Downloaded From

Home , Listeriosis

Antibodies Highly Effective in SCID Mice During Infection by the Intracellular Bacterium Ehrlichia chaffeensis Are of Picomolar Affinity and Exhibit Preferential This information is current as Epitope and Isotype Utilization of September 29, 2021. Julia Shu-yi Li, Frederick Chu, Andrew Reilly and Gary M. Winslow J Immunol 2002; 169:1419-1425; ; doi: 10.4049/jimmunol.169.3.1419 Downloaded from http://www.jimmunol.org/content/169/3/1419

References This article cites 39 articles, 24 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/169/3/1419.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 29, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2002 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Antibodies Highly Effective in SCID Mice During Infection by the Intracellular Bacterium Ehrlichia chaffeensis Are of Picomolar Afﬁnity and Exhibit Preferential Epitope and Isotype Utilization1

Julia Shu-yi Li,* Frederick Chu,† Andrew Reilly,† and Gary M. Winslow2*†

Although often considered to be ineffective against intracellular bacteria, Abs, in the absence of lymphocytes, have been shown previously to protect SCID mice from lethal infection by the obligate intracellular bacterium Ehrlichia chaffeensis, even when administered well after infection has been established. To identify characteristics of Abs that are critical for host defense during this intracellular infection, a panel of Ehrlichia-specific mAbs was generated and analyzed. Among 100 Abs recovered, 39 recognized an amino-terminal hypervariable region of an outer membrane protein (OMP), demonstrating that the OMPs are both Downloaded from antigenically variable and immunodominant. A subset of 16 representative OMP-specific Abs was further examined to identify characteristics that were essential for in vivo efficacy. The highly effective Abs recognized a linear epitope within the first hypervariable region of OMP-1g. Only IgG were found to be effective, and among the effective IgG, the following hierarchy was IgG2b. The most striking characteristics of the highly effective Abs were their picomolar binding ؍ observed: IgG2a > IgG3 affinities and long binding t1/2. Thus, although epitope recognition and isotype use may contribute to efficacy, high affinity may be a critical characteristic of Abs that can act effectively during this intracellular bacterial infection. The Journal of Immunology, http://www.jimmunol.org/ 2002, 169: 1419Ð1425.

uman monocytic ehrlichiosis is an acute febrile illness fective IgG2a and IgG3 outer membrane protein (OMP)3-specific caused by Ehrlichia chaffeensis, an obligate intracellular Abs (5), but it was not clear to what extent properties such as H bacterium that resides in monocytes and macrophages epitope, isotype, and affinity contributed to Ab efficacy. The (for review, see Ref. 1). The immune response to E. chaffeensis is present study extends previous work by evaluating the in vivo incompletely characterized, but T cells most likely play an impor- efficacy of a large panel of OMP-specific mAbs. The findings re- tant role in host defense (2–4). However, Abs also can contribute veal important characteristics of Abs that are most likely critical by guest on September 29, 2021 to host defense against this intracellular pathogen. Previous studies for effectiveness during this intracellular infection. in the mouse have demonstrated that passive transfer of either immune serum or mAbs, in the absence of lymphocytes, can pro- Materials and Methods vide long-term protection to susceptible SCID mice, even when the Mice Abs are administered therapeutically (2, 5). Ab-mediated immu- C57BL/6 and BALB/c-scid mice were obtained from The Jackson Labo- nity in this model was associated with partial to apparently com- ratory (Bar Harbor, ME), or were bred in the Wadsworth Center Animal plete elimination of the bacteria from liver tissue within as early as Care Facility, under institutional guidelines for animal care and use. Six- to 4 days of Ab administration. Abs were also shown to be effective 12-wk-old, sex-matched mice were routinely infected via the peritoneum with 1–2 ϫ 106 E. chaffeensis-infected DH82 cells, as described previously in immunocompetent mice during this intracellular infection (2). (6). Institutional animal care and use guidelines do not permit the use of The mechanism of humoral immunity is not known, but may in- death as an experimental endpoint in these studies, so body weight mea- volve opsonization of bacteria that are exposed to Abs during in- surements were used to monitor animal health and Ab efficacy (see below). tercellular transfer, or perhaps novel mechanisms. Immunity is not The infected animals were sacrificed when moribund, as characterized by achieved by passive neutralization, and most likely involves the a lack of mobility, hunched posture, ruffled fur, and a pronounced loss of body weight (Ͼ30% loss of initial weight). Tissue samples were harvested interaction of Abs with host cells and host cell receptors. and stored at Ϫ80°C before further analysis. One approach to identifying the relevant mechanism of humoral Hybridoma production and Ab purification immunity is to identify characteristics of Abs that are critical for efficacy in vivo. A previous study identified and characterized ef- Three fusions for hybridoma production were performed independently. C57BL/6 mice were infected via the peritoneum with E. chaffeensis-infected DH82 cells, in the absence of adjuvant. For the first and second *Department of Biomedical Sciences, School of Public Health, State University of fusions, mice were infected two to four times at 2- to 4-wk intervals. For New York, Albany, NY 12201; and †Wadsworth Center, New York State Department the third fusion, mice were infected once, and the fusion was performed 2 of Health, Albany, NY 12208 wk later. Splenocytes were harvested from the infected mice and fused to Received for publication January 29, 2002. Accepted for publication May 29, 2002. the myeloma cell line SP2/0, using standard protocols. The hybridoma supernatants were screened for reactivity to E. chaffeensis by immunoflu- The costs of publication of this article were defrayed in part by the payment of page orescence assay, as described previously (2), and hybridomas that pro- charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. duced specific Abs were expanded and subcloned by limiting dilution. H and L chain Ab isotypes were determined by ELISA using isotype-specific 1 This work was supported in part by U.S. Public Health Service Grant polyclonal reagents (Southern Biotechnology Associates, Birmingham, 5R29CA69710-02. 2 Address correspondence and reprint requests to Dr. Gary Winslow, Wadsworth Center, 120 New Scotland Avenue, Albany, NY 12208. E-mail address: 3 Abbreviations used in this paper: OMP, outer membrane protein; HVR, hypervari- [email protected] able region; MIWD, mean integrated weight difference; RU, resonance unit.

