Veterinary Parasitology, 44 ( 1992 ) 3-13 Elsevier Science Publishers B.V., Amsterdam
3
The development of a recombinant Babesia vaccine
I . G . W r i g h t a, R. C a s u a, M . A . C o m m i n s a, B.P. D a l r y m p l e a, K . R . G a l e a, B.V. G o o d g e r a, P . W . R i d d l e s a, D.J. W a l t i s b u h l a, I. A b e t z ", D . A . B e r r i e a, Y. B o w l e s a, C. D i m m o c k ~, T. H a y e s a, H. K a l n i n s b, G . L e a t c h a, R. M c C r a e ~, P.E. M o n t a g u e c, I.T. N i s b e t b, F. P a r r o d i a, J . M . P e t e r s ", P.C. S c h e i w e a, W . S m i t h a, K. R o d e - B r a m a n i s a a n d M . A . W h i t e ~ aCS1R 0 Division of Tropical Animal Production, Long Pocket Laboratories, Private Bag, No. 3 P. 0., Indooroopilly, Qld. 4068, Australia bCommonwealth Serum Laboratories, 45 Poplar Road, Parkville, Vic. 3052, Australia cPitman-Moore Australia, Princes Highway, Werribee, Vic. 3030, Australia
ABSTRACT Wright, I.G., Casu, R., Commins, M.A., Dalrymple, B.P., Gale, K.R., Goodger, B.V., Riddles, P.W., Waltisbuhl, D.J., Abetz, I., Berrie, D.A., Bowles, Y., Dimmock, C., Hayes, T., Kalnins, H., Leatch, G., McCrae, R., Montague, P.E., Nisbet, I.T., Parrodi, F., Peters, J.M., Scheiwe, P.C., Smith, W., Rode-Bramanis, K. and White, M.A., 1992. The development of a recombinant Babesia vaccine. Vet. Parasitol., 44:3-13. Crude extracts of Babesia bovis parasites were shown to induce levels of protection in susceptible cattle equivalent to that resulting from natural infection. The crude material was systematically fractionated and tested in numerous sequential vaccination/challenge experiments in adult cattle. Antigens in protective fractions were then purified by affinity chromatography with monoclonal antibodies. Three highly protective (more than 95% reduction in parasitaemias) antigens were thus identified. None of these antigens was immunodominant; a number of immunodominant antigens were identified and all were immunosuppressive and/or non-protective. The three protective antigens were cloned and expressed as either/?-galactosidase or glutathione-Stransferase (GST) fusion proteins. Two of these, GST-12D3 and GST-11 C5, when used in combination were almost as protective as has been previously shown for the commercially available live attenuated vaccine. A short fragment of a third antigen (21B4) has also been shown to be protective. In two of the antigens, repetitive segments have been shown to be non-protective while the third antigen ( 12D3 ) does not contain repetitive domains. Homologues of these antigens exist in other Babesia species and it is anticipated that these may be candidate antigens for protective vaccines against those species. Correspondence to: I.G. Wright, CSIRO Division of Tropical Animal Production, Long Pocket Laboratories, Private Bag, No. 3 P.O., Indooroopilly, Qld. 4068, Australia.
© 1992 Elsevier Science Publishers B.V. All rights reserved 0304-4017/92/$05.00
4
I.G. W R I G H T ET AL.
INTRODUCTION
Babesia bovis is a tick-transmitted haemoprotozoan which is a pathogen of major significance in cattle. The parasite causes large economic losses in Africa, southeast Asia, Australia and South America. A live attenuated vaccine has been developed in Australia (Callow et al., 1979 ) but it is little used outside that country. The vaccine has a number of limitations, including short shelf life (5-7 days at 5 °C), reversion to virulence and to transmission by the tick vector and the possibility of co-transmission of other infectious agents, especially viruses (Wright and Riddles, 1989 ). As well as the live attenuated vaccine, cell culture-derived exoantigens have also been extensively studied in an attempt to develop non-living vaccines. In the case of B. bovis, the microaerophilous stationary phase (MASP) method for the continuous propagation of the parasite has resulted in the production of large amounts of exoantigens in the culture medium (Levy and Ristic, 1980). However, reports of the efficacy of this material are quite varied, with one group claiming a high degree of protection (Montenegro-James et al., 1989) while another (Timms et al., 1983) reported that this material induced no protection whatsoever. Short-term MASP cultures of Babesia bigem ina have also been reported (Montenegro-James et al., 1989); the exoantigens induced good levels of protection. A modified MASP system has also been used to develop a commercial Babesia canis vaccine in France (Moreau et al., 1988). This product 'Pirodog' is used extensively in that country and gives about 80% protection when dogs are revaccinated annually (Y. Moreau, personal communication, 1990). Because of the inherent problems that exist with the live vaccine and the variable results which have been obtained with MASP-derived material, our group has been researching for a number of years the possibility of developing a stable, highly protective, reproducible and cheap vaccine using recombinant DNA technologies. This has been based on a pragmatic approach whereby protective antigens are identified by in vivo vaccination/challenge experiments rather than by immunochemical identification alone. The work has now reached the development stage whereby defined antigens are undergoing trials to meet registration requirements prior to commercial release. This paper gives an account of the work that led up to this point. IMMUNITY TO BABESIA INFECTIONS
