Msp2 variation in Anaplasma phagocytophilum in vivo does not stimulate T cell immune responses or interferon-γ production

Authors

  • Kyoung-Seong Choi,

    1. Division of Medical Microbiology, Department of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD, USA
    2. College of Life Science and Natural Resources, Department of Animal Science, Sangju National University, Sangju, Korea
    Search for more papers by this author
  • Diana G. Scorpio,

    1. Department of Comparative and Molecular Pathobiology, The Johns Hopkins University School of Medicine, Baltimore, MD, USA
    Search for more papers by this author
  • Nicole C. Barat,

    1. Division of Medical Microbiology, Department of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD, USA
    Search for more papers by this author
  • J. Stephen Dumler

    1. Division of Medical Microbiology, Department of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD, USA
    Search for more papers by this author

  • Editor: Patrik Bavoil

Correspondence: J. Stephen Dumler, Division of Medical Microbiology, Department of Pathology, The Johns Hopkins University School of Medicine, 720 Rutland Avenue, Ross 624, Baltimore, MD 21205, USA. Tel.: +410 955 8654; fax: +443 287 3665; e-mail: sdumler@jhmi.edu

Abstract

Msp2 is Anaplasma phagocytophilum's immunodominant protein. Antigenic variability with msp2 gene conversion may drive differential immunopathology with infection by bacteria of different in vitro passage intervals. We examined msp2 transcript variation and its relationship to histopathology, T-cell and antibody responses in mice infected with differentially passaged A. phagocytophilum. Hepatic inflammation peaked on day 2–4 with low passage bacteria and on day 4–7 with high passage bacteria infection. Nineteen msp2 variant transcripts were identified. The low and high passage inocula shared four, but differed in one and two msp2 transcript variants, respectively. After infection, three and two msp2 variants were only identified in low or high passage infected mice. However, per mouse, msp2 variant profiles were unique with no evident expression program. In low and high passage bacteria-infected mice, splenocytes proliferated to whole A. phagocytophilum at day 7–10, diminishing thereafter. Weak mitogenic responses to whole bacteria were detected in mock and infected mice at d0 and sporadically thereafter. Essentially no lymphoproliferation or IFN-γ production resulted from stimulation by six Msp2 hypervariable region proteins, although antibodies were detected to all, including cross-reactions. Differential A. phagocytophilum Msp2 expression is unrelated to T-cell response and unlikely to induce the cellular immunopathology underlying disease manifestations.

Introduction

Human granulocytic anaplasmosis (HGA) is a tick-borne disease caused by Anaplasma phagocytophilum. Anaplasma phagocytophilum strains from different geographic regions vary in their antigenic contents (Dumler et al., 1995; Asanovich et al., 1997). The basis of antigenic variability is thought to relate to differential expression of immunodominant surface proteins encoded by a multigene family (Msp2, p44, 44 kDa) that are characterized by conserved sequences flanking a hypervariable region (HVR) predicted to contain T and B cell epitopes (Caspersen et al., 2002; Scorpio et al., 2004). Msp2 s of A. phagocytophilum are phylogenetically most similar to major surface protein-2 (MSP2) of Anaplasma marginale (Palmer et al., 1994; Zhi et al., 1997, 1999; IJdo et al., 1998; Barbet et al., 2001; Brayton et al., 2002; Caspersen et al., 2002); A. marginale MSP2 antigenic variation leads to persistent, relapsing infection in ruminants. However, Msp2 may have a role as and adhesin for A. phagocytophilum binding to mammalian cells (Park et al., 2003), and its differential expression in mammals and ticks support a nonimmunological role in infection pathogenesis (IJdo et al., 2002; Zhi et al., 2002).

Anaplasma phagocytophilum Msp2 transcription and expression changes with increasing lengths of in vitro propagation (Scorpio et al., 2004). Horses experimentally infected by A. phagocytophilum passaged in vitro develop clinical manifestations either typical of natural virulent disease when infected with low passage bacteria, or with significantly diminished clinical signs when infected with high passage bacteria (Pusterla et al., 2000). The murine model of HGA is imperfect because infection does not cause any clinical signs in immune competent mice, yet the development of histopathologic lesions, such as inflammatory liver lesions, that mimic those in infected humans and horses indicates that this model is an important tool for investigating A. phagocytophilum pathogenesis (Bunnell et al., 1999; Lepidi et al., 2000; Borjesson & Barthold, 2002). Moreover, because histopathologic lesions in mice are directly linked to the production of interferon (IFN)-γ (Martin et al., 2001), and because Msp2 is the major immunological target of the host immune response (Asanovich et al., 1997; IJdo et al., 1997), a link between expression of specific Msp2 s, immunologic stimulation, and disease manifestations or histopathologic lesions (in the murine model) were hypothesized. In this study, T cell stimulation was assayed by specific A. phagocytophilum Webster strain Msp2 hypervariable domain proteins in mice inoculated with low and high passage bacteria. This was performed to determine whether responses in the mouse model of HGA to varying Msp2 T cell epitopes occur, with the concept that if present, such T cell responses could account for IFN-γ production, immunopathologic responses, and ultimately clinical manifestations or pathogen virulence.

