INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Heartwater is caused by the rickettsia Cowdria ruminantium (7). The pathogen is transmitted exclusively by ticks of the genus Amblyomma to wild and domestic ruminants in sub-Saharan Africa and the Caribbean (31). Heartwater is also a threat to livestock on the American mainland, where potential vectors are present but where the disease is currently absent (1). Cowdriosis is controlled primarily through immunization by infection with virulent blood and subsequent treatment with antibiotics to prevent a serious course of the disease (38-40). This procedure has obvious drawbacks, including the possibility of unwanted transmission of other pathogens and its unsuitability for use in countries where potential vectors are present but which do not harbor the disease (30). There is thus a real need for an improved vaccine.

Studies have demonstrated that animals can be protectively immunized with culture-attenuated (18) or inactivated (23, 26) organisms, suggesting that the development of a subunit vaccine may be feasible. Vaccination with culture-attenuated organisms nevertheless has disadvantages in that cross-protection against different isolates is not complete. It would therefore be necessary to attenuate several strains to cover the entire antigenic repertoire. In addition, only certain strains can be attenuated by in vitro passage, which limits the use of this method in vaccination (18). On the other hand, immunization with inactivated C. ruminantium in adjuvant could allow many different strains to be incorporated into a vaccine.

It has been shown that peripheral blood mononuclear cells (PBMC) from animals rendered resistant to challenge by vaccination with inactivated organisms contain C. ruminantium-specific, major histocompatibility complex class II-restricted, gamma interferon-producing, CD4⁺ T lymphocytes (36). Short term CD4⁺-T-cell lines generated by using C. ruminantium lysates respond strongly to whole lysates but not to the recombinant 32-kDa protein (major antigenic protein 1 [MAP1]) or the 21-kDa protein (MAP2) (35). When these cell lines were stimulated with soluble C. ruminantium proteins fractionated by fast-performance liquid chromatography, a single peak of proliferation, which included proteins of between 20 and 32 kDa, was observed (37). Flow cytometric analysis of PBMC showed no significant change in the immune cell population after vaccination and boosting with inactivated organisms. Nevertheless, significant changes occurred after challenge, including an initial progressive depletion of CD4⁺, CD8⁺, and T-cell subsets and a rise in numbers of monocytes together with strong activation. This was followed by an increase in CD8⁺ T lymphocytes (25). The last finding is in accordance with previous studies with a murine model which led the authors to suggest that CD8⁺ T cells play a major role in immunity to heartwater (10-12).

Recently, Mwangi and coworkers (28) demonstrated that cattle immunized against heartwater by infection and treatment generated T-cell responses specific for two immunodominant recombinant antigens of C. ruminantium, namely, MAP1 and MAP2. Proliferation of PBMC was also elicited in vitro by either autologous infected endothelial cells or infected monocytes but not by elementary bodies. The endothelial cells required pretreatment with T-cell growth factors to induce class II major histocompatibility complex expression prior to infection and their subsequent use as stimulators of PBMC. These proliferative responses were characterized by a mixture of CD4⁺, CD8⁺, and T cells and strong expression of gamma interferon, tumor necrosis factors alpha and beta, and interleukin-2 (IL-2) (27).

In another approach, a naked-DNA vaccine containing the map1 gene of C. ruminantium was shown to be able to protect between 23 and 88% of immunized mice (29). The best-characterized proteins of C. ruminantium are the above-mentioned MAP1 (19, 33) and MAP2 (24) proteins, as well as the GroEL heat shock protein (22), the genes of which have been cloned (4, 22, 24, 42). No studies to date have reported on additional antigens being involved in stimulating protective immunity. For the development of a subunit vaccine, it may nonetheless be important to identify additional C. ruminantium proteins which stimulate T lymphocytes and, in turn, to relate these to protective responses. Protective malarial parasite proteins have been identified by vaccination trials with fractions obtained by continuous-flow electrophoresis (CFE) of Plasmodium chabaudi adami shizont proteins (21). Since T cells are required for protective immunity in malaria (3, 15), this finding suggested that potentially protective T-cell antigens could be identified by this technique. Using a similar procedure, Brown and coworkers (5, 6, 34) have successfully identified several antigens of Babesia bovis merozoites that stimulated proliferation of T-cell lines and clones. To identify the proteins involved in recall T-cell responses in immune cattle, C. ruminantium was therefore fractionated by CFE, and each fraction was tested for its ability to stimulate lymphocyte proliferation in vitro. In an attempt to stimulate the natural infective process, the PBMC used in this study were obtained from animals rendered immune by infection and treatment. In a parallel study, four cattle were immunized with inactivated C. ruminantium to determine whether their PBMC recognize similar proteins in proliferation assays.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