AL). Abs were purified from hybridoma culture supernatants by fast per- Prism 7700 Sequence Detection System (PE Applied Biosystems). A stan- formance liquid chromatography, using protein A- or G-Sepharose (for dard curve was generated using data from analysis of standardized quan- IgGs; Amersham Pharmacia Biotech, Piscataway, NJ) or IgM purification tities of a gel-purified DNA fragment that contained nt 1–619 of the 16S- columns (for IgM; Amersham Pharmacia Biotech), following the instruc- rDNA gene of E. chaffeensis. Quantitation of the template standards was tions of the manufacturer. The concentrations of purified Abs were deter- performed using a Picogreen DNA quantitation kit (Molecular Probes, Eu- mined by measurement of absorbance at 280 nm using a spectrophotometer gene, OR). The data were analyzed using the software supplied by the (Pharmacia Biotech, Cambridge, U.K.). Unpublished data indicated that a manufacturer (PE Applied Biosystems). single dose of 50 ␮g mAb Ec56.5, administered on day 10 postinfection, was sufficient to mediate nearly complete bacterial clearance within 4 days. Evaluation of Ab efficacy in vivo Abs were administered via the peritoneum in weekly doses of 200 ␮gto facilitate protection in the long-term studies. Ab efficacy was evaluated in long-term studies by measurement of body weight changes. Purified Abs (200 ␮g) were injected via the peritoneum ELISA into BALB/c-scid mice on day 7 postinfection, and at weekly intervals thereafter. The mice were monitored for disease and morbidity throughout Epitope analyses were performed as described previously (5). Purified the experiments, and the changes in body weight from day 0 were recorded. rOMP-1g and truncated rOMPs were adsorbed overnight to 96-well mi- In each experiment, buffer was administered as a negative control, and the crotiter plates (Dynex Technologies, Chanitilly, VA) at a concentration of protective Ab Ec56.5 was used as a positive control. In some experiments, 3 ␮g/ml in PBS, and peptides were adsorbed overnight in sodium carbonate uninfected SCID mice were also included in the analyses. buffer (pH 9.6) at a concentration of 10 ␮g/ml. The microtiter plates were To quantitate Ab efficacy data, and to compare data from multiple ex- blocked with 1% nonfat dry milk in PBS. Bound Abs were detected using periments, the value of mean integrated weight difference (MIWD) was alkaline phosphatase-conjugated anti-mouse Ig secondary Abs (Southern determined for each Ab-treated group to indicate the in vivo efficacy of Biotechnology Associates), followed by p-nitrophenyl phosphate (Sigma- each Ab. The body weight changes in groups of Ab-treated vs untreated

Aldrich, St. Louis, MO). The absorbance was read at 405 nm with a Ther- infected mice were ﬁrst normalized by integrating the body weight data, Downloaded from moMax microplate reader (Molecular Devices, Sunnyvale, CA). over time, up to the point that the untreated mice were deemed moribund. The difference in integrated values between the Ab-treated and untreated Ab serum t1/2 measurements control groups was then divided by the number of days to morbidity, to obtain the MIWD values (in grams) for each group of Ab-treated mice. The For determination of Ab t in serum, uninfected BALB/c-scid mice were 1/2 weight data integration was performed using KaleidaGraph data analysis/ administered a single dose of 200 ␮g Ab via the peritoneum, and 100 ␮l graphing application software (Synergy Software, Reading, PA). blood samples were withdrawn retroorbitally 1, 4, and 7 days after Ab

administration. Sera were obtained after centrifugation of blood samples at http://www.jimmunol.org/ 4000 rpm for 20 min. The concentrations of mAb in the sera were deter- Results mined by ELISA, using predetermined concentrations of purified Abs as Production of E. chaffeensis mAbs standards. Abs used as standards for quantitation were identical with those used for in vivo administration. For production of mAbs for study of the humoral immune response, immunocompetent C57BL/6 mice were infected via the Affinity measurements peritoneum with E. chaffeensis (the Arkansas isolate), spleen cells Affinity measurements were performed with the BIAcore 3000 instrument were isolated at various times thereafter, hybridomas were gener- (BIAcore, Uppsala, Sweden). Purified rOMP-1g and purified rabbit anti- ated, and Abs were screened by immunofluorescence assay. A total mouse Fc␥ were covalently bound to the flow cell surfaces of CM5 sensor of 100 E. chaffeensis-specific hybridomas was recovered and char- chips (BIAcore), using an amine coupling kit supplied by the manufacturer. by guest on September 29, 2021 HEPES-buffered saline (10 mM HEPES, 150 mM NaCl, 3.4 mM EDTA, acterized. The hybridomas described in this study were recovered and 0.05% surfactant P20, pH 7.4) was used as the standard running buffer. from three separate hybridoma fusions. Two fusions were per- Serial dilutions of mAb (1000, 500, 250, 125, 62.5 nM) were injected over formed following repeated infections (fusions 1 and 2), and a third four flow cell surfaces containing either rOMP-1g, anti-Fc␥, or a blank 14 days after a single infection (fusion 3; Table I). Two Abs from control treated with coupling reagents, or an unmodified blank control. The the panel have been described previously, Ec56.5 (IgG2a) and injection volume was 30 ␮l, with flow rate of 15 ␮l/min. Ab dissociation in running buffer was monitored for 4.7 min, and the flow cells were re- Ec18.1 (IgG3). Both Abs were effective in infected SCID mice, generated after each assay with 100 mM HCl/50 mM glycine, pH 2.4. The mediating significant clearance of the infection from liver tissues responses were measured in resonance units (RU). Association and disso- within 4 days of administration (5). ciation rates were calculated from RU data using the BIAevaluation soft- Mice and humans make immunodominant responses to E. ware supplied by the manufacturer. chaffeensis OMPs (2, 7), so the Abs were first analyzed to deter- Quantitative PCR analyses mine the frequency of OMP-specific Abs. Accordingly, 40% of the Abs recovered from three independent fusions recognized OMP- The Ehrlichiae were quantitated by quantitative real-time PCR using the SYBR Green PCR Reagent kit (PE Applied Biosystems, Foster City, CA), 1g, an Ag known to be expressed by the Arkansas isolate used in and the following E. chaffeensis-specific 16S rDNA oligonucleotide prim- these studies (Table I). IgM and all subclasses of IgG were recov- ers: 5Ј-AACACATGCAAGTCGAACGG-3Ј (sense) and 5Ј-CCCCCGCAG ered, except IgG1, even though OMP-specific IgG1 polyclonal se- GGATTATACA-3Ј (antisense) at a concentration of 100 nM. Genomic rum responses were identified in C57BL/6 mice (data not shown). DNA (5 ng) was amplified for 40 cycles (95°C/15 s; 60°C/60 s) using 0.025 U/␮l AmpliTaq Gold DNA polymerase (PE Applied Biosystems), in The IgGs were predominant among those Abs that were OMP specific (75%), and IgM were predominant among the non-OMP- reaction buffer containing 3 mM MgCl2 and 200 nM dNTPs, in a volume of 25 ␮l. Reaction products were monitored in real time using the ABI specific Abs (57%; data not shown).