Materials and methods

Experimental animals

Female C57BL/6 mice (6-week-old) were purchased from The Jackson Laboratory (Bar Harbor, Maine). All animals were maintained and used in strict accordance with the guidelines issued by the Johns Hopkins University School of Medicine Animal Care and Use Committee.

Anaplasma phagocytophilum culture and infection of mice

Anaplasma phagocytophilum was maintained in RPMI 1640 medium supplemented with 5% FBS and 2 mM l-glutamine until 100% of the cells contained morulae. Cultures were coordinated so that low (passage 7–8) and high (passage 17–22) passage materials were available for inoculation. On the day of inoculation, p8 and p22 infected and uninfected HL-60 cells were centrifuged (200 g, 10 min) to concentrate the cells, and then the cell pellet was resuspended in serum-free RPMI 1640 medium. Groups of 28 mice corresponding to the two infected and the mock-infected HL-60 cell groups (total 84 mice) were injected intraperitoneally with 1 mL (1 × 106 mL−1) of heavily infected (>90%) cells. Mice were evaluated for clinical signs such as ruffled fur, reduced activity, failure to eat or drink, a hunched posture and lack of interaction with cage mates.

Histopathologic analyses

Four mice of each group (low passage, high passage and mock-infected) were necropsied at seven time points; day 0 (1 h after inoculation), 2, 4, 7, 10, 14 and 21. The liver was harvested from each mouse and fixed in zinc fixation solution (BD Pharmingen, San Diego, CA) for paraffin embedding and H&E staining. Parametric statistics (means and Student's t-tests) were not used owing to the uncertainty regarding whether histopathologic severity is normally distributed in these mice. Thus, hepatic histopathologic changes were assessed for severity using robust nonparametric tests after ranking individual mouse livers with regard to inflammation severity based on the size, number and density of inflammatory infiltrates/focal lesions, and the degree of necrosis and/or apoptosis; this results in the assignment of rank numbers for each mouse, ranging from the least severe (rank number 1) to most severe (rank number 84). Under these circumstances, groups with greater histopathologic severity would have higher median ranks, and variability in severity could be approximated by maximum and minimum ranks for each group. All evaluations were performed with investigators blinded to the identity of mouse treatment status and with consensus of two microscopists; in general, independent review did not result in changes in rank by more than one or two positions. To generate daily medians, the histopathologic severity ranks of mice from each interval were reranked and compared using nonparametric Mann–Whitney tests; P values <0.05 were considered significant.

Preparation of whole and component A. phagocytophilum antigens

Low and high passage A. phagocytophilum Webster strain-infected HL-60 cells were used. Heavily infected cells (>95%, infection rate) were centrifuged (200 g, 15 min) and the cell pellet was washed with phosphate-buffered saline (PBS). Cell-free bacteria were prepared by sonication lysis (Branson Sonifier 250, Danbury, CT; output control=2 for 30 s) of heavily infected cells that were then centrifuged (12 000 g for 20 min) and washed twice with PBS. The final pellet was resuspended in sterile PBS and protein content was measured by the bicinchoninic assay method (Pierce Biotechnology, Inc., Rockford, IL). The low and high passage bacteria purified in this way were used as a stimulant for the lymphoproliferation and cytokine assays.

Cloning, expression and purification of six recombinant Msp2 hypervariable region proteins

The construction of recombinant plasmids was performed using the Invitrogen Gateway Technology System (Invitrogen, Carlsbad, CA). Primers were designed containing specific attB sequences to amplify only the hypervariable region of 22 distinct msp2 genes previously found expressed by A. phagocytophilum Webster strain, WMSP2 (GenBank accession no. AF443405), WMSP10 (AF443400), WMSP17 (AF443396), WMSP19 (AF443404), RW1 (AY253530) and RWMSP11 (AF443403) (Caspersen et al., 2002; Scorpio et al., 2004) (predicted amino acid sequences are shown in Fig. 1). After amplification and purification, the amplicons were cloned into pDONR vector (BP reaction) and transformed into Escherichia coli DH5α, as per the Gateway Technology system instructions. The recombinant plasmid was then recovered and the insert transferred into the pDEST17 vector by the LR reaction, all according to the manufacturer's instructions. The pDEST17 vector allows in-frame cloning to obtain an msp2 HVR N-terminal 6X his tag fusion that is amplified by transformation into E. coli DH5α, and then into E. coli BL21-AI (Invitrogen). After overnight growth, recombinant protein was expressed following 4 h incubation in LB medium supplemented with l-arabinose (Sigma, St Louis, MO) to induce protein expression. After the bacteria were harvested, cells were lysed in binding buffer (20 mM Tris-HCl, 500 mM NaCl, 5 mM imidazole, 1 mM β-mercaptoethanol, 6 M guanidine HCl) with lysozyme (200 μg mL−1), and the recombinant fusion proteins were purified from the supernatants by column chromatography using HIS-trap nickel columns on an AKTA Prime liquid chromatography system (GE Healthcare, Piscataway NJ). Renaturation to attempt proper refolding was performed by replacing guanidine with urea and then slowly reducing the urea concentration in-column. After pilot studies, the recombinant proteins were collected based upon absorbance at 280 nm, concentrated, desalted and evaluated for purity by observing a single band with Coomassie blue and/or protein immunoblot staining. All recombinant proteins were shown free of endotoxin using the E-TOXATE™ Kit Limulus amoebocyte lysate assay (Sigma). Six clones from which high quality pure recombinant protein was prepared were used for further studies.