In vitro cultivation of C. ruminantium. The Welgevonden isolate of C. ruminantium was cultured either in bovine saphenic vein (BSV) endothelial cells (prepared from animal B9191) or in a calf endothelial cell line (E₅), as described previously (2). The Welgevonden isolate was chosen because the pathological patterns caused by this isolate resemble those caused by the Ball3 isolate presently used in the live vaccine (32). Additionally, unlike the Ball3 isolate, the Welgevonden isolate is pathogenic to mice, permitting viability tests to be done with these animals. The Welgevonden isolate elicits total immunity against more isolates than does the Ball3 isolate (13).

Preparation of crude C. ruminantium infected or uninfected BSV cell culture lysates. Crude extracts of Welgevonden isolate-infected or uninfected cell cultures were prepared as described previously (33). Briefly, uninfected or infected (harvested when maximally infected with organisms) BSV cell lysates were centrifuged for 30 min at 10,000 × g. The resulting pellet was suspended in phosphate-buffered saline (PBS) (0.14 M NaCl, 1 mM KH₂PO₄, 8 mM Na₂HPO₄ · 12H₂O, and 3 mM KCl, pH 7.4). The lysates were stored at 20°C and used for fractionation of antigens by CFE.

Preparation of semipure inactivated C. ruminantium lysates from cell cultures. Semipure inactivated organisms were prepared as described previously (36). Differential centrifugation (DC) was done with either maximally infected E₅ or BSV cell cultures by centrifuging first at 1,000 × g for 10 min. The resultant supernatant was centrifuged at 30,000 × g for 30 min. The pellet was resuspended in PBS containing sodium benzylpenicillin (0.12 mg/ml) and streptomycin sulfate (0.198 mg/ml). The suspension was subjected to five freeze-thaw cycles performed with liquid nitrogen and stored at 20°C. In order to confirm that the preparation contained inactivated C. ruminantium, mice were inoculated intravenously with 0.2 ml of the same lysate per mouse. The inactivated C. ruminantium lysate prepared from infected E₅ cell cultures was used for immunization of cattle, and the inactivated C. ruminantium lysate prepared from infected BSV cell cultures was used as antigen in proliferation assays.

Protein determination. Protein concentrations were determined by the Bio-Rad (Hercules, Calif.) protein microassay with bovine serum albumin as a standard.

Experimental cattle. All animals were initially seronegative to C. ruminantium, B. bovis, and Theileria mutans as determined by indirect fluorescent-antibody assays and to Anaplasma as determined by competition inhibition enzyme-linked immunoassay.

(i) Immunization by infection and treatment. Animals B9191 (Fresian) and B1359 (Nguni) were inoculated intravenously with 5 ml of a Welgevonden isolate infective sheep blood stabilate. The animals were monitored daily and treated on the third day of a rising febrile reaction by intramuscular injection with long-acting oxytetracycline (Liquamycin LA; Pfizer) at a dose of 20 mg/kg of body weight. The animals were challenged with the same batch of homologous stabilate at the following intervals: B9191 at 1 month and 3 years and B1359 at 6 months and 8 months after immunization. The animals were monitored daily for temperature. As a positive control to verify the viability and virulence of the stabilate, mice were inoculated intravenously with 0.2 ml of the same stabilate per mouse. Postmortem examinations were done on the mice to determine the cause of death. The presence of C. ruminantium in the mice was confirmed by immunohistochemical identification of the organism in formalin-fixed tissue sections (17).

(ii) Immunizations with inactivated C. ruminantium. Four Nguni cattle (B809, B821, B1351, and B775) were injected with 15 µg of inactivated C. ruminantium per ml in Montanide ISA50 adjuvant (Seppic, Paris, France). One animal (B816) received only adjuvant as a control. The animals were injected twice at an interval between 14 and 16 weeks.

