Natural Brucella melitensis Infection and Rev. 1 Vaccination Induce Specific Brucella O-Polysaccharide Antibodies Involved in Complement Mediated Brucella Cell Killing

Vaccination against brucellosis using live attenuated strains is the primary approach in protecting livestock against the disease through a strong cellular immune response. Attenuated vaccine strains also induce serum anti-Brucella antibodies, mostly against Brucella O-polysaccharide, but their role in protection against the disease remains unclear. In this study, we show that Brucella OPS serum antibodies after vaccination or natural infection could kill Brucella in vitro as shown by the serum bactericidal activity (SBA) assay. We used serum samples of Rev. 1 vaccinated sheep that were negative or positive for Brucella OPS antibodies by either one of complement fixation test (CFT), microplate agglutination test (MAT) and ELISA, or sera of naturally infected sheep positive by CFT. We found a significant increase in the killing ability of sera 30 days after intraocular vaccination with Rev. 1 as compared with pre-vaccination. SBA was significantly higher in sera containing Brucella OPS IgG antibodies in comparison with sera lacking such antibodies (p < 0.001 against 16M & Rev. 1 strains). All 10 sera of convalescent sheep demonstrated significant killing ability against the 16M B. melitensis field strain. Specific OPS antibodies participate in the in vitro complement mediated Brucella killing suggesting a potential role in protection against the disease through this mechanism and relevance of developing OPS-based Brucella vaccines.


Introduction
Brucellosis severely affects animal and human health [1]. Brucella species belong in the Gram-negative group, partly classified according to their natural affiliation to distinct animal species. To date, 12 species have been validly published but only three, B. melitensis, B. abortus and B. suis associated with small ruminant, bovid and swine brucellosis, respectively, are the most distinct human pathogens. Moreover, small ruminants that are natural hosts of B. melitensis, establish a critical vehicle in transmitting the disease to humans, and partly to bovines. Brucella species cause abortion storms in their natural animal hosts followed by excessive shedding of the bacteria via contaminated placenta, fetus membranes and fluids, which leads to spreading of free Brucella organisms in the farm and environment [2]. Brucellosis is transmitted to humans through their contact with infected animals or with contaminated material such as aborted placenta or following consumption of contaminated raw milk and dairy products. In humans, brucellosis is frequently confused with flu-like symptoms but may exacerbate into complications such as endocarditis, meningitis and osteo-articular or epididymo-orchitis manifestations, and may rarely be lethal [3].
Controlling the disease in the livestock population is the most effective approach in preventing human cases and reducing the overall burden of disease in both the animal and human health systems. An ideal brucellosis control program consists of prophylactic vaccination of the livestock with live attenuated strains, smooth B. abortus S19 strain or rough B. abortus RB51 in cattle and B. melitensis Rev. 1 in small ruminants, respectively [4][5][6]. A test and cull policy after serological confirmation might complement successful eradication campaigns where possible, as owner compensation is mandatory in such cases [7]. Such a dual control policy has been implemented successfully in Israel throughout the years, resulting in minimizing the problem of B. melitensis [8,9] but failure to sustain the program due to budgetary concerns has hampered these achievements.
The host first encounters Brucella by mediating an innate bactericidal response against organisms residing extracellularly with a successive development of an effective protective cellular response [10]. The live vaccine strains resembling their field strains in engaging the host immune system only after entrance establish a Th1 immune response involving humoral and cell-mediated B and T cells [11][12][13]. Importantly, these vaccines protect against abortion and therefore reduce bacterial spread in the farm, thereby decreasing the economic and health burden in the farm. However, despite achieving effective vaccination coverage with the live strains, Israeli dairy herds and many small ruminants have been consistently infected with B. melitensis field strains inferring on lack of protection against bacterial entrance to the host [9].
Here, we hypothesized that anti-Brucella OPS antibodies might fulfil the missing fragment in sustaining herd protection. Previous mice experiments have resulted with controversies regarding the specific contribution serum antibodies might play in host protection against Brucella challenges [11,[14][15][16][17][18]. In light of the unresolved disagreement, we have revisited the question whether anti-Brucella OPS serum antibodies could be protective against brucellosis through complement mediated killing of Brucella cells during the extracellular stage of infection.
In this study we show that anti-Brucella OPS antibodies developed post-vaccination or following natural field infection kill Brucella in vitro as revealed by the serum bactericidal activity assay [19,20]. These findings open new horizons to developing a successful Brucella OPS-conjugate vaccine as recently suggested by Bundle and McGiven [18,21,22] that could induce high levels of such functional antibodies.  Serum samples from infected sheep were collected from animals in a bacteriologically confirmed infected flock. Ocular Rev. 1 vaccination of female ewe-lambs was conducted as part of the Israeli Veterinary Services vaccination program in one flock known to be brucellosis free, and sera samples were obtained from the animals before vaccination at day 0, and after vaccination at day 15 and day 30. All serum samples were stored at 4 • C until their use. Anti-Brucella OPS serological titers were tested by complement fixation test (CFT, titer conversion to IU = ×1.6) and microplate agglutination test (MAT, titer conversion to IU = ×1.905) [24].