Figure 1.

 Alignment of predicted amino acid sequences for the Msp2 hypervariable region recombinant proteins used for in vitro lymphoproliferation and antibody studies. (*), Fully conserved amino acid residue; (:), strongly conserved residue; (.), weakly conserved residue.

In vitro splenocyte proliferation assay and IFN-γ assays

Splenocytes (2 × 105 cells mL−1) obtained at necropsy from mice in each experimental group on days 0, 7, 10, 14 and 21 were resuspended in RPMI 1640 containing 10% FBS and 1 × penicillin/streptomycin, and seeded in duplicate into 96-well culture plates. After 24 h of incubation, duplicate wells of a 96-well plate containing the cell suspension from each spleen were stimulated by adding to individual wells the following: sterile medium, anti-CD3ɛ (4 and 1 μg mL−1; BD Pharmingen, San Diego, CA), ConA (5 and 1 μg mL−1; Sigma), purified whole A. phagocytophilum at p7 and p17 (10 and 1 μg mL−1), HL-60 cell lysate (10 μg mL−1), and three different concentrations (50, 5 and 1 μg mL−1) of the six recombinant Msp2 hypervariable region proteins (WMSP2, WMSP10, WMSP17, WMSP19, RW1 and RWMSP11) previously found to be expressed among low and/or high passage A. phagocytophilum (Caspersen et al., 2002; Scorpio et al., 2004). After 72 h incubation, the supernatant was harvested from each well and frozen at −20°C until analyzed. The remaining cells were used in cell proliferation assays (BrdU Cell Proliferation Assay kit; Calbiochem, San Diego, CA) as per the manufacturer's recommendations. BrdU incorporation into cells was measured at dual wavelengths of 450 and 540 nm. Owing to the lower sensitivity and signal-to-noise ratio of this nonradiometric assay as compared with 3H-thymidine incorporation (Wemme et al., 1992), OD for each splenocyte preparation was compared to that obtained when cells were incubated with medium only (for recombinant proteins), or with HL-60 cell lysate (for whole A. phagocytophilum antigens). To ascertain significant lymphoproliferative activity, duplicate wells were averaged, and the mean results for mice of the same group at each time interval were compared by one-sided paired Student's t-tests; P<0.05 was considered significant proliferation. For IFN-γ assays, supernatants were tested in duplicate using the Mouse Cytokine/Chemokine LINCOplex Kit (Linco Research, Inc., St Charles, MO), or using a murine IFN-γ ELISA kit (R&D Systems Inc., Minneapolis, MN), both as per manufacturers' instructions.

Antibodies to whole A. phagocytophilum and recombinant hypervariable region proteins

ELISA was used to detect the production of antibodies reactive with HVR epitopes on WMSP2, WMSP10, WMSP17, WMSP19, RW1 and RWMSP11. The six individual purified recombinant Msp2 HVR proteins were diluted in 0.1 M sodium carbonate buffer (pH 9.6) and coated on to 96-well plates overnight at 4°C. The following day, plates were washed three times with PBS/Tween 20 (PBST; Sigma) and blocked with 1% bovine serum albumin (BSA, Sigma) in PBS for 1 h at 37°C. After washing the plates, the day 0 baseline and day 21 plasma samples from mice infected with low and high passage A. phagocytophilum were diluted 1:100 in PBST and incubated for 1 h at 37°C. The plates were washed three times with PBST, and after adding alkaline phosphatase-conjugated antimouse IgG (Kirkegaard & Perry Laboratories; KPL, Gaithersburg, MD), they were incubated for 1 h at 37°C. After thorough plate washing, BluePhos Microwell Phosphatase Substrate (KPL) solution was added to each plate and all were read at a wavelength between 620 and 650 nm. All reactions were adjusted for background in wells that received buffer instead of plasma. Antibody production was determined significant when OD of reactions with plasma from infected mice exceeded the cutoff value of the mean OD +3 SD of reactions with plasma from uninfected mice.

Plasma was assayed for antibodies to whole A. phagocytophilum among infected mice by indirect immunofluorescent assay (IFA) as described (Bunnell et al., 1999; Martin et al., 2001) using A. phagocytophilum Webster strain-infected HL-60 cells as antigen. Plasma samples were screened at a 1:80 dilution and titrated to endpoint if reactive.

Preparation of polyclonal antibodies to six Msp2 HVR recombinant proteins

To prepare antibodies reactive with each Msp2 HVR protein, two mice were immunized with each recombinant protein according to the mouse protocol recommended by the manufacturers of the Ribi Adjuvant System (Sigma). Mice were test bled and exsanguinated when antibodies were detectable by ELISA, as above. These reagents were prepared to assess specificity of antibodies toward hypervariable region epitopes of Msp2.