Immunity to B. bovis is lifelong, once an animal recovers from a primary infection (Mahoney et al., 1973, 1979b). Sterile immunity does not occur in B. bovis infections and infected cattle undergo a continuous series of parasite recrudescences (Mahoney et al., 1973). These may well serve to stimulate immune responses continuously, and high antibody levels can be detected 4
DEVELOPMENT OF A RECOMBINANT BABESIA VACCINE
5
years after infection. In B. bigemina infections, however, sterile immunity does occur (Mahoney et al., 1973 ) and a strong protective response is elicited without parasite recrudescences. In contrast to B. bovis infections, antibodies can generally not be detected 18-21 months after infection even though the animals are protected against subsequent infection. The nature of the immune response in these different biological phenomena is not understood. An important finding was made by Mahoney (1967a) and later by Mahoney et al. (1979a) who demonstrated that passive transfer of IgG from animals made hyperimmune by repeated B. bovis infections could protect susceptible cattle against challenge by that parasite. These studies also strongly supported the idea that effector cells were necessary for protection, as, in the absence of effector cells, the IgG fraction had no direct effect on B. bovis parasites. The IgM antibodies were not involved in protection. This indicated that it was highly likely that stimulation of B cells and effector cells by a vaccine should result in protective immunity. Recently in our laboratory (Jacobson et al., 1992) we have developed an in vitro opsonisation assay using mouse macrophages, to assess the influence of anti-B, bovis antisera on erythrocytes infected with B. bovis and have confirmed that effector cells play an essential role in protective immunity (Mahoney et al., 1979a). It is interesting that Mahoney et al. (1980) implicated complement as a factor in the immune response to babesiosis, but Jacobson et al. (1992), using in vitro phagocytosis did not. I NDUC T I ON OF IMMUNITY WITH PARASITE EXTRACTS
Immunity to B. bovis by vaccination with killed organisms was first demonstrated by Mahoney (1967b) and later by Mahoney and Wright (1976). These studies showed that a degree of protection was induced which was similar to that of naturally acquired immunity. This indicated that it was feasible to develop a killed vaccine against this parasite. The systematic fractionation of B. bovis extracts to identify protective components by our group began around 1980. Initially, a whole range of chromatographic and electrophoretic procedures were utilised at each stage of purification. The fractions were then tested in calves in vaccination/challenge trials. A total of about 80 such trials was conducted. The protective status of a particular antigen was mainly determined by the pathophysiological response of the vaccinated animal upon challenge (Goodger et al., 1981 ). A large number of pathophysiological parameters were identified which were accurate indices of the severity of the disease. These included the presence in the serum of cryofibrinogen, haptoglobin, immunoconglutinin and haemoglobin and reduced levels of conglutinin, haemolytic complement, C3, coagulation factors, fibronectin, as well as lymphocytes and thrombocytes. Collectively, marked changes in these parameters indicated that
6
I.(3. WRIGHT ET AL.
the animal had not been protected by vaccination, and that particular antigen was not further studied. The converse was equally true. In studies where little or no change occurred with these pathophysiological indices, a strong protective response was confirmed. This pragmatic approach to antigen identification led to a number of highly protective fractions being isolated. These fractions were then subjected to further subfractionation using affinity chromatography with monoclonal and polyclonal antisera or with various ligands including protease inhibitors as well as various electrophoretic procedures. Three antigens were purified and identified for intensive study. The first of these was an antigen designated 15B1 (after the identifying monoclonal antibody (MAb) A15B1 ) (Wright et al., 1983, 1985). This antigen was subsequently redesignated 12D3 (after the more stable IgG1 isotype MAb which cross-reacted with this molecule). This antigen was of low molecular weight and was present in low concentration. Two injections, at monthly intervals, of 10/~g of 12D3 with adjuvant, reduced parasitaemias by over one order of magnitude (as compared with controls when the cattle were challenged with an heterologous strain of B. bovis) (Riddles et al., 1992 ). A second fraction, designated fl, was found to be highly protective (Goodger et al., 1985b). The components of the fl fraction were separated by affinity chromatography using the MAbs produced against this fraction. Eight fractions were subsequently tested by vaccination/challenge trials in calves before a protective antigen was identified. This antigen was designated 11C5 after the MAb W 11 C5, was of high molecular weight and was also present in low concentration. A third antigen, designated protease, was first identified by substrate electrophoresis whereby proteolytic activity was demonstrated in acrylamide gels in which a substrate such as gelatin was incorporated (Commins et al., 1985 ). The gel band was excised and shown to induce strong protection in calves. A range of MAbs was then raised against this gel band fraction and used