59 83 49

12 (2nd challege) 51 87 9613 (healthy/sacrificed) 45 74 —

Rh-ig8 (ARM) 0 35 39 702 23 38 674 12 43 756 20 48 82

8 (challenge) 23 43 9910 22 55 10412 16 50 104

13 (healthy/sacrificed) 15 39 —Rh-iv3b 0 28 121 164

14 (morbid/death) 1,694 1,499 305

aData are in international units per liter (IU/L). Aminotransferases (AST, ALT) and gamma-glutamyltransferase (GGT) reference range data are18–82 (AST), 22–87 (ALT), and 28–86 (GGT) IU/L [Wolford et al., 1986]. Bolded data are those outside the normal reference range.

426 Rodas et al.

Analysis of Lymphocyte Subsets andActivation-Induced Cell Death (AICD)

Flow cytometry was used to determine changes inwhite blood cell subsets after infection. PBMC wereparaformaldehyde-fixed and antibody-stained to iden-tify CD3þ, CD4þ, CD8þ, andCD20þ lymphocyte subsets.FITC-conjugated monoclonal antibody against CD4 orCD8 (Antigenix America, Franklin Square, NY), orFITC-conjugated anti-CD3 or CD20 (Becton-Dickinson,Mountain View, CA), were used with relevant isotypecontrols as described [Pauza et al., 1997; Lukashevichet al., 2002]. Samples were analyzed on a FACScan flowcytometer (Becton-Dickinson), and data were processedusing FlowJo 2.7.8 software (Three star, 1997–1998).Absolute counts were determined by multiplying thepercentage for each subset by the absolute lymphocytecount obtained from clinical hematology data.

Activation-induced cell death (AICD) was measuredas the frequency of apoptosis induced in different cellsubsets after PHA stimulation and was detected bystaining with 7-amino actinomycin D as described[Lecoeur et al., 1997]. Briefly, 2�106 PBMC/well in a6-well plate were cultured for 24 hr in RPMI with 10%FBS with or without 1 mg/ml of PHA (Sigma). After theincubation, PBMCwere labeled for surfacemarkers andthen incubated for 20 min at 48C in PBS that contained20 mg/ml of 7-AAD (Sigma, St. Louis,MO). Sampleswerewashed in PBSþ2% FBS containing 20 mg /ml of non-fluorescent actinomycin D (AD, Sigma) and fixed in thesame buffer containing 1% paraformaldehyde. Sampleswere analysed 15 min later (10,000 events/sample)using the FL-3 channel to detect 7-AAD staining.

Antibody Titers and Plaque-ReductionNeutralization (PRN) Assays

Antibody titersweremeasured byELISAas describedpreviously [Lukashevich et al., 2003].Neutralizing anti-bodies weremeasured by a plaque reduction neutraliza-tion assay (PRN). This assay used the same protocol asthe plaque assays except that each plasma dilution waspre-incubated with a constant number of viral pfu for1 hr at 378C. The neutralization titer is the highestdilution of plasma sample able to reduce the number ofplaques to 50% of the positive control. To determinewhether plaque reduction could be enhanced by comple-ment, we followed a published technique [Meyer et al.,2002]. Once the virus–plasmamixtures were incubatedfor 1 hr at 378C, a 1/20 dilution of titered guinea pigcomplement was added and samples were incubated foran additional 45 min at 378C. This mixture was addedto cell monolayers and incubated for 1 hr more at 378Cbefore the semisolid agar was added. The assay wasanalyzed as described for the plaque reduction assay.

Cytotoxic T Lymphocyte (CTL) Assays

CTL assays followed published protocols [Trivediet al., 1996; Yin et al., 1999] with slight modifications.

Briefly, effector cells were peripheral blood mono-nuclear cells (PBMC) that had been cryopreservedin RPMI 1640 (Gibco, Grand Island, NY) containing10% dimethyl sulfoxide (tissue culture grade: Sigma,St. Louis, MO) and 20% fetal bovine serum (FBS, Gibco,Grand Island, NY). The PBMC used as effectors werecomprised of lymphocytes (70–90%; CD2þ) and mono-cytes (10–30%; CD2�). PBMC for these monkeys were21–26% CD4þ, 24–62% CD8þ, and 9–28% CD20þ. Inpreparation for the CTL assays, effector cells werethawed, washed three times with RPMI 1640þ10%FBS, and cultured at 378C for 3 days at 2� 106 cellsper ml in RPMI 1640 with 20% FBS, penicillin/streptomycin (100 U/ml and 100 mg/ml respectively,Gibco), L-glutamine (2 mM, Gibco), Concanavalin A(ConA, 1 mg/ml, Sigma), and interleukin-2 (IL-2, 20U/ml,Roche Diagnostics, Mannheim, Germany). After this,cellswerewashedand culturedagainwithout theConA,addingnewIL-2 every other day for 7days. TargetswereautologousB-lymphoblastoid linesmade by transforma-tion with Herpes papio (a gift from N. Letvin, HarvardMedical School) and clones of transformed cells wereobtained by limiting dilution as we described [Yin et al.,1999]. Target cells were cultured for 6 hr at 378C eitheruninfected or infected with 3 moi of VVwt, VV-NP, orVV-GP. After incubation, target cells were washedwith RPMIþ10% FBS and labeled for 2 hr with [51Cr]Na2CrO4 (Amersham Pharmacia Biotec, Irvine, CA),200 mCi/2�106 cells. Then cells were washed threetimes and resuspended in RPMIþ 10% FBS at 2�106 cells/ml.Hundredmicrolitres of these cells or 2�104

cells, were plated per well in a round-bottomed 96well plate in triplicate for each effector:target ratio.Three ratios were used in this assay, 50:1, 25:1, and12.5:1. Spontaneous lysis varied from 10 to 20% andspecific release was calculated as follows: [(experimen-tal lysis�spontaneous lysis)/(maximal lysis�sponta-spontaneous lysis)]� 100. Specific lysis values abovemean and standard deviation of background lysis usinguninfected targets or VV-infected targets were consid-ered positive.