RT-PCR for A. phagocytophilum msp2 transcripts

Total cellular RNA was isolated using the QIAamp RNA Blood mini kit (QIAGEN Inc., Valencia, CA) from the blood of low and high passage A. phagocytophilum-infected and mock-infected (uninfected HL-60 cell inoculated) mice on days 2, 7, 14 and 21 postinfection. RT-PCR was performed with Supersript One-Step RT-PCR and Platinum Taq (Invitrogen) according to the instructions of the manufacturer. Broad range primers for amplification of all msp2 cDNAs were created from the alignment of 10 A. phagocytophilum msp2 sequences in GenBank, and used as previously described (Caspersen et al., 2002; Scorpio et al., 2004). The reactants were subjected to electrophoresis in 1% agarose gels and visualized by ethidium bromide staining.

Cloning and sequencing of RT-PCR products

RT-PCR products were cloned into the pGEM-T Easy Vector (Promega, Madison, WI) followed by transformation into E. coli DH5α, and then plated onto LB agar containing ampicillin 50 μg mL−1. Plasmids were purified (Wizard SV 96 Plasmid DNA Purification System, Promega) and assessed for insert size after EcoRI digestion, and when the appropriate size insert was detected (∼550 bp), 10–20 from each cloning reaction were submitted for sequencing in one direction only with the goal of comparing at least 10 msp2 transcripts from each mouse at each interval. Sequences were then edited to remove vector and incorporated primers using BioEdit (Hall, 1999); ambiguous nucleotides were resolved by inspecting the chromatograms or by repeating sequencing reactions. After editing, all sequences were aligned with A. phagocytophilum Webster strain msp2 reference sequences (Page, 1996; Scorpio et al., 2004) using clustalx (Thompson et al., 1997). The numbers of clones that aligned with each reference or into a new clade were identified to document the diversity of msp2 transcription that emerged as infection progressed in vivo with reference to low or high passage bacteria. The diversity of msp2 transcripts was then stratified by low or high passage inoculation and by day of infection.

Results

Histopathologic severity of A. phagocytophilum infection

In accordance with all prior studies of A. phagocytophilum infection in mice, no clinical signs of infection were observed at any time interval. However, as previously demonstrated, histopathologic lesions comprising few to numerous focal accumulations of lymphocytes, macrophages and sometimes neutrophils, with and without apoptotic cells, were uniformly detected in infected mice in hepatic lobules and periportal regions. Histopathologic severity peaked earlier (day 2–4) in low passage than in high passage A. phagocytophilum-infected mice, and resolved by day 7, vs. a peak at days 4–7 for high passage A. phagocytophilum infection that resolved by day 10 (Fig. 2). The differences between low and high passage infection were significant only on day 7 (P=0.014).

Figure 2.

 Hepatic histopathologic severity among mice infected by low or high passage A. phagocytophilum, or mock-infected by inoculation with HL-60 cell lysate. Results are shown as median ranking for each group at each time interval; error bars represent the maximum and minimum rank for each group at each time point; asterisks designate a significant difference between the median of the infected vs. mock-infected groups after reranking at the same time interval.

Anaplasma phagocytophilum msp2 transcript diversity in vivo

Anaplasma phagocytophilum msp2 transcripts were detected in mice inoculated with low passage A. phagocytophilum-infected HL-60 cells at every time point after day 0, except day 21, but only up to day 10 in high passage A. phagocytophilum-infected mice (Fig. 3). Several attempts were made to obtain 10 sequenced clones per mouse per time interval, although this was not achieved for several intervals (Fig. 3), probably owing to low levels of bacteremia after day 7. All mock-infected mice were negative by RT-PCR and also remained without clinical signs.

Figure 3.

 RT-PCR sequencing analysis to determine changes in transcription of msp2 by hypervariable region by in vitro passage, by mouse and by day after inoculation. Anaplasma phagocytophilum Webster strain msp2 loci are indicated along the top; the total number of cloned transcripts examined for each mouse per time interval is shown in the far right-hand column. Results are indicated in relative numbers of transcripts identified among cloned RT-PCR products: clear boxes indicate no transcripts detected, gray boxes indicate 1–2 transcripts detected, and black boxes indicate three or more transcripts detected. The rows corresponding to day 0 represent the relative transcript content in the inoculum.

