Isotyping of Immunoglobulin G Responses of Ruminants and Mice to Live and Inactivated Antigens of Cowdria ruminantium the Causative Agent of Cowdriosis in Ruminants

: The Immunoglobulin IgG 1 and IgG 2 isotype immune responses of domestic ruminants and mice to Cowdria. ruminantium live infection or by immunization with inactivated organisms were determined by the enzyme linked immunosorbent assay and Western blotting. Immunization of goats with inactivated elementary bodies (IEBs) led to a predominant IgG 1 isotype response. This indicated that a Th2 response was induced. After challenge, the IgG isotype responses were mixed whereby both IgG 1 and IgG 2 antibodies were detected. Two goats that survived virulent challenge had a predominant IgG 2 isotype response. In cattle live infection by natural challenge or experiment led to a predominant IgG 1 isotype response. Immunization of cattle with IEBs however led to mixed IgG responses characterized by similar IgG 1 and IgG 2 ratios. In the mouse live infection led to a predominant IgG 2 isotype response. This indicated the mouse developed a true Th1 type cell mediated immune response when inoculated with live organisms. Immunization with inactivated organisms on the other hand led to a dominant IgG 1 response. It is evident from this work that the immune responses of ruminants and mice to C. ruminantium are different and that using mice as the experimental model for immune responses to Cowdria ruminantium . is not the appropriate. (Asian-Aust. J. Anim. Sci. 2003. Vol 16, No. 4 : 541-548)


INTRODUCTION
Heartwater (Cowdriosis) is an infectious virulent tickborne disease of domestic and wild ruminants caused by the rickettsial organism Cowdria ruminantium. Animals recovering from natural or experimental infection develop antibodies to this organism (Neitz et al., 1986;Semu et al., 1992). Detection of the antibody responses of ruminants naturally infected with live C. ruminantium or those inoculated with inactivated elementary bodies (EBs) are carried out by a number of serological tests. These include the indirect fluorescent antibody test (IFAT) (Du Plessis et al., 1987). The enzyme-linked immunosorbent assay (ELISA) either as an indirect assay (Soldan et al., 1993) or as a competitive assay (Jongejan et al., 1991). Western blotting (Rossouw et al., 1990) has been used to identify antigenic components of the organism.
Four immunoglobulin classes occur in the ruminant, i.e IgG, IgM, IgA and IgE. Within the IgG class, are two major subclasses IgG 1 and IgG 2 of which IgG 1 is the predominant in sheep (McQuire, et al., 1979). Immunization of sheep with protein antigens generates antibody responses of both isotypes (Bird et al., 1995). Immunization of mice with soluble protein antigens leads to two types of CD4 + T helper lymphocyte responses referred to as T helper1 (Th1) and T helper2 (Th2) that are differentiated by the cytokines they produced (Mosmann and Coffman, 1989). In Th1 type responses CD4 + T cells produce interferon gamma (IFN-γ), interleukin-2 (IL-2) and tumor necrosis factor alpha (TNF-α). Their effect is to promote cell-mediated immunity and support production of IgG 2 antibodies. Immunity to intracellular bacteria is thought to be mediated principally by cell-mediated immunity. The IgG isotype antibody, which is associated with these infections, would therefore be mainly IgG 2 .
C. ruminantium is an intracellular rickettsial organism, which is found in the cytoplasm of endothelial cells. Protective immune response to this organism has been suggested to be predominantly cell-mediated since passive transfer of antibodies do not confer protection against challenge. The objective of this study was to examine the IgG 1 and IgG 2 responses of cattle, sheep, goats and mice to C. ruminatium elememntary body antigens following experimental infection by infection and treatment (I/T), field challenge, and immunization with inactivated elementary bodies (IEBs) or recombinant antigens. This was to establish the predominant IgG isotype produced after natural or experimental challenge and by immunization with inactivated organisms. This would indicate the type of immune response induced in the host and further elucidate if the immune response to C. ruminantium is principally cell mediated or mixed type response.

