Major histocompatibility complex class I and II antigens frequencies in Baloch ethnic group living in the southeast region of Iran

Major histocompatibility complex class I and II expression onmacrophages containing a virulent strain of Brucella abortus

measured using green £uorescent protein-expressingbrucellae and £ow cytometry

E. Murphy a, G.T. Robertson b, M. Parent a, S.D. Hagius c, R.M. Roop II b,P.H. Elzer c, C.L Baldwin a;�

a Department of Veterinary and Animal Sciences, Paige Laboratory, University of Massachusetts, Amherst, MA 01003, USAb Department of Microbiology and Immunology, Louisiana State University Medical Center, 1501 Kings Highway, Shreveport, LA 71130-3932, USA

c Department of Veterinary Science, LSU AgCenter, 111 Dalrymple Building, Baton Rouge, LA 70803-6002, USA

Received 3 August 2001; received in revised form 1 March 2002; accepted 20 March 2002

First published online 8 May 2002

Abstract

Immune responses appropriate for control of an intracellular pathogen are generated in mice infected with Brucella abortus, shown bythe ability of T cells to adoptively transfer resistance to naive mice. The infection nevertheless persists for months. It was hypothesizedthat one factor in maintaining the infection despite the presence of immune T cells was suboptimal expression of major histocompatibilitycomplex (MHC) molecules on macrophages containing brucellae. This would allow B. abortus to elude detection by the host’s immunesystem. To test this, B. abortus organisms expressing green fluorescent protein (GFP-Brucella) were constructed and three-color flowcytometry used to evaluate MHC expression on macrophages following in vitro or in vivo infection. When infected in vitro, the levels ofMHC class I and class II expression on J774 macrophages containing GFP-Brucella were the same or higher than on macrophageswithout GFP-Brucella in the same cultures. Similarly, the MHC expression was higher on GFPþ peritoneal exudate cells followinginfection or phagocytosis of heat-killed GFP-Brucella than it was on uninfected peritoneal exudate cells. Following in vivo infection ofmice the level of MHC class I and II expression on GFPþ cells in their spleens (the main site of infection) also tended to be as high as orhigher than that on the GFP-negative cells. The only in vivo GFPþ cells that showed a decreased MHC expression was a population ofsplenic Mac1þ cells recovered from interferon-Q gene-disrupted mice at the time of their death due to an overwhelming number of bacteriaper spleen. Overall, it was concluded that decreased MHC expression is not a general principle associated with brucella infection ofmacrophages and thus not likely to contribute to maintenance of the chronic infection. ; 2002 Federation of European MicrobiologicalSocieties. Published by Elsevier Science B.V. All rights reserved.

Keywords: Macrophage; Green £uorescent protein; Intracellular bacterium; Brucella

1. Introduction

Brucella spp. are Gram-negative, intracellular bacteriathat cause serious chronic infections in humans and live-stock. The disease in humans, known as undulant fever,causes recurrent high fever, cachexia, lethargy, arthritisand splenomegaly. Brucellosis in ruminants can result inabortion because of the presence of their preferred carbon

source erythritol in the pregnant uterus and association ofthe bacteria with trophoblasts [1]. Infection of both hu-mans and animals can be lifelong.Resistance to infection is dependent upon interferon-Q

(IFN-Q) [2^6] which activates macrophages for anti-brucel-la activity [7], largely by enhancing production of reactiveoxygen intermediates [8]. It is known that both CD4 andCD8 T cells mediate resistance to infection during the ¢rstweek [9^12] and that they are more important than anti-bodies [13]. Both subpopulations of T cells could poten-tially contribute to protection through production of IFN-Q. In addition, CD8 T cells may have an additional role inprotection through lysis of infected macrophages [11].

0928-8244 / 02 / $22.00 ; 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.PII: S 0 9 2 8 - 8 2 4 4 ( 0 2 ) 0 0 3 0 8 - 5

* Corresponding author. Tel. : +1 (413) 545 3167;Fax: +1 (413) 545 6326.

E-mail address: [email protected] (C.L. Baldwin).

FEMSIM 1409 24-6-02

FEMS Immunology and Medical Microbiology 33 (2002) 191^200

www.fems-microbiology.org

Therefore, presentation of brucellae antigens on both ma-jor histocompatibility complex (MHC) class I and MHCclass II molecules to CD8 and CD4 T cells, respectively,must be necessary for controlling and/or clearing the in-fection.Brucellae are known to survive in phagocytic cells, in-