Although the HZ strain of A. phagocytophilum contains at least 100 msp2 paralogs (Dunning Hotopp et al., 2006), and at least 37 distinct msp2 paralogs have been identified in the Webster strain (Scorpio et al., 2004), we detected a total of only 19 distinct msp2 variant transcripts from all animals over the 21 days of these experiments (Fig. 3). When compared with the previous A. phagocytophilum Webster strain msp2 paralogs (Caspersen et al., 2002; Scorpio et al., 2004), five new, unique hypervariable region transcripts were identified among infected mice. These included three found only among mice infected with low passage, one found only among mice infected with high passage, and one msp2 variant found in both mice infected with low or high passage bacteria. Individual mice contained from as few as one to as many as seven msp2 variant transcripts, illustrating the diversity of infection among all mice (Fig. 3). The low passage A. phagocytophilum inoculum expressed five distinct msp2 transcript variants. After infection of mice by low passage A. phagocytophilum, four variant msp2 transcripts continued to be expressed in vivo through day 7, and 12 msp2 transcript variants not present in the inoculum were detected in vivo from days 2 to 14. New msp2 variant transcripts were detected on each of days 2, 7 and 14. Similar results were observed in the high passage inoculum and thereafter in infected mice, except that infection was not detected after day 7 (Fig. 3). During this interval, all six msp2 variant transcripts originally expressed in the high passage inoculum and nine new msp2 variant transcripts were detected in vivo. Other than the rapid clearance of bacteria by days 14–21, no reproducible pattern of msp2 transcription was noted in mice infected by low or high passage bacteria, or early or late after infection.

T cell lymphocyte and antibody responses to A. phagocytophilum and recombinant Msp2 hypervariable region proteins

For in vitro T cell assays, we utilized recombinant proteins derived from six Msp2 HVRs previously found to be transcribed among low and high passage A. phagocytophilum Webster strain (Caspersen et al., 2002; Scorpio et al., 2004). One of these was encoded by the predominant transcript previously detected in low passage A. phagocytophilum (Scorpio et al., 2004), and was also highly transcribed in the low passage bacterial inoculum used here. Of the other Msp2 HVR proteins used, three were also highly transcribed in mice infected with either low or high passage A. phagocytophilum suggesting that they too would be appropriate antigens to assay Msp2 HVR-specific T cell responses.

Lymphoproliferative activity after exposure to mitogens

Overall, splenocytes from all animals proliferated when exposed to ConA or anti-CD3 at every interval examined, although there was considerable variation in response to ConA or anti-CD3 dose, within groups, and over the time course of the experiments (Fig. 4a). Considerably more lymphoproliferation to ConA and anti-CD3 was detected among splenocytes from infected animals on day 7 (P=0.017), and this was greater for animals infected with low passage A. phagocytophilum than for high passage organisms (P=0.04), suggesting that splenocytes from infected animals were already activated or perhaps exposed to an additional mitogenic factor, and that this factor may be differentially present in low passage bacteria. The difference in proliferation to ConA between infected and mock-infected animals persisted into day 10, and resolved by day 14.

Figure 4.

Figure 4.

 Lymphoproliferative assays. Representative proliferation profiles of splenocytes from low (Ap-lp) and high passage (Ap-hp) A. phagocytophilum-infected and mock-infected (HL-60 cells only) mice at various intervals postinoculation after in vitro (a) mitogen stimulation (ConA, anti-CD3) or medium control, (b) whole cell high (p17) or low (p7) passage A. phagocytophilum antigens, and (c) representative Msp2 HVR recombinant proteins. The asterisk (*) represents significant (P<0.05) differences from either medium controls (for mitogens and recombinant proteins) or from HL-60 cell lysates (for whole A. phagocytophilum antigens); error bars represent SE of the mean for each group of three to four mice. Note proliferative activity in (a) ConA- and anti-CD3 stimulated cells and among cells at all intervals and among cells from infected animals stimulated with (b) A. phagocytophilum whole cell antigens, especially on days 7–10. Weak proliferation is also demonstrated among cells from infected and uninfected mice on day 0 and among uninfected mice thereafter. Note the lack of any proliferation after stimulation with HL-60 cell lysates alone. Proliferation after stimulation by recombinant Msp2 hypervariable region proteins (c) was rarely present (compare with medium only control in 4a). Full results for stimulation by three of the six recombinant proteins at three different stimulation concentrations are shown; the remaining results were similar except no proliferation was identified.

Figure 4.

Figure 4.

 Lymphoproliferative assays. Representative proliferation profiles of splenocytes from low (Ap-lp) and high passage (Ap-hp) A. phagocytophilum-infected and mock-infected (HL-60 cells only) mice at various intervals postinoculation after in vitro (a) mitogen stimulation (ConA, anti-CD3) or medium control, (b) whole cell high (p17) or low (p7) passage A. phagocytophilum antigens, and (c) representative Msp2 HVR recombinant proteins. The asterisk (*) represents significant (P<0.05) differences from either medium controls (for mitogens and recombinant proteins) or from HL-60 cell lysates (for whole A. phagocytophilum antigens); error bars represent SE of the mean for each group of three to four mice. Note proliferative activity in (a) ConA- and anti-CD3 stimulated cells and among cells at all intervals and among cells from infected animals stimulated with (b) A. phagocytophilum whole cell antigens, especially on days 7–10. Weak proliferation is also demonstrated among cells from infected and uninfected mice on day 0 and among uninfected mice thereafter. Note the lack of any proliferation after stimulation with HL-60 cell lysates alone. Proliferation after stimulation by recombinant Msp2 hypervariable region proteins (c) was rarely present (compare with medium only control in 4a). Full results for stimulation by three of the six recombinant proteins at three different stimulation concentrations are shown; the remaining results were similar except no proliferation was identified.

Figure 4.

Figure 4.

 Lymphoproliferative assays. Representative proliferation profiles of splenocytes from low (Ap-lp) and high passage (Ap-hp) A. phagocytophilum-infected and mock-infected (HL-60 cells only) mice at various intervals postinoculation after in vitro (a) mitogen stimulation (ConA, anti-CD3) or medium control, (b) whole cell high (p17) or low (p7) passage A. phagocytophilum antigens, and (c) representative Msp2 HVR recombinant proteins. The asterisk (*) represents significant (P<0.05) differences from either medium controls (for mitogens and recombinant proteins) or from HL-60 cell lysates (for whole A. phagocytophilum antigens); error bars represent SE of the mean for each group of three to four mice. Note proliferative activity in (a) ConA- and anti-CD3 stimulated cells and among cells at all intervals and among cells from infected animals stimulated with (b) A. phagocytophilum whole cell antigens, especially on days 7–10. Weak proliferation is also demonstrated among cells from infected and uninfected mice on day 0 and among uninfected mice thereafter. Note the lack of any proliferation after stimulation with HL-60 cell lysates alone. Proliferation after stimulation by recombinant Msp2 hypervariable region proteins (c) was rarely present (compare with medium only control in 4a). Full results for stimulation by three of the six recombinant proteins at three different stimulation concentrations are shown; the remaining results were similar except no proliferation was identified.

Lymphoproliferative activity when exposed to whole A. phagocytophilum antigens

When stimulated by purified whole low passage A. phagocytophilum, weak lymphoproliferative activity of splenocytes (P=0.006) was observed among cells from some infected and uninfected animals on day 0, indicating endogenous mitogenic activity for the whole A. phagocytophilum antigen preparations (Fig. 4b). In contrast, no lymphoproliferation to HL-60 cell lysates was observed at any time, whether among splenocytes from infected or mock-infected animals. Lymphoproliferation after stimulation by whole A. phagocytophilum low or high passage antigens peaked on day 7 among splenocytes from both infected and mock-infected animals. The proliferative activity on day 7 did not differ between animals infected by low or high passage A. phagocytophilum (P>0.149), but was three to five-fold greater among infected than mock-infected animals in most comparisons (P=0.046 to 0.003). Lymphoproliferative activity generally diminished after day 7, but was still two to three times higher in splenocytes from both high and low passage A. phagocytophilum-infected than mock-infected animals on day 10, was still present at day 14, and resolved to levels observed with HL-60 cell lysate stimulation by day 21.

Lymphoproliferative activity when exposed to recombinant Msp2 HVR proteins

Lymphoproliferation among splenocytes stimulated by the six recombinant HVR proteins was almost uniformly absent (Fig. 4c), with very weak activity being detected only on days 0 and 14 among splenocytes from high passage A. phagocytophilum-infected mice stimulated with 1 μg mL−1 of WMSP19 Msp2 HVR protein or 50 μg mL−1 of RWMSP11 Msp2 HVR protein, respectively. This is remarkable given that splenocytes from the mock-infected and high or low passage-infected animals were stimulated in duplicate with six distinct Msp2 HVR proteins in three concentrations covering almost three magnitudes at five different times over 21 days.

In vitro cytokine and chemokine production

Owing to the low levels of lymphoproliferative activity, only limited splenocyte cultures were studied for in vitro production of IFN-γ and other cytokines/chemokines after stimulation by A. phagocytophilum antigens. Consistent with the lymphoproliferative data, low levels of IFN-γ were detected in splenocyte cultures from mice infected with either low or high passage A. phagocytophilum maximally on day 7, ranging up to threefold higher than in the HL-60 cell-stimulated controls, but not exceeding 16.1 pg mL−1. The expression of MIP-1α (Ccl3; 3090 pg mL−1; 14-fold increase) and KC (Cxcl1; 368 pg mL−1; 15-fold increase) in the splenocytes demonstrated their ability to respond to the A. phagocytophilum antigenic stimuli. Similarly, IFN-γ was not detected in splenocyte supernatants of 16 mice infected by high passage A. phagocytophilum. However, it was detected at low levels (35.7 pg mL−1; 2.2-fold higher than control) in one of 16 mice infected by low passage A. phagocytophilum, and only on day 7 when stimulated with a single recombinant HVR protein (WMSP19).

Detection of A. phagocytophilum and rMsp2 HVR antibodies

All infected mice seroconverted when tested using whole A. phagocytophilum antigens. Moreover, the majority of infected, but not mock-infected mice developed serologic reactions to each of the six rHVR proteins (Fig. 5). Antibody responses to rHVR proteins were more often detected among mice infected with low passage A. phagocytophilum. Antibody responses to rHVR proteins RW1 and RWMSP11 were detected, although neither of the transcripts corresponding to these proteins was identified in mice during the in vivo infections. This phenomenon was also confirmed among the antibodies prepared against each recombinant Msp2 HVR protein, although homologous reactions yielded higher antibody titers than did heterologous reactions in five of six instances (data not shown). Cross reactions did occur among most of the recombinant HVR proteins; however, homologous reactions were typically twice the OD observed for the strongest heterologous reactions. The exception to this was with recombinant HVR protein RW1, for which all reactions were very weak.