Lymphocyte Proliferation Assay

Fresh PBMC were adjusted to 106 cells per ml inRPMIþ10% FBS. The proliferation assay [Djavaniet al., 2001], was performedwith the followingmodifica-tions: 100 ml of cells/well were plated in triplicate foreach of the different stimuli using round-bottomed 96well plates. Each well received an additional 100 ml ofRPMI for negative controls, an additional 100 ml ofRPMIþPHA (10 mg/ml, Sigma, St. Louis, MO) for pro-liferation to mitogen, or an additional 100 ml of RPMIþLCMV antigen (WE or ARM) that was 3�105 pfu priorto inactivation. Plates were set up in duplicate to becollected 3 and 6 days after incubation. Each well waspulsed with 1 mCi of [3H] thymidine (NEN Life ScienceProducts, Boston, MA) and supernatants were harvest-ed 12 hr later using a 96-well glass fiber filtermat onan automatic cell harvester (Tomtec harvester 96).

Lethal Hemorrhagic Fever 427

Thymidine incorporationwasmeasuredby liquidscintil-lation counting using an automatic microbeta Wallacplate reader (Trilux, 1450Microbeta).Counts pe minute(cpm) were recorded and the stimulation index (SI) wasdetermined as themean number of 3H cpm incorporatedin the presence of antigen divided by the mean cpmincorporated in the presence ofmediumalone. Anythingless than an SI of 5 was considered background.

RESULTS

Outcomes in Monkeys InfectedIntragastrically With LCMV

Of four monkeys inoculated i.g. with LCMV-WE, twoshowed no signs of disease, and two developed disease,(Table I). We previously reported on i.g.-infectedmonkeys that displayed no disease signs but containedviral nucleic acids [Lukashevich et al., 2002]. Of the two‘‘healthy’’ monkeys in this study, one displayed elevatedliver enzymes (Table II, Rh-ig8b, 2weeks after infection)and both Rh-ig7a and Rh-ig8b developed LCMV-specificCTL responses that were present in at least 12 assayseach, using at least three different blood draws each,before lethal challenge (see CMI results). Of the twodiseased animals, one had a rapidly fatal disease coursesimilar to Rhesus macaques inoculated intravenously[Lukashevich et al., 2002], while the other was pre-viously described as having a transient viremia and atransient elevation of liver enzymes [Lukashevich et al.,2003]. We add here that the transiently ill animal,Rh-ig7b, experienced a 25% weight loss and a burst ofcirculating neutrophils followed by a burst of circulatingmonocytes during its illness (Fig. 1).

Intragastric inoculation with LCMV-WE did not pro-vide uniform protection against a lethal challenge withthe same virus. Of four animals inoculated i.g. withWE,two showed no signs of disease and failed to resistchallenge and onedied before being challenged (Table I).Only the animal that recovered from acute WE diseasewas protected against a subsequent homologous chal-lenge. Two additional animals inoculated with LCMV-ARM, Rh-ig8, and Rh-iv3, were both protected against aheterologous challenge with LCMV-WE, and only oneof these, Rh-iv3, showed neutrophilia. In this study,neither acute viremia nor neutrophilia was correlatedwith protection against lethal challenge.

Our previous studies showed a reduction in plateletnumbers after i.v. inoculation with 103 and 106 pfu ofLCMV-WE [Lukashevich et al., 2002], but no changesafter i.g. inoculation with 106 pfu of the same strain.In human cases of Lassa, platelet numbers are onlymoderately depressed but their function is almost com-pletely abolished by an unknown circulating inhibitor[Fisher-Hoch, 1993]. No significant platelet changeswere noticed before and after non-lethal i.g. inoculationof monkeys with LCMV. Lethal i.g. infections weresimilar to lethal i.v. infections with regards to reducedplatelet numbers.

It is known that elevated levels of aspartate amino-transferase (AST)> 150 IU/L are associated with lethal

disease in Lassa fever [McCormick et al., 1986b]. As wedescribed earlier [Lukashevich et al., 2003], a monkeythat recovered from disease reached a peak of 184 IU/L2 weeks after infection whereas the monkeys that diedreached AST levels of around 1,000. All monkeys thatsurvived lethal challenge in this study, Rh-ig7b (WE),Rh-iv3 (ARM), and Rh-ig8 (ARM), did not deviate

Fig. 1. Hematological data. Blood cell counts were obtained over thecourse of our study for two monkeys that failed to survive lethalchallenge and for threemonkeys that survived.A:MonkeyRh-ig7bhadbeen i.g. inoculated with 107 pfu of LCMV-WE and subsequentlysurvived challenge with 103 pfu LCMV-WE i.v. Note the rise inneutrophils during disease and the monocytosis at the end of disease.B: Monkey Rh-ig7a, initially inoculated with 107 pfu WE and Rh-ig8binoculatedwith 108 pfuWEdid not survive the lethal challenge (103 pfuof LCMV-WE). A sharp rise in lymphocytes and neutrophils is depictedfor Rh-iv3 (WE) that succumbed after the first inoculation. Oneadditional monkey, Rh-8a (not shown) also succumbed immediatelyafter the first inoculation (108 pfuWE) and experienced a similar rise inlymphocytes and neutrophils. C: Monkey Rh-iv3 inoculated i.v. with103 pfu ARM, and monkey Rh-ig8 inoculated with 108 pfu ARM bothsurvived lethal challenge. Note the spike in neutrophils at week 4 forRh-iv3. Neither monkey experienced disease.

428 Rodas et al.

fromnormalAST,alanineaminotransferase (ALT), org-glutamyltransferase (GGT) levels after challenge,whereas those animals that succumbed developed ASTand ALT levels around 1,000. Although normal GGTlevel is 40–80 IU/L, GGT levels were 160–300 IU/L atnecropsy for lethal cases. It is notable that two out of thethree lethal-challenge survivors (ARM inoculated mon-keys) experienced an increased GGT of 100 IU/L at 8–12 weeks after the lethal challenge (Table II).