Table I. Characteristics of E. chaffeensis-speciﬁc mAbs

Ag Speciﬁcity Isotype Distribution

OMP Other ␮ ␥3 ␥2b ␥2a Fusion No. % (no.) % (no.) % (no.) % (no.) ␥1 % (no.) % (no.)

1 15 46.7 (7) 53.3 (8) 40.0 (6) 26.7 (4) 0 20.0 (3) 13.3 (2) 2 24 54.2 (13) 45.8 (11) 54.2 (13) 0 (0) 0 8.3 (2) 37.5 (9) 3 61 32.8 (20) 67.2 (41) 41.0 (25) 19.7 (12) 0 34.4 (21) 4.9 (3) Total 100 40.0 (40) 60.0 (60) 44.0 (44) 16.0 (16) 0 26.0 (26) 14.0 (14) The Journal of Immunology 1421

Table II. Epitope analysis of OMP-specific Abs thetic peptides derived from HVR1. Seventy-two percent (29 of 40) of the OMP-specific Abs recognized the OMP-1g peptide 61– Ag % No. 90, and 32% (13 of 40) recognized nested peptide containing residues 65–78 (Table II). OMP-1g 100 40 ⌬3a 97.5 39 The observation that among the 39 Abs that recognized OMP-1g ⌬2/3 97.5 39 HVR1 only 29 recognized peptide 61–90 suggested that two or ⌬1/2/3 0 0 more distinct epitopes were found in HVR1. Moreover, among the c ⌬1 7.5 3 29 Abs that recognized peptide 61–90, only 13 recognized the OMP-1d 5.0 2 nested peptide 65–78. This suggested that additional differences in Peptide 61–90b 72.5 29 Peptide 65–78 32.5 13 fine specificity or affinity, or Ag conformation, abrogated recognition of the nested peptide by some Abs. The epitope analyses a Indicates HVR truncation. b Numbering indicates amino acid residues within OMP-1g. nevertheless demonstrated that the majority of the Abs, obtained c Two Abs, Ec54.1 and Ec95.2, exhibited unexplained binding to OMP⌬1. from three independent immunizations, recognized identical or closely related epitopes in HVR1 of OMP-1g, and revealed that the murine humoral immune response to the Arkansas isolate was Highly restricted epitope distribution of OMP-1g Abs dominated not only by recognition of OMP-1g, but also by rec- The predominance of OMP Abs in the humoral response, as indi- ognition of restricted epitopes within HVR1. cated by the hybridoma analyses, was consistent with that observed in studies of the polyclonal Ab response, which suggested Evaluation of Ab efficacy Downloaded from that the specificity of the Ab panel was indeed representative of the To identify Abs that could mediate host defense, 16 of the 40 overall humoral response during infection of C57BL/6 mice. Fur- OMP-specific Abs were chosen for in vivo study, including the ther characterization of the OMP-specific Abs was performed, be- two previously described Abs, Ec56.5 and Ec18.1 (5). Abs were cause OMP-specific Abs were shown previously to provide pro- chosen for in vivo analyses on the basis of their class (IgM and tection from lethal infection in SCID mice (5). IgG), subclass (IgG2a, IgG2b, and IgG3), and epitope specificity