to further enrich the antigen by affinity chromatography. One of these MAbs (T2 I B4) identified an antigen of 60-70 kDa which had strong proteolytic activity and which also induced protection when injected into calves. This antigen, like the other two, was only present in minute amounts. All three antigens are present as soluble proteins and all appear to be secreted by the parasite. Furthermore, preliminary analyses of the proteins indicate that they all play a functional rather than a structural role in the parasite's life cycle. Subsequently, crude extracts ofB. bigemina have been used to induce strong protective responses against challenge with that organism (Wright et al., 1987 ). Similar immunity to subsequent challenge with Babesia ovis organisms has also been induced in sheep injected with crude B. ovis extracts (A1abay et al., 1987). In the course of these two studies some cross-protection was elicited by B. bigernina extracts against B. bovis challenge and by B. bovis
DEVELOPMENT OF A RECOMBINANT BABES1A VACCINE
7
extracts against B. ovis challenge, indicating that a degree of antigenic commonality exists between the various species. THE ROLE OF IMMUNODOMINANTANTIGENSIN THE IMMUNE RESPONSE During the course of this work numerous abundant, i m m u n o d o m i n a n t antigens were isolated. These included the 70 kDa immunodiffusion (ID) antigen (Goodger et al., 1986), the dominant 200 kDa antigen (B.V. Goodger, unpublished data, 1988) and an antigen of more than 1500 kDa (the 3C1 antigen) (Wright et al., 1983 ). In all instances these antigens, whilst inducing high antibody titres, were non-protective. Furthermore, in the case of the 3C 1 antigen, the protective response to the 15B 1/ 12D3 antigen was abolished when these two antigens were used together. A second group of so-called Babesia antigens has also been identified (Goodger et al., 1985a). These antigens were recognised on Western blots as dominant bands when probed with B. boris-positive antisera. However, they were also recognised avidly by sera from animals that had undergone a nonspecific acute inflammatory reaction and to a lesser extent, normal bovine sera also cross-react with them. They are thus isoantigens resulting from the inflammation induced by Babesia or by other agents. In vaccination trials they were shown to be non-protective. This finding highlights the fallacy that screening with antibodies per se must identify protective parasite antigens. The finding that these i m m u n o d o m i n a n t antigens are non-protective further reinforces the fact that the only way to identify protective antigens is the arduous fractionation/vaccination/challenge methodology using an array of fractionation procedures. This may be time consuming and expensive; it is also a successful procedure. It is interesting to speculate on why these parasites produce immunodominant proteins a n d / o r proteins with multiple repeat epitopes. In all likelihood, these antigens have evolved to 'decoy' the host's protective i m m u n e response (Goodger et al., 1986). The fact that these antigens are non-protective a n d / or immunosuppressive lends credence to this. It is also of interest that protective functional antigens such as 21 B4 and 11C5 also contain repetitive nonprotective domains. It is possible that these too have evolved in order to divert the host's response away from the conserved, functionally vital portion of the antigen, thus ensuring the survival of the parasite. This aspect of the protective response is only now beginning to be studied and should become a priority research area in future. Enhancement of the protective response to these functional molecules preferably using only those fragments which play an essential role in the parasite's life cycle is likely to emerge as the most important i m m u n e modulatory strategy in future. This is certainly a priority of our group.
8
I.G. WRIGHT ET AL.
INDUCTION OF IMMUNITY WITH RECOMBINANT ANTIGENS
Few studies have been reported where recombinant antigens have been used to vaccinate cattle against B. bovis. Using hyperimmune serum to screen a cDNA library, Gill et al. ( 1987 ) and T i m m s et al. ( 1988 ) isolated a fragment of a large i m m u n o d o m i n a n t antigen which was found to be non-protective in cattle. There are also two reports of the successful vaccination of cattle with recombinant B. bovis antigens (Gale et al., 1990; Riddles et al., 1990). Both of these antigens have been the subject of intensive animal trials at CSIRO over the past 3 years. The first protective antigen cloned and expressed at CSIRO was the 11 C5 antigen (Gale et al., 1990). Initially this was expressed as a ~-galactosidase (/~-gal) fusion protein in 2GT 11. It was found that 10 pg of the fusion peptide (approximately 280 kDa) in either crude Escherichia coli lysate or as a purified acrylamide gel band induced high levels of protection in cattle. Parasitaemias of vaccinated cattle, when challenged with a heterologous strain of B. bovis, were suppressed by more than 95% relative to controls. Five of six controls required treatment while only one of six vaccinated cattle was treated. Injection of/?