Of the seven monkeys listed in Table I, plasmaviremia was only detectable in monkeys that experi-enceddisease:Rh-ig7b (WE),Rh-ig8a (WE), andRh-iv3b(WE). RT-PCR of PBMC confirmed the presence of viralRNA for these same three monkeys but not for anyothers. Infection was indicated for the four non-viremicmonkeys by their immune responses (all fourmonkeys),resistance to challenge (two monkeys), or blood chemis-tries (one monkey).

Flow Cytometry of Lymphocyte SubsetsAfter LCMV-ARM Inoculations

We tried to detect alterations in lymphocyte subsetsafter infection with LCMV-ARM (i.v. or i.g.) and afterchallenge with the WE strain. The absolute number ofCD3þ cells went up substantially in the intravenouslyinoculated monkey by 2 weeks after infection (Fig. 2A).This increase corresponds to a significant increase in thenumber of CD8þ T cells. However, the opposite patternwas seen in the i.g.-inoculated animal. Samples werealways run in triplicate in order to obtain a standarddeviation. By the time of challenge (Fig. 2) neither of thetwo animals exhibited large changes in the absolutenumber of cell subsets, although the number of CD3þ

was slightly lower in both before the challenge, seem-ingly a consequence of the decrease in CD8þ T cells.We also noted a CD8þ T cell increase (Fig. 2A,B) aftera second challenge. In that case, not only absoluteCD8þ but also absolute CD4þ numbers increasedsubstantially.

Neutralizing Antibodies do not ExplainSurvival of Monkeys Exposed to Lethal

Challenge With LCMV-WE

ThePRNassaywas carried out for all themonkeysbutonly the three monkeys that survived the challenge areshown. Only two of the three monkeys that survivedlethal challengewith LCMV-WE, had neutralizing anti-bodies as detected by PRN assays (Fig. 3). The additionof complement to the assay only slightly enhanced theneutralizing capability of the plasma samples in vitro.

Cell-Mediated Immunity Is Correlated WithProtection From Lethal Challenge

CTL assays were performed to detect anamnesticresponses for one lethal-challenge survivor and twonon-surviving monkeys. Target cells for chromium-releaseassays were autologous B cell lines and effector cellswere cryopreserved PBMC stimulated with Con A andIL-2 (seeMaterials andMethods). Monkey Rh-ig7b that

Fig. 2. Flow cytometry.Only theLCMV-ARM-infected animalswereanalyzed for fluctuations in lymphocyte subsets since LCMV-WE-infected animals were analyzed previously [Lukashevich et al., 2002].CD3þ cells are roughly the sum of CD4þ and CD8þ T lymphocytes, andCD20þ are B lymphocytes. The search for particular changes was onlyextended to samples taken 2 weeks before or 2 weeks after theinoculation, and 2 weeks before and after the 1st and 2nd challenge.

Fig. 3. Plaque-reductionneutralization assay (PRN). This assaywascarried out for all themonkeys included in this study, but only the threemonkeys that survived the challenge are shown to illustrate thecontribution of humoral immune responses in protecting against alethal challenge. Note that complement-assisted PRN is similar to, butmore strongly neutralizing than, complement un-assisted PRN.Whereas monkey Rh-ig8 (ARM) never had any neutralizing titer, theother two animals had good neutralizing titers. The antibody titer isexpressed as the log10 of the plasma dilution that inhibited 50% of theplaques compared to the viral control. The time of challenge is depictedby an arrowhead (~). No other animals showed any neutralizing titers.

Lethal Hemorrhagic Fever 429

survived the lethal i.v. challenge with LCMV-WE, had astrong, specific CTL response against targets infectedwith the recombinant vaccinia viruses, VV-NP or VV-GP. Two monkeys that did not survive challenge, Rh-ig7a andRh-ig8b, had relatively poor but LCMV-specificCTL responses (Fig. 4).

Monkey Rh-ig7b showed high CTL responses to GP at8weeks after infection, 4 weeks after the first challenge,and 1 week after the second challenge; whereas thestrongestCTLresponses toNPwere observedat 8weeksafter infection, 2 weeks after the first challenge, and1 week after the second challenge. Nonspecific cytotoxi-

city against the uninfected and the vaccinia virus-infected cells was usually under 10% specific lysis.Although the assays were performed at effector: targetratios of 50:1, 25:1, and 12.5:1, only results from the 50:1and 25:1 ratios are shown (Fig. 4) because these ratiosgave the strongest responses. Low but reproducibleLCMV-specific cytotoxic activities were also observed intwo other WE-inoculated monkeys (Rh-ig7a and Rh-ig8b) that did not survive the challenge. In the case ofRh-ig7a, specific lysis against GP and NP peaked at10weeksafter infectionand forRh-ig8b, these responsesdid not begin until week 12 after infection. Responses

Fig. 4. Chromium-release CTL assays. The percent specific lysis was calculated as described in theMaterials and Methods section and is depicted as a number on each bar associated with the differentantigenic stimuli. CTLactivity over the course of infection is shown for challenge-surviving animalRh-ig7b(A), and for twoanimals that succumbed,Rh-ig7a (B) andRh-ig8b (C).Only the results for the 50:1 and25:1effector:target ratio are shown.

430 Rodas et al.

in these monkeys were never more than 21% specificlysis above background levels and were apparently notenough to save them from lethal challenge.