Fractionation of crude C. ruminantium-infected and uninfected BSV cell lysates by CFE. CFE of crude C. ruminantium-infected or uninfected BSV cell lysates was performed with a Prep-Cell Apparatus (Bio-Rad) as described previously (5) with the following modifications. Approximately 10 mg of protein was solubilized in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer, boiled, and electrophoresed under reducing conditions on 28-mm (internal diameter) cylindrical gels consisting of either 7 or 15% acrylamide. Fractions of 2.5 ml were eluted and collected over the course of 15 h (7% acrylamide gels; 180 fractions) or 7.5 h (15% acrylamide gels; 175 fractions). The fractions were stored at 70°C. The fractions were precipitated by adding 8 times the volume of ice-cold acetone, incubated at 20°C for 16 h, and centrifuged at 10,000 × g for 10 min. The acetone was aspirated, and the precipitates were resuspended in 70% cold ethanol and centrifuged at 10,000 × g for 30 min. After the ethanol was aspirated, the pellets were air dried, suspended in PBS containing antibiotics, and stored at 70°C. To examine the potential toxicity of the fractions on T-cell proliferation, the effect of a randomly selected fraction, at concentrations ranging from 3 to 24% (vol/vol), on either IL-2- or concanavalin A-induced PBMC proliferation was tested.

SDS-PAGE analysis of fractionated C. ruminantium antigens. A volume of 20 µl of every 10th precipitated fraction was analyzed by SDS-PAGE. A minigel system with acrylamide gels of either 7% (7% Prep-Cell fractions) or 12% (15% Prep-Cell fractions) was used. Protein bands were visualized by silver staining.

Lymphocyte proliferation assays. Proliferation assays were carried out in duplicate or triplicate wells of half-area flat-bottom 96-well plates (Costar) at 37°C in a humidified atmosphere containing 5% CO₂ for 5 days (3 days with concanavalin A) as described previously (5). Each well (total volume, 100 µl) contained complete RPMI 1640 medium, responder cells added at a final concentration of 4 × 10⁶ PBMC/ml, and various concentrations of C. ruminantium-infected or uninfected BSV cell antigens (0.03 to 10 µg of antigen/ml). Fractionated antigens were included at final concentrations of 5% (vol/vol). Proliferation was determined by measuring the incorporation of 1 µCi of [methyl-³H]thymidine added during the final 18 h of the assay. The cells were harvested, and the radioactivity was counted in a scintillation counter. Results are presented as a stimulation index (SI) ± standard deviation (SD), where SI = mean counts per minute of test sample/mean counts per minute of unstimulated control. The unstimulated control was PBMC in medium. Unless otherwise stated, an SI of 2 was considered to be an indication of antigen-specific proliferation. The one-tailed Student t test was used to determine the levels of significance between the uninfected and infected BSV cell CFE fractions.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Proliferative responses elicited by different rickettsial preparations. Different C. ruminantium preparations were tested for their ability to elicit an optimum proliferative response in the PBMC obtained from a naive animal, two immune animals (B9191 and B1359), and three animals immunized with inactivated organisms (B821, B775, and B1351). C. ruminantium was partially purified from BSV cells by (i) DC, (ii) PSIAC, and (iii) PDGC. All of these antigen preparations were assayed with PBMC in triplicate wells on the same day. The highest proliferative response was obtained with antigen prepared by DC followed by PSIAC and PDGC (Table 1). DC was therefore subsequently routinely used to purify C. ruminantium. When differentially centrifuged uninfected BSV cell lysates were used as antigen in proliferation assays with PBMC from immunized animals, proliferation was consistently observed to be lower than that obtained with the infected preparations (Table 1). However, higher proliferation values were obtained with PBMC from cattle immunized with inactivated organisms and assayed with differentially centrifuged uninfected BSV cell lysates. This may be due to the presence of endothelial cell debris in the differentially centrifuged preparations used for immunizing these cattle. When differentially centrifuged uninfected BSV cell lysates were used as antigen in proliferation assays with PBMC from nine naive cattle, no proliferation was detected (SI  1.7) (results not shown). These results showed that uninfected BSV cell antigen did not induce alloreactive responses in PBMC from unrelated immunized or naive cattle.