-gal alone had no effect on the outcome of the disease. Subsequently, the antigen was recloned into the pGEX plasmid as a glutathione-Stransferase (GST) fusion peptide (Smith and Johnson, 1988). This vector had the advantage of high expression rates (the antigen was about 20% of the total protein expressed) and in principle, enabled the production of pure preparations using glutathione agarose affinity chromatography. This GST11C5 fusion peptide was later shown to be as protective as the/~-gal 11C5 antigen. A further trial with this antigen indicated that large amounts (more than 50 pg) of the antigen actually suppressed the protective response. A series of adjuvant trials were conducted and it was shown that commercial grade saponin (from Quillaja bark) induced a protective response similar to that of Freund's Complete Adjuvant (FCA). A later trial showed that 1-2 mg of the adjuvant Quil A induced even stronger levels of protection than did saponin itself. Consequently, Quil A has become the adjuvant of choice. Recently, a trial was conducted in which the small unique C terminal region of the GST11 C5 antigen and the large region containing numerous repeat sequences were tested individually, as well as in various combinations. Although the data were not conclusive, the unique region appeared to induce high levels of protection when used in combination with small amounts of the repeat region, whereas the repeat region when used alone was not protective. Neither fragment when used alone was as effective as the whole fusion protein itself. The second protective antigen to be cloned and expressed at CSIRO was the 12D3 antigen (Riddles et al., 1992). This was produced by screening a cDNA library with a synthetic oligonucleotide based on seven amino acids in
DEVELOPMENT
OF A RECOMBINANT
9
BABESIA V A C C I N E
the N terminal portion of the native 12D3 antigen. This antigen was also initially cloned as a fl-gal fusion but then subcloned into pGEX vectors. The subsequent GST- 12D3 fusion protein was identified by Western blotting with the MAb A 12D3. This protein had a babesial component identical in length to the native 12D3 protein (38 kDa). This fusion peptide was insoluble and was purified by first centrifuging the insoluble material and then solubilising it using sodium dodecyl sulphate (SDS) and 2-mercaptoethanol (2-ME): 50 /~g of this material was then mixed with FCA and used to vaccinate cattle. Immunised animals had parasitaemias one order of magnitude lower than controls when challenged with a heterologous strain of B. bovis. In a series of experiments, 100% of controls were treated, whereas only 16% of the GST12D3 vaccinated animals required treatment. As with the 11 C5 antigen, the 12D3 antigen was shown in later trials to perform optimally with Quil A as adjuvant. Pilot scale production has been carried out for both GST-11 C5 and GST12D3 fusion proteins: 30 1 fermenter cultures of recombinant E. coli were grown and then processed by ion-exchange chromatography or SDS/2-ME solubilisation to obtain purified GST-11 C5 or GST-12D3, respectively. Using these processes, gram quantities of the recombinant antigens have been prepared for vaccination trials. A trial utilising both 11C5 and 12D3 fusion proteins in saponin was performed. On challenge, the parasitaemias in the vaccinated group were 99% lower than in controls. All the controls, but none of the vaccinated animals required treatment (Fig. 1 ). This combination was then tested under field conditions. Forty percent of control animals and 10% of vaccinated animals
/
30
o 25 ,~O "~
20
---.
Control ,,
.....
~
( )
12D3-11C5
/
Non-treated
~
( ,5~
~2 1~ x ~~ ] 0 i/1
.~ ~
Q.
~
~ 0
i
~
6
7
~ ...... 8
9
~. . . . . . 10
r ' ~ 11
~5 5i 12
Days
Fig. 1. Mean daily parasitaemiasof adult cattle vaccinatedwith the GST-11C5 and GST-12D3 antigens, and of unvaccinatedcontrols after challengewith blood inoculationof 1× 108B. bovis virulent heterologousorganisms.
10
I.G. WRIGHT ET AL.
required treatment after receiving tick-transmitted B. bovis infections. The mean maximum parasitaemias were 2154 _+3065 ~tl-~ blood and 232 _+345 #1 -~ blood (Fig. 2) and the mean maximum haematocrit falls were 45.5 + 10.6% and 30.1 _+14.7% for the control and vaccinated groups respectively. Data from animals upon initial vaccination with the live attenuated vaccine were 1205 _+ 1548 parasites ~1-~ blood and 25.8 _+9.7% mean maximum haematocrit fall; 5% of animals required treatment. This initial field trial, which used 70 animals, was extremely encouraging, especially as the results obtained with two recombinant antigens were almost comparable to the results obtained with initial vaccination with the live vaccine. It has been reported that up to 5% of animals vaccinated with the live vaccine require treatment upon exposure to field challenge (Wright, 1990 ). In order to achieve the degree of protection induced by naturally acquired infection, we are further researching other antigens to combine with the existing bivalent 11C5-12D3 vaccine. One of these is the protease-associated antigen (21 B4 ). A short cDNA construct has been identified using MAb T21 B4 and cloned. This was subcloned into pGEX and the resulting soluble, highyield fusion protein (20% of total protein) was purified by glutathione agarose chromatography. Twenty-five micrograms of the fusion peptide in Quil A adjuvant induced significant protection in cattle. The mean maximum parasitaemias were lower than in controls ( 1112 _+395 ~1- ~blood vs. 3331 _+ 1810 ~1-~ blood) (R. Casu, unpublished data, 1990). This antigen is now being further evaluated as a longer construct and in addition, the short fragment is also being tested in combination with the bivalent 11 C5-12D3 vaccine.