cluding macrophages [14,15], and their residence in thisprotected niche is believed to be responsible for thechronic nature of brucellosis in people and animals. Itwas hypothesized here that a decrease in expression ofMHC molecules occurs following infection of macro-phages thereby contributing to the lack of immune-medi-ated clearance during the plateau phase of the infectionthat lasts for weeks in mice [16]. Viruses use a number ofdi¡erent mechanisms for decreased expression of MHCincluding inhibition of peptide transport across the endo-plasmic reticulum [17,18] and disruption of the Jak/Statsignaling pathway [19]. Moreover, recently it has beenshown that bacteria may use similar mechanisms for sur-vival. That is, Mycobacterium tuberculosis escapes immunedetection by decreasing the expression of the antigen-pre-senting molecule CD1 on human cells [20], enabling it topersist and evade detection by double negative T cells. Thepresent study examined the e¡ect of intracellular brucellaeon MHC expression by macrophages using Brucella abor-tus strain 2308 expressing green £uorescent protein (GFP)and three-color £ow cytometric evaluation to monitorMHC expression on the Brucella-containing macrophages.

2. Materials and methods

2.1. Construction of the GFP-Brucella

GFP-Brucella was generated by introducing a GFP-ex-pressing derivative of the broad-host-range stable cloningvector pBBR1MCS [21] into virulent B. abortus 2308. Thiswas accomplished by joining the coding region for GFPwith the promoter region of the kanamycin antibiotic re-sistance gene, aphA (PaphA), through the use of overlap-extension PCR [22]. The primer set (a) 5P-atgcatGGAT-GAATGTCAGCTACT-3P and (b) 5P-CTTCTCCTTTAC-TCATGCGAAACGATCCTCATC-3P was used to ampli-fy PaphA from pBS-Kan [23], while the primer set (c)5P-GATGAGGATCGTTTCGCATGAGTAAAGGAGA-AG-3P and (d) 5P-ctgcagTTATTTGTATAGTTCATCCA-TGCC-3P was used to amplify S65-T GFP from pS65T-C1(Clontech Laboratories, Inc., Palo Alto, CA, USA). Theresultant PCR products contained short overlapping re-gions of homology (the heterologous sequences are under-lined in each oligonucleotide) which were extended andampli¢ed using the outside primers denoted as ‘a’ and‘d’ in a second round of ampli¢cation. The resultantPaphA-GFP product was cloned by blunt end ligationinto the EcoRV site of pBluescript II KSþ (Stratagene,La Jolla, CA, USA) and subsequently liberated by diges-

tion with NsiI and PstI, whose recognition sequences wereincorporated into the primers ‘a’ and ‘d’ respectively. Gelpuri¢ed PaphA-GFP was then ligated to the NsiI site ofpJS14 [24] yielding the recombinant broad-host-rangeplasmid, pBBR1MCS-6. Partial digestion of pBBR1-MCS-6 with NcoI and complete digestion with HpaI re-sulted in the removal of 319 bp of the GFP coding se-quence including the S65-T GFP chromophore. This wasreplaced with a similar sequence encoding the enhancedS65-T, F64-L GFP chromophore [25] isolated frompTII-6 (kindly provided by Lucy Robinson, LouisianaState University Health Sciences Center). The resultantbroad-host-range plasmid, pBBR1MCS-6Y, was then in-troduced into B. abortus 2308 via electroporation, andrecombinant clones were recovered by plating on Schaed-ler blood agar supplemented with 15 Wg ml31 chloram-phenicol to select for vector-encoded antibiotic resistance.Presence of the GFP-containing plasmid was con¢rmed by£ow cytometric analysis of bacteria. B. abortus 2308 or-ganisms containing GFP are referred to as ‘GFP-Brucella’henceforth.

2.2. Infection of mice and post-infection analyses

Male BALB/c, C57BL/10 and C57BL/6J-Ifgtml@ (IFN-Qgene knockout mice) mice were purchased from JacksonLaboratories (Bar Harbor, ME, USA). Mice were infectedintraperitoneally at around 12 weeks of age with approx-imately 5U104 CFU (colony forming units) of either pa-rental B. abortus strain 2308 or GFP-Brucella, and sacri-¢ced at the times indicated. Mouse spleens were removed,weighed and mechanically disrupted to yield a suspensionof splenocytes. Serial dilutions of splenocytes were platedon Brucella agar or Schaedler blood agar plates (BectonDickinson, Cockeysville, MD, USA) containing 15 Wgml31 chloramphenicol (Sigma, St. Louis, MO, USA) andthe colony forming units of brucellae were enumeratedafter incubation at 37‡C in 5% CO2 for 3 days. Forsome experiments, aliquots of splenocytes were stainedby immuno£uorescence as described below.