Proliferative responses during the course of immunization. Following immunization by infection and treatment, both cattle (B9191 and B1359) were found to be immune to challenge as determined by the lack of a febrile response and symptoms of the disease. All mice injected with this live blood stabilate used for cattle challenge died. The symptoms before death, the time to death (±12 days), and the postmortem findings all indicated that the mice were infected with C. ruminantium. In addition, the presence of C. ruminantium in tissue sections of the mice was confirmed histopathologically. These findings indicated that the stabilate was indeed virulent. Proliferative responses during the course of immunization are illustrated in Fig. 1. PBMC tested prior to immunization or challenge did not respond to C. ruminantium antigens when tested at a range of protein concentrations varying from 0.03 to 10 µg/ml. In contrast, PBMC from both immune animals responded specifically to C. ruminantium antigens after challenge. The proliferative response peaked at 1 µg/ml (results not shown). Antigen concentrations of 2.5 µg/ml inhibited both concanavalin A- and IL-2-induced responses (results not shown). After B9191 was challenged for the second time, C. ruminantium-specific lymphocyte proliferation peaked 4 weeks after challenge (31,764 ± 8,702 cpm; 164 weeks after immunization) but was not measurable 2 weeks later (Fig. 1a). By comparison, proliferation assays with PBMC from B1359 indicated an antigen-specific response 1 week after the first challenge (52,099 ± 3,899 cpm; 28 weeks after immunization) and 3 weeks after the second challenge (48,626 ± 4,992 cpm; 41 weeks after immunization) and was not measurable a week later (Fig. 1b). When PBMC from these cattle were assayed again 1 to 2 years later, they showed antigen-specific proliferation with an SI of 5.5. The duration of the PBMC response is not known.

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Proliferative responses of PBMC from animals B9191 (a) and B1359 (b) at various intervals after immunization by infection and treatment. The PBMC were cultured for 6 days with differentially centrifuged lysates of C. ruminantium-infected BSV cells at a concentration of 1 µg/ml in either duplicate or triplicate wells. The mean counts per minute for the PBMC controls were 2,065 ± 2,194 (B9191) and 3,943 ± 3,090 (B1359). The arrows indicate the times of challenge. Results are presented as SI ± SD.

View larger version (191915K): [in this window] [in a new window]   FIG. 2.   Proliferative responses of PBMC from animals B809 (a), B821 (b), and B816, B1351, and B775 (c; , , and , respectively) at various intervals after immunizing with inactivated C. ruminantium. The PBMC were cultured for 6 days with differentially centrifuged lysates of C. ruminantium-infected BSV cells at a concentration of 1 µg/ml in either duplicate or triplicate wells. The mean counts per minute for the PBMC controls were 15,707 ± 18,666 (B816), 4,120 ± 6,930 (B809), 2,529 ± 2,710 (B821), 8,948 ± 11,612 (B1351), and 6,724 ± 6,063 (B775). The arrows indicate the times of boosting. Results are presented as SI ± SD.

Fractionation of C. ruminantium-infected and uninfected BSV proteins by CFE and the proliferative responses they stimulate. In an attempt to identify which of the C. ruminantium proteins were responsible for the T-cell responses described above, Welgevonden-infected and uninfected BSV cell culture lysates were fractionated under reducing conditions by CFE. Welgevonden-infected BSV cell culture proteins were fractionated at between 50 and 168 kDa on a 7% gel and at between 11 and 38 kDa on a 15% gel (Fig. 3). In a similar profile BSV cell culture proteins were fractionated at between 50 and 123 kDa on a 7% gel and at between 11 and 34 kDa on a 15% gel (results not shown). There were artifacts caused by silver staining at approximately 66 kDa on both the 7 and 15% gels. The other protein bands at fraction 150 on the 7% gel and fractions 5, 15, 25, 35, 65, and 75 on the 15% gel may be due to breakdown products. A total of 700 fractions were collected and prepared for lymphocyte proliferation assays by acetone precipitation. A randomly chosen Prep-Cell fraction (with a molecular mass of approximately 100 kDa), when tested at concentrations of 3 to 24% (vol/vol), did not inhibit a concanavalin A-induced proliferative response (results not shown). Based on these findings together with previous results (5), a final concentration of 5% (vol/vol) was selected for testing with C. ruminantium-immune PBMC.