40
~,~
3O
~ 0
~) ~ ~
PO
~ ~ ~.
~o
3
The development of a recombinant Babesia vaccine
I . G . W r i g h t a, R. C a s u a, M . A . C o m m i n s a, B.P. D a l r y m p l e a, K . R . G a l e a, B.V. G o o d g e r a, P . W . R i d d l e s a, D.J. W a l t i s b u h l a, I. A b e t z ", D . A . B e r r i e a, Y. B o w l e s a, C. D i m m o c k ~, T. H a y e s a, H. K a l n i n s b, G . L e a t c h a, R. M c C r a e ~, P.E. M o n t a g u e c, I.T. N i s b e t b, F. P a r r o d i a, J . M . P e t e r s ", P.C. S c h e i w e a, W . S m i t h a, K. R o d e - B r a m a n i s a a n d M . A . W h i t e ~ aCS1R 0 Division of Tropical Animal Production, Long Pocket Laboratories, Private Bag, No. 3 P. 0., Indooroopilly, Qld. 4068, Australia bCommonwealth Serum Laboratories, 45 Poplar Road, Parkville, Vic. 3052, Australia cPitman-Moore Australia, Princes Highway, Werribee, Vic. 3030, Australia
ABSTRACT Wright, I.G., Casu, R., Commins, M.A., Dalrymple, B.P., Gale, K.R., Goodger, B.V., Riddles, P.W., Waltisbuhl, D.J., Abetz, I., Berrie, D.A., Bowles, Y., Dimmock, C., Hayes, T., Kalnins, H., Leatch, G., McCrae, R., Montague, P.E., Nisbet, I.T., Parrodi, F., Peters, J.M., Scheiwe, P.C., Smith, W., Rode-Bramanis, K. and White, M.A., 1992. The development of a recombinant Babesia vaccine. Vet. Parasitol., 44:3-13. Crude extracts of Babesia bovis parasites were shown to induce levels of protection in susceptible cattle equivalent to that resulting from natural infection. The crude material was systematically fractionated and tested in numerous sequential vaccination/challenge experiments in adult cattle. Antigens in protective fractions were then purified by affinity chromatography with monoclonal antibodies. Three highly protective (more than 95% reduction in parasitaemias) antigens were thus identified. None of these antigens was immunodominant; a number of immunodominant antigens were identified and all were immunosuppressive and/or non-protective. The three protective antigens were cloned and expressed as either/?-galactosidase or glutathione-Stransferase (GST) fusion proteins. Two of these, GST-12D3 and GST-11 C5, when used in combination were almost as protective as has been previously shown for the commercially available live attenuated vaccine. A short fragment of a third antigen (21B4) has also been shown to be protective. In two of the antigens, repetitive segments have been shown to be non-protective while the third antigen ( 12D3 ) does not contain repetitive domains. Homologues of these antigens exist in other Babesia species and it is anticipated that these may be candidate antigens for protective vaccines against those species. Correspondence to: I.G. Wright, CSIRO Division of Tropical Animal Production, Long Pocket Laboratories, Private Bag, No. 3 P.O., Indooroopilly, Qld. 4068, Australia.
© 1992 Elsevier Science Publishers B.V. All rights reserved 0304-4017/92/$05.00
4
I.G. W R I G H T ET AL.
INTRODUCTION
Babesia bovis is a tick-transmitted haemoprotozoan which is a pathogen of major significance in cattle. The parasite causes large economic losses in Africa, southeast Asia, Australia and South America. A live attenuated vaccine has been developed in Australia (Callow et al., 1979 ) but it is little used outside that country. The vaccine has a number of limitations, including short shelf life (5-7 days at 5 °C), reversion to virulence and to transmission by the tick vector and the possibility of co-transmission of other infectious agents, especially viruses (Wright and Riddles, 1989 ). As well as the live attenuated vaccine, cell culture-derived exoantigens have also been extensively studied in an attempt to develop non-living vaccines. In the case of B. bovis, the microaerophilous stationary phase (MASP) method for the continuous propagation of the parasite has resulted in the production of large amounts of exoantigens in the culture medium (Levy and Ristic, 1980). However, reports of the efficacy of this material are quite varied, with one group claiming a high degree of protection (Montenegro-James et al., 1989) while another (Timms et al., 1983) reported that this material induced no protection whatsoever. Short-term MASP cultures of Babesia bigem ina have also been reported (Montenegro-James et al., 1989); the exoantigens induced good levels of protection. A modified MASP system has also been used to develop a commercial Babesia canis vaccine in France (Moreau et al., 1988). This product 'Pirodog' is used extensively in that country and gives about 80% protection when dogs are revaccinated annually (Y. Moreau, personal communication, 1990). Because of the inherent problems that exist with the live vaccine and the variable results which have been obtained with MASP-derived material, our group has been researching for a number of years the possibility of developing a stable, highly protective, reproducible and cheap vaccine using recombinant DNA technologies. This has been based on a pragmatic approach whereby protective antigens are identified by in vivo vaccination/challenge experiments rather than by immunochemical identification alone. The work has now reached the development stage whereby defined antigens are undergoing trials to meet registration requirements prior to commercial release. This paper gives an account of the work that led up to this point. IMMUNITY TO BABESIA INFECTIONS