The E. chaffeensis OMPs together form a family of 21 related (Table III). Abs of common isotype were grouped operationally, http://www.jimmunol.org/ proteins that differ largely in three short regions, called hypervari- based on their ability or inability to recognize HVR1 peptide 61– able regions (HVRs) (8, 9). To identify the epitopes of the panel of 90. Purified Abs (200 ␮g) were injected via the peritoneum into OMP-specific Abs, a series of truncated rOMP-1g Ags, each lack- susceptible immunodeficient BALB/c-scid mice after infection had ing one or more of the three HVRs (OMP⌬3, ⌬2/3, ⌬1/2/3, and been established (7 days postinfection), and at weekly intervals ⌬1), were analyzed by ELISA, as described previously (5). All but thereafter. The mice were monitored for disease and morbidity one of the OMP-specific Abs (39 of 40) recognized rOMP-1g Ags throughout the experiments. Effective Abs have been shown pre- that lacked HVRs 3 and 2 (OMP⌬3, ⌬2/3), but most failed to viously to mediate partial to apparently complete bacterial clear- recognize those lacking HVR1 (OMP⌬1/2/3 or ⌬1; Table II). ance from liver tissue within as early as 4 days post-Ab adminis- These observations indicated that most Abs recognized an epitope tration, whereas isotype-matched irrelevant control Abs or PBS by guest on September 29, 2021 within the first HVR. Moreover, most Abs also failed to recognize had no effect (5). Fully quantitative assays had been unavailable, a related OMP, OMP-1d (Table II), indicating that these Abs most however, so to refine the quantitation of Ab efficacy in the long-term likely recognized residues in HVR1 that differed between OMP-1g treated BALB/c-scid mice used in this study, a quantitative real-time and OMP-1d. This pattern of epitope specificity was similar to that PCR assay for E. chaffeensis 16S rDNA was developed and used (for described previously for the Abs Ec56.5 and Ec18.1, so additional details, see Materials and Methods). Weekly treatment of mice with analyses of the remaining OMP Abs were performed using syn- Ec56.5 led to a 3500-fold reduction in bacterial colonization in the

Table III. Properties of effective and ineffective OMP HVR1-speciﬁc Abs

Recognition of OMP-1g Peptide MIWDc (g) a b Isotype 61–90 mAb Morbidity Mean (range) KD (M) Binding t1/2 (h) IgG2a ϩ Ec56.5 Ϫ 2.6 (1.7–4.7) 3.0 ϫ 10Ϫ11 128 ϩ Ec54.1 Ϫ 1.5 8.0 ϫ 10Ϫ12 683 ϩ Ec65.1 Ϫ 3.0 6.0 ϫ 10Ϫ11 85 Ϫ Ec53.7 ϩ 0.9 2.2 ϫ 10Ϫ8 17 Ϫ Ec57.3 ϩ 1.1 (1.1–1.2) 5.8 ϫ 10Ϫ8 13 Ϫ Ec14.1 ϩ 0.7 1.6 ϫ 10Ϫ6 19 IgG3 ϩ Ec18.1 ϩ 0.6 1.1 ϫ 10Ϫ8 0.35 Ϫ Ec16.1 ϩ 0.3 3.1 ϫ 10Ϫ8 12 Ϫ Ec17.9 ϩϪ0.3 4.2 ϫ 10Ϫ7 0.013 IgG2b ϩ Ec104.2 ϩ 1.2 1.2 ϫ 10Ϫ8 0.20 ϩ Ec105.4 ϩ 0.7 2.1 ϫ 10Ϫ8 0.20 Ϫ Ec13.1 ϩϪ0.6 ND ND Ϫ Ec26.1 ϩϪ0.1 1.2 ϫ 10Ϫ6 19 IgM ϩ Ec69.3 ϩϪ0.3 1.1 ϫ 10Ϫ7 0.050 ϩ Ec70.5 ϩϪ0.9 1.2 ϫ 10Ϫ7 0.083 Ϫ Ec12.7 ϩ 0.2 7.2 ϫ 10Ϫ7 0.017

a In vivo serum t1/2 of Ab in SCID mice were 3 days for IgG2a, 2 days for IgG3, and 1.2 days for IgG2b. b Indicates whether the treated mice exhibited morbidity at any time during the experiment. Morbidity was characterized by lack of mobility, hunched posture, and a pronounced loss of body weight (Ͼ30% loss of initial weight). c MIWD was determined as described in Materials and Methods. One to five repetitions were performed for each Ab. 1422 EFFECTIVE E. chaffeensis OMP Abs liver (Fig. 1), supporting our previous study. Bacterial infection was or only partially effective (e.g., Ec18.1, Ec104.2; Fig. 2, a and c). also decreased 100- to 400-fold in the spleen, peritoneal exudate, and Mice treated with the highly effective Abs remained healthy. Par- peripheral blood (Fig. 1). These data indicate that the Abs had pro- tially effective Abs delayed, but did not prevent, the onset of mor- found systemic effects on bacterial colonization in the animals that bidity (Fig. 2, Table III). had undergone periodic Ab treatment. To quantitate Ab efficacy data, and to compare data from mul- As an alternative to the PCR analyses, Ab efficacy was also tiple experiments, body weight changes in groups of Ab-treated vs evaluated in long-term studies by measurement of body weight untreated infected mice were normalized. This was performed by changes. Previous studies demonstrated significant weight loss integrating the body weight data, over time, up to the point that the (Ͼ30% loss of initial weight) preceded the onset of morbidity in untreated mice were deemed moribund. The difference in inte- infected, untreated SCID mice, and Ab treatment delayed or re- grated values between the Ab-treated and untreated control groups versed the weight loss and accompanying disease (5). Weight loss was then divided by the number of days to morbidity, to obtain the was generally observed beginning 7–10 days postinfection, prob- MIWD (in grams) for each group of Ab-treated mice (Table III). ably due to the time required to establish infection. Measurements Highly effective Abs, initially identified by their ability to prevent of body weight changes were performed in the present studies to morbidity in SCID mice, exhibited MIWD values greater than 1.5 g. evaluate Ab efficacy because this method provided a noninvasive Partially effective Abs, which delayed morbidity and prolonged sur- and readily obtainable assay of Ab efficacy. Body weight changes vival, exhibited MIWD values between 0.5 and 1.5 g. Abs that ex- observed in four independent and representative experiments are hibited MIWD values less than 0.5 g were considered ineffective. shown in Fig. 2. In each experiment, buffer was administered as a The data also revealed that among IgG Abs of identical subclass, negative control because previous studies demonstrated that irrel- higher efficacy was correlated with recognition of the OMP HVR1 Downloaded from evant isotype-matched Abs were ineffective. Ec56.5, which was peptide 61–90. Within this group, the IgG2a were highly effective, previously shown to be effective, was used as a positive control. followed by the partially effective IgG3 and IgG2b. None of the The data in Fig. 2 are expressed as the relative change in body IgM was effective. Differences in Ab efficacy among Abs of dif- weight from day 0, to control for differences in starting body ferent isotypes were not due to significant differences in serum t1/2, weight among the mice. In some experiments, uninfected SCID as measurements revealed similar t1/2 of 1.2–3 days for IgG2a, mice were also included in the analyses. IgG2b, and IgG3 in SCID mice (Table III). The serum t1/2 of the http://www.jimmunol.org/ Some Abs, such as Ec56.5 and Ec54.1 (both IgG2a), were Abs measured in this study in SCID mice were much shorter than highly effective in vivo, blocking or even reversing the weight loss those measured in normal mice (4–8 days for IgG2a, IgG2b, and and morbidity observed in untreated SCID mice (Fig. 2, a and b). IgG3) (10). It is known that the rate of IgG metabolism is inversely Other Abs were ineffective (e.g., Ec12.7, Ec13.1; Fig. 2, a and d), correlated to the serum concentration of IgG (11), so the shorter serum