View larger version (91K): [in this window] [in a new window]   FIG. 3.   SDS-PAGE and silver staining analysis of C. ruminantium-infected BSV cell lysates fractionated by CFE. Crude C. ruminantium-infected BSV cell lysates were fractionated by CFE on either 7% (a) or 15% (b) acrylamide gels. The fractions were precipitated with acetone and resuspended in 500 µl of PBS. A volume of 20 µl of every 10th precipitated fraction was subjected to analytical SDS-PAGE (7% acrylamide for the 7% CFE fractions and 12% acrylamide for the 15% CFE fractions) and visualized on the silver-stained gels. The relative mobilities of the low-molecular-mass standards (lanes L) and high-molecular-mass standards (lanes H) are indicated on the left of each panel in kilodaltons.

(i) Proliferative responses of PBMC from animal B9191 induced by CFE fractions. Fractions from C. ruminantium-infected and uninfected BSV cells were assayed with PBMC collected from animal B9191 at week 291 after immunization. The SI_neg was determined to be 1 ± 0.1 and 2 ± 0.1 for the 7 and 15% polyacrylamide CFE fractions, respectively. An examination of the fractions derived from the CFE with a 7% polyacrylamide gel revealed that those inducing C. ruminantium-specific proliferation were localized to the first 24 fractions. These had molecular masses of 74 kDa. In addition, a further three discrete pools of higher-molecular-mass proteins (50 to 67, 81 to 111, 115 to 140, and 151 to 162 kDa) also induced proliferation (Fig. 4a). When a similar preparation was fractionated on a 15% polyacrylamide gel to resolve the low-molecular-mass proteins, the stimulatory fractions ranged from 11 to 38 kDa (Fig. 4b). Due to the high SI_neg of 2 obtained for the 15% polyacrylamide CFE, an SI 4 was considered as antigen specific for this preparation. Ten groups of fractions had SIs of 4. The molecular masses in these groups ranged from 11 to 24 kDa and from 26 to 38 kDa.

View larger version (29K): [in this window] [in a new window]   FIG. 4.   Proliferative response of PBMC from animal B9191 to C. ruminantium-infected and uninfected BSV cell lysates fractionated by CFE. C. ruminantium-infected or uninfected BSV cells were electrophoresed on either 7% (a) or 15% (b) acrylamide gels. Proteins were eluted from the gels, collected as 2.5-ml fractions, precipitated with acetone, and resuspended in PBS. Open circles, fractions from control uninfected BSV cell lysates were pooled at six per sample and assayed in triplicate wells for stimulation of PBMC collected 291 weeks after immunization. Closed circles, fractions from infected BSV cell lysates were pooled at two per sample and assayed in triplicate wells for stimulation of PBMC collected 291 weeks after immunization. The mean counts per minute for the PBMC controls were 1,378 ± 361. The results are presented as SI. *, P  0.05; #, P  0.1. The approximate molecular masses of the fractions inducing C. ruminantium-specific proliferation are shown above the charts.

(ii) Proliferative responses of PBMC from animal B1359 induced by CFE fractions. Fractions from C. ruminantium-infected BSV cells were assayed with cryopreserved PBMC collected 41 weeks after immunization of B1359. In a similar way, lymphocytes from this animal responded specifically to C. ruminantium fractions with relatively low molecular masses of approximately 11 to 23 kDa and 26 to 27 kDa (Fig. 5b). In addition, only one pool of high-molecular-mass proteins of approximately 85 to 90 kDa also induced lymphocyte proliferation (Fig. 5a). In contrast, the SI_neg was determined to be 0.4 ± 0.1 and 0.7 ± 0.2 for the 7 and 15% CFE fractions, respectively. The low SI_neg values obtained indicated that no alloreactive responses were induced by the BSV cells in CFE fractions.