Indirect ELISA for Brucella-OPS Antibodies
Besides conducting the standard serological tests, sera obtained in the vaccine experiment were evaluated for presence of anti-Brucella OPS antibodies by ELISA using the commercially available kit ID Screen ® Brucellosis Serum Indirect Multi-Species (IDVET, Amman, Jordan). Results were expressed in OD values measured at 450 nm as indicated by manufacturer.

Serum Bactericidal Activity
Fresh bacterial cultures of B. melitensis 16M and B. melitensis Rev. 1 passage 101 Elberg strain were grown in Tryptic Soy broth (TSB), for 2 days at 37 • C in presence of 5% CO 2 atmosphere. Bacterial suspensions were centrifuged at 10,000× g for 10 min and cell pellets were suspended in PBS-containing 0.5 mM MgCl 2 and 0.15 mM CaCl 2 , and cells were serially diluted to reach a cell concentration of 10 4 CFU per ml. Each serum sample was tested in four separate experiments, in duplicates using Oswald et al. (1990) method with minor modifications [25]. Shortly, sheep serum was heat-inactivated at 56 • C for 30 min, and 50 µL were mixed with 30 µL of the master Brucella dilution in 96-well flat-bottomed polystyrene microtiter plates (Nunc™ Thermo scientific, Rochester, NY, USA). Cells were incubated in presence of 5% CO 2 atmosphere, keeping the plate rotated at 100 rpm for 30 min at 37 • C, which allowed Brucella cells to interact with antibodies. After 30 min, 40 µL of human complement (Quidal, San Diego, CA, USA) was added to the Brucella-serum reaction mixture and incubation at 37 • C was continued for additional 1.5 h. Then, a 25 µL aliquot from each well (estimated number of 62 CFU in the untreated suspension) was plated in duplicates on Serum-Dextrose tryptic soy agar plates and incubated for 2 days at 37 • C with 5% CO 2 for CFU counting. Two control experiments were conducted, a Brucella cell suspension incubated in presence of buffer without serum and a Brucella cell suspension incubated in presence of complement but without serum, respectively. Because results showed only a minimal cell killing by the two systems, around 2-3% (data not shown), we used the complement reaction mixture as our negative control in further calculations [11,26]. The percentage of cell killing (killing efficiency) was calculated as follows.
o f CFU a f ter incubation with serum and complement mean no. o f CFU a f ter incubation only with complement