t1/2 of IgG subclasses in SCID mice may be due to faster catabolism of Abs in mice lacking endogenous Ig. Despite the shorter t1/2 in vivo, serum Ab concentrations in SCID mice ranged, for Ec56.5 (IgG2a), from 33 ␮g/ml (212 nM) on day 1 to 9 ␮g/ml (57.7 nM) on day 7 post-Ab administration, so reasonably high concentrations of the ad- by guest on September 29, 2021 ministered Abs were most likely available during the weekly interval between Ab administrations (data not shown).

Effective Abs were of extraordinarily high affinity Although the above data suggested that both isotype and epitope were important correlates for Ab efficacy, it was also possible that affinity differences were critical. To evaluate the contribution of

Ab affinity, equilibrium-binding constants and binding t1/2 were determined by surface plasmon resonance, using highly purified rOMP-1g (Fig. 3). The results revealed striking affinity differences among the Ab panel. The most highly effective Abs, all IgG2a,

exhibited picomolar afﬁnity constants and binding t1/2 of 3–28 days (Table III). In contrast, Abs of lower or no efﬁcacy exhibited

affinities in the micromolar to nanomolar range, and binding t1/2 FIGURE 1. Reduction of bacterial load in long-term Ab-treated infected that ranged from minutes to hours. Therefore, high affinity was a BALB/c-scid mice. BALB/c-scid mice were infected via the peritoneum with critical characteristic of the highly effective Abs. 2 ϫ 106 E. chaffeensis-infected DH82 cells on day 0. The mice were administered PBS (F) or mAb Ec56.5 (E), beginning day 7 postinfection, and at weekly intervals thereafter. The mice were sacrificed on day 28 postinfection, Discussion at which time the PBS-injected mice were judged to be moribund. Samples of Effective Abs recognized OMP HVR1 liver, spleen, and peritoneal exudate from single mice, and pooled peripheral blood from each group of mice were analyzed using a real-time quantitative This study demonstrates that mice make dominant Ab responses to PCR assay for E. chaffeensis 16S rDNA. The copy number of E. chaffeensis E. chaffeensis OMPs, as suggested from previous studies of the 16S rDNA was enumerated in 5 ng total DNA extracted from each sample. polyclonal Ab responses in both mice and humans (2, 7). Forty These correspond to the following bacterial loads in each tissue from PBS- percent of the Abs recovered from three independent hybridoma 9 9 7 treated mice, 2.1 ϫ 10 (liver), 1.5 ϫ 10 (spleen), 2.6 ϫ 10 (peritoneal fusions recognized OMPs. Abs that recognized proteins other than ϫ 7 ϫ 5 exudate), and 8.3 10 (blood/ml); and Ec56.5-treated mice, 6.0 10 (liv- OMPs were also recovered, but for the most part were low affinity er), 8.7 ϫ 106 (spleen), 6.9 ϫ 104 (peritoneal exudate), and 1.1 ϫ 106 (blood/ ml). No PCR amplification product was detected in the control reactions con- IgM of undetermined specificity, and in preliminary experiments taining either no template or equivalent amounts of genomic DNA obtained were not effective in vivo. Therefore, OMP recognition plays a from uninfected animals. It was previously demonstrated that isotype-matched major role in Ab-mediated host defense during this intracellular irrelevant control Abs had no effect on bacterial load (5). bacterial infection. The Journal of Immunology 1423 Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021 FIGURE 2. In vivo efficacy of OMP-1g HVR1-specific mAbs. Infected BALB/c-scid mice were administered PBS, or the indicated mAbs, beginning 7 days postinfection, and at weekly intervals thereafter. The body weights of the mice were monitored, and the relative changes in body weight from day 0 are shown. In one experiment, uninfected mice were also included as controls. Each group contained three mice, and the average body weight change is shown. Each panel represents a separate experiment. Error bars indicate positive and negative standard deviation. †, Morbidity was observed, as characterized by lack of mobility, hunched posture, and a pronounced loss of body weight (Ͼ30% loss of initial weight).