2.3. Infection of phagocytic cells

For in vitro studies, the murine macrophage cell lineJ774A.1, peritoneal exudates containing macrophagesfrom BALB/c or C57BL/10 mice or neutrophils from bo-vine blood were utilized. Neutrophils were obtained frombovine blood by centrifuging over Ficoll-Hypaque gra-dients by standard techniques, collecting the lower fractioncontaining granulocytes and erythrocytes and hypotoni-cally lysing the erythrocytes. The remaining leukocytepopulation consisted of approximately 90% neutrophilsas determined by cytospin smears and Giemsa staining.Peritoneal exudate cells were obtained from 8^12 weekold virus-free male or female mice that had been main-tained in barrier housing by injecting 1 ml of 5% proteose

FEMSIM 1409 24-6-02

E. Murphy et al. / FEMS Immunology and Medical Microbiology 33 (2002) 191^200192

peptone (Difco Laboratories, Detroit, MI, USA) intraperi-toneally and harvesting cells 5 days later as described [8].Peritoneal exudate cells and J774 cells were plated in 24-well culture plates after suspension in RPMI 1640 with10% heat-inactivated fetal bovine serum, 5U1035 M2-mercaptoethanol, and 50 Wg ml31 L-glutamine using3U105 cells/well for the peritoneal exudates and 1.5U105

cells/well for J774 macrophages (the lower concentration isused for the cell line since they may divide in culture).Cells were allowed to adhere to plastic for 2 h at 37‡Cfollowing which the non-adherent cells were removed. Theadherent cells were cultured with live or heat-killedB. abortus 2308 or heat-killed Escherichia coli at a ratioof approximately 100 bacteria to one macrophage for 2 hby centrifuging the bacteria onto the macrophages at1800Ug for 20 min, incubating at 37‡C for 1 h 40 minand then washing three times to remove the majority ofextracellular bacteria, bringing the total ‘infection period’to approximately 2 h. Following this, the remaining extra-cellular brucellae were killed by addition of gentamicin(50^100 Wg ml31) to the cultures for 2 h. The cells werethen washed twice and cultured in medium containing 12.5Wg ml31 gentamicin for the additional time indicated forthe experiment. At the appropriate time points, a sampleof the cells was lysed by adding 0.2% deoxycholate anddilutions of the lysate plated onto agar plates to determinethe number of intracellular CFU. Cells for analysis by£ow cytometry were removed from the wells by scrappingwith a rubber policeman, so that surface proteins such asMHC molecules were not removed, and stained by immu-no£uorescence as described below.

2.4. Immuno£uorescence and £ow cytometry

Cells for immunostaining were incubated on ice for 1 hin medium containing 10% heat-inactivated horse serum toinhibit non-speci¢c binding of monoclonal antibodies(mAb) to Fc receptors. Cells were stained with R-phyco-erythrin (PE)-conjugated rat anti-mouse CD11b (Mac1)mAb (PharMingen, San Diego, CA, USA). The levels ofMHC class I and II expression were determined by stain-ing with biotinylated anti-mouse class I MHC mAb (mAbM1/42) or biotinylated anti-mouse class II MHC mAb(mAb M5/114), provided by Eric Martz, University ofMassachusetts. The binding of biotin-labeled mAb wasvisualized by reacting with the secondary reagent Strepta-vidin Cy-Chrome (PharMingen, San Diego, CA, USA).After staining, cells were ¢xed with 4% paraformaldehydefor 10 min at room temperature and permeabilized over-night with 0.1% saponin in phosphate bu¡ered saline and4% paraformaldehyde to kill intracellular brucellae. Asample of the so-treated cells was cultured in brucellabroth for 7 days and the lack of bacterial growth deter-mined by turbidity measurements and culture on agarplates to ensure killing before £ow cytometric analysis.Flow cytometric analysis was performed using a FACS

Caliber (Becton Dickinson, CA, USA) and data were col-lected for 1^5U104 cells/sample. When analyzed by £owcytometry, brucellae-containing cells could be visualizedby the presence of the GFP and are referred to as‘GFPþ cells’ throughout.

2.5. Fluorescence microscopy analysis

In initial experiments, infected macrophages were ¢xedwith 4% paraformaldehyde as described in Section 2.4 andcells examined for the presence of the GFP using a NikonUV microscope (DonSanto Corporation, Natick, MA,USA) using 1000U magni¢cation and oil immersion.Also in initial experiments, extracellular GFP-Brucellawere killed by heating to 60‡C for 1 h or by treatingwith 1% hydrogen peroxide or 200 Wg ml31 of gentamicinovernight at 37‡C and then analyzed for £uorescence bymicroscopy.

3. Results

3.1. Growth of GFP-Brucella in vitro and in vivo

Experiments were conducted to ensure that the insertionof GFP into B. abortus 2308 did not a¡ect the ability ofthe brucellae to infect cells or establish chronic infectionsin mice. When neutrophils (Fig. 1A) or macrophages (Fig.1B) were infected in vitro with GFP-Brucella there was atypical pattern of killing and/or bacterial replication thatdid not signi¢cantly di¡er from that observed with theparental strain B. abortus 2308. Similarly, there was nodi¡erence in the ability of GFP-Brucella to infect and es-tablish a chronic infection in mice as determined by thenumber of CFU recovered from spleens between 1 and8 weeks post infection (Fig. 1C).