Statistical Analysis
Graphpad Prism (Graphpad Prism. Available online: https://www.graphstats.net/ graphpad-prism (accessed on 19 July 2021) was used for statistical analysis. Bonferroni's multiple comparison tests was used to analyze significance of differences of Brucella cell killing between serum samples of non-vaccinated and vaccinated ewes according to postvaccination time periods two ways ANOVA unweighted analysis was employed (Bonferroni post hoc test was used to analyze significance of differences between serum activity without or in presence of anti-Brucella OPS antibodies. Unpaired t-test was used to identify the Brucella killing by post-infection serum samples. B. melitensis Rev. 1 strain vaccination is used routinely in Israel in young replacement ewe-lambs as a full dose (1-2 × 10 9 cfu) ocular application. This vaccination approach was introduced in the 1980s as a method of differentiating infected from vaccinated animals (DIVA) solution to minimize post-vaccine humoral antibody titers [7]. About 40% of the animals have been found negative in the MAT and CFT tests (unpublished data). We chose this platform to specifically test our hypothesis regarding development of complement mediated anti-Brucella OPS serum bactericidal activity (SBA) of antibodies induced by vaccination. We used B. melitensis strain 16M, a type strain of the genus, and strain B. melitensis Rev. 1, Elberg passage 101, 1970, a reference vaccine type strain, as readout cells, addressing the potential attenuation of strain Rev. 1 in this pathway.
Three time periods were chosen, day 0 before vaccination, and 15 and 30 days postvaccination, respectively. Day 0 expressing SBA before animal vaccination, e.g., in naïve animals with background serum components in comparison to post-vaccination time periods that anticipate contribution of serum-developed anti-Brucella OPS antibodies to Brucella cell killing.
As shown in Table 1, our initial experiment included four animals which were tested before and after vaccination with Rev. 1 as mentioned above. None had detectable antibody titers by both MAT and CFT before vaccination. Serum number 9914 was negative for anti-Brucella OPS by CFT and MAT during the whole experiment. In contrast, serum number 9912 had positive MAT titers on day 30 but was CFT-negative along this period, and serum numbers 9913 and 9921 that showed MAT titers at early days after vaccination (day 15) became responsive also in the CFT test at day 30. Cell killing of both strains increased according to the post-vaccination time period but the extent of cell killing between individual serum samples and time periods varied. SBA of the consecutive sera of 3 of 4 animals was significantly higher on days 15 and 30 post vaccination against both Rev. 1 and 16M strains (p < 0.05 16M; p < 0.01 Rev. 1, Bonferroni's multiple comparison test). Serum number 9914 that lacked anti-Brucella OPS antibodies along the whole experiment showed nevertheless a strong SBA activity even at time 0 that is before vaccination. As shown in Supplementary Material Figures S1 and S2, this result was unexpected, as other naïve serum samples only killed Brucella cells at background levels assumably associated with nonspecific cross-reactive antibodies. Accordingly, one might link the higher SBA activity of serum sample 9914 with presence of non-Brucella reactive serum components.  significantly higher against 16M than against Rev. 1 (p < 0.001, Bonferroni post hoc test). As depicted in Table 1, at this time after vaccination, 3 out of 4 and 2 out of 4 sera, had elevated IgM and IgG antibody titers against Brucella OPS, respectively.
Neg stands for Negative. Figure 1 summarizes the mean value of the collective SBA activities of the 4 sera samples against Rev. 1 and 16M strains at the 3 timings related to vaccination. Pre-vaccination naïve sera (day 0) had an SBA of around 10% cell killing against both Brucella strains, 16M and Rev.1. As post-vaccination time period increased, there was increased killing activity by the serum samples against both strains (p ≤ 0.0001, two-way ANOVA). The mean SBA was highest on day 30 post-vaccination with percentages of cell killing significantly higher against 16M than against Rev 1 (p < 0.001, Bonferroni post hoc test). As depicted in Table  1, at this time after vaccination, 3 out of 4 and 2 out of 4 sera, had elevated IgM and IgG antibody titers against Brucella OPS, respectively. In an attempt to better understand the importance of antibody development to the extent of SBA, we performed a second experiment, using the same treatments as in the above but measuring antibody isotype titers by adding indirect ELISA (i-ELISA) to the antibody tests and increasing the number of vaccinated animals to eight.
As shown in Table 2  In an attempt to better understand the importance of antibody development to the extent of SBA, we performed a second experiment, using the same treatments as in the above but measuring antibody isotype titers by adding indirect ELISA (i-ELISA) to the antibody tests and increasing the number of vaccinated animals to eight.
As shown in Table 2  As shown in Table 2, CFT and MAT negative sera were still shown positive by i-ELISA, indicating the higher sensitivity of the i-ELISA method in comparison to the standard serology. In fact, CFT and SAT have been adjusted a priori to be able to perform as qualitative DIVA methods which excludes reading of vaccine elicited antibodies. Figures 2 and 3 show that serum samples before vaccination lacked anti-Brucella OPS antibodies but were endowed with a basic SBA level which varied among the serum samples between 8 to 13% of cell killing between strain 16M ( Figure 2) and Rev. 1 (Figure 3). SBA against both Rev.  Figure S3 summarizes the results of individual sera samples when combined into two groups, those which lacked i-ELISA titers and in correlation were not endowed with SBA in contrast to the other serum samples which possessed i-ELISA titers and were endowed with SBA, respectively. Bonferroni two-way ANOVA analysis showed the significant killing difference (p > 0.001) against B. melitensis 16M and B. melitensis Rev.1 with anti-Brucella OPS antibodies sera in comparison to sera-lacking anti-Brucella-OPS antibodies.