View larger version (23K): [in this window] [in a new window]   FIG. 5.   Proliferative response of PBMC from animal B1359 to C. ruminantium-infected and uninfected BSV cell lysates fractionated by CFE. C. ruminantium-infected and uninfected BSV cell lysates were electrophoresed on either 7% (a) or 15% (b) acrylamide gels. Proteins were eluted from the gels, collected as 2.5-ml fractions, precipitated with acetone, and resuspended in PBS. Open circles, fractions from control uninfected BSV cell lysates were pooled at six per sample and assayed in triplicate wells for stimulation of PBMC collected 111 weeks after immunization. Closed circles, fractions from C. ruminantium-infected BSV cell lysates were pooled at six per sample and assayed in triplicate wells for stimulation of PBMC collected 41 weeks after immunization. The mean counts per minute for the PBMC controls were 2,434 ± 2,134. The results are presented as SI. *, P  0.05; #, P  0.1. The approximate molecular masses of the fractions inducing C. ruminantium-specific proliferation are shown above the charts.

(iii) Proliferative responses of PBMC from animals B809, B821, B1351, B775, and B816 induced by CFE fractions. Fractions from C. ruminantium-infected BSV cells were assayed with PBMC collected 1 week before and between 18 and 25 weeks after immunizations commenced. Fractions from control uninfected BSV cells were assayed with PBMC collected between 41 and 57 weeks after immunizations commenced. The proliferative responses of PBMC from ox B809 induced by CFE fractions are represented in Fig. 6 to illustrate the type of responses obtained. The proliferative responses of PBMC from animals B821, B1351, B775, and B816 induced by CFE fractions of C. ruminantium-infected BSV cells are summarized in Table 2. No antigen-specific proliferation was detected with PBMC from naive animals (before immunization with inactivated organisms) and mean SIs of 2 were obtained (results for B809 are shown in Fig. 6; results for the remaining cattle are not shown). An SI_neg of <1.0 was observed for all cattle, except the control animal, when PBMC were tested with 7 and 15% polyacrylamide CFE fractions from uninfected endothelial cells (results for B809 are shown in Fig. 6; results for the remaining cattle are not shown). The low SI_neg values obtained indicated that no alloreactive responses were induced by the BSV cells in CFE fractions. PBMC from cattle immunized with inactivated organisms responded specifically and significantly to antigen fractions containing the following proteins: for B809, 11 to 31, 60 to 80, 85 to 134, and 145 to 148 kDa; for B821, 11 to 26 and 60 to 74 kDa; for B1351, 11 to 38 and 55 to 167 kDa; and for B775, 11 to 38, 55 to 152, and 157 to 167 kDa (Fig. 6 and Table 2). The SI_neg for the control animal B816 was determined to be 3.6 ± 0.9 and 4 ± 0.5 for the 7 and 15% polyacrylamide CFE fractions respectively. The high SI_neg of 4 obtained was taken into consideration, and an SI of 8 was considered to be antigen specific for this preparation. The high SI values obtained for this control animal (immunized with adjuvant alone) may have been a result of an adjuvant effect. However, six groups of fractions which induced significant proliferation were identified and corresponded to 13, 18, 26, 27, 29, and 30 kDa (Table 2).

View larger version (23K): [in this window] [in a new window]   FIG. 6.   Proliferative response of PBMC from animal B809 to C. ruminantium-infected and uninfected BSV cell lysates fractionated by CFE. C. ruminantium-infected or uninfected BSV cell lysates were electrophoresed on either 7% (a) or 15% (b) acrylamide gels. Proteins were eluted from the gels, collected as 2.5-ml fractions, precipitated with acetone, and resuspended in PBS. Squares, fractions from infected BSV cell lysates were pooled at six per sample and assayed in triplicate wells for stimulation of PBMC collected 1 week before immunization (naive animal). Open circles, fractions from control uninfected BSV cell lysates were pooled at six per sample and assayed in triplicate wells for stimulation of PBMC collected 57 weeks after immunization. Closed circles, fractions from infected BSV cell lysates were pooled at six per sample and assayed in triplicate wells for stimulation of PBMC collected 18 weeks after immunization. The mean counts per minute for the PBMC controls were 6,505 ± 11,797. The results are presented as SI. *, P  0.05; #, P  0.1. The approximate molecular masses of the fractions inducing C. ruminantium-specific proliferation are shown above the charts.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This is the first report of fractionation of C. ruminantium proteins by CFE, as well as of the identification and characterization of the molecular masses of antigens that induce proliferation of PBMC obtained from cattle rendered immune by infection and treatment or immunized with inactivated organisms. The high resolution afforded by CFE allowed relatively fine discrimination of the immunostimulatory proteins. Seven outbred cattle were used in this study, resulting in the identification of a common region (11 to 23 kDa) that induced proliferation of their PBMC. Proteins with molecular masses of 20 to 23 kDa identified in our assays fall within a range of proteins (fractionated by fast-performance liquid chromatography) that had previously been identified to induce proliferation of a T-cell line derived from cattle immunized with killed C. ruminantium (37). It is interesting that fractions containing molecular masses of the major immunodominant proteins of C. ruminantium, of approximately 21 kDa (MAP2) (24), 27 kDa (33), and 32 kDa (MAP1) (19, 33), stimulated PBMC proliferation in our assays. Similar T-cell responses were also recently observed for recombinant forms of the 21-kDa (MAP2) and the 32-kDa (MAP1) proteins (28). The 32-kDa (MAP1) protein has also been implicated in protection in an immunization trial in mice (29). Our results confirm previous reports (37) that low-molecular-mass proteins of 32 kDa may be important in stimulating the cellular immune response, since they predominated in our stimulatory fractions.