Fine specificity analyses indicated that nearly all of the OMP three independent fusions, were characterized. It is also unlikely Abs studied recognized an identical or closely related epitope at that the hybridoma-screening technique selectively identified the amino terminus of OMP-1g HVR1. Ab recognition of the HVR1-specific Abs. Thus, although HVR1 is highly antigenic, and HVR1, a region that is highly polymorphic among E. chaffeensis is therefore likely to be exposed on the bacterial surface, polymor- OMPs, supports the notion that OMP genetic diversity and anti- phic HVR2 and HVR3 may not be targets for recognition by Abs. genic recognition are correlated (12). Our previous study demon- Thus, some genetic diversity might be required for functions un- strated that the highly effective Ab Ec56.5 required for recognition related to immune evasion (13). a glutamine at position 70 (5), which is substituted by a lysine residue in 10 of 21 putative OMPs of the Arkansas isolate (9). Ec56.5 did not recognize a peptide containing a glutamine to lysine Characteristics of highly effective Abs substitution at position 70 (J. S. Li and G. M. Winslow, unpub- Three parameters were evaluated to identify properties of the Abs lished data), indicating that some of the E. chaffeensis OMPs that were highly effective in vivo: affinity, epitope, and isotype. would not be predicted to be bound by this effective Ab. Genetic Although it is not known whether high affinity is essential, it was variation in OMPs is therefore a potential means to generate an- the most important correlate of efficacy. The three highly effective tigenic diversity, and may allow the bacteria to evade the humoral Abs exhibited picomolar affinities and binding t1/2 on the order of immune response in the natural host. The data presented in this days. Abs of no or low efficacy exhibited micromolar to nanomolar study suggest that OMP antigenic variation and subsequent im- affinities, and typically much shorter binding t1/2. Nevertheless, mune evasion did not occur in the mouse, however, because im- epitope recognition also appeared to influence Ab efficacy. Among munity was maintained in the long-term Ab-treated animals. Abs of identical isotype, higher efficacy was correlated with rec- Given that the OMP HVRs are highly polymorphic and that ognition of the OMP peptide 61–90. Abs that did not bind peptide HVR1 is antigenic, it was therefore surprising that none of the 61–90 presumably recognized a conformational determinant OMP Abs under study in this investigation recognized HVR2, and within HVR1. It is unclear why epitope recognition may influence at most one Ab recognized HVR3. This result was unlikely to be Ab efficacy in this model, but a possible explanation is that high due to insufficient sampling, because 40 OMP Abs, obtained from affinity Abs were only generated against the linear determinant. 1424 EFFECTIVE E. chaffeensis OMP Abs

Mechanisms of humoral immunity during intracellular infections It is not yet understood how Abs mediate bacterial clearance during E. chaffeensis infection. Several studies of other intracellular pathogens have provided evidence that Abs can protect against many important intracellular bacteria, fungi, and protozoa (reviewed in Ref. 17), including Mycobacterium tuberculosis (18), Listeria monocytogenes (19, 20), Salmonella typhimurium (21, 22), Brucella abortus (23), Legionella pneumophila (24), Crypto- coccus neoformans (25, 26), and Toxoplasma gondii (27). The mechanisms of humoral immunity during intracellular infection, where they are known, are highly pathogen dependent. Abs might affect the growth of some intracellular pathogens within the host cell, as has been observed during L. monocytogenes infection, in which Abs can act within infected macrophages to neutralize list- eriolysin O (20). Studies of Bartonella grahamii suggest that Abs may prevent the intercellular transfer and/or subsequent invasion of intracellular bacteria (28). Alternatively, uptake of Ab-opso- nized pathogens might induce an oxidative burst in phagocytes Downloaded from (29), or promote phagosome-lysosome fusion (30). Immune complexes of Abs and microbes or microbial products may also acti- vate macrophages, via Fc receptors, and this may result in the elimination of intracellular pathogens, such as C. neoformans (31, 32), L. monocytogenes (33), and Leishmania major (34), through the production of reactive oxygen or nitrogen intermediates. It is not known which, if any, of these mechanisms are relevant during http://www.jimmunol.org/ ehrlichial infection. It is also not clear why picomolar affinity was an apparently FIGURE 3. Surface plasmon resonance measurements. A sensogram critical characteristic of highly effective Abs during E. chaffeensis demonstrating the binding of a representative Ab (Ec56.5) to immobilized rOMP-1g is shown in a. Ab binding, in RU, is shown for a range of Ab infection. High affinity IgG2a have been shown to be highly neu- concentrations (62.5–1000 nM). The start of injection (I) and disassocia- tralizing during influenza infection (35), so it is possible that Abs tion (D) phases is indicated. A plot of equilibrium binding vs Ab concen- act to opsonize Ehrlichiae released from infected cells. Perhaps the tration is shown for four representative Abs in b. very long binding t1/2 that are correlated with high affinity are