Antigens
The antigens used for Western blotting and ELISA tests were from elementary bodies (EBs) of C. ruminantium (Welgevonden stock). The organisms were propagated in bovine endothelial cell as described by Pow et al. (1993). Two recombinant antigens, (Ags) a 58 kDa Heat shock protein (58 kDa Hsp) and its subclone a 35 kDa protein were also used. The 35 kDa recombinant protein was expressed in E. coli and purified by a standard method then used for ELISA tests to detect IgG 1 and IgG 2 responses of mice and sheep immunized with the 58kDa Hsp recombinant Ag.
Sera : Sera from five groups of goats, two groups of sheep, three groups of cattle and three groups of mice were examined for Cowdria specific IgG antibodies by Western blotting and ELISA.
Caprine sera : Sera from five groups of goats (C1, to C5 with six goats each) were tested. The goats in the first group (C1) received primary inoculations of inactivated elementary bodies (IEBs) only and those in the second group (C2) were inoculated 2 times with IEBs 21 days apart. The antigens used for primary immunisation were mixed with equal volumes of Freund complete adjuvant (FCA) and inoculated intramuscularly. Antigens used for booster inoculations were mixed with equal volumes of Freund incomplete adjuvant (FIA) or Phosphate buffered saline (PBS) and inoculated 21 days later. Goats in the third group (C3) received detergent extracted elementary body (EB) antigens for both primary and booster inoculations.
Sera were collected sequentially from each group at days 0, 7, 14, 21, 28, and day 34 post inoculation (p.i.). All the goats in the above groups were challenged by intravenous inoculation of 5 ml of virulent blood stabilate of the Welgevonden stock at 36 days pi. Then detection of Cowdria specific IgG 1 and IgG 2 antibodies was carried out on sera collected after immunization (day 34 pi) and sera collected at day 64 pi (day 28 post challenge) from 4 survivors two each from group C1 (G614, G668) and C2 (G476, G601).
Sera in the fourth group (group C4) were collected from two infection treatment (I/T) goats. Two serum samples collected on day 28 and day 80 after infection were examined for Cowdria specific IgG 1 and IgG 2 antibodies. Sera in the fifth group (group C5) were from 2 goats (G107, G108) immunized with IEB but not challenged.
Ovine sera : Two groups of sheep sera were tested. In the first group (S1) 6 sera were tested. Three sera each originated from three sheep infected with the Mara and Nonile stock of C. ruminantium respectively. They were obtained at day 0, 14 and 28 after infection. In the second group (S2) were sequential sera obtained on day 0, 14, 21, 28, 64, and day 143 pi from 6 sheep immunized with 58 kDa Heat shock protein (Hsp) recombinant antigen. These sera were examined for Cowdria specific IgG 1 and IgG 2 responses to recombinant 35 kDa sub-clone antigen of the 58 kDa Hsp of C.ruminantium with their respective controls.
Bovine sera : Sera from three groups of cattle were examined for Cowdria specific IgG 1 and IgG 2 antibodies as follows: Group B1. These sera were raised by experimental infection using blood stabilates. Five cattle from which the sera were obtained were immunised by I/T with 3 stocks of C. ruminantium, Ball 3 and Mara stocks (2 cattle each) and Kwanyanga (one cow).
Group B2. There were 10 sera in this group collected from cattle (B196, B199, B200 to B206, B210 and B211) in a heartwater endemic area of South Africa. Group B3. In the third group were thirteen sera from Kenyan cattle. Nine of the sera were obtained from 3 cattle (S295, S296, and S304) immunized with IEBs of the Gardel stock of C. ruminantium then challenged with virulent homologus organism. They were collected on day 0, 36, pi and day 31 post challenge (pc) respectively. The remaining 4 sera were obtained from two control cattle (S234 and S264) collected on day 0 and day 11 pc.
Mouse sera : Sera from 4 I/T out-bred TO white mice and another 5 mice immunized with IEBs were collected at day 34 pi were examined for IgG 1 and IgG 2 antibody responses to C. ruminantium. Additionally pooled sequential mouse sera (day 0, 14, 21 and 34 pi) obtained from mice which had been immunized with a recombinant 58kDa Hsp were also examined for IgG 1 and IgG 2 antibodies to a recombinant 35kDa subclone protein of C. ruminantium by ELISA.
Monoclonal antibodies : The monoclonal antibodies (MoAb) used to determine sheep and goat IgG 1 and IgG 2 responses were rat anti-sheep IgG 1 (IRS) and IgG 2 (2RS). Their optimal working dilutions was 1/20 in phosphate buffer saline containing 0.02%Tween 20 to which was added 4% normal rabbit serum in (PBST/4%NRS). For the mice tests, rat-anti-mouse IgG 1 and IgG 2 monoclonal antibodies (Serotech, UK) were used. For the bovine tests, sheep anti-bovine IgG 1 HRP conjugated polyclonal serum was obtained from Bethyl Laboratories (USA) used at a dilution of 1/50,000 to detect IgG 1 . Mouse anti-bovine IgG 2 obtained from Sigma (Sigma UK) was used at a dilution of 1/10,000 to detect bovine IgG 2. All the working dilutions for the bovine ELISA were carried out in phosphate buffered saline containing 0.02%Tween 20 to which was added 4% normal goat serum in (PBST/4%NGS)).

