Q-VAX^® was purchased from Seqirus (bioCSL, Melbourne, Australia). Equivalent C. burnetii WCV was generously provided by the Dr. James Samuel lab at Texas A&M University. C. burnetii NMI clone 7 (RSA 493) was grown in selective ACCM-2 media and inactivated with 2% formalin for 48 h for the WCV. Recombinant proteins were expressed in Escherichia coli BL21 cells and purified by multi-column chromatography and endotoxin removal procedure (<0.1 EU/ng) by GenScript (Piscataway, NJ, USA). C. burnetii NMI RSA 493 was purchased from BEI Resources (Manassas, VA, USA) and quantified via qPCR with genomic DNA extraction via High Pure PCR Template Preparation Kit (Roche, Basel, Switzerland) and PowerUp SYBR Green Master Mix (Thermo Fisher Scientific, Waltham, MA, USA) for infection. Primers amplify a 74-base pair fragment in C. burnetii com1 [FAF216: 5’ GCACTATTTTTAGCCGGAACCTT 3’, RAF290: 5’ TTGAGGAGAAAAACTGGATTGAGA 3’] (22). AddaVAX™ (squalene oil-in-water emulsion) and CpG ODN 1018 (TLR9 agonist) were purchased from InvivoGen (San Diego, CA, USA) and Integrated DNA Technologies (Coralville, IA, USA), respectively. CpG-ODNs were dissolved in sterile water at 1 mM as stock. MPLA (a TLR4 agonist) was purchased from Avanti Polar Lipids Inc. (Alabaster, AL, USA). TLR7 agonist 2Bxy, an imidazoquinoline derivative, was purified by either high -performance liquid chromatography or gel extraction and confirmed by MALDI-TOF or electrospray ionization-mass spectrometry (23).

C57BL/6 female mice (6–12 weeks) and Hartley guinea pigs (300–400 g) were obtained from Charles River Laboratories (Wilmington, MA, USA). Animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of California, Irvine, and the Animal Care and Use Review Office (ACURO) of the U.S. Army Medical Research and Materiel Command (USAMRMC). Mice were housed in standard cages with enrichment at Animal Biosafety Level 2 (ABSL-2), and guinea pigs were housed in approved ABSL-3 facilities. For immunogenicity studies, mice were anesthetized in induction chambers with inhaled isoflurane/O₂ and passive scavenging with F/air canisters. Guinea pigs were anesthetized with an intraperitoneal (IP) injection of 100 mg/kg ketamine and 10 mg/kg xylazine in phosphate-buffered saline (PBS). Where indicated, mice were immunized either subcutaneously (SC) at the base of the tail or intramuscularly (IM) in the semitendinosus and semimembranosus muscles of the hind limb. Guinea pigs were immunized intramuscularly in the caudal thigh. Blood was collected from the submandibular vein in mice and the lateral saphenous vein in guinea pigs with 25 G hypodermic needles (Medline, Northfield, IL, USA), and collected into Microvette CB 300 lithium heparin (Sarstedt, Newton, NC, USA), and BD Microtainer PST tubes with lithium heparin (BD, Franklin Lakes, NJ, USA), respectively.

Gene identification and nomenclature are based throughout this study on the complete genome sequence published by Seshadri et al. (24). C. burnetii proteome microarrays were produced as described previously (25–27). Briefly, proteins from the C. burnetii NMI strain RSA 493 proteome were expressed from purified plasmids in an E. coli-based cell-free in vitro transcription translation system (IVTT) (Biotechrabbit GmbH, Hennigsdorf, Germany). IVTT reactions were printed onto nitrocellulose-coated glass AVID slides (Grace Bio-Labs Inc., Bend, OR, USA) using an Omni Grid 100 microarray printer (Genomic Solutions). Plasma was diluted 1:100 in protein array blocking buffer (GVS, Sanford, ME, USA) and incubated with 0.1 mg/ml of a His-tag-containing peptide (HHHHHHHHHHGGGG) (Biomatik, Wilmington, DE, USA) at room temperature for 30 min to block anti-His antibodies generated by the immunizations. Afterward, the arrays were incubated overnight at 4°C with gentle rocking. Arrays were washed three times with TBS-0.05% Tween 20 (T-TBS) and then incubated with goat anti-mouse immunoglobulin G (IgG)-biotin, IgG1-biotin, or IgG2c-biotin (1:200 in array blocking buffer) (Jackson ImmunoResearch, West Grove, PA, USA) for 1 h at room temperature with gentle rocking. Following another set of T-TBS washes, bound antibodies were detected with streptavidin-conjugated Qdot^®655 or Qdot^®800 (Thermo Fisher Scientific, Waltham, MA, USA), diluted 1:250 in array blocking buffer for 1 h at room temperature with gentle rocking. Arrays were washed three times with T-TBS, and the slides were rinsed thoroughly with water and then air-dried by centrifugation at 500 g for 10 min. Images were acquired, and spot fluorescence intensities were quantified using the ArrayCAM™ Imaging System (Grace Bio-Labs, Bend, OR, USA). Signal intensities (SIs) for each antigen on the array were background-corrected by subtracting sample-specific T-PBS buffer signals from purified protein spot signals.

Down-selected proteins were expressed in vitro in 200 μl IVTT reactions as described above. Each protein is expressed with an N-terminal 10× polyhistidine and C-terminal HA epitope tags. After 16 h reaction at 21°C, proteins were captured via ×10 His tags by adding 100 μl to nickel-charged His-TrapSpin resin (GE Healthcare Life Sciences). The resin consists of ~34 μm-diameter Sepharose beads with a binding capacity of 750 μg of His-tagged protein per ml. Flow-through was reapplied to columns and then washed to remove non-bound IVTT material. The successful capture of proteins to the beads was monitored by printing microarrays of IVTT, flow-through, and bead wash on nitrocellulose-coated slides ( Supplementary Figure S2).

Spleens were harvested from mice 10 days after they were boosted via the IP route, and erythrocyte-depleted splenocyte suspensions were prepared for T cell recall assay (IFNγ ELISpot) as previously described (28). Purified antigens were titrated in the assay (final concentrations of 10, 5, 2.5, and 0 μg/ml). Spleen cells from naive mice were assayed in parallel as a control for potential mitogenic activity of the recall antigens. Assays were performed in T cell medium (TCM) comprising Iscove’s Modified Dulbecco’s Medium (IMDM), containing 5 × 10^-5 M β-mercaptoethanol, 100 IU/ml penicillin, 100 μg/ml streptomycin, and 10% heat-inactivated fetal calf serum. After 18 h of incubation, the assay supernatants were collected for multiplex cytokine screening using the LEGENDplex kit (BioLegend Inc., San Diego, CA, USA) according to the manufacturer’s instructions before the ELISpot was processed. Spots were quantified in an ImmunoSPOT^® ELISpot plate reader (Cellular Technology Limited, Cleveland, OH, USA).

Guinea pigs were sensitized with an SC administration of Q-VAX and rested for 2 weeks. RFID microchip Transponders (BMDS Avidity Science, Waterford, WI, USA) were delivered SC for identification and to monitor weight and temperature change. To elicit hypersensitivity responses, four or six approximately 2–3 cm areas were shaved using electric clippers on the right and left flanks. Vaccine candidates, WCV, or PBS sham was then injected intradermally at each of the shaved sites. Temperature, weight, and reaction sites were monitored daily for 2 weeks. Each vaccine candidate was evaluated in four guinea pigs.