Lymphocyte proliferation assays were performed forthree challenge-surviving and one non-surviving mon-key. Fresh PBMC were cultured in vitro to determinetheir proliferative capacity in the presence of media,PHA, or LCMVantigen (Fig. 5). Proliferationwas deter-mined by tritium incorporation. We observed thatPBMC from Rh-ig7b (WE), Rh-iv3 (ARM), and Rh-ig8(ARM) proliferated against autologous antigens,LCMV-ARM for theARM-infectedmonkeys andLCMV-WE for the WE-infected monkeys. Results were similarfor proliferation against heterologous antigens (WE-infected to ARM and vice versa) indicating that some ofthe helper epitopes are similar on the two viruses.Proliferative responses of circulating PBMC peak at 2–4 weeks after inoculation, and then again at 4 weeksafter challenge.TheSIwashighest inRh-iv3 (ARM) thathad been inoculated intravenously unlike the other twosurvivingmonkeys. Proliferative responses for two non-surviving monkeys (Rh-ig7a and Rh-ig8b) were mon-itored for the day of challenge and they were both belowSI¼5 (not shown). Proliferation data required freshPBMC and could not be obtained from cryopreservedblood because fewantigen-presenting cells survived andresponses to mitogens were low. In general, the SI forPHA-treated PBMC was between 100 and 1,000, peak-ing at day 3 (not shown) and SI for antigen was 5–100peaking at day 6 (Fig. 5A).

Activation-Induced Cell Death (AICD) Fails toExplain Decreased Proliferative Capacity of

PBMC After Injection of Antigen

Proliferative capacities of PBMC decreased immedi-ately after the first and second i.v. challenges withLCMV-WE for the three survivors (Fig. 5). There are atleast two explanations for this drop in proliferativecapacity after an antigenic boost: (1) TH cells were beingdepleted from the circulation by being sequesteredto virus-infected tissues, and (2) viral antigens madelymphocytes susceptible to AICD as seen in AIDS[Wallace et al., 1999]. To explore this latter possibilityforRh-ig8 (ARM) andRh-iv3 (ARM)weperformedaflowcytometric analysis of different lymphocyte subsetsafter activation with PHA (Fig. 6). PBMC were incu-bated with or without 1 mg/ ml of PHA, and then werestained with 7-AAD as an early marker for apoptosis[Wallace et al., 1999]. In general, early PBMC sampleshad higher dramatic increases in cell death as a resultof PHA activation (Fig. 5). In conclusion there was noAICD associated with our antigen inoculations thatcould be compared to the phenomenon observed in therhesus/AIDS model.

DISCUSSION

Murine studies in our laboratory showed that LCMVcould infect and disseminate via the i.g. route and thati.g. infection could protect mice from lethal intacerebral(ic) challenge [Rai et al., 1996, 1997]. In later studies, we

Fig. 5. Lymphocyte proliferation assay for rhesus macaques that survived lethal challenge.FreshPBMCwere incubatedwithantigenic stimuli during the course of infection. Theupper right corner ofeach panel indicates the antigenic stimulus: LCMV-WE for 3 days (A), LCMV-WE for 6 days (B), LCMV-ARM for 3 days (C) and LCMV-ARM for 6 days (D). PBMC tested were from monkeys Rh-ig7b (WE-infected), Rh-iv3 (ARM-infected) or Rh-ig8 (ARM-infected). Stimulation index (SI) is the thymidineincorporation in the presence of test antigen divided by the thymidine incorporation in the presence ofmedium alone. Non-surviving monkeys only exhibited background levels and are not included in thisfigure. The time of challenge is depicted by an arrowhead (~).

Lethal Hemorrhagic Fever 431

used themacaquemodel to describe thedisease course ofa virulent LCMV infection after i.g. inoculation [Luka-shevich et al., 2002]. We found that 103 pfu i.v. was areproducibly lethal dose, and that 107–108 pfu by the i.g.route could sometimes be lethal. This dose is similar tothe virus ingested by consuming a single infected rodentspleen, for example [Rai et al., 1997]. In addition to theacute disease cases, we were able to study one case inwhich amonkey recovered fromdisease and two cases inwhich monkeys were infected with an avirulent strainand became immune to disease. It was surprising thatprotectionaftermucosal infectionwasobservedso rarely(only two out of fivemonkeyswere protected). In naturalsettings repeated mucosal exposures probably improvechances for developing protective immunity.

Of five mucosally inoculated monkeys, only two be-came viremic and a third became seropositive (thoughnot viremic). Evidence that the remaining two were in-fected came primarily fromCTL assays inwhichLCMV-specific lytic responses were measured: for monkeyRh-ig7a, specific lysis exceeded background on fourseparate occasions before challenge, and formonkeyRh-ig8b specific lysis exceeded background on two occasionsbefore challenge responses could begin. Ample prece-dent exists for arenavirus infections that are only de-tectable by cell-mediated immunity. For example, mice

given low i.g. doses of LCMV lacked virological orserological evidence of infection yet resisted intracer-ebral challenge with LCMV [Rai et al., 1996], and thisresistance is due to virus-specific cell-mediated immu-nity [Doherty and Zinkernagel, 1975]. In LCMV trans-genic mice, CTL responses could be detected in theabsence of detectable gene expression [Oldstone et al.,1991]. In AIDS, there are numerous examples ofmucosally infected subjects with no detectable virus orantibodybutwith virus-specific cell-mediated immunity[Clerici et al., 1992; Salvato et al., 1993; Trivedi et al.,1996]. Our experiences with cell-mediated immunity inSIV- or LCMV-infected monkeys convinced us that allseven of the animals reported here were infected by theinitial inoculation.

Elevated levels of liver enzymes, particularly amino-transferases, correlate well with disease progression inboth monkeys and in Lassa fever patients [McCormicket al., 1986b; Fisher-Hoch and McCormick, 1987;Fisher-Hoch, 1993]. In this study, all three animals thatexperienced disease after inoculation had high amino-transferases, butonlyoneof the fouranimals thathadnodisease experienced abnormal aminotransferase levels.In our previous publications [Lukashevich et al., 2003],we suggest that increases in IL-6 and soluble TNF-receptors aremarkers of infection and potentially useful

Fig. 6. No activation-induced cell death (AICD). Apoptosis was experimentally determined basedon the progressive loss ofmembrane permeability through the incorporation of 7-aminoactinomycinD. Theextent of apoptosis after PHA-stimulation of PBMC was examined to explain the apparent in vitroproliferation inhibition observed in the samples taken after each inoculationwith virus. CD4þ, CD8þ, andCD20þ subsets were examined from samples cryo-preserved 2 weeks before infection, 2 weeks afterinfection, 2 weeks before the first challenge, 2 weeks after the first challenge, 2 weeks before the secondchallenge, or two weeks after the second challenge. Panel A shows the % of 7AADþ cells from Rh-iv3(LCMV-ARM-infected) and panel B shows the % of 7AADþ cells from Rh-ig8 (LCMV-ARM-infected).There areno increases inPHA-stimulated cell death to explain thedissapearance of proliferative responsesafter antigenic challenges.

432 Rodas et al.

in prognosis, but these markers did not reach abnormallevels in any animals that did not experience disease[Lukashevich, unpublished].

Three out of seven monkeys survived all challengesand, of these three, two experienced neutrophilia at4 weeks after the primary infection. Neutrophilia in ourmonkey experiments was associated with a delayeddisease or a subclinical infection, and was not essentialfor surviving lethal challenge. The neutrophil-attract-ing chemokine, IL-8, was suppressed during in vitrostudies with Lassa fever virus, but was not suppressedin parallel cultures with the avirulent Mopeia virus[Lukashevich et al., 1999]. In monkeys [Lukashevichet al., 2003] and in human beings [Mahanty et al., 2001]high levels of IL-8 during infection are associated witha better prognosis. This is consistent with the inter-pretation that the neutrophil burst we observed in twomonkeys could have contributed positively to theirsurvival.

Flow cytometry of lymphocyte subsets showed fewremarkable fluctuations. In previous studies we notedlymphocytosis upon death for onemonkey and at week 2for a subclinically infected animal [Lukashevich et al.,2002]. Changes in circulating populations ofwhite bloodcells are most likely driven by high levels of viralantigen. We noted a rise in CD3þ cells due to a rise inCD8þ cells 2 weeks after infection of Rh-iv3 (ARM) but afall in the same cells during the same period in Rh-ig8(ARM). In human EBV infections, in cases of inflamma-tion, and in acute murine LCMV infection, increases intotal CD8þ cells can be linked to clonal expansions ofvirus-specificT cells [Evenet al., 1995; Silins et al., 1998;Sourdive et al., 1998]. It is possible the expansion ofCD8þ cells observed inRh-iv3wasdrivenbyantigenandlinked to its protection from lethal challenge. No similarexpansionwas observed for another challenge-protectedanimal that had been mucosally infected, but it is likelythat expansion in that animal only occurred amongstmucosal immunocytes, such as intraepithelial lympho-cytes and mesenteric lymph node lymphocytes, andthese were not sampled.

Since only 30% of LAS-infected individuals everexperience acute disease [McCormick et al., 1986a,b],an important question has been ‘‘How did the other 70%escape disease?’’ Our effort to analyze immune re-sponses in surviving monkeys addresses this question.In our study, three (of five) monkeys survived lethalchallenge, and these were also the monkeys with thehighest cell-mediated immunity (CMI). Lassa vaccinestudies and surveys of Lassa patients provided circum-stantial evidence that CMI is required for protectionfrom lethal challenge or for recovery. First, LAS pa-tients, convalescing primates, and guinea pigs oftenlacked neutralizing antibody responses [Jahrling et al.,1980, 1982;Fisher-Hochetal., 2000, 2001].Neutralizingantibodies often appear late in infection and the transferof ‘‘immune’’ plasma is rarely effective for clearing virus,cross-protection, or recovery from disease in animal orhuman trials [Jahrling and Peters, 1984; McCormicket al., 1986a]. Second, gamma-irradiated preparations

of purified LAS virions elicited antibodies inmonkeys toall three major structural proteins, N, G1, and G2, butdid not protect againstLAS challenge [McCormick et al.,1992]. Third, infectious virus vaccines that are known toelicit CMI (such as LAS-expressing alphavirus repli-cons, vaccinia recombinants, Salmonella recombinants,or Mopeia) elicit protective immunity in animals[Fisher-Hoch et al., 1989, 2000; Auperin, 1993; Pushkoet al., 1997; Djavani et al., 2000]. Fourth, adoptivetransfer of effector spleen cells canmediate recovery andcross-protection from LAS infection in a mouse model[Peters et al., 1987; Lukashevich, 1992]. Finally, it hasbeen demonstrated that LAS-seropositive persons froman endemic area have very strongCD4þT cell responsesagainst LAS NP [ter Meulen et al., 2000]. Thus, by theaccumulation of circumstantial evidence and by extra-polation from the mouse model, CMI is considered es-sential for survival or recovery from hemorrhagic fever.

Humoral responses also have a role in protection fromarenaviral hemorrhagic fevers. Neutralizing activity asmeasured by PRN does not reflect antibody-dependentcellular cytotoxicity, antibody-mediated antigen uptakeby phagocytes, or antibody-mediated antigen presen-tation, all of which modulate virus infection in vivo[Villinger et al., 2003]. The adoptive transfer studies inthe mouse were contradicted by studies in a more re-levant guinea pig model, in which splenocyte transferresulted in death for 5 of 5 challenged animals whereasthe transfer of immune plasma protected 4 of 5 chal-lenged animals [Peters et al., 1987]. Neutralizingplasma from monkeys with titers above 4 logs couldprotect Lassa-infected cynomolgous macaques, andwere especially effective when given with ribavirin[Jahrling and Peters, 1984; Jahrling et al., 1984]. Con-valescent plasma from Lassa patients has also beenshown to be protective in guinea pigs when the titer ofneutralizing antibodies exceeds 2 logs, however thislevel is difficult to obtain from human plasma [Jahrling,1983]. Our experience with immune responses in therhesus/AIDSmodel and now in the rhesus/LCMVmodelindicates that a good CMI is always attended by a goodhumoral response, and that neutralizing antibodies arerarely correlated with protection. In the absence ofcostly immune depletion studies and adoptive transferstudies as have been performed in the rhesus/AIDSmodel [Schmitz et al., 1999, 2003], it is impossible todetermine whether one or both responses are critical forimmunity. It seems plausible that the humoral responsealone cannot protect, that high antibody responses areindicative of good helper and CMI responses, and thatCMI is most critical for survival.