critical for efficient opsonization and/or immune complex forma- by guest on September 29, 2021 tion. The data might also be comparable with studies of murine Ab Alternatively, fine specificity differences might modulate Ab ac- responses against several viruses, such as influenza (36) and Ebola tivities mediated by other host serum components, such as com- (37). In addition to being highly neutralizing for viral particles, plement (14). IgG2a is the most efficient isotype at fixing complement (38) and Although the highly effective Abs were IgG2a, the requirement for binding to Fc receptors on macrophages (39) and NK cells (40). for isotype remains unresolved, because Abs of picomolar affinity The similar effectiveness of humoral immunity in ehrlichial and were not recovered among other subclasses. The requirement for viral infections suggests that similar mechanisms may be involved. isotype can be best addressed using a family of IgG subclasses The data presented in this study provide further support for our sharing identical V regions, but a complete family of isotype- observations that Abs, in the absence of lymphocytes, can be switched variants was not recovered. However, an IgG2b isotype highly effective during this intracellular bacterial infection. Elici- switch variant of a partially effective IgG3 (Ec18.1) did not exhibit tation of effective Ab responses might therefore be an important increased efficacy (J. S. Li and G. M. Winslow, unpublished data), goal of prophylactic or therapeutic vaccine development for hu- supporting the data indicating that Abs of these isotypes are man monocytic ehrlichiosis, and perhaps related rickettsial equally effective. The observation that highly effective Abs were diseases. recovered as IgG2a suggests a relationship between affinity mat- uration and isotype switching, although additional studies would be required to resolve this possibility. The recovery of partially Acknowledgments effective IgG3 contrasts with data from studies of Ab efficacy dur- We thank Dr. David Woodland (Trudeau Institute, Saranac Lake, NY) for ing intracellular infection by Cryptococcus neoformans, in which critical reading of the manuscript, Pamela Scott Adams (Trudeau Institute) IgG3 were found to be ineffective (15). Thus, fundamentally dif- for assistance with the real-time PCR analyses, Ulrich Rudofsky (Wads- ferent mechanisms of humoral immunity may be involved in the worth Center) and Lorin Young (State University of New York) for assis- two intracellular infections. tance with BIAcore analysis, and Melissa Reilly for excellent technical The requirement that effective Abs exhibit high affinities might assistance. We also thank the Wadsworth Center Immunology Core Facil- ity, Animal Care Facility, and the Computational Molecular Biology and also offer one possible explanation for the apparent lack of a role Statistics Core Facility. for humoral immunity in many studies of host defense during other intracellular bacterial infections (16). Perhaps Abs used in adop- tive transfer experiments failed to provide protection against other References intracellular pathogens because Abs of appropriately high affinity 1. Dumler, J. S., and D. H. Walker. 2001. Tick-borne ehrlichioses. Lancet Infect. were not generated, or were present in insufficient quantities. Thus, Dis. 1:21. 2. Winslow, G. M., E. Yager, K. Shilo, E. Volk, and F. K. Chu. 2000. Antibody a possible involvement of Abs during other intracellular bacterial mediated elimination of the obligate intracellular bacterial pathogen Ehrlichia infections may require further evaluation. chaffeensis during active infection. Infect. Immun. 68:2187. The Journal of Immunology 1425

3. Feng, H.-M., and D. H. Walker. 1999. Mechanisms of immunity to Ehrlichia 22. Ornellas, E. P., R. J. Roantree, and J. P. Steward. 1970. The specificity and chaffeensis in mice. In Rickettsiae and Rickettsial Diseases at the Turn of the importance of humoral antibody in the protection of mice against intraperitoneal Third Millenium. D. Raoult and P. Brouqui, eds. Elsevier, Paris, p. 145. challenge with complement-sensitive and complement-resistant Salmonella. 4. Ganta, R. R., M. J. Wilkerson, C. Cheng, A. M. Rokey, and S. K. Chapes. 2002. J. Infect. Dis. 121:113. Persistent Ehrlichia chaffeensis infection occurs in the absence of functional ma- 23. Cloeckaert, A., I. Jacques, P. De Wergifosse, G. Dubray, and J. N. Limet. 1992. jor histocompatibility complex class II genes. Infect. Immun. 70:380. Protection against Brucella melitensis or Brucella abortus in mice with immu- 5. Li, J. S., E. Yager, M. Reilly, C. Freeman, G. R. Reddy, F. K. Chu, and noglobulin G (IgG), IgA, and IgM monoclonal antibodies specific for a common G. Winslow. 2001. Outer membrane protein specific monoclonal antibodies pro- epitope shared by the Brucella A and M smooth lipopolysaccharides. Infect. tect SCID mice from fatal infection by the obligate intracellular bacterial patho- Immun. 60:312. gen Ehrlichia chaffeensis. J. Immunol. 166:1855. 24. Brieland, J. K., L. A. Heath, G. B. Huffnagle, D. G. Remick, M. S. McClain, 6. Winslow, G., E. Yager, K. Shilo, D. N. Collins, and F. K. Chu. 1998. Infection M. C. Hurley, R. K. Kunkel, J. C. Fantone, and C. Engleberg. 1996. Humoral of the laboratory mouse with the intracellular pathogen Ehrlichia chaffeensis. immunity and regulation of intrapulmonary growth of Legionella pneumophila in Infect. Immun. 66:3892. the immunocompetent host. J. Immunol. 157:5002. 7. Chen, S.-M., L. C. Cullman, and D. H. Walker. 1997. Western immunoblotting 25. Mukherjee, J., M. D. Scharff, and A. Casadevall. 1992. Protective murine mono- analysis of the antibody responses of patients with human monocytotropic ehr- clonal antibodies to Cryptococcus neoformans. Infect. Immun. 60:4534. lichiosis to different strains of Ehrlichia chaffeensis and Ehrlichia canis. Clin. 26. Mukherjee, S., S. C. Lee, and A. Casadevall. 1995. Antibodies to Cryptococcus Infect. Dis. 4:731. neoformans glucuronoxylomannan enhance antifungal activity of murine macro- 8. Ohashi, N., Y. Rikihisa, and A. Unver. 2001. Analysis of transcriptionally active phages. Infect. Immun. 63:573. gene clusters of major outer membrane protein multigene family in Ehrlichia 27. Mineo, J. R., I. A. Khan, and L. H. Kasper. 1994. Toxoplasma gondii: a mono- canis and E. chaffeensis. Infect. Immun. 69:2083. clonal antibody that inhibits intracellular replication. Exp. Parasitol. 79:351. 9. Yu, X., J. W. McBride, X. Zhang, and D. H. Walker. 2000. Characterization of 28. Koesling, J., T. Aebischer, C. Falch, R. Schulein, and C. Dehio. 2001. Cutting the complete transcriptionally active Ehrlichia chaffeensis 28 kDa outer mem- edge: antibody-mediated cessation of hemotropic infection by the intraerythro- brane protein multigene family. Gene 248:59. cytic mouse pathogen Bartonella grahamii. J. Immunol. 167:11. 10. Vieira, P., and K. Rajewsky. 1988. The half-lives of serum immunoglobulins in 29. Canning, P. C., B. L. Deyoe, and J. A. Roth. 1988. Opsonin-dependent stimula- tion of bovine neutrophil oxidative metabolism by Brucella abortus. Am. J. Vet. adult mice. Eur. J. Immunol. 18:313. Downloaded from Res. 49:160. 11. Pollock, R. R., D. L. French, J. P. Metlay, B. K. Birshtein, and M. D. Scharff. 30. Malik, Z. A., G. M. Denning, and D. J. Kusner. 2000. Inhibition of Ca2ϩ sig- 1990. Intravascular metabolism of normal and mutant mouse immunoglobulin naling by Mycobacterium tuberculosis is associated with reduced phagosome- molecules. Eur. J. Immunol. 20:2021. lysosome fusion and increased survival within human macrophages. J. Exp. Med. 12. Yu, X. J., J. McBride, and D. H. Walker. 1999. Genetic diversity of the 28- 191:287. kilodalton outer membrane protein gene in human isolates of Ehrlichia chaffeen- 31. Chaturvedi, V., B. Wong, and S. L. Newman. 1996. Oxidative killing of Cryp- sis. J. Clin. Microbiol. 37:1137. tococcus neoformans by human neutrophils: evidence that fungal mannitol pro- 13. Allsopp, M. T., C. M. Dorfling, J. C. Maillard, A. Bensaid, D. T. Haydon, tects by scavenging reactive oxygen intermediates. J. Immunol. 156:3836. H. van Heerden, and B. A. Allsopp. 2001. Ehrlichia ruminantium major antigenic 32. Mozaffarian, N., J. W. Berman, and A. Casadevall. 1995. Immune complexes http://www.jimmunol.org/ protein gene (map1) variants are not geographically constrained and show no increase nitric oxide production by interferon-␥-stimulated murine macrophage- evidence of having evolved under positive selection pressure. J. Clin. Microbiol. like J774.16 cells. J. Leukocyte Biol. 57:657. 39:4200. 33. Boockvar, K. S., D. L. Granger, R. M. Poston, M. Maybodi, M. K. Washington, 14. Feng, J. Q., K. Mozdzanowska, and W. Gerhard. 2002. Complement component J. B. Hibbs, Jr., and R. L. Kurlander. 1994. Nitric oxide produced during murine c1q enhances the biological activity of influenza virus hemagglutinin-specific listeriosis is protective. Infect. Immun. 62:1089. antibodies depending on their fine antigen specificity and heavy-chain isotype. 34. Green, S. J., M. S. Meltzer, J. B. Hibbs, Jr., and C. A. Nacy. 1990. Activated J. Virol. 76:1369. macrophages destroy intracellular Leishmania major amastigotes by an L-argi- 15. Yuan, R., A. Casadevall, G. Spira, and M. D. Scharff. 1995. Isotype switching nine-dependent killing mechanism. J. Immunol. 144:278. from IgG3 to IgG1 converts a nonprotective murine antibody to Cryptococcus 35. Huber, V. C., J. M. Lynch, D. J. Bucher, J. Le, and D. W. Metzger. 2001. Fc neoformans to a protective antibody. J. Immunol. 154:1810. receptor-mediated phagocytosis makes a significant contribution to clearance of 16. Schaible, U. E., H. L. Collins, and S. H. E. Kaufmann. 1999. Confrontation influenza virus infections. J. Immunol. 166:7381.