Detection of IgG 1 and IgG 2 isotype responses of goats, sheep and mice and cattle infected with live or immunized with inactivated C. ruminantium antigens by indirect ELISA
The method used to determine isotype responses was a modification of that described by Soldan et al., (1993). Optimal working dilutions of each monoclonal antibody (MoAb) and antigens were determined by checkerboard titration.
Antigens : The antigen used in ELISA tests to detect IgG 1 and IgG 2 responses of goats, sheep and mice were from the Welgevonden stock of C. ruminantium prepared as described by Soldan et al., (1993). The optimal dilution of this antigen used to coat the ELISA plates was 1/1000 in coating buffer (0.05M carbonate/bicarbonate buffer pH 9.6). For the bovine IgG 1 and IgG 2 tests the Gardel stock of C. ruminantium was used at a dilution of 1/4,000 in coating buffer for the Kenyan bovine sera while the Welgevonden stock was used at a dilution of 1/6,000 for the South African bovine sera.
Buffers : The blocking buffer for the goat, sheep and mouse tests was PBS containing 0.02%Tween 20 to which was added 4% normal rabbit serum in (PBST/4%NRS). While the blocking buffer for the bovine tests was PBST to which 4% normal goat serum was added. The washing buffer for all the tests was PBST containing 0.9% sodium chloride.
The ELISA tests carried as follows: Ninety six (96) well flat-bottomed microtiter ELISA plates (Immulon 1, Dynatech Laboratories) were coated with 50 µl of respective ELISA antigen in coating buffer The plates were covered with cling film and incubated at 4°C overnight. Then the contents were discarded and the plates were washed three times (3 min/wash) in 0.9% sodium chloride containing 0.05% (v/v) Tween 20 (0.9% NaCl/PBST). The plates were blocked with 100 µl per well of PBST/4% NRS for all sheep, goat and mice tests. For the bovine tests, the plates were blocked with 100 µl per well PBST/4% NGS. All the tests were incubated for 1 hour at room temperature and the buffer was discarded. Then, fifty microlitters (50 µl) of goat, sheep, and mice sera diluted to 1/50 in blocking buffer (PBST/4%NRS) were added to respective wells. The bovine sera were diluted to 1/800 in blocking buffer (PBST/4%NGS) then 50 µl were added to all the wells and the wells were topped up 100 µl with respective blocking buffer.
Detection of goat and sheep IgG 1 and IgG 2 responses was achieved by adding 50 µl of rat anti-sheep IgG 1 (1RS) and IgG 2 (2RS) monoclonal antibody (diluted to 1/20) to respective wells prior to incubation at 37°C for 1 h. While for the mice tests, rat anti-mouse IgG 1 and IgG 2 monoclonal antibodies (Serotech, UK) were used. After incubation the plates were washed 3 times as above. This was followed by the addition of 50 µl of goat anti-rat IgG whole molecule horse radish peroxidase (Sigma) conjugate (diluted 1/1,000) to all wells and incubated at 37°C for 1 h. Detection of bovine IgG 1 was achieved by using a 1/50,000 dilution of sheep anti-bovine IgG 1 horse radish peroxidase (HRP) conjugated ployclonal antibodies (Bethyl Labs UK) and detection of bovine IgG 2 was achieved by adding goat-antibovine IgG whole molecule HRP conjugate (Sigma) diluted to 1/2,000.
The plates were washed 3 times and 50 µl of the peroxidase substrate tetramethyl benzidine (TMB, Kirkegaard and Perry laboratories) was added to all wells. The reaction was stopped after 15 minutes by addition of 50 µl of 0.2M sulfuric acid. The absorbency (Optical density/ODs) at 450 nm was read using a plate reader (Multiscan Plus, Version 2.03, Labsystems). The mean of 2 tests was calculated, then the OD of the negative control was subtracted from this value to remove background absorbency. The results were expressed as the mean OD values of 2 tests