Proliferative responses toLCMVantigenwereobserv-ed in all three challenge-surviving monkeys but notin monkeys that succumbed to lethal challenge. It wascurious that proliferative capacity frequently decreasedfollowing inoculation with more virus; so we tested thehypothesis that decreases in SI were due to AICD asobserved inAIDS [Wallace et al., 1999]. Thiswas not thecase, however, since lymphocyte subsets were just assusceptible to cell death before and after activation with

Lethal Hemorrhagic Fever 433

PHA (Fig. 6). There is an alternative hypothesis that wedid not pursue in this study, but which finds support inmurine studies. Klenerman et al. [2002], found thatafter infection, virus-specific memory T cells were se-questeredprimarily in solid organs suchaskidney, liver,lungs, and spleen, and less than 20% of these T cellsremained in the circulation. This finding raises the pos-sibility that inoculations with virus can drive the majo-rity of the proliferative response out of the circulationand into infected organs.

The significance of this study is primarily in char-acterizing the diversity of outcomes arising from anarrowly defined set of inoculi. Such outcomes run thegamut of what would be expected in nature afterinfection with hemorrhagic fever virus. The most fertilearea of research for the future will be to define theearliest responses to infection that can beused to predictprogression to the various outcomes.

ACKNOWLEDGMENTS

This work was supported by National Institutes ofHealth (NIH) grants RR138980 and AI5252367 (to ISL)and by NIH grants AI53620 and AI53619 (to MSS).We are grateful for excellent animal care facilities andpathology support from Dr. Joseph Bryant andDr. Harry Davis.

REFERENCES

Ahmed R, Salmi A, Butler LD, Chiller JM, Oldstone MB. 1984.Selection of genetic variants of lymphocytic choriomeningitis virusin spleens of persistently infected mice. Role in suppression ofcytotoxic T lymphocyte response and viral persistence. J Exp Med16:521–540.

Auperin DD. 1993. Construction and evaluation of recombinant virusvaccines for Lassa fever. In: Salvato MS, editor. The Arenaviridae.New York: Plenum Press, 259–280.

Clerici M, Giorgi JV, Chou CC, Gudeman VK, Zack JA, Gupta P, HoHN, Nishanian PG, Berzofsky JA, Shearer GM. 1992. Cell-mediated immune response to human immunodeficiency virus(HIV) type 1 in seronegative homosexual men with recent sexualexposure to HIV-1. J Infect Dis 165:1012–1019.

Danes L, Benda R, Fuchsova M. 1963. Experimentalni inhalacninakaza opic druhu macacus cynomolgus a macacus rhesus viremlymphocitarni choriomeningitidy (kmenem WE). Bratisl Lek Listy43:71–79 (in Czech.).

DjavaniM, Lukashevich IS, SalvatoMS. 1998. Sequence comparison ofthe large genomic RNA segments of two strains of lymphocyticchoriomeningitis virus differing in pathogenic potential for guineapigs. Virus Genes 12(2):149–153.

Djavani M, Cheng Y, Xia L, Lukashevich IS, Pauza CD, Salvato MS.2000.Murine immune responses tomucosally deliveredSalmonellaexpressing Lassa fever virus nucleoprotein. Vaccine 18:1543–1554.

Djavani M, Yin C, Lukashevich IS, Rodas J, Rai SK, Salvato MS. 2001.Mucosal immunization with S. typhimurium expressing Lassavirus NP cross-protects mice from lethal challenge with LCMV.J Hum Virol 4:103–108.

Doherty PC, Zinkernagel RM. 1975. Capacity of sensitized thymus-derived lymphocytes to induce fatal lymphocytic choriomeningitisis restricted by the H-2 locus. J Immunol 114:30–33.

Even J, Lim A, Puisieux I, Ferradini L, Dietrich PY, Toubert A,Hercend T, Triebel F, Pannetier C, Kourilsky P. 1995. T-cellrepertoires in healthy and diseased human tissues analysed byT-cell receptor beta-chain CDR3 size determination: Evidence foroligoclonal expansions in tumours and inflammatory diseases. ResImmunol 146(2):65–80.

Fisher-Hoch SP. 1993. Arenavirus pathophysiology, In: Salvato MS,editor. The Arenaviridae. New York, NY: Plenum Press. pp 299–323.

Fisher-Hoch SP,McCormick JB. 1987. Pathophysiology and treatmentof Lassa fever. Curr Top Microbiol Immunol 134:231–239.

Fisher-Hoch SP, McCormick JB. 2001. Towards a human Lassa fevervaccine. Rev Med Virol 11:331–341.

Fisher-Hoch SP, McCormick JB, Auperin D, Brown BG, Castor M,Perez G, Ruo S, Conaty A, Brammer L, Bauer S. 1989. Protection ofrhesus monkeys from fatal Lassa fever by vaccination with arecombinant vaccinia virus containing the Lassa virus glycoproteingene. Proc Natl Acad Sci USA 86:317–321.

Fisher-Hoch SP, Hutwagner L, Brown B, McCormick JB. 2000.Effective vaccine for lassa fever. J Virol 74:6777–6783.

HindsPW,YinC,SalvatoMS,PauzaCD.1997.Pertussis toxin-inducedleukocytosis in Rhesus macaques. J Med Primatol 25:1–7.

Jahrling PB. 1983. Protection of Lassa virus-infected guinea pigs withLassa-immune plasma of guinea pig, primate, and human origin.J Med Virol 12:93–102.

Jahrling PB, Peters CJ. 1984. Passive antibody therapy of Lassa feverin cynomolgus monkeys: Importance of neuralizing antibody andLassa virus strains. Infect Immunol 44:528–533.

Jahrling PB,Hesse RA, EddyGA, JohnsonKM,Callis RT, StephenEL.1980. Lassa virus infection of rhesus monkeys: Pathogenesis andtreatment with ribavirin. J Infect Dis 141:580–589.

Jahrling PB, Smith S, Hesse RA, Rhoderick JB. 1982. Pathogenesis ofLassa virus infection in guinea pigs. Infect Immunol 37:771–778.