3.2. Evaluation of cells containing GFP-Brucella by£ow cytometry

To evaluate macrophages with GFP-Brucella by £owcytometry, peritoneal exudate cells from BALB/c andC57BL/10 mice as well as the J774 macrophage cell linewere infected using several bacteria to cell ratios. Flowcytometric analysis readily distinguished macrophageswith GFP-Brucella (see Fig. 2 for an example) and showedthere was a proportional increase in the percentage ofGFPþ macrophages as the ratio of bacteria to macro-phages used for infection was increased (Table 1). Thisoccurred with all three cell populations examined. An in-creased percentage of GFPþ cells in the samples also cor-related with an increase in CFU detected in the cultures(data not shown). Infected J774 macrophages were eval-uated with and without trypsinization to ensure the GFPdetected represented intracellular bacteria. Since no di¡er-ence was found in the percentage of GFPþ macrophages

FEMSIM 1409 24-6-02

E. Murphy et al. / FEMS Immunology and Medical Microbiology 33 (2002) 191^200 193

(Table 1) or in £uorescence intensity (data not shown)when the macrophages were treated with trypsin, it wasconcluded that the bacteria did not remain adherent ex-tracellularly following processing for £ow cytometry.Thus, further experiments were conducted without trypsi-nization.Experiments were conducted to determine whether kill-

ing of the brucellae extinguished the GFP £uorescence,

and thus whether the GFPþ cells re£ected only thosewith live brucellae. GFP-brucellae were treated with hy-drogen peroxide, shown previously to be an importantanti-brucella agent generated by macrophages [8], andevaluated for £uorescence by microscopy. Killing the bac-teria by hydrogen peroxide did not diminish GFP £uores-cence, nor did killing by heating or antibiotic treatment(data not shown), suggesting GFP £uorescence would notbe extinguished when intracellular bacteria were killed inmacrophages. Thus, it was concluded that the £ow cyto-metric analysis could identify cells that contained GFP-Brucella but could not be used to determine the propor-tion with live bacteria. This was con¢rmed by the follow-ing experiment. It is known that the number of CFU de-creases at least 10-fold in brucella-infected macrophagesbetween 0 and 24 h after infection and then increasesbetween 24 and 48 h [7,8]. Analysis of CFU and £owcytometric analyses of infected BALB/c peritoneal macro-phages carried out at these times showed no decrease inthe percentage of GFPþ cells to correlate with the decreasein CFU (Fig. 3).

3.3. MHC expression on GFP-Brucella-containing cellsinfected in vitro

J774 macrophages and peritoneal exudate cells wereinfected with GFP-Brucella and MHC expression de-termined using three-color £ow cytometric analysis. At48 h after infection, a time when the bacteria were in-creasing in number within the cells in culture (see Fig.3A for example), the peak of MHC class I expression onJ774 cells was the same on the GFPþ macrophages ason the GFP3 macrophages in the culture (log10 1.5; see

Table 1Flow cytometric analysis of macrophages infected in vitro with GFP-Brucella

Ratio of bac-teria to mac-rophages usedfor infection

% GFPþ macrophagesa

BALB/cmacrophages

C57BL/10macrophages

J774macrophages

TrypsinizedJ774macrophagesb

25:1 4 6 5 550:1 9 9 nd nd100:1 22 15 30 nd1000:1 nd nd 65 62

nd, not done.aFlow cytometric analyses of macrophages infected in vitro with GFP-Brucella. Analysis was conducted 1 h after the 2-h infection period wascompleted. The infections using 1000:1 ratio were performed twice; the100:1 ratio experiment was done three times; the 50:1 and 25:1 ratioexperiments were done once each. For those ratios done more thanonce the data shown are representative.bIn these experiments, the J774 infected macrophages were trypsinizedbefore analysis to remove any extracellular brucellae adherent to themembranes.

Fig. 1. Killing, survival and growth of GFP-Brucella in (A) neutrophils,(B) BALB/c peritoneal macrophages infected in vitro and (C) in BALB/c mice, compared to non-GFP-containing Brucella organisms. CFU re-covered from phagocytes at the times indicated in A and B are reportedas the percentage of those recovered immediately following infection.The mean and S.E.M. of replicate cultures are shown and are represen-tative of four experiments performed. In C, the CFU per spleen forgroups of ¢ve mice is shown as the mean and S.E.M.