Detection of IgG 1 and IgG 2 isotype responses of immunized goats to C. ruminantium by Western blotting
Elementary bodies of the Welgevonden stock of C.ruminantium were used as antigens. SDS-PAGE and Western blotting was carried out as described by Lally et al (1995) with the following modifications to enable specific isotype responses to be detected: Optimal working dilutions of each monoclonal antibodies (MoAb) used in Western blot analysis were established to be 1/10 for 2RS and 1/20 for 1RS. Following blocking of the membranes and incubation with primary sera, the test blots were reacted with monoclonal antibodies to sheep IgG 1 and IgG 2 . They were then incubated for 60 minutes and color was developed in the usual way. The Molecular masses of the bands were estimated by a standard curve then the blots were photographed for permanent record.

Analysis of statistics
The student's t test was used to analyze for differences between IgG 1 and IgG 2 antibody responses in sera from IEB immunized goats which survived challenge with virulent homologous organism.
After live virulent challenge of the IEBs immunized goats, the level of IgG 2 isotype as measured by OD value in ELISA, rose to between 7 and 10 times their pre-challenge levels and IgG 1 ODs in post challenge sera increased by a factor of 3 to 6 times. Analysis for differences between IgG 2 responses pre-challenge and post-challenge indicate that the levels of IgG 2 had increased after challenge (p<0.05). The rise in IgG 2 had the result that the differences in IgG 1 and IgG 2 levels were no longer significantly different (p=0.06). After challenge the ratio between IgG 1 and IgG 2 responses had changed to between 2:1 to 1:1.6.

IgG 1 and IgG 2 antibodies in post immunization and post-challenge sera from IEBs immunized goats recognize specific antigens of the elementary body in Western blots
To confirm that IgG 1 and IgG 2 responses following IEB immunization were to Cowdria antigens and not to endothelial cell antigens, which contaminate IEB preparations, Western blotting was performed. Sera collected at day 28 from an I/T goat (G74) reacted strongly with EB antigens of molecular masses of 21 kDa, 24 kDa, 28kDa and 32kDa. A few antigens of higher molecular masses reacted faintly. These reactions were however only with IgG 1 (Figure 2, lanes 2 and 4) but not IgG 2 . The IgG 1 and IgG 2 responses of sera collected on day 34 pi from an IEB goat (G668, group C2) were the same as those of the I/T goat C4 (Figure 2 lanes 6 and 7). The very significant   Lenes 1 and 10 SDS low range molecular weight protein standards; Western blots were reacted with sera lenes 2 amd 3, IgG 1 positive day 28 sera from I/T goat (G74), lenes 4 and 5 day 28 I/T serum negative for IgG 2 , lanes 6 and 7 pre-challenge day 34 sera from IEB goat (G668) tested for IgG 1 ; lane 8 and 9 prechallenge sera from G668 tested IgG 2 . Note the apparent absence of IgG 2 in both I/T and pre-challenge IEB sera.