The results presented here were obtained by studies undertaken with PBMC. As the animals B9191 and B1359 were immunized with live C. ruminantium, both replicating and circulating organisms should thus be present in these animals. In addition the organism has been shown to occur in various cell types, including macrophages, monocytes, Küpfer cells, reticulum cells of the lymph nodes, fibroblasts, and connective tissues, as well as cells of the spleen, brain, pancreas, and heart (7-9, 16). Therefore, the immune responses may be taking place locally (e.g., in the lymph nodes and/or spleen) during the periods when no detectable proliferative responses were obtained with circulating peripheral lymphocytes. On the other hand, animals that have been immunized with inactivated C. ruminantium always had responsive circulating lymphocytes present in the blood (36). Gale and coworkers (14) observed a highly variable PBMC proliferative response in cattle immune to Anaplasma marginale, but sensitized T cells were readily detected in their spleens. It has been suggested that sensitized T cells home into lymphoid tissue from the circulation via the expression of specific cell surface molecules. Therefore, a study of the responses in other immune compartments such as the lymph nodes or the spleen may give a more defined picture of the immune response to C. ruminantium during these periods. Furthermore, antigen-specific proliferation was later detected in PBMC from our cattle, and this may be explained by the return of such lymphocytes into circulation. Mwangi and coworkers (27) similarly failed to detect a proliferative response before challenge, with PBMC collected from animals rendered immune by infection and treatment. Proliferation was detected only when autologous endothelial cells were pretreated with T-cell growth factors prior to infection with live organisms, fixed, and then used as stimulators. Their results suggest that antigen processing and presentation by infected endothelial cells or monocytes may be essential for the induction of specific T-cell responses during infection. This may be an alternative explanation for the periods during which no proliferation was observed in our assays.

An important consideration in vaccine design is whether a single parasite antigen will elicit the appropriate protective humoral and cell-mediated responses or whether these responses will need to be activated by many distinct antigens (21). Further studies should therefore not only investigate common proteins recognized by memory T cells from genetically different cattle but also characterize epitopes conserved between isolates. Different cross-protection patterns (13, 20) and antigenic diversity between isolates (20, 33) indicate the potential for polymorphism of proteins and T- and B-cell epitopes. The Welgevonden isolate was used in this study since it induces cross-protective immunity against more isolates than any other and is also highly virulent (13). Comparative studies performed with isolates against which the Welgevonden isolate is not cross-protective may identify additional proteins of immunological importance. The combination of infection and treatment resembles natural infection, but the dose administered is likely to be higher than that acquired in the field. Nevertheless, in this study identification of proteins by PBMC from cattle immunized in this way led to the identification of candidate vaccine antigens.

In summary, our results indicate that, as with P. chabaudi adami and B. bovis, fractionation of organisms by CFE provides a way to identify potential vaccine antigens of C. ruminantium. The use of sensitized primary polyclonal lymphocytes permits rapid and simple screening of CFE fractions for the proteins that stimulate specific immune responses (43). In this way, C. ruminantium proteins with molecular masses of 11, 12, 14 to 17, and 19 to 23 kDa with the potential to play a role in protection were identified, and these proteins will be further characterized.