Immunity to B. bovis is lifelong, once an animal recovers from a primary infection (Mahoney et al., 1973, 1979b). Sterile immunity does not occur in B. bovis infections and infected cattle undergo a continuous series of parasite recrudescences (Mahoney et al., 1973). These may well serve to stimulate immune responses continuously, and high antibody levels can be detected 4
DEVELOPMENT OF A RECOMBINANT BABESIA VACCINE
5
years after infection. In B. bigemina infections, however, sterile immunity does occur (Mahoney et al., 1973 ) and a strong protective response is elicited without parasite recrudescences. In contrast to B. bovis infections, antibodies can generally not be detected 18-21 months after infection even though the animals are protected against subsequent infection. The nature of the immune response in these different biological phenomena is not understood. An important finding was made by Mahoney (1967a) and later by Mahoney et al. (1979a) who demonstrated that passive transfer of IgG from animals made hyperimmune by repeated B. bovis infections could protect susceptible cattle against challenge by that parasite. These studies also strongly supported the idea that effector cells were necessary for protection, as, in the absence of effector cells, the IgG fraction had no direct effect on B. bovis parasites. The IgM antibodies were not involved in protection. This indicated that it was highly likely that stimulation of B cells and effector cells by a vaccine should result in protective immunity. Recently in our laboratory (Jacobson et al., 1992) we have developed an in vitro opsonisation assay using mouse macrophages, to assess the influence of anti-B, bovis antisera on erythrocytes infected with B. bovis and have confirmed that effector cells play an essential role in protective immunity (Mahoney et al., 1979a). It is interesting that Mahoney et al. (1980) implicated complement as a factor in the immune response to babesiosis, but Jacobson et al. (1992), using in vitro phagocytosis did not. I NDUC T I ON OF IMMUNITY WITH PARASITE EXTRACTS
Immunity to B. bovis by vaccination with killed organisms was first demonstrated by Mahoney (1967b) and later by Mahoney and Wright (1976). These studies showed that a degree of protection was induced which was similar to that of naturally acquired immunity. This indicated that it was feasible to develop a killed vaccine against this parasite. The systematic fractionation of B. bovis extracts to identify protective components by our group began around 1980. Initially, a whole range of chromatographic and electrophoretic procedures were utilised at each stage of purification. The fractions were then tested in calves in vaccination/challenge trials. A total of about 80 such trials was conducted. The protective status of a particular antigen was mainly determined by the pathophysiological response of the vaccinated animal upon challenge (Goodger et al., 1981 ). A large number of pathophysiological parameters were identified which were accurate indices of the severity of the disease. These included the presence in the serum of cryofibrinogen, haptoglobin, immunoconglutinin and haemoglobin and reduced levels of conglutinin, haemolytic complement, C3, coagulation factors, fibronectin, as well as lymphocytes and thrombocytes. Collectively, marked changes in these parameters indicated that
6
I.(3. WRIGHT ET AL.
the animal had not been protected by vaccination, and that particular antigen was not further studied. The converse was equally true. In studies where little or no change occurred with these pathophysiological indices, a strong protective response was confirmed. This pragmatic approach to antigen identification led to a number of highly protective fractions being isolated. These fractions were then subjected to further subfractionation using affinity chromatography with monoclonal and polyclonal antisera or with various ligands including protease inhibitors as well as various electrophoretic procedures. Three antigens were purified and identified for intensive study. The first of these was an antigen designated 15B1 (after the identifying monoclonal antibody (MAb) A15B1 ) (Wright et al., 1983, 1985). This antigen was subsequently redesignated 12D3 (after the more stable IgG1 isotype MAb which cross-reacted with this molecule). This antigen was of low molecular weight and was present in low concentration. Two injections, at monthly intervals, of 10/~g of 12D3 with adjuvant, reduced parasitaemias by over one order of magnitude (as compared with controls when the cattle were challenged with an heterologous strain of B. bovis) (Riddles et al., 1992 ). A second fraction, designated fl, was found to be highly protective (Goodger et al., 1985b). The components of the fl fraction were separated by affinity chromatography using the MAbs produced against this fraction. Eight fractions were subsequently tested by vaccination/challenge trials in calves before a protective antigen was identified. This antigen was designated 11C5 after the MAb W 11 C5, was of high molecular weight and was also present in low concentration. A third antigen, designated protease, was first identified by substrate electrophoresis whereby proteolytic activity was demonstrated in acrylamide gels in which a substrate such as gelatin was incorporated (Commins et al., 1985 ). The gel band was excised and shown to induce strong protection in calves. A range of MAbs was then raised against this gel band fraction and used to further enrich the antigen by affinity