IgG 1 and IgG 2 responses of sheep immunized by infection and treatment and those inoculated with recombinant 58kDa antigen determined by ELISA
Infection of sheep with live virulent organisms in blood resulted in a predominant IgG 1 response and an apparent absence IgG 2 response. The IgG 1 and IgG 2 responses of sheep (group S2) immunized with recombinant 58 kDa Hsp were also characterized by a predominant IgG 1 response with a ratio of at least 2:1 and no detection of IgG 2 in 3 out of 6 animals.

IgG 1 and IgG 2 responses of cattle following live infection by experiment (I/T) or by field exposure
The Cowdria specific IgG 1 and IgG 2 responses to experimental infection of 7 cattle (group B1) with 3 different stocks of C. ruminantium were characterized by a dominant IgG 1 although IgG 2 was detectable above background. The IgG 1 : IgG2 ratio had a median of 5:1 (range 1:1 to 8:1, Table 1). The IgG 1 : IgG2 ratios of field sera from South Africa (group B2) were characterized by IgG 1 : IgG2 ratios greater than 1 (Table 2) with a range from 3:1 to 56:1 (mean 5:1). Except for one animal which had a ratio of 1:2. These ratios were similar to those obtained following experimental infection (Tables 1 and 2).

IgG 1 and IgG 2 responses of cattle following immunization with IEBs
Immunization of cattle with IEBs of C. ruminantium led to production of both similar amounts of IgG 1 and IgG 2 (Table 3, Figure 4). The level of IgG 2 was much higher after immunization with IEBs than that obtained in cases of the experimental infection and field sera (Figure 4, Table 3). The IgG 1 : IgG2 ratios in sera of three immunized cattle (S295, S296, S304) were the same before challenge and after challenge although there was an increase in OD values of both isotypes after challenge. In contrast to goats the increase in OD value of the two IgG isotypes was of a similar magnitude (Figure 4), and did not show greater increase in IgG 2 . The mean OD values of sera obtained before immunization of the immunized group and those from two control cattle (S234, S264) remained low and their IgG 1 : IgG 2 ratios were same or IgG 1 ratio was higher than 1.

IgG 1 and IgG 2 responses of mice to C. ruminantium after immunization with I/T, with IEBs or recombinant antigens
Immunization of mice with live C. ruminantium was observed to result in a dominant IgG 2 isotype response, in contrast to that obtained in live infection of ruminants. Sera obtained on day 34 pi. From 4 mice infected with live EBs had higher IgG 2 OD values in comparison to sera from mice inoculated with IEBs and collected at day 34 after immunization ( Figure 5). The IgG 1 : IgG 2 ratios of the sera from I/T mice was 1:2 and that of sera obtained after immunization with inactivated EB was 2:1. The antibody responses of mice immunized with recombinant 58 kDa Hsp were characterized by predominantly of IgG 1 isotype as determined using the 35 kDa antigen (Figure 6).

DISCUSSION
Immunization of goats with inactivated elementary bodies led to the development of stronger IgG 1 than IgG 2 responses pre-challenge. After challenge the ELISA OD values of both isotypes were 7 and 10 times those of prechallenge levels. The increase in IgG 2 titers after challenge was such that in two of four goats which survived challenge, the IgG 2 OD values were higher than those of IgG 1 , indicating that this isotype was being preferentially produced. The shift in antibody isotype from IgG 1 to IgG 2 or an increase in IgG 2 after challenge indicated that a Th1 type response was stimulated by the challenge infection. In a Th1 type response CD4 + lymphocytes secrete the cytokines IFN-γ, IL-2 and TNF-α, promoting cell-mediated immunity and IgG 2 production. These findings are in agreement with the evidence that stimulation of bovine B cells sorted for IgM with IFN-γ induces secretion of IgG 2 in vitro (Estes et al., 1995). In contrast immunization of cattle with IEB stimulated production of both IgG 1 and IgG 2 , and challenge of the immunized cattle led to a further increase in the titters of both isotypes. The differences in IgG isotype responses in goats and cattle to C. ruminantium may be associated with the greater resistance of cattle to heartwater (Uilenberg, 1983;Camus et al., 1996).
Sera from two I/T goats (Group 4, G69, G74) infected with the Welgevonden stock and treated to prevent death, had higher levels of IgG 1 responses. The reason for these exceptions is not clear. However it suggests that C. ruminantium may be inducing production of IL-4 or IL-10 at the beginning of infection which drives the immune responses to a Th2 type antibody responses rather than a Th1 type IgG 2 immune response.
Cattle immunized by I/T (group B1) and field sera (group B2) had dominant IgG 1 responses. It is possible that live infection and immunization by I/T does not lead to a classical Th1 response but to an IL-4/IL-10 driven response since bovine IL-4 upregulates production of IgG 1 , IgM and IgE in the presence of a variety of costimulators lgG 1 responses ( ) and IgG2 responses ( ). Results are express as the mean OD 450 nm of 2 tests. Error bars show standard error of the mean. (Estes et al., 1995). Immunization of cattle with IEBs induced a strong IgG 2 response in addition to IgG 1 , which suggests that a mixed Th1 type and Th2 type response was induced by inactivated organisms in contrast to live, or natural (tick mediated) infection. The results suggest live Cowdria shift responses towards Th2. The fact that natural infection and I/T may lead to an IL-4 driven Th2 type responses indicates that C. ruminantium may circumvent the host immune system by directing it towards a Th2 type response. However, I/T leads to solid immunity, which indicates that, other effector mechanisms such as cellmediated immune responses mediated by cytotoxic T cell, IFN-and NK cells are probably involved in protective immunity. Other workers have found that cytotoxic T cell populations are induced by Cowdria infection but only at a suprisingly late stage (Ben Said, Personal Communication). The IgG 1 response of mice immunized with inactivated EBs was characterized by higher titters of IgG 1 whereas that of I/T mice had higher IgG 2 titters. These findings are in agreement with those of Du Plessis et al. (1991) who showed that following I/T immunity is cell mediated and that CD8 + (Lyt2 + ) T cells were responsible for protection as indicated by adoptive transfer of immune cells to unimmunized mice. The situation in ruminants appears different in that in natural infections and I/T leads to a Th2 type response characterized by a dominant IgG 1 response.
The role of antibodies in immunity to heartwater is not clear. Experimental transfer of serum or gamma globulins by in vivo or in vitro neutralization tests has given variable results (Du Plessis, 1993) found no correlation between antibody titters and immunity to heartwater in calves. Furthermore Martinez et al. (1993) observed that sera from survivors of a challenge experiment after immunization did not neutralize C. ruminantium infection of endothelial cell cultures in vitro. However the same authors observed that serum from immune mice and cattle inhibited adhesion and entry of endothelial cell cultures by C. ruminantium. In other experiments, Byrom et al., (1993) used mouse serum with or without complement or purified antibodies administered simultaneously into mice with C. ruminantium during incubation or clinical reaction. They observed each treatment did confer immunity or alter the course of C. ruminantium infection in mice. In contrast addition of complement to immune serum which was subsequently mixed with infectious C. ruminantium (Kumm) inhibited their infectivity (Du Plessis, 1993). The Kumm stock however appears to have a tropism for macrophages and this may be responsible for the different results.
Live immunization by infection and treatment leads to development of very low IgG 2 antibodies in the ruminant whereas it leads to development of a high concentration in the mouse. These results indicates that the two species respond differently to live infections of C. ruminantium and therefore the use of mice to study immune responses to C. ruminantium does not seem to be appropriate.