between intracellular bacteria and the immune system. Adv. Immunol. 71:267. 36. Gerhard, W., K. Mozdzanowska, M. Furchner, G. Washko, and K. Maiese. 1997. by guest on September 29, 2021 17. Casadevall, A. 1998. Antibody-mediated protection against intracellular patho- Role of the B-cell response in recovery of mice from primary influenza virus gens. Trends Microbiol. 6:102. infection. Immunol. Rev. 159:95. 18. Teitelbaum, R., A. Glatman-Freedman, B. Chen, J. B. Robbins, E. Unanue, 37. Wilson, J. A., M. Hevey, R. Bakken, S. Guest, M. Bray, A. L. Schmaljohn, and A. Casadevall, and B. R. Bloom. 1998. A mAb recognizing a surface antigen of M. K. Hart. 2000. Epitopes involved in antibody-mediated protection from Ebola Mycobacterium tuberculosis enhances host survival. Proc. Natl. Acad. Sci. USA virus. Science 287:1664. 95:15688. 38. Neuberger, M. S., and K. Rajewsky. 1981. Activation of mouse complement by 19. Edelson, B. T., P. Cossart, and E. R. Unanue. 1999. Paradigm revisited: antibody monoclonal antibodies. Eur. J. Immunol. 11:1012. provides resistance to Listeria infection. J. Immunol. 163:4087. 39. Heusser, C. H., C. L. Anderson, and H. M. Grey. 1977. Receptors for IgG: 20. Edelson, B. T., and E. R. Unanue. 2001. Intracellular antibody neutralizes Lis- subclass specificity of receptors on different mouse cell types and the definition teria growth. Immunity 14:503. of two distinct receptors on a macrophage cell line. J. Exp. Med. 145:1316. 21. Eisenstein, T. K., L. M. Killar, and B. M. Sultzer. 1984. Immunity to infection 40. Kipps, T. J., P. Parham, J. Punt, and L. A. Herzenberg. 1985. Importance of with Salmonella typhimurium: mouse-strain differences in vaccine- and serum- immunoglobulin isotype in human antibody-dependent, cell-mediated cytotoxic- mediated protection. J. Infect. Dis. 150:425. ity directed by murine monoclonal antibodies. J. Exp. Med. 161:1.