Figure 5.

 Antibody reactivity with recombinant Msp2 hypervariable region proteins on day 21 after infection with low and high passage A. phagocytophilum. Relative antibody quantity is shown as OD The recombinant Msp2 HVR protein antigens tested are on the z-axis; the numbers on the x-axis refer to individual mice in each of the mock-infected, and low and high passage A. phagocytophilum-infected groups. Note the diversity and intensity of antibody responses that occur with infection per individual mouse.

Discussion

Msp2 is the major immunodominant protein on the surface of A. phagocytophilum (Dumler et al., 1995; Asanovich et al., 1997; Zhi et al., 1997). Its structure, beyond its molecular size (40–49 kDa) and predicted amino acid sequence, is not known. The gene sequences of the multiple genomic msp2 paralogs possess highly conserved nucleotides that flank a hypervariable region predicted to encode T- and B-cell epitopes (Murphy et al., 1998; Zhi et al., 2002). In addition, Msp2 s of A. phagocytophilum clinical isolates are heterogeneous, depending in part on geographical origin (Asanovich et al., 1997). Evidence exists to support a role for the related ruminant species A. marginale MSP2 in antigenic variation, immune evasion and persistence in reservoir hosts, mostly mediated by antibody production (Palmer et al., 1994; Dumler et al., 1995; Greig et al., 1996; Asanovich et al., 1997; Zhi et al., 1999). Moreover, the hypervariable regions for this bacterium contain important epitopes that can stimulate T cell immune responses (Brown et al., 1998; Brown et al., 2001, 2003, 2004). However, similar data for immune evasion, persistence and T cell responses related to Msp2 do not exist for A. phagocytophilum. In fact, A. phagocytophilum infection in humans, horses and dogs that develop severe febrile infections is generally very brief, lasting from days to weeks at most, indicating that a role for Msp2 variation in persistence among these nonpermissive hosts is unlikely (Madigan, 1993; Aguero-Rosenfeld et al., 1996; Greig et al., 1996; Bakken & Dumler, 2000; Lotric-Furlan et al., 2004; Dumler et al., 2005).

Disease from A. phagocytophilum infection likely results from induction of proinflammatory responses, dominated by the production of IFN-γ. This is evident with infection of IFN-γ knockout mice that do not develop histopathologic lesions yet have significant increases in pathogen load, in contrast to wild-type mice (Akkoyunlu & Fikrig, 2000; Martin et al., 2001). Msp2 is a prime candidate for stimulating immune and proinflammatory responses (Zhi et al., 1997, 1998), as it is the dominant A. phagocytophilum component recognized by the humoral immune system in all infected animals. Disease in experimentally infected horses is modified depending upon the length of in vitro passage of A. phagocytophilum (Pusterla et al., 2000), and the only known A. phagocytophilum structure modified with in vitro passage is Msp2 (Caspersen et al., 2002; Scorpio et al., 2004). As a result, we hypothesized that differentially expressed T cell epitopes in the hypervariable regions of A. phagocytophilum Msp2 s could also lead to changing proinflammatory immune responses able to cause immunopathologic injury. Thus, to determine their capacity for driving immune responses, we assayed in vitro T cell proliferation to isolated Msp2 hypervariable region proteins among splenocytes obtained from mice during in vivo infection.

Although more than 100 msp2 paralogs have been identified in the A. phagocytophilum HZ strain genome (Dunning Hotopp et al., 2006), we detected only 19 unique msp2 variant transcripts in mice infected for <3 weeks with the A. phagocytophilum Webster strain, confirming expression of a limited repertoire over short intervals (IJdo et al., 2002; Wang et al., 2004; Lin & Rikihisa, 2005). Some prior studies indicated that msp2 gene conversion in specific hosts follows a ‘program’ leading to msp2 transcription that would support antigenic variation and immune evasion (IJdo et al., 2002; Wang et al., 2004; Lin & Rikihisa, 2005); however, we found no evidence of such a response among animals inoculated with low or high passage organisms, or over the relatively short interval of 2–3 weeks for which infection lasted in mice (Buitrago et al., 1998; Wang et al., 2004). The data more strongly support random generation of A. phagocytophilum populations expressing msp2 variant transcripts (Lin & Rikihisa, 2005), with no particular variant able to promote survival or persistence, and without bias toward specific variant expression after high or low passage infection. The fact that only 20% of the potential genomic set of msp2 transcripts was detected in mice does not entirely exclude the possibility of a reproducible pattern of transcriptional progression that could vary by host, as the remaining set of paralogs could be expressed under conditions not achieved by the mouse model. However, these findings further corroborate similar studies in horses where an even greater diversity of msp2 transcripts was identified (D.G. Scorpio, manuscript in preparation). These findings occur despite the confirmation of dichotomous clinical manifestations with high and low passage A. phagocytophilum infection in horses (Pusterla et al., 2000), and with differential histopathologic responses in mice, as demonstrated here.