chromatography. One of these MAbs (T2 I B4) identified an antigen of 60-70 kDa which had strong proteolytic activity and which also induced protection when injected into calves. This antigen, like the other two, was only present in minute amounts. All three antigens are present as soluble proteins and all appear to be secreted by the parasite. Furthermore, preliminary analyses of the proteins indicate that they all play a functional rather than a structural role in the parasite's life cycle. Subsequently, crude extracts ofB. bigemina have been used to induce strong protective responses against challenge with that organism (Wright et al., 1987 ). Similar immunity to subsequent challenge with Babesia ovis organisms has also been induced in sheep injected with crude B. ovis extracts (A1abay et al., 1987). In the course of these two studies some cross-protection was elicited by B. bigernina extracts against B. bovis challenge and by B. bovis
DEVELOPMENT OF A RECOMBINANT BABES1A VACCINE
7
extracts against B. ovis challenge, indicating that a degree of antigenic commonality exists between the various species. THE ROLE OF IMMUNODOMINANTANTIGENSIN THE IMMUNE RESPONSE During the course of this work numerous abundant, i m m u n o d o m i n a n t antigens were isolated. These included the 70 kDa immunodiffusion (ID) antigen (Goodger et al., 1986), the dominant 200 kDa antigen (B.V. Goodger, unpublished data, 1988) and an antigen of more than 1500 kDa (the 3C1 antigen) (Wright et al., 1983 ). In all instances these antigens, whilst inducing high antibody titres, were non-protective. Furthermore, in the case of the 3C 1 antigen, the protective response to the 15B 1/ 12D3 antigen was abolished when these two antigens were used together. A second group of so-called Babesia antigens has also been identified (Goodger et al., 1985a). These antigens were recognised on Western blots as dominant bands when probed with B. boris-positive antisera. However, they were also recognised avidly by sera from animals that had undergone a nonspecific acute inflammatory reaction and to a lesser extent, normal bovine sera also cross-react with them. They are thus isoantigens resulting from the inflammation induced by Babesia or by other agents. In vaccination trials they were shown to be non-protective. This finding highlights the fallacy that screening with antibodies per se must identify protective parasite antigens. The finding that these i m m u n o d o m i n a n t antigens are non-protective further reinforces the fact that the only way to identify protective antigens is the arduous fractionation/vaccination/challenge methodology using an array of fractionation procedures. This may be time consuming and expensive; it is also a successful procedure. It is interesting to speculate on why these parasites produce immunodominant proteins a n d / o r proteins with multiple repeat epitopes. In all likelihood, these antigens have evolved to 'decoy' the host's protective i m m u n e response (Goodger et al., 1986). The fact that these antigens are non-protective a n d / or immunosuppressive lends credence to this. It is also of interest that protective functional antigens such as 21 B4 and 11C5 also contain repetitive nonprotective domains. It is possible that these too have evolved in order to divert the host's response away from the conserved, functionally vital portion of the antigen, thus ensuring the survival of the parasite. This aspect of the protective response is only now beginning to be studied and should become a priority research area in future. Enhancement of the protective response to these functional molecules preferably using only those fragments which play an essential role in the parasite's life cycle is likely to emerge as the most important i m m u n e modulatory strategy in future. This is certainly a priority of our group.
8
I.G. WRIGHT ET AL.
INDUCTION OF IMMUNITY WITH RECOMBINANT ANTIGENS
Few studies have been reported where recombinant antigens have been used to vaccinate cattle against B. bovis. Using hyperimmune serum to screen a cDNA library, Gill et al. ( 1987 ) and T i m m s et al. ( 1988 ) isolated a fragment of a large i m m u n o d o m i n a n t antigen which was found to be non-protective in cattle. There are also two reports of the successful vaccination of cattle with recombinant B. bovis antigens (Gale et al., 1990; Riddles et al., 1990). Both of these antigens have been the subject of intensive animal trials at CSIRO over the past 3 years. The first protective antigen cloned and expressed at CSIRO was the 11 C5 antigen (Gale et al., 1990). Initially this was expressed as a ~-galactosidase (/~-gal) fusion protein in 2GT 11. It was found that 10 pg of the fusion peptide (approximately 280 kDa) in either crude Escherichia coli lysate or as a purified acrylamide gel band induced high levels of protection in cattle. Parasitaemias of vaccinated cattle, when challenged with a heterologous strain of B. bovis, were suppressed by more than 95% relative to controls. Five of six controls required treatment while only one of six vaccinated cattle was treated. Injection of/?-gal alone had no effect on the outcome of the disease. Subsequently, the antigen was recloned into the pGEX plasmid as a glutathione-Stransferase (GST) fusion peptide (Smith and Johnson, 1988). This vector had the advantage of high expression rates (the antigen was about 20% of the total protein expressed) and in principle, enabled the production of pure preparations using glutathione agarose affinity chromatography. This GST11C5 fusion peptide was later shown to be as protective as the/~-gal 11C5 antigen. A further trial with this antigen indicated that large amounts (more than 50 pg) of the antigen actually suppressed the protective response. A series of adjuvant trials were conducted and it was shown that commercial grade saponin (from Quillaja bark) induced a protective response similar to that of Freund's Complete Adjuvant (FCA). A later trial showed that 1-2 mg of the adjuvant Quil A induced even stronger levels of protection than did saponin itself. Consequently, Quil A has become the adjuvant of choice. Recently, a trial was conducted in which the small unique C terminal region of the GST11 C5 antigen and the large region containing numerous repeat sequences were tested individually, as well as in various combinations. Although the data were not conclusive, the unique region appeared to induce high levels of protection when used in combination with small amounts of the repeat region, whereas the repeat region when used alone was not protective. Neither fragment when used alone was as effective as the whole fusion protein itself. The second protective antigen to be cloned and expressed at CSIRO was the 12D3 antigen (Riddles et al., 1992). This was produced by screening a cDNA library with a synthetic oligonucleotide based on seven amino acids in
DEVELOPMENT
OF A RECOMBINANT
9
BABESIA V A C C I N E
the N terminal portion of the native 12D3 antigen. This antigen was also initially cloned as a fl-gal fusion but then subcloned into pGEX vectors. The subsequent GST- 12D3 fusion protein was identified by Western blotting with the MAb A 12D3. This protein had a babesial component identical in length to the native 12D3 protein (38 kDa). This fusion peptide was insoluble and was purified by first centrifuging the insoluble material and then solubilising it using sodium dodecyl sulphate (SDS) and 2-mercaptoethanol (2-ME): 50 /~g of this material was then mixed with FCA and used to vaccinate cattle. Immunised animals had parasitaemias one order of magnitude lower than controls when challenged with a heterologous strain of B. bovis. In a series of experiments, 100% of controls were treated, whereas only 16% of the GST12D3 vaccinated animals required treatment. As with the 11 C5 antigen, the 12D3 antigen was shown in later trials to perform optimally with Quil A as adjuvant. Pilot scale production has been carried out for both GST-11 C5 and GST12D3 fusion proteins: 30 1 fermenter cultures of recombinant E. coli were grown and then processed by ion-exchange chromatography or SDS/2-ME solubilisation to obtain purified GST-11 C5 or GST-12D3, respectively. Using these processes, gram quantities of the recombinant antigens have been prepared for vaccination trials. A trial utilising both 11C5 and 12D3 fusion proteins in saponin was performed. On challenge, the parasitaemias in the vaccinated group were 99% lower than in controls. All the controls, but none of the vaccinated animals required treatment (Fig. 1 ). This combination was then tested under field conditions. Forty percent of control animals and 10% of vaccinated animals
/
30
o 25 ,~O "~
20
---.
Control ,,
.....
~
( )
12D3-11C5
/
Non-treated
~
( ,5~
~2 1~ x ~~ ] 0 i/1
.~ ~
Q.
~
~ 0
i
~
6
7
~ ...... 8
9
~. . . . . . 10
r ' ~ 11
~5 5i 12
Days
Fig. 1. Mean daily parasitaemiasof adult cattle vaccinatedwith the GST-11C5 and GST-12D3 antigens, and of unvaccinatedcontrols after challengewith blood inoculationof 1× 108B. bovis virulent heterologousorganisms.
10
I.G. WRIGHT ET AL.
required treatment after receiving tick-transmitted B. bovis infections. The mean maximum parasitaemias were 2154 _+3065 ~tl-~ blood and 232 _+345 #1 -~ blood (Fig. 2) and the mean maximum haematocrit falls were 45.5 + 10.6% and 30.1 _+14.7% for the control and vaccinated groups respectively. Data from animals upon initial vaccination with the live attenuated vaccine were 1205 _+ 1548 parasites ~1-~ blood and 25.8 _+9.7% mean maximum haematocrit fall; 5% of animals required treatment. This initial field trial, which used 70 animals, was extremely encouraging, especially as the results obtained with two recombinant antigens were almost comparable to the results obtained with initial vaccination with the live vaccine. It has been reported that up to 5% of animals vaccinated with the live vaccine require treatment upon exposure to field challenge (Wright, 1990 ). In order to achieve the degree of protection induced by naturally acquired infection, we are further researching other antigens to combine with the existing bivalent 11C5-12D3 vaccine. One of these is the protease-associated antigen (21 B4 ). A short cDNA construct has been identified using MAb T21 B4 and cloned. This was subcloned into pGEX and the resulting soluble, highyield fusion protein (20% of total protein) was purified by glutathione agarose chromatography. Twenty-five micrograms of the fusion peptide in Quil A adjuvant induced significant protection in cattle. The mean maximum parasitaemias were lower than in controls ( 1112 _+395 ~1- ~blood vs. 3331 _+ 1810 ~1-~ blood) (R. Casu, unpublished data, 1990). This antigen is now being further evaluated as a longer construct and in addition, the short fragment is also being tested in combination with the bivalent 11 C5-12D3 vaccine.
40
~,~
3O
~ 0
~) ~ ~
PO
~ ~ ~.
~o