Serology and Serum Bactericidal Activity (SBA) of Post-Infection Sera
We also examined SBA in serum samples obtained from B. melitensis naturally infected sheep, selecting animals according to their positive CFT serology of ≥1:20 which indicated presence of IgG anti-Brucella antibodies in the serum. Because animals in the field are identified bona fide, without information on their serum activity before exposure to Brucella, we used baseline SBA values from the SBA data mentioned above among sheep before vaccination Figure 1. As can be seen in Figure 4, serum samples of infected animals demonstrated complement mediated cell killing against both strains, all showing a higher cell killing activity than their baseline response (horizontal broken line). In this test, killing was more effective against strain 16M than against vaccine strain Rev.1 (unpaired t-test, p ≤ 0.05) reiterating the observation discussed in Figure 1.     field are identified bona fide, without information on their serum activity before exposure to Brucella, we used baseline SBA values from the SBA data mentioned above among sheep before vaccination Figure 1. As can be seen in Figure 4, serum samples of infected animals demonstrated complement mediated cell killing against both strains, all showing a higher cell killing activity than their baseline response (horizontal broken line). In this test, killing was more effective against strain 16M than against vaccine strain Rev.1 (unpaired t-test, p ≤ 0.05) reiterating the observation discussed in Figure 1.

Discussion
Use of Rev. 1 vaccine has always been conceived effective in protection against brucellosis due to elicitation of the cellular arm of the immune response. In contrast, development of antibodies has been considered as interfering with DIVA diagnosis of vaccinated herds [18,21,27,28]. This led scientists to favor the ocular application of the vaccine throughout vaccination campaigns, and our manuscript addresses this point by showing that ocular vaccination still elicits antibodies which are not detectable by standard serological tests, whereas LPS ELISA can identify them. We have hypothesized that anti-Brucella-OPS antibodies elicited following Rev. 1 sheep vaccination, or otherwise sheep infection, would be functional in host protection. To test this hypothesis, two sets of Unpaired T-test shows overall significant difference between 16M and Rev. 1 cell killing (p < 0.05) (with Bonferroni correction for multiple comparisons), (* p < 0.05) shows the significant difference in the killing activity of individual sera. Horizontal line depicts percentage of Brucella cell killing attributed to innate immunity (A generalized innate killing activity of sheep serum samples calculated as a mean value of killing results of the four serum samples prior to animal vaccination). Killing activity above this level is considered, therefore, to be contributed by any of anti-Brucella antibodies present in the reaction, including non-anti-OPS antibodies.