FEMSIM 1409 24-6-02

E. Murphy et al. / FEMS Immunology and Medical Microbiology 33 (2002) 191^200194

Fig. 4) although the mean £uorescence was higher for theGFPþ macrophages. This also occured for class II expres-sion.Further experiments compared the MHC expression on

uninfected peritoneal exudate cells and those that hadphagocytosed either live or heat-killed GFP-Brucella usingperitoneal cells. Fewer cells were GFPþ than with the J774macrophage cell line, indicating a lower proportion of in-fected cells at 48 h. This is consistent with previous resultsshowing brucellae survive better and grow to higher num-bers in J774 cells than in peritoneal exudate cells [26]. Themean £uorescence intensity for both class I and class IIMHC on GFPþ peritoneal cells following culture with liveor heat-killed GFP-Brucella for 48 h was slightly greaterthan that on cultures of peritoneal exudate cells that had

not been exposed to brucellae or those allowed to phago-cytose E. coli (Fig. 5). Peritoneal exudate cells from theBALB/c and C57BL/10 mice were also compared with oneanother since these mice are more susceptible or resistantto chronic infection, respectively [16]. There was no gen-eralized decrease in the level of MHC class I or II expres-sion on GFPþ cells following infection with live GFP-Bru-cella relative to that on the GFP3 cells for either of themouse strains, suggesting that this is not a mechanism thatcontributes to susceptibility (data not shown).

3.4. MHC expression on GFP+ cells following infection invivo

Further experiments evaluated cells infected in vivo.

Fig. 3. Peritoneal macrophages from BALB/c mice were infected with GFP-Brucella and evaluated for (A) CFU by culturing on agar plates and (B) the% GFPþ cells detected by £ow cytometry after culture for the times indicated. The data are for replicate cultures and the mean and S.E.M. are shownand are representative of three experiments performed.

Fig. 2. Detection of J774 macrophages containing GFP-Brucella. The left-hand panel shows uninfected macrophages while the right-hand panel showscells in the macrophage culture 48 h after infection. Mac1þ macrophages are found in the top quadrants shown while GFPþ cells are in the right-handquadrants.

FEMSIM 1409 24-6-02

E. Murphy et al. / FEMS Immunology and Medical Microbiology 33 (2002) 191^200 195

Fig. 5. Peritoneal exudate cells from BALB/c mice were (A) uninfected, (B) containing heat-killed E. coli, (C) infected with GFP-Brucella, or (D) con-taining heat-killed GFP-Brucella. They were evaluated for expression of MHC class I (left-hand panels) and MHC class II (right-hand panels) moleculesafter 48 h of culture. Only the GFPþ cells within the cultures are shown in panels C and D.

Fig. 4. Cells were infected with GFP-Brucella and evaluated for expression of MHC class I (left-hand panels) and MHC class II (right-hand panels)molecules on GFP3 (lighter gray histogram) or on GFPþ cells (darker gray histogram) using the J774 macrophage cell line evaluated after 48 h of cul-ture. The GFPþ cells were 42% of the total cells in the culture and the mean £uorescence intensity for GFPþ cells was 102 for class I and 111 for classII while for the GFP3 cells it was 26 for class I and 32 for class II MHC.

FEMSIM 1409 24-6-02

E. Murphy et al. / FEMS Immunology and Medical Microbiology 33 (2002) 191^200196

Splenocytes from BALB/c mice that had been infected1 and 3 weeks previously were analyzed by three-color£uorescence. Mac1 was used as a marker of macrophageswithin the splenic leukocyte population. It should be notedthat there were 1^10% neutrophils within this Mac1þ pop-

ulation [6] and approximately 5% Mac1lo cells (data notshown) that may have been dendritic cells [27] were in-cluded. It was found that the Mac1þ/GFPþ splenic leuko-cytes had levels of MHC class I and II expression at leastas high as that on the Mac1þ/GFP3 splenocytes recovered

Fig. 6. Expression of MHC class I and II molecules on GFP-Brucella splenic leukocytes from 2308-GFP-infected BALB/c mice at (A) 1 week or (B)3 weeks post infection. Left-hand panels were stained for class I MHC; right-hand panels were stained for MHC class II. The white histograms repre-sent GFP3 cells, while the black represent the GFPþ cells.

Fig. 7. Expression of MHC class I and II molecules on splenic macrophages from GFP-Brucella infected IFN-Q knockout C57BL/6 mice at (A) 5 weekspost infection or (B) at the time of death (6.5 weeks post infection). Left-hand panels were stained for class I MHC; right-hand panels were stained forMHC class II. The white histograms are for GFP3 cells, while the black are for the GFPþ cells.