Skin sites from guinea pigs were fixed in 10% neutral buffered formalin for at least 72 h at room temperature. For the hypersensitivity experiments, three sections were cut from the shaved areas containing the epidermis to the underlying abdominal or intercostal muscle. Tissues were submitted to AML Laboratories (Jacksonville, FL, USA) for processing, embedding, and sectioning at 5 μm before staining with hematoxylin and eosin (H&E). Histopathology slides were deidentified and evaluated by an American College of Veterinary Pathologists (ACVP) board-certified pathologist. Histopathologic scoring was performed on a 0–5 scale.

Guinea pigs were SC administered RFID microchip transponders in the back of the neck for identification purposes and to monitor weight and temperature for the duration of the experiment. They were then administered candidate vaccines, WCV, or PBS (sham) in 100 -μl sterile PBS by intramuscular injection in the semitendinosus and semimembranosus muscles. A boost vaccine was given in the opposite hind limb 2 weeks later. Blood and plasma were collected from the lateral saphenous vein on days 10, 21, 28, 35, and 42 post-prime. Guinea pigs were intratracheally infected with 5 × 10⁵ genomic equivalents (GEs) of C. burnetii NMI RSA 493 after resting for 7 weeks post-prime. For this, animals were anesthetized with an IP injection of 100 mg/kg ketamine and 10 mg/kg xylazine in PBS. A MADgic pediatric laryngotracheal mucosal atomizer device (Teleflex, Morrisville, NC, USA) was inserted to administer the bacteria intratracheally in 100 μl of PBS. Guinea pigs were monitored daily, and weight and temperature measurements were taken. Four to five guinea pigs were utilized for each experimental group.

Statistical analyses were performed with GraphPad Prism v9.4.0 (GraphPad Software, La Jolla, CA, USA). Results were compared using one-way or two-way ANOVA with Dunnett’s or Tukey’s correction for multiple comparisons. Differences were considered significant if p-value ≤ 0.05 (*), ≤ 0.01 (**), ≤ 0.001 (***), or ≤ 0.0001 (****).

It was of considerable interest to first profile antibodies (Abs) and T cells in response to Q-VAX to identify recognized antigens as subunit vaccine candidates. Ab (IgG) profiling was performed in C57BL/6 mice administered a single dose via SC or intramuscular routes using proteome microarrays (Figure 1A). Plasma was collected on days 0, 28, 49, and 69. Only one protein, CBU1910, was consistently recognized, with other reactive proteins giving a scattered, more stochastic recognition pattern. The breadth of the Q-VAX Ab profile was unexpectedly narrow, given that there are 1,815 different proteins in the proteome. There were no obvious differences in Ab profiles between the intramuscular and SC routes of immunization.

To attempt to broaden the Ab profile and identify additional candidate vaccine targets, mice primed via the SC route were boosted with Q-VAX, weighed, and carefully observed daily for changes in behavior. No adverse events were seen, and animals were bled at the end point 14 days later (day 83). After boosting, IgG signals increased only to existing reactive antigens, namely, CBU1910, with no increase in breadth of response (Figure 1A). These data show that the antibody breadth in response to administration of Q-VAX is remarkably narrow, even after boosting.

Given the limited serological breadth induced by Q-VAX, we broadened the search to include serological profiling in cases of natural infection. For this, we used published and unpublished in-house protein microarray studies from human and animal Q-fever studies as well as mass spectrometry identification of seroreactive proteins from C. burnetii, as reviewed in 2013 by Vranakis et al. (25, 26, 29–34). Antigens were ranked based on reactivity on the protein microarrays, and a total score was assigned based on these two metrics (Table 1). Higher priorities were given to those antigens with a predicted transmembrane domain, as these are more likely to be surface proteins and have easily accessible epitopes for antigen presentation. This list initially comprised CBU0612, CBU1910, CBU0891, CBU0307, and CBU0664 as “Tier 1” antigens.

Where possible, we attempted to produce “tag-less” proteins to minimize incorporation of off-target epitopes. Thus, CBU0891, CBU0307, and CBU0664 proteins were expressed with the SUMO-His tag that was cleaved from the protein before use (35). However, CBU0612 and CBU1910 were expressed with Sortase recognition motifs (either C-terminal LPXTG Sortag or an N-terminal GGG tag) with the intention of using the Sortase tag to couple proteins to TLR agonists (36). The expression and solubility of CBU0891 and CBU0307 were initially low but improved by subsequently expressing truncated versions lacking transmembrane domain and signal peptide, and signal peptide, respectively. CBU0612-Sortag and CBU1910-Sortag were subsequently modified with a C-terminal polyhistidine (termed CBU0612-SorHis and CBU1910-SorHis) to evaluate the utility of tris-NTA for coupling proteins to TLRs in place of Sortase (37). Overall, proteins expressed with His-SUMO (CBU0891, CBU0307) showed only modest solubility. Protein CBU0664 showed poor solubility and was not pursued further.

An additional 12 antigens (Tier 2) were selected and tested in pilot experiments for E. coli expression and solubility. All 12 Tier 2 proteins were expressed as GST-GG fusion proteins separated by a TEV protease cleavage site to improve solubility. The GG motif was included, so cleavage of His-GST-GG-TEV yields protein with an N-terminal GGG tag available for downstream coupling reactions. Pilot studies showed that only HisGST-CBU0545 remained soluble after GST cleavage. Of the 12 Tier 2 proteins, seven were suitable for scale -up (CBU0545, GST-CBU1398, GST-CBU0630, GST-CBU0718, HisGST-CBU1094, GST-CBU1513, and HisGST-CBU0611), whereas five proteins remained either poorly expressed or insoluble (CBU0092, CBU1143, CBU1627, CBU1260, and CBU0198) and were not investigated further. The final list of purified Tier 1 and Tier 2 proteins used for immunogenicity studies is indicated in Table 1.

To determine if Q-VAX induced a T cell response against any of the purified Tier 1 and Tier 2 proteins, we performed a T cell recall assay on splenocytes of mice that were immunized IP with Q-VAX. Splenocytes were harvested 10 days after immunization, pooled from n = 3 mice, and incubated with purified recall protein for 18 h (Supplementary Figure S1). Antigen-specific IFNγ responses were seen to GST-CBU1398, CBU0891, CBU0545, and CBU1910-SorHis, with CBU1910-SorHis being dominant. Antigens that induced low levels of IFNγ in the T cell recall response and were difficult to purify were excluded from further experiments. The T cell recall experiment was repeated with a subset of six lead candidate antigens (Table 1), with a representative experiment shown in Figure 1B. Overall, the data indicate that CBU1910 is dominant in the antibody response to Q-VAX, which is accompanied by a robust IFNγ response in the recall assay (p < 0.0001).

In parallel with protein purification described above, immunogenicity studies were also performed with antigens expressed from pXi plasmids in transcription/translation (IVTT) reactions (as used for custom protein microarrays) and captured onto Ni-Sepharose beads. This method, which exploits the ability of IVTT to express any protein in soluble form and obviates the need for traditional protein purification, allows many potential target antigens to be screened directly for immunogenicity in vivo. Loading of protein to the beads was confirmed by printing the protein beads on arrays and probing arrays with anti-epitope tag antibodies (Supplementary Figure S2). We then administered each individually formulated with AddaVAX to mice (n = 3 per group) via the SC route. One group of animals was administered purified CBU1910 protein for comparison. Plasma was collected at days 0, 10, and 28 and analyzed for antigen-specific IgG responses on protein microarrays displaying both IVTT-expressed and purified proteins, as shown in Figures 2A, B. Robust antibody responses were evident to three of the proteins tested —CBU1910, CBU0612, and CBU0891. The Abs produced against the IVTT expression products were both IgG1 and IgG2c, while those engendered by purified protein (CBU1910-SorHis) were IgG1 (Th2) only. Increased IgG2c responses correspond to the Th1 T cell response, which is defined by proinflammatory cytokine production and macrophage activation promoting intracellular bacterial killing (38). IgG1 corresponds to a Th2 response, which is appropriate for upregulating granule release in basophils and other granulocytes. Previous studies have demonstrated that protective immunity against C. burnetii is driven by a Th1 response, as evident by elevated IFNγ, TNFα, and IgG2c antibodies (39–41). We attribute this to the purified proteins lacking immunostimulatory components when compared to the IVTT expression products. The increase in IgG signals over time is shown in Figure 2B.

A T cell recall assay was performed with the same mice boosted with the corresponding adjuvanted purified antigens via the IP route on day 28, and spleens were harvested 10 days later (Figure 2C). Minor nonspecific or mitogenic background activity was noticed against purified CBU0545 in splenocytes from mice immunized with beads incubated in IVTT without DNA template; the remaining proteins elicited no response. Antigen-specific IFNγ responses were seen to all the priming antigens except for CBU0891. In the case of CBU0545, the recall response remained high at all antigen concentrations tested, revealing a robust antigen-specific response above any nonspecific effects. The weakest IFNγ response was seen against CBU0891. T cell cross-reactivities were noted; for example, mice immunized with CBU0612-SorHis showed measurable cross-reactivity for CBU0545 and CBU0307. These were non-reciprocal and appeared independent of shared tags, and sequence alignments did not reveal any obvious reasons for these patterns. Nevertheless, the serology and T cell data were consistent with the immunogenicity of all eight proteins when administered individually. Overall, the data indicate that the candidate vaccine antigens can engender Ab and/or T cell responses in C57BL/6 mice.

We next evaluated the immunogenicity-enhancing effects of including TLR agonists and the squalene oil-in-water emulsion AddaVAX. For this, we formulated purified CBU1910 in a combination adjuvant, IVAX-1, which comprises of TLR4 agonist MPLA, TLR9 agonist CpG oligodeoxynucleotide (ODN) 1018, and AddaVAX (7, 42). Formulations containing purified CBU1910 and IVAX-1 were administered to C57BL/6 mice as a single dose and compared to control groups Q-VAX and PBS (Table 2 and Figure 3A). One group was given two doses 14 days apart to observe the benefits of a prime/boost model versus a single dose. Interestingly, including a boost did not significantly increase IgG responses when compared to single-dose groups. Another experimental group contained NMI LPS to emulate Q-VAX more closely. LPS has been shown by other groups to be a prospective antigen, with adoptive transfer of immune sera or splenocytes protecting naive mice (43).

Plasma was collected on days 9, 14, 28, and 42 and evaluated on protein microarrays for antibodies against CBU1910 (Figure 3B). Interestingly, IgG signals had lower intensities in the group containing the NMI LPS. Plasma from day 42 was analyzed for IgG1 and IgG2c to determine Th2 vs. Th1 responses, respectively (Figure 3C). Mice immunized with CBU1910 without adjuvant skewed very heavily toward an IgG1/Th2 response, while two of the five animals in the Q-VAX group polarized toward IgG2c/Th1. Experimental groups composed of CBU1910 with adjuvants IVAX-1, IVAX-1 prime/boost, or IVAX-1 + NMI LPS exhibited a balanced IgG1/IgG2c response and higher overall responses when compared to the unadjuvanted group (p = 0.0151, p = 0.0008, and p = 0.7565 for IgG1; p < 0.0001, p < 0.0001, and p = 0.4710 for IgG2). We observed a significant dampening response with lower IgG2c signal intensities in the group containing the NMI LPS when compared to the CBU1910 + IVAX-1 group (p < 0.0001).

In the T cell immunogenicity screen, animals were primed with a single dose and euthanized 9 days later. One group was administered a boost 14 post-prime for comparison (Figure 3A). Splenocytes were stimulated for 18 h with CBU1910 antigen, and IFNγ produced from the stimulation was captured and quantified by both ELISpot (Figure 3D) and cytokine bead assay (Figure 3E). The prime/boost group with CBU1910 and IVAX-1 generated more IFNγ spot-forming cells than the other groups (p = 0.0380), and the group given only the protein antigen by itself generated 0 spots. The addition of IVAX-1 to antigens trended toward an increase in IFNγ production when compared to antigen alone (p = 0.9806), although the addition of NMI LPS seemed to dampen the response, similar to the IgG response. The response of the adjuvanted group with NMI LPS is weaker than that of Q-VAX (p = 0.8776). The cytokine bead assay results also reflected the IFNγ T -cell recall results in that the group that received a prime/boost of CBU1910 and IVAX-1 adjuvant induced the highest IFNγ and IL-2 response, corresponding to a Th1 response (Figure 3E). IFNγ and IL-2 production was significantly higher in the prime/boost group compared with the single-dose immunization (p < 0.0001 and p = 0.0186, respectively). The Th2 cytokine levels generated in response to recall antigen stimulation in all cases were low. Overall, these data show that the addition of the IVAX-1 adjuvant enhanced the IgG, T cell recall, and cytokine response when compared to soluble antigen alone and sham-vaccinated groups. The inclusion of NMI LPS lowers immunogenicity in reducing IgG and IFNγ production.

The Hartley guinea pig is considered a more relevant animal model than the mouse for the development of Q-fever vaccines due to its high susceptibility to respiratory pathogens with its ability to develop fever and other visible pathological changes (44–46). The formulations used in the guinea pig immunogenicity studies are indicated in Table 3. Group 1 contains TLR7 agonist 2Bxy and was a formulation from a previous study (27). 2Bxy is an imidazoquinoline derivative that promotes CD8⁺ T cell activity (47). We evaluated reactogenic responses to the vaccine candidates using an intradermal assay in guinea pigs. For this, guinea pigs were sensitized to Q-VAX via SC injection followed by resting for 2 weeks (48, 49). Animals were shaved in their flanks, and vaccine formulations were then administered to the exposed skin. Experimental animals (n = 4) received intradermal immunizations of the six candidate vaccines, and control animals (n = 7) received WCV and PBS immunizations. Weights and temperatures were monitored for 14 days (Supplementary Figure S3A). There was no significant difference between experimental and control groups, and all animals steadily gained weight over time.

At the end of the 14-day period, the animals were euthanized and the injection sites were sectioned for H&E staining (Figure 4A). Pathology in each section was scored on a scale from 0 to 5, with 5 being the most severe. Histopathology evaluation was parsed into four morphological categories: necrosis/suppuration, degeneration of collagen, fibrosis/granulation tissue, and mononuclear cell infiltration (Figure 4B and Supplementary Figure S3B). Necrosis and suppuration are weighted the most heavily in the parsed scores, with Group 1 and WCV demonstrating the most (p = 0.0253 and p = 0.0149). Group 1, containing six antigens, TLR7 agonist, and IVAX-1, showed marked cellular inflammation composed primarily of macrophages with occasional central foci of necrosis and cellular debris (abscesses) or collagen degeneration. There is a moderate amount of fibrosis surrounding the inflammation, and foci of hemorrhage are present within the lesions. Multifocally, adjacent skeletal muscle myofibers are degenerate. These observations are comparable with findings in the WCV sites. The presence and severity of foci of necrosis in the remaining experimental groups were not significantly greater than those in the PBS group. Group 2 exhibited noticeable degeneration of collagen (p < 0.0001) compared to the PBS control group, while all the remaining experimental groups were comparable to the WCV group. In all experimental groups, there was moderate fibrosis within the subcutis as well as mononuclear cell infiltration (Supplementary Figure S3B).

We also wanted to confirm whether the antigens identified as immunogenic in the mouse model were similarly immunogenic in the guinea pig. Animals were administered multivalent protein vaccines with the IVAX-1 adjuvant, boosted 14 days later, and rested for 7 weeks. Plasma was collected from vaccinated guinea pigs at regular intervals and assessed on the protein microarray platform for antigen-specific antibody responses (Figure 5A). Consistent with mouse studies, Q-VAX generated antibodies to CBU1910 only, which increased over time. All six experimental groups elicited durable CBU1910-specific IgG responses that persisted to day 42 post-immunization (p < 0.0001, p < 0.0001, p = 0.0027, p = 0.0012, p < 0.0001, and p = 0.3828). Interestingly, the introduction of NMI LPS dampened the IgG response to antigens CBU0891, CBU0612, and CBU0545. IgG responses against CBU0307 were not detectable in any of the groups despite being an antigen in two of the six experimental groups. This corroborated with Figure 2 when mice were immunized with individual antigens. The vaccine formulation group containing only CBU1910 and IVAX-1 exhibited mild cross-reactivity with CBU0612, but this is likely due to the purified proteins containing the same Sortag tag used in the purification process.

IgG1 and IgG2 responses were also measured on a protein microarray to assess Th1/Th2 biases (Figure 5B). Based on the subtyping data, all experimental vaccine formulations generated stronger CBU1910-specific IgG2 responses than Q-VAX (p = 0.0692, p = 0.0043, p = 0.0057, p = 0.0070, p = 0.0068, and p = 0.0445). Generally, signal intensities for IgG2 were higher than IgG1, indicating that the formulations all skewed more toward Th1 than Th2 response. Again, introducing the NMI LPS dampened the response for both IgG1 and IgG2 against CBU0612 and CBU0891. Additionally, IgG against NMI LPS was evaluated (Supplementary Figure S4). Even though LPS has been added in groups 4 and 6, the only observable response to NMI LPS is from the Q-VAX group, with signals concentrated in IgG2.

Guinea pigs used for immunogenicity screening were challenged 7 weeks post-prime with 5 × 10⁵ genomic equivalents of C. burnetii NMI RSA 493. Animals were monitored for weight and temperature change over a 14-day period post-infection. WCV (an in-house produced vaccine equivalent to Q-VAX) and sham-PBS were used as positive and negative controls, respectively.

Sham-vaccinated animals exhibited a marked decrease in body weight between days 9 and 12 and increased body temperature between days 7 and 10 when compared to the other groups (Figures 6A, B). WCV positive control animals showed steady increases in body weight and maintained consistent body temperature for the duration of the 2-week challenge. When observing changes in body weight, the groups that performed most similarly to WCV include groups 2, 4, and 6 with no significant difference on any of the days (Figure 6A). This includes both groups that include NMI LPS as an antigen. Change in temperature shows groups 1, 4, and 6 with no marked significance on any of the days when compared to WCV (Figure 6B). The monovalent group containing NMI LPS, group 6, outperforms the quadrivalent group lacking LPS, group 3, in both changes in body weight and temperature, which exhibits significance on days 10 and 8–10 in those categories, respectively (body weight d10 p = 0.0298; body temperature d8 p = 0.0002, d9 p = 0.0002, and d10 p = 0.0079). Meanwhile, the monovalent group 6 demonstrates no significant difference when compared to WCV for the entire 14-day duration. From this, we can conclude that the groups containing six antigens, groups 1 and 2, showed similar trends with little difference when compared to positive control WCV changes in body temperature and weight, and the inclusion of NMI LPS, in groups 4 and 6, provides significant additional protection.

The animals were terminated after the 14-day observation period. Splenomegaly, a hallmark of Q fever, was significantly lower in four of the six vaccine candidate groups when compared to PBS sham (Figure 6C) (p = 0.0013, p = 0.0022, p = 0.5071, p = 0.0055, p = 0.7927, and p = 0.0165). The top performers were groups 1, 2, and 4 and were determined by being the most comparable to WCV. Group 6 also showed a significant reduction in splenomegaly despite containing a single protein antigen.