Discussion
Use of Rev. 1 vaccine has always been conceived effective in protection against brucellosis due to elicitation of the cellular arm of the immune response. In contrast, development of antibodies has been considered as interfering with DIVA diagnosis of vaccinated herds [18,21,27,28]. This led scientists to favor the ocular application of the vaccine throughout vaccination campaigns, and our manuscript addresses this point by showing that ocular vaccination still elicits antibodies which are not detectable by standard serological tests, whereas LPS ELISA can identify them. We have hypothesized that anti-Brucella-OPS antibodies elicited following Rev. 1 sheep vaccination, or otherwise sheep infection, would be functional in host protection. To test this hypothesis, two sets of experiments were designed on the basis of studying development of anti-OPS antibodies by both CFT and MAT serology in the first set and anti-Brucella-OPS indirect ELISA serology in the other set. We have furthered the experiment by testing first development of SBA in correlation with Rev. 1 ocular vaccination and secondly, in sera obtained after natural Brucella infection. The serum agglutination test (MAT), complement fixation test CFT and i-ELISA were used to indicate presence of IgM (MAT) and IgG antibodies (CFT and i-ELISA) in the serum samples) [24].
Because Rev. 1 is delivered in Israel by the ocular route, aiming at reducing antibody titers, we specifically selected sheep that best represent their antibody response to Rev. 1 vaccination through CFT, MAT analyses in the first set of experiment and including i-ELISA in the second test, assuming i-ELISA would reveal anti-Brucella-OPS antibodies which MAT and CFT are missing, respectively. We chose two B. melitensis strains as read out analytes, a typical strain of the genus (B. melitensis strain 16M), and its counter live attenuated vaccine strain Rev.1 Elberg, passage 101, 1970, best representing the authentic Rev. 1 clone first developed by Elberg in the 1950s [29], aiming also at revealing how its attenuation has affected cell susceptibility to SBA.
From the Rev. 1 pre-and post-vaccination sera testing, our data have reiterated a past notion that anti-Brucella serum antibodies participate in Brucella cell killing. In this study we have clearly shown the correlation between development of elevated antibody levels, mostly IgG, against Brucella-OPS in sheep and percentage of cell killing in both cases of Rev. 1 ocular vaccination or following natural infection, respectively. We have observed, however, that in the vaccine experiment, CFT-positive serum samples showed significantly lower SBA against Rev. 1 cells than against strain 16M (Figure 1, p < 0.001; Bonferroni post hoc test) possibly explained by differences in the cell envelop between the vaccine strain and 16M, a typical strain of the genus [30]. One may thus take it one step further to hypothesize that vaccine strain Rev. 1 attenuation may have involved changes in its OPS structure.
Comparing between vaccine-and infection-induced antibodies provided a unique opportunity to test the functional responsiveness of the immune response against both Brucella strains. B. melitensis is naturally affiliated to small ruminants and B. abortus to cattle. Nevertheless, both strains share cross reactive antibodies against their OPS antigen making serology irrelevant in determining which was the etiological agent involved in an infection. Despite using the ocular vaccine, aimed at reducing the humoral response of the immune system, antibodies are developed by the host.
We demonstrate that anti-Brucella-OPS antibodies induced by vaccination or natural infection are functional and may contribute to preventing infection by activating a complement mediated cell-killing mechanism. Until now, development of anti-Brucella-OPS antibodies has been considered a flaw in the vaccination policy due to hampering DIVA serological surveillance studies following Rev. 1 vaccination campaigns. To overcome this problem, enforcing a stringent vaccination policy where only young replacement animals are vaccinated by a single-lifetime inoculation and flock coverage is confirmed would be necessary.
Similar to Rev. 1 vaccination, our results now show that natural infection also accounts for development of functional anti-Brucella OPS antibodies that participate in SBA against Brucella (Figure 4). However, in the case of infection it is understood that the field strain persists in the flock and animal culling is required to stop the spread of the disease to other animals.
Our results indicate an opportunity to improve vaccination effectiveness against brucellosis by increasing development of anti-Brucella-OPS bactericidal antibodies using a DIVA established vaccine, first proposed by Bundle and McGiven (18). It has been shown that conjugate vaccines against other bacterial pathogens such as Hemophilus influenzae type b, S. pneumoniae or Shigella can induce stronger binding and functional antibody responses as compared to the corresponding natural infections [31][32][33]. The development and use of a conjugate Brucella vaccine should, therefore, be sought as an ultimate solution to maintaining high coverage of herd immunity. Rev. 1 vaccination of replacement animals should nevertheless remain a major strategic tool in controlling the disease by strengthening individual immunity among the animals through prevention of abortions thus minimizing the risks of building up chronic brucellosis in the flock.

Conclusions
Rev. 1 vaccination as well as B. melitensis infection elicit functional anti-Brucella-OPS antibodies which kill Brucella cells in vitro by the classical complement mediated pathway. Rev. 1 vaccine, which is a live attenuated strain, survives in the host for a critical period required for the buildup of the adapted cellular arm of the immune response. However, this vaccine does have adverse effects; it causes abortions if used in pregnant ewes, it elicits persistent antibody titers which hamper DIVA diagnosis and it is zoonotic.
In view of our results, we assume that endowing the host with anti-Brucella-OPS antibodies could augment the innate immune response against brucellosis, therefore aiding