Jahrling PB, Peters CJ, Stephen EL. 1984. Enhanced treatment ofLassa fever by immune plasma combined with ribavirin incynomolgus monkeys. J Infect Dis 149:420–427.

Jahrling PB, Frame JD, Smith SB, Monson MH. 1985. Endemic Lassafever in Liberia. III. Characterization of Lassa virus isolates. TransR Soc Trop Med Hyg 79:374–379.

Klenerman P, Cerundolo V, Dunbar PR. 2002. Tracking T cells withtetramers: New tales from new tools. Nat Rev Immunol 2:263–272.

Lecoeur H, Ledru E, Prevost MC, Gougeon ML. 1997. Strategies forphenotyping apoptotic peripheral human lymphocytes comparingISNT, annexin-V, and 7-AAD cytofluorometric staining methods.J Immunol Methods 209:111–123.

Lukashevich IS. 1992. Generation of reassortants between Africanarenaviruses. Virology 188:600–605.

Lukashevich IS, Maryankova R, Vladyko AS, Nashkevich N, Koleda S,Djavani M, Horejsh D, Voitenok NN, Salvato MS. 1999. Lassa andMopeia virus replication in human monocytes/macrophages and inendothelial cells: Different effects on IL-8 and TNF-alpha geneexpression. J Med Virol 59:552–560.

Lukashevich IS,DjavaniM,Rodas JD, Zapata JC,UsborneA,EmersonC, Mitchen J, Jahrling PB, Salvato MS. 2002. Hemorrhagic feveroccurs after intravenous, but not after intragastric, inoculation ofRhesus macaques with lymphocytic choriomeningitis virus. J MedVirol 67:171–186.

Lukashevich IS, Tikhonov I, Rodas JD, Zapata JC, Yang Y, DjavaniM,Salvato MS. 2003. Arenavirus-mediated liver pathology: Acutelymphocytic choriomeningitis virus infection ofRhesusmacaques ischaracterized by high-level interleukin-6 expression and hepato-cyte proliferation. J Virol 77:1727–1737.

Mackett M, Smith GL, Moss B. 1984. General method for productionand selection of infectious vaccinia virus recombinants expressingforeign genes. J Virol 49:857–864.

Mahanty S, Bausch DG, Thomas RL, Goba A, Bah A, Peters CJ, RollinPE. 2001. Low levels of interleukin-8 and interferon-inducibleprotein-10 in serum are associated with fatal infections in acuteLassa fever. J Infect Dis 183:1713–1721.

McCormick J. 1990. Arenaviruses. In: Fields BN, Knipe DM, HowleyPM, editors. Field’s virology. 2nd edn. NewYork: Lippincott-RavenPress. pp 1245–1267 (Ch. 44).

McCormick JB, Fisher-Hoch SP. 2002. Lassa fever. Curr Top MedMicrobiol Immunol 262:75–109.

McCormick JB, King IJ, Webb PA. 1986a. Lassa fever: Effectivetherapy with ribavirin. N Engl J Med 314:202–226.

McCormick JB, Walker DH, King IJ, Webb PA, Elliott LH, WhitfieldSG, Johnson KM. 1986b. Lassa virus hepatitis: A study of fatalLassa fever in humans. Am J Trop Med Hyg 35:401–407.

McCormick JB, Webb PA, Krebs JW, Johnson KM, Smith ES. 1987. Aprospective study of the epidemiology and ecology of Lassa fever.J Infect Dis 155:437–442.

McCormick JB, Mitchell SW, Kiley MP, Ruo M, Fisher-Hoch S. 1992.InactivatedLassa virus elicits a non-protective immune response inrhesus monkeys. J Med Virol 37:1–7.

434 Rodas et al.

Meyer K, Basu A, Przyiecki CT, Lagging LM, Bisceglie AM, Conley AJ,Ray R. 2002. Complement-mediated enhancement of antibodyfunction for neutralization of pseudotype virus containing hepatitisC virus E2 chimeric glycoprotein. J Virol 76:2150–2158.

Montali RJ, ScangaCA,PernikoffD,WessnerDR,WardR,HolmesKV.1993. A common source outbreak of callitrichid hepatitis in captivetamarins and marmosets. J Infect Dis 167:946–950.

Montali RJ, Connolly BM, Armstrong DL, Scanga CA, Holmes KV.1995. Pathology and immunohistochemistry of callitrichid hepati-tis, an emerging disease of captive New World primates caused bylymphocytic choriomeningitis virus. Am J Pathol 147:1441–1449.

Oldstone MB. 2002. Biology and pathogenesis of lymphocytic chor-iomeningitis virus infection. Curr Top Microbiol Immunol 263:83–117.

Oldstone MB, Nerenberg M, Southern P, Price J, Lewicki H. 1991.Virus infection triggers insulin-dependent diabetes mellitus in atransgenic model: Role of anti-self (virus) immune response. Cell19:319–331.

Pauza CD, Hinds PW II, Yin C, McKechnie TS, Hinds SB, Salvato MS.1997. The lymphocytosis promoting agent pertussis toxin affectsvirus burden and lymphocyte distribution in the SIV-infectedrhesus macaque. AIDS Res Hum Retroviruses 13:87–95.

Peters CJ, Jahrling PB, Lui CT, Kenyon RH, McKee KT, Barrera-OroJG. 1987. Experimental studies of arenaviral hemorrhagic fevers.Curr Top Microbiol Immunol 134:5–68.

Pushko P, Parker M, Ludwig GV, Davis NL, Johnston RE, Smith JF.1997. Replicon-helper systems from attenuated Venezuelan equineencephalitis virus: Expression of heterologous genes in vitro andimmunization against heterologous pathogens in vivo. Virology239:389–401.

Rai SK, Cheung DS, Wu MS, Warner TF, Salvato MS. 1996. Murineinfection with lymphocytic choriomeningitis virus following gastricinoculation. J Virol 70:7213–7218.

Rai SK,Micales BK,WuMS, Cheung DS, Pugh TD, Lyons GE, SalvatoMS. 1997. Timed appearance of lymphocytic choriomeningitis virusafter gastric inoculation of mice. Am J Pathol 151:633–639.

Riviere Y, Ahmed R, Southern PJ, Buchmeier MJ, Oldstone MBA.1985. Genetic mapping of lymphocytic choriomeningitis viruspathogenicity: Virulence in guinea pigs is associated with the LRNA segment. J. Virol 55:704–708.

Salvato MS, Shimomaye EM. 1989. The completed sequence oflymphocytic choriomeningitis virus reveals a unique RNA struc-ture and a gene for a zinc finger protein. Virology 173:1–10.

Salvato M, Borrow P, Shimomaye E, Oldstone MB. 1991. Molecularbasis of viral persistence: A single amino acid change in theglycoprotein of lymphocytic choriomeningitis virus is associatedwith suppression of the antiviral cytotoxic T-lymphocyte responseand establishment of persistence. J Virol 65:1863–1869.

SalvatoM,EmauP,MalkovskyM,SchultzK,MacKenzieD, JohnsonE,Pauza CD. 1993. Cellular immune responses in Rhesus macaqueswith low dose SIV infection. J Med Primatol 23:125–130.

Schmitz JE, Kuroda MJ, Santra S, Sasseville VG, Simon MA, LiftonMA, Racz P, Tenner-Racz K, DalesandroM, Scallon BJ, Ghrayeb J,Forman MA, Montefiori DC, Rieber EP, Letvin NL, Reimann KA.1999. Control of viremia in Simian immunodeficiency virusinfection by CD8þ lymphocytes. Science 283:857–860.

Schmitz JE, Kuroda MJ, Santra S, Simon MA, Lifton MA, Lin W,Khunkhun R, Piatak M, Lifson JD, Grosschupff G, Gelman RS,

Racz P, Tenner-Racz K, Mansfield KA, Letvin NL, Montefiori DC,Reimann KA. 2003. Effect of humoral immune responses oncontrolling viremia during primary infection of rhesus monkeyswith simian immunodeficiency virus. J Virol 77:2165–2173.

Silins SL, Cross SM, Krauer KG, Moss DJ, Schmidt CW, Misko IS.1998. A functional link for major TCR expansions in healthy adultscaused by persistent Epstein–Barr virus infection. J Clin Invest102:1551–1558.

Sourdive DJ, Murali-Krishna K, Altman JD, Zajac AJ, Whitmire JK,Pannetier C, Kourilsky P, Evavold B, Sette A, Ahmed R. 1998.Conserved T cell receptor repertoire in primary andmemory CD8 Tcell responses to an acute viral infection. J Exp Med 188:71–82.

terMeulenJ,Lukashevich I, SidibeK, InapoguiA,MarxM,DorlemannA, Yansane ML, Koulemou K, Chang-Claude J, Schmitz H. 1996.Hunting of peridomestic rodents and consumption of their meat aspossible risk factors for rodent-to-human transmission of Lassavirus in the Republic of Guinea. Am J Trop Med Hyg 55:661–666.

ter Meulen J, Badusche M, Kuhnt K, Doetze A, Satoguina J, Marti T,Loeliger C, Koulemou K, Koivogui L, Schmitz H, Fleischer B,Hoerauf A. 2000. Characterization of human CD4þ T-cell clonesrecognizing conserved and variable epitopes of the Lassa virusnucleoprotein. J Virol 74:2186–2192.

Trivedi P, Horejsh D, Hinds SB, Hinds PW II, Wu MS, Salvato MS,Pauza CD. 1996. Intrarectal transmission of Simian immunodefi-ciency virus in Rhesus macaques: Selective amplification and hostresponses to transient or persistent viremia. J Virol 70:6876–6883.

Villinger F, Mayne AE, Bostik P, Mori K, Jensen PE, Ahmed R, AnsariAA. 2003. Evidence for antibody-mediated enhancement of simianimmunodeficiency virus (SIV) Gag antigen processing and crosspresentation in SIV-infected Rhesus macaques. J Virol 77:10–24.

WallaceM,WatermanPM,MitchenJL,DjavaniM,BrownC,TrivediP,Horejsh D, Dykhuizen M, Kitabwalla M, Pauza CD. 1999.Lymphocyte activation during acute Simian/human immunodefi-ciency virus SHIV(89.6PD) infection in macaques. J Virol 73:10236–10244.

Whitton JL, Southern PJ, Oldstone MBA. 1988. Analysis of thecytotoxic T lymphocyte responses to glycoprotein andnucleoproteincomponents of lymphocytic choriomeningitis virus. Virology 162:321–327.

Wolford ST, Schroer RA, Gohs FX, Gallo PP, Brodeck M, Falk HB,RuhrenR. 1986.Reference rangedatabase for serumchemistry andhematology values in laboratory animals. J Toxicol Environ Health18:161–88.

Wunderlich J, Schearer G. 1994. Unit 3: In vitro assays for lymphocytefunction. In: Colligan JE, Kruisbeek AM, Margulies DH, ShevachEM, Strober W, editors. Current Protocols in Immunology Vol 1,Section 3.11. New York: John Wiley and Sons, Inc.

Yin C, WuMS, Pauza CD, Salvato MS. 1999. High major histocompat-ibility complex-unrestricted lysis of simian immunodeficiency virusenvelope-expressing cells predisposes macaques to rapid AIDSprogression. J Virol 73:3692–701.

Zinkernagel RM, Doherty PC. 1974. Restriction of in vitro T cell-mediated cytotoxicity in LCMV with syngenic or semiallogenicsystem. Nature 248:701–702.

Zinkernagel RM, Haenseler E, Leist T, Cerny A, Hengartner H, AltageA. 1986. T cell-mediated hepatitis inmice infectedwith lymphocyticchoriomeningitis virus. Liver cell destruction by H-2 class-I-restricted. J Exp Med 162:2125–2141.

Lethal Hemorrhagic Fever 435