FEMSIM 1409 24-6-02

E. Murphy et al. / FEMS Immunology and Medical Microbiology 33 (2002) 191^200 197

from the infected mice at both 1 and 3 weeks after infec-tion (Fig. 6).In a second type of experiment, IFN-Q gene-disrupted

C57BL/6 mice were employed since these mice have pre-viously been shown to have extremely high numbers ofsplenic CFU and to ultimately succumb to brucellosis asa result [6]. Again, the expression of MHC class I and IIon the Mac1þ/GFPþ macrophages at 5 weeks post infec-tion was again at least as high on the Mac1þ/GFP3 sple-nocytes (Fig. 7). While both class I and class II MHClevels decreased on some of the Mac1þ/GFPþ and a pop-ulation of the Mac1þ/GFP3 cells when the mice weremoribund at 6.5 weeks post infection, their decrease wasnot greater on the Mac1þ/GFPþ (Fig. 7).

4. Discussion

Immune responses appropriate for control of an intra-cellular pathogen are generated in mice infected with thevirulent strain 2308 of B. abortus, as shown by the abilityof their T cells to adoptively transfer resistance to naivemice [9]. The infection nevertheless persists for months[16]. We hypothesized that the persistence was due to sub-optimal presentation of antigen to the T cells in vivo as aresult of decreased MHC class I and II molecules on mac-rophages containing brucellae. Decreased expression ofMHC molecules by antigen-presenting cells would allowB. abortus to elude or diminish the probability of detectionby the host’s immune system, similar to mechanisms usedby viruses [18] and mycobacteria [20].Experiments that evaluated expression of MHC mole-

cules on the J774 macrophage cell line and peritoneal exu-date cells that had been infected in vitro found no obviousdecrease in MHC class I or II expression 48 h followingphagocytosis of brucellae (GFPþ cells) relative to that oncells that did not contain brucellae (GFP3 cells) eitherfrom the same cultures or in a separate culture. Similarly,no lower MHC expression was apparent among Mac1þ/GFPþ splenocytes obtained following in vivo infection ofBALB/c or C57BL/6 IFN-Q knockout mice. These resultsare in contrast to studies evaluating the e¡ect of E. coliand mycobacteria on a human macrophage cell line wherea decrease in MHC class I and II molecules occurred ini-tially although they recovered by 8 days post infection[28].When evaluating the peritoneal exudate cells by £ow

cytometry using Mac1 expression as a criterion it is pos-sible that a small percentage of cells other than macro-phages were included, such as neutrophils and dendriticcells. Peritoneal exudates contain approximately 2% neu-trophils and although dendritic cells tend to have lowMac1 expression and to be non-phagocytic [27], we cannotrule out their inclusion in the analysis since there were afew Mac1lo cells in the population analyzed. Similarly, innormal B. abortus-infected C57BL/6 mice neutrophils

comprise approximately 10% of the splenic Mac1þ popu-lation throughout the infection of normal C57BL/6 mice[6] while in the IFN-Q knockout mice they constitute ap-proximately 1% at the later stages of infection [6], the timewhen we examined the splenocytes in this study. A smallpercentage of those splenocytes may also have been den-dritic cells as identi¢ed by the Mac1lo expression [27].Thus, while these cell populations would have been in-cluded in the analysis of the Mac1þ splenocytes, the im-portant question of whether the presence of brucellae incells with the potential to present antigen, regardless ofwhether they are macrophages or dendritic cells, is coinci-dent with loss of MHC expression was still answered. Thatis, there was no apparent decrease in MHC expressioncoincident with the presence of GFP although a decreaseocurred on both GFPþ and GFP3 Mac1þ splenic cellsfrom moribund IFN-Q knockout mice. This is when veryhigh numbers of brucellae per spleen are recovered and themice die from symptoms similar to endotoxemia or ca-chexia and the spleens are friable [6]. It is possible thatthere was an increase in proteolytic enzyme concentrationduring this time due to cell death and that these enzymescleaved MHC class I and II molecules from cell surfaces.Alternatively, or in addition, the cytokines produced dur-ing this acute in£ammatory response may have generallydecreased cellular MHC expression. However, we can stillconclude that the presence of brucellae did not enhancethe decrease.As an aside, it is notable that the IFN-Q knockout mice

had the same percentage of GFP-containing macrophagesin their spleens as the normal mice following infection.This occurred even though the total number of brucel-lae/spleen is about 30-fold greater in the C57BL/6 IFN-Qgene-disrupted mice and there are approximately twice asmany macrophages in their spleens [6]. Thus, either thebacteria were largely extracellular or the number of bac-teria per macrophage was greatly increased.The ability to tract bacteria in vivo during an infection

has obvious advantages. Conventionally this has beenachieved by analyzing tissue sections from infected ani-mals [1]. When histology is combined with the use of anti-bodies that identify cell types or molecules, much can belearned about the pathology associated with disease.Nevertheless, the development of methods to insert genescoding for £uorescent molecules into bacteria augmentedour ability to analyze events in infected animals. Not onlycan bacteria be tracked in tissues and cells but, because itis possible to evaluate host cells from infected animals by£ow cytometric analysis, quantitative data regarding thesecells may be obtained after they have been subjected to theenvironmental changes that occur in vivo during the in-fection. We feel con¢dent that the results using GFP-Bru-cella represent those that occur during infection with bru-cellae not containing the GFP plasmid since no di¡erencesin the ability of the bacteria to infect cells or mice wasnoted. Moreover, evaluations of immune parameters in-

FEMSIM 1409 24-6-02

E. Murphy et al. / FEMS Immunology and Medical Microbiology 33 (2002) 191^200198

cluding induction of antibodies to brucellae and delayedtype hypersensitivity responses were the same when GFP-Brucella was used to infect mice (P.H. Elzer and S.D.Hagius, unpublished data). One shortcoming noted inthis study was that detection of GFP as a measure ofthe presence of brucellae did not allow distinction betweenlive and dead intracellular bacteria. Even so, because theremaining bacteria in macrophages begin to replicate afterthe initial killing event [7], it is likely that many if not mostof the GFPþ macrophages contained some live organisms.In conclusion, no correlation of decreased MHC expres-

sion levels with the presence of GFP was found except in asubpopulation of in vivo infected macrophages at the timeof death of IFN-Q gene-disrupted mice. A decrease inMHC molecules at this time in the IFN-Q gene-disruptedmice was almost certainly irrelevant to the survival of themice and thus we conclude that a decrease in MHC ex-pression is not likely to contribute to the chronic infectionassociated with brucellosis. Finally, the results indicatedthat there was no obvious di¡erence with BALB/c andC57BL/10 peritoneal cells with regard to MHC expressionwhen infected in vitro even though they di¡er in inherentresistance to chronic infection, the C57BL/10 being moreresistant [16]. Thus, it seems unlikely that the di¡erence inresistance relates to a greater tendency for MHC mole-cules to decrease on BALB/c cells that are infected withbrucellae or in their ability to present brucellae antigens.

Acknowledgements

This work was supported by USDA/NRI competitivegrants program #9303615 and 9501995. We wish to thankSeth Blumerman for assistance with the £ow cytometricpro¢les.

References

[1] Anderson, T.D. and Cheville, N.F. (1986) Ultrastructural morpho-metric analysis of Brucella abortus-infected trophoblasts in experi-mental placentitis. Am. J. Pathol. 124, 226^237.

[2] Stevens, M.G., Pugh Jr., G.W. and Tabatabai, L.B. (1992) E¡ects ofQ-interferon and indomethacin in preventing Brucella abortus infec-tions in mice. Infect. Immun. 60, 4407^4409.

[3] Khan, Y. and Cheers, C. (1993) Endogenous Q interferon mediatesresistance to Brucella abortus infection. Infect. Immun. 61, 4899^4901.

[4] Fernandes, D.M. and Baldwin, C.L. (1995) IL-10 downregulates pro-tective immunity to Brucella abortus. Infect. Immun. 63, 1130^1133.

[5] Fernandes, D.M., Jiang, X., Jung, J.H. and Baldwin, C.L. (1996)Comparison of T cell cytokines in resistant and susceptible mice in-fected with virulent Brucella abortus strain 2308. FEMS Immunol.Med. Microbiol. 16, 193^203.

[6] Murphy, E.A., Sathiyaseelan, J., Parent, M., Zou, B. and Baldwin,C.L. (2001) Interferon-Q is crucial for survival of Brucella abortusinfections in resistant C57BL/6 and susceptible BALB/c mice. Immu-nology 103, 511^518.

[7] Jiang, X. and Baldwin, C.L. (1993) E¡ects of cytokines on intra-cellular growth of Brucella abortus. Infect. Immun. 61, 124^134.

[8] Jiang, X., Leonard, B., Bensen, R. and Baldwin, C.L. (1993) Macro-phage control of Brucella abortus : role of reactive oxygen intermedi-ates and nitric oxide. Cell. Immunol. 151, 309^319.

[9] Araya, L.N., Elzer, P.H., Rowe, G.E., Enright, F.M. and Winter,A.J. (1989) Temporal development of protective cell-mediated andhumoral immunity in BALB/c mice infected with Brucella abortus.J. Immunol. 143, 3330^3337.

[10] Mielke, M.E.A. (1991) T cell subsets in granulomatous in£ammationand immunity to L. monocytogenes and B. abortus. Behring Inst.Mitt. 88, 99^111.

[11] Oliveira, S.C. and Splitter, G.A. (1995) CD8 type 1 CD44hiCD45RBlo T lymphocytes control intracellular Brucella abortusinfection as demonstrated in major histocompatibility complexclass I and class II de¢cient mice. Eur. J. Immunol. 25, 2551^2557.

[12] Pavlov, H., Hogarth, M., McKenzie, I.F.C. and Cheers, C. (1982) Invivo and in vitro e¡ects of monoclonal antibody to Ly antigens onimmunity to infection. Cell. Immunol. 7, 127^138.

[13] Araya, L.N. and Winter, A.J. (1990) Comparative protection ofmice against virulent and attenuated strains of Brucella abortus bypassive transfer of immune T cells or serum. Infect. Immun. 58, 254^256.

[14] Baldwin, C.L. and Roop II, R.M. (1998) Brucella infections andimmunity. In: Opportunistic Intracellular Bacteria and Immunity(Paradise, L.J., Friedman, H. and Bendinelli, M., Eds.), PlenumPress, New York, pp. 255^279.

[15] Cheers, C. and Pagram, F. (1979) Macrophage activation duringexperimental murine brucellosis : a basis for chronic infection. Infect.Immun. 23, 197^205.

[16] Montarez, J.A. and Winter, A.J. (1986) Comparison of living andnonliving vaccines for Brucella abortus in BLAB/c mice. Infect. Im-mun. 53, 245^251.

[17] Fruh, K., Kwangseog, A., Djaballah, H., Sempe, P., van Endert,P.M., Tampe, R., Peterson, P.A. and Yang, Y. (1995) A viral inhib-itor of peptide transporters for antigen presentation. Nature 375,415^418.

[18] York, I.A., Roop, C., Andrews, D.W., Riddell, S.R., Graham, F.L.and Johnson, D.C. (1994) A cytosolic herpes simplex virus proteininhibits antigen presentation to CD8þ T lymphocytes. Cell 77, 525^535.

[19] Brander, C., Suscovich, T., Lee, Y., Nguyen, P.T., O’Connor, P.,Seebach, J., Jones, N.G., Van Gorder, M., Walker, B.D. and Scad-den, D.T. (2000) Impaired CTK recognition of cells latently infectedwith Kaposi’s sarcoma-associated herpes virus. J. Immunol. 165,2077^2083.

[20] Stenger, S., Niazi, K.R. and Modlin, R.L. (1998) Down-regulation ofCD1 on antigen-presenting cells by infection with Mycobacteriumtuberculosis. J. Immunol. 161, 3582^3588.

[21] Elzer, P.H., Kovach, M.E., Phillips, R.W., Robertson, G.T., Peter-son, K.M. and Roop II, R.M. (1995) In vivo and in vitro stability ofthe broad-host-range cloning vector pBBR1MCS in six Brucella spe-cies. Plasmid 33, 51^57.

[22] Horton, R.M., Cai, Z.L., Ho, S.N. and Pease, L.R. (1990) Genesplicing by overlap extension: tailor-made genes using the polymerasechain reaction. BioTechniques 8, 528^535.

[23] Kovach, M.E., Elzer, P.H., Hill, D.S., Robertson, G.T., Farris,M.A., Roop II, R.M. and Peterson, K.M. (1995) Four newderivatives of the broad-host-range cloning vector pBBR1MCS,carrying di¡erent antibiotic-resistance cassettes. Gene 166, 175^176.

[24] Robertson, G.T., Kovach, M.E., Allen, C.A., Ficht, T.A. and RoopII, R.M. (2000) The Brucella abortus Lon functions as a generalizedstress response protease and is required for wild-type virulence inBALB/c mice. Mol. Microbiol. 35, 577^588.

FEMSIM 1409 24-6-02

E. Murphy et al. / FEMS Immunology and Medical Microbiology 33 (2002) 191^200 199

[25] Cormack, B.P., Valdivia, R. and Falkow, S. (1996) FACS-optimizedmutants of the green £uorescent protein (GFP). Gene 173, 33^38.

[26] Baldwin, C.L., Jiang, X. and Fernandes, D.F. (1993) Macrophagecontrol of Brucella abortus : in£uence of cytokines and iron. TrendsMicrobiol. 1, 99^104.

[27] Steinman, R.M. (1991) The dendritic cell system and its role in im-munogenicity. Annu. Rev. Immunol. 9, 271^296.

[28] De Lerma Barbaro, A., Tosi, G., Valle, M.T., Megiovanni, A.M.,Sartoris, S., D’Agostino, A., Soro, O., Mingari, M.C., Canonica,G.W., Manca, F. and Accolla, R.S. (1999) Distinct regulation ofHLA class II and class I cell surface expression in the THP-1 macro-phage cell line after bacterial phagocytosis. Eur. J. Immunol. 29, 499^511.

FEMSIM 1409 24-6-02

E. Murphy et al. / FEMS Immunology and Medical Microbiology 33 (2002) 191^200200