The absence of significant lymphoproliferative responses to isolated recombinant Msp2 HVR proteins further confirms these findings. Owing to the difficulty required in preparing more than 100 recombinant proteins, we were unable to assess T cell responses to all potential Msp2 hypervariable region proteins or linear T cell epitopes. Nor was it possible to prove that recombinant proteins are adequate for evaluating T cell responses as implemented here; however, the clear reactivity of the recombinant HVR proteins with antibodies postinfection, and the usual requirement for only small linear peptide epitopes for T cell reactions, argues that this approach is appropriate and likely reflects a lack of significant T cell response toward epitopes in these regions and in this model system. Moreover, the lack of any substantial or sustained lymphoproliferative responses or IFN-γ production to four of the HVR proteins encoded by transcripts expressed at high levels among infected animals, lends further support to the concept that differential msp2 expression is unrelated to significant T cell immune response, including an immunopathologic response. These findings contrast with the detection of vigorous CD4 responses generated to epitopes in A. marginale MSP2 HVR, although infection or vaccination may occasionally suppress CD4 responses as well (Brown et al., 2004; Abbott et al., 2005). The reason for the discrepancy between CD4 responses stimulated by these two related species is unclear, but could be the result of the ability of A. phagocytophilum to avoid innate immune control (von Loewenich et al., 2004), limiting early access of antigens for presentation and blunting rapid CD4 response. It is possible that the recombinant Msp2 HVR proteins used here had some suppressive effect on in vitro lymphoproliferation; however, this is very unlikely, as the overall proliferative activity among splenocytes stimulated by the recombinant proteins did not differ from one another or from that observed with medium only or with HL-60 cell lysates. Overall, these data provide strong evidence that T cell responses to Msp2 HVR epitopes are meager and unlikely to be responsible for inducing the immunopathologic findings that underlie the murine model histopathology.

Although T cell responses to Msp2 HVR epitopes may not be substantial, all infected animals demonstrated antibody production, including significant reactivity with the recombinant hypervariable region proteins. This latter finding demonstrates the capacity of Msp2 to induce immune reactions largely dominated by antibody production. In fact, the identification of antibodies directed toward HVR proteins for which no corresponding transcript was detected in vivo suggests two alternatives: (1) cross-reactivity between B cell epitopes in the Msp2 HVRs or (2) antibody-mediated selection against specific clones that bear certain Msp2 HVRs. The studies here provide evidence of the former cross-reactivity hypothesis as an explanation, but do not exclude the latter possibility. This observation is deserving of much more study, as (1) it could be associated with immune control of certain clones (Wang et al., 2004); (2) a similar mechanism for restricting A. marginale clonal growth is demonstrated (French et al., 1999); and (3) passive immunization with antibody partially protects against A. phagocytophilum challenge (Sun et al., 1997).

An important unresolved question then remains: what components of A. phagocytophilum are responsible for driving the differential cellular immune responses observed in horses and mice, and is this related to adaptive or innate immunity? Although we are unable to provide support for Msp2 as the immune driving force for T cell reactions, significant levels of lymphoproliferation were observed in splenocytes obtained on day 7 postinfection when stimulated by cell-free A. phagocytophilum, whether derived from low or high passage cultures. In fact, compared to uninfected HL-60 cell lysates, responses to cell-free A. phagocytophilum occurred among naïve splenocytes and splenocytes from both infected and uninfected mice on all days. The responses were most dramatic on days 7–10 postinfection among infected animals, and were higher among mice infected with low passage A. phagocytophilum, corresponding to a more rapid induction of histopathologic lesions. It is likely that the peak response at day 7 and its waning thereafter corresponds to antigen-specific stimulation superimposed on mitogenic activity endogenous to A. phagocytophilum. This implies that A. phagocytophilum cells contain components that nonspecifically stimulate lymphocyte proliferation and cytokine production, and this activity is quantitatively greater in low passage bacteria and facilitates stronger total cellular immune responses. This observation supports prior differences observed in vivo among horses infected by low and high passage A. phagocytophilum (Pusterla et al., 2000), and supports an important role for innate immunity in the induction of immunopathologic injury with infection (Scorpio et al., 2006).

These findings also support previous observations in which whole A. phagocytophilum cells were shown to activate NF-κB nuclear translocation via Toll-like receptor-2 (TLR2) (Choi et al., 2004), with the detection of significant IFN-γ expression in C57BL/6 mouse plasma as early as 4 h postinfection (Martin et al., 2000), and with demonstrated activation of innate immune pathways after A. phagocytophilum infection in vivo (Scorpio et al., 2006). Although these results provide compelling data that immune responses directed toward the changing expression of A. phagocytophilum Msp2 cannot explain the variable inflammatory tissue injury that is IFN-γ-mediated, further study should help by identifying the key component(s) of A. phagocytophilum that induce(s) the innate immune responses and immunopathology.

Acknowledgements

This work was supported by grant R01 AI41213 from the National Institutes of Allergy and Infectious Diseases.

Ancillary