Dysregulation of TNF-α and IFN-γ expression is a common host immune response in a chronically infected mouse model of melioidosis when comparing multiple human strains of Burkholderia pseudomallei

In a previous study, BALB/c mice were exposed to different clinical isolates of B. pseudomallei (Table 1) to determine their 21 day aerosol LD₅₀ to assess their relative virulence [29]. In the present study we examined survivors when they became available after the 21 day aerosol exposure study and examined their sera and spleens to evaluate their immune response to the infection (see Additional file 1: Figure S1). We examined 4–7 surviving mice after exposure to each strain from the same colony forming unit (CFU) exposure group whenever possible. Although there were differences in the time post-infection (PI) that we examined survivors after they were initially exposed (22–70 days PI), we still found apparently uninfected and chronically infected mice in many cases from the same exposure group. The number of survivors in each exposure group was influenced by the virulence of the B. pseudomallei strain that was evaluated [29]. We focused our studies on sera and spleens from the exposed mice because of the large number of samples that had to be processed and examined, and the spleens were a good representative organ of the course of infection because they became infected soon (1–2 days) after exposure to the pathogen [29, 30]. We validated the mice that were chronically infected by isolation of the pathogen from their spleens after the initial 21 day LD₅₀ study.

Figure 1 shows examples of some spleens examined from survivors from the aerosol exposure studies showing heterogeneity in infection. We examined survivors after they were exposed to 2 CFU of B. pseudomallei 1026b, 30 days PI (Fig. 1a). We recovered 40, 5.13 × 10⁷, and 3.61 × 10⁷ CFU from spleens shown in panels 1, 2, and 3, respectively. In Fig. 1b we examined survivors after exposure to 3 CFU of B. pseudomallei MSHR305, 35 days PI. In this case we recovered 0, 8.60 × 10⁶, and 7.68 × 10⁷ CFU from spleens shown in panels 1, 2, and 3, respectively. In Fig. 1c, we examined survivors after exposure to 2851 CFU of B. pseudomallei 1106a, 49 days PI. We recovered 0, 1.08 × 10⁸, and 1.48 × 10⁸ CFU from spleens in panels 1, 2, and 3, respectively. In summary, we recovered spleens from aerosol exposed mice that were heterogeneous in appearance that could be placed into 3 general categories based on size, CFU recovered, and the appearance of a pyogranulomatous lesion. These categories were: Gp 2, normal appearing spleens, 0 CFU recovered; Gp 3, enlarged/swollen spleens without a visable pyogranulomatous lesion, and 0 to few CFU recovered; and Gp 4, enlarged spleens with a visable pyogranulomatous lesion(s), and a large number of CFU recovered (> 1000) (chronically infected). Spleens from naïve mice that served as unexposed controls for comparison were placed into Gp 1.

Table 2 is a summary of the different categories of spleens recovered and analyzed from BALB/c mice exposed to each of the 11 individual clinical strains of B. pseudomallei with their LD₅₀ to show relative virulence and how they relate to the other strains [29]. Generally, with mice that were exposed to the most virulent B. pseudomallei strains MSHR5855 and HBPUB10134a, with reported LD₅₀ of 0.35 and 0.99 CFU, respectively, we saw more surviving mice with normal appearing spleens than mice with chronically infected spleens (presence of pyogranulomatous lesion(s) and isolated CFU) because most mice in the latter group that were acutely infected did not survive past 21 days PI. On the other hand, with mice exposed to the least virulent strain (B. pseudomallei 1106a (LD₅₀ 4266 CFU) we saw more chronically infected mice. We also listed the number of mice that survived in the group(s) of mice from which we obtained mice that we evaluated. We analyzed the immune state of each individual mouse after exposure to the different strains of B. pseudomallei (see below). (Although we were not able to examine mice after aerosol exposure to B. pseudomallei HBPUB10134a because they died by 21 days PI,) We had previously reported on the acute infection of BALB/c mice by B. pseudomallei HB10134a [29, 32].

We wanted to examine the antigen specific antibody level in survivors from the 21 day aerosol study after exposure to the different clinical strains of B. pseudomallei to evaluate the host humoral immune response to the pathogen. We examined the levels of B. pseudomallei specific IgG and the subclasses IgG1 and IgG2a in sera of BALB/c mice by ELISA that survived aerosol exposure to different strains of B. pseudomallei. Table 3 shows the antibody response in sera from mice arranged into groups according to the characteristics of the spleens that we recovered after the 21 day aerosol study. Results of antibody titers from sera that were repeated but not acceptable, for example if the reading was inhibited, or not complete were not included.

We saw little difference in the antibody IgG titer of BALB/c mice with normal appearing spleens (Gp 2) compared to that of naïve mice in Gp.1. In mice that we recovered enlarged spleens (Gp3) we saw a significant increase in the antibody IgG titer (P = 0.0002) and the subclasses IgG1 (P = 0.0002) and IgG2a titers (P = 0.0019) compared to naïve mice. There was a further increase in the IgG titer when we examined the sera from mice with chronically infected spleens (Gp 4). When we compared the ratio of the subclass IgG2a/IgG1 responses for Gp 3 and Gp 4, we found that mice with spleens in Gp 3 had a modest Th1-like response, while the ratio for the subclass response for mice with spleens in Gp 2 and Gp 4 were more of an equal Th1- to Th2-like response.

To evaluate the cellular host response in BALB/c mice after exposure to different strains of B. pseudomallei, we used flow cytometry to examine the cellular composition of spleens from chronically infected survivors. We assessed the amount of CD4⁺ and CD8⁺ T cells, B cells, monocytes/macrophages, NK cells, and granulocytes in spleens of surviving BALB/c mice and compared them to the cellular composition of spleens from naïve mice (Gp 1) (Fig. 2). Results of flow cytometry readings that were repeated when possible but not acceptable, such as readings were inhibited, or not complete were not included. There was little change in the cellular composition of spleens from BALB/c mice placed into Gp 2 (normal appearing, n = 32) compared to spleens from naïve mice. However, we saw small but notable increases in the percentages of CD8⁺ T cells and NK cells compared to the amounts in naïve spleens (P = 0.0114 and P = 0.0028, respectively). In enlarged spleens from BALB/c mice survivors belonging to Group 3 (n = 8) we observed a modest increase in the relative amounts of monocyte/macrophages (P < 0.0001), NK cells (P = 0.0085), and granulocytes (P < 0.0001) compared to spleens from naïve mice. In infected spleens (Gp 4) with visible pyogranulomatous lesions (n = 25), we found a substantial increase (P < 0.0001) in the percentage of monocyte/macrophages [29.26% (1.12)], NK cells [13.26% (1.20)], and granulocytes [23.25% (1.13)].

We wanted to examine the host inflammatory immune response in chronically infected survivors after exposure to different strains of B. pseudomallei. We thus examined the presence of inflammatory cytokines/chemokines in sera and spleen extracts from the aerosol exposed survivors after the 21 day aerosol study by Luminex technology (Additional file 2: Table S1). Results of Luminex readings that were repeated when possible but not acceptable, such as readings were not within the standard, or not complete were not included. We assessed 20 cytokines/chemokines in the sera of selected survivors, and we observed no significant increases in the level of cytokines/chemokines in sera from mice with normal appearing spleens (Gp 2, n = 32) compared to that in sera from naïve mice (Gp 1, n = 10). Of the cytokines/chemokines we examined in sera of BALB/c mice with enlarged spleens (Gp. 3, n = 8) none were elevated over that found in sera from naive mice (Gp. 1). In contrast, some cytokines/chemokines were significantly elevated (9/17 shown) in sera from mice with chronically infected spleens (Gp. 4, n = 23) compared to amounts in serum from naive mice (Gp. 1). These were IFNγ, IL-1α, IL-1β, IL-4, IL-6, IL-12, KC, MIG, and MIP-1α. Those cytokines/chemokines found in sera that were increased greater than 2-fold over that in naïve mice are shown in Fig. 3A (8/20 examined).

In contrast to sera, we saw a substantial increase in the number and amount of cytokines/chemokines present in spleen extracts from BALB/c mice that survived the 21 day aerosol study (see Additional file 2: Table S1). In spleen extracts prepared from normal appearing spleens in Gp 2 (n = 32), 3/17 cytokines/chemokines were modestly increased over that in spleen extracts from naïve mice in Gp. 1 (IL-1α, IL-1β, and MIG). In spleen extracts prepared from enlarged spleens in Gp. 3 (n = 8) there was a notable increase in 5/17 of the cytokines/chemokines (IL-1α, IL-1β, MIG, MIP-1a, and VEGF). In contrast, in spleen extracts prepared from chronically infected mice in Gp.4 (n = 23), the levels of 14/17 cytokines/chemokines were markedly increased (Additional file 2: Table S1). We showed in Fig. 3B the cytokines/chemokines (11) that were greater than 2-fold elevated over that in spleen extracts from naïve mice. They were GM-CSF, IFN-γ, TNF-α, IL-1α, IL-1β, IL-6, IP-10, KC, MIG, MIP-1α, and VEGF. Those that were not more than 2-fold elevated over that in naïve mice although were significantly increased were IL-2, IL-5, and MCP-1. The major elevated inflammatory cytokines/chemokines in both sera and spleen extracts were IFN-γ IL-1α, IL-1β, IL-6, KC, MIG, and MIP-1α (see Additional file 2: Table S1; and Fig. 3A, B).

When we examined the correlation of Th2-like cytokines (IL-4, IL-5, IL-10) to the Th1-like IFNγ in sera of Gp 3 mice, there was a modest Th1-like response (1.81-fold, the average of the fold-increase of the 3 cytokines divided into the fold-increase of IFNγ), which was similar to the IgG2a/IgG1 ratio (2.21) (see Additional file 2: Table S1 and Table 3, respectively, although we did not see this response in extracts from mice in Gp 3. When we examined the same relationship of cytokines in the sera of Gp 4 mice, we saw an overall higher Th1-like response (4.6-fold), which was more positive than the Th1/Th2 ratio of the antibody subclass response (1.21). At the same time when the relationship of these cytokines to the subclass response was examined in the spleen extracts from chronically infected Gp 4 mice, the Th1-like response was even overall higher (10.9-fold). Thus overall the cellular immune response in mice chronically infected with B. pseudomallei was noticeably toward a Th1-like response compared to what we observed with the antibody response.

We saw in B. pseudomallei chronically infected mice (Gp. 4) that we could recover high CFU from spleens, and we also could detect the expression of significant amounts of the proinflammatory cytokines IL-1α, IL-1β, TNF-α, and IFN-γ in their spleens. We wanted to investigate the cellular location of the pathogen relative to the location where the proinflammatory cytokines were expressed in the chronically infected spleens. To reveal the presence of B. pseudomallei in situ in these spleens, we probed the spleens with a high titer rabbit antibody that reacted with the exopolysaccharide of B. pseudomallei [33‐35]. A total of 15 sections of different spleens were stained with hematoxylin and eosin (H&E), and 16 spleen sections were probed with an anti-B. pseudomallei rabbit antibody (naïve, normal appearing, and chronically infected spleens from B. pseudomallei strains 22, K96243, or 1106a). In Fig. 4 we show representative spleens from BALB/c mice that are (A) a spleen from a naïve BALB/c mouse, (B) a normal appearing spleen (from B. pseudomallei Bp22 exposed), and (C and D) two chronically infected spleens (from B. pseudomallei 1106a exposed). In the normal appearing spleen (B) we did not see any exopolysaccharide positive staining material (Fig. 4B, panels 3 and 4) by immunohistochemical (IHC) analysis. These are similar to results that we saw when spleens from naïve mice were examined (Fig. 4A). In the examples of infected spleens from chronically infected BALB/c mice (Fig. 4C and D) we saw immunoreactive staining exopolysaccharide material associated with pyogranulomatous lesions that were nodular and largely multifocal that were primarily composed of a mixture of neutrophils, macrophages, and lymphocytes often with a thin surrounding layer of epithelioid macrophages (Fig. 4C and D, panels 3 and 4). We choose to show the spleen in Fig. 4D because it represented a good example of an early pyogranuloma lesion, compared with the spleen in Fig. 4C which showed a gross pyogranuloma, even though the mice were both exposed to Bp 1106a. We also observed the pathogen in multifocal pyogranulomatous lesions encased within a single large capsule-like structure that may displace in some cases more than half the spleen, in addition to separate individual pyogranulomatous lesions (see Fig. 4C).

We then wanted to determine the location of the proinflammatory cytokines relative to the location of the pathogen in spleens from exposed mice. In the following study and those below, we showed by different methods that we could detect the expression of proinflammatory cytokines, specifically TNF-α and IFN-γ expression in splenocytes from chronically infected mice. Because of the difficulty in obtaining reliable antibodies to probe protein targets in formalin-fixed, paraffin-embedded tissue, we used an in situ RNA hybridization method to locate the expression of proinflammatory cytokines (IL-1α, IL-1β, TNF-α, IFN-γ) in spleens from B. pseudomallei exposed mice [36]. We examined a total of 12 spleens with each set of 4 different cytokine RNA probes: 2 naïve spleens; 5 normal appearing spleens (exposed to B. pseudomallei Bp22 or K96243); or 5 infected spleens (exposed to B. pseudomallei K96243 or 1106a). In Fig. 5B and C we show examples of in situ RNA hybridization to detect the expression of IL-1α, IL-1β, TNF-α, and IFN-γ in two B. pseudomallei 1106a infected BALB/c spleens, 49 days PI. In Fig. 5A, panels 1–4, we show the results of the in situ RNA hybridization with the 4 RNA probes with a spleen from a naïve mouse where no concentrated cluster or pattern of a positive signal was seen with the probes. In Fig. 5B and C, panels 1–3, we saw at least two types of staining with the RNA probes for IL-1α, IL-1β, and TNF-α. First, there was a positive signal that appeared to be predominately located within the outer edge of an encapsulated structure that contained multifocal pyogranulomatous lesions. The multifocal pyogranulomatous lesions within the large encapsulated structure did not stain with any of cytokine RNA probes tested although they were positive for the presence of the exopolysaccharide (see Fig. 4C). Some epithelioid macrophages within the fibrotic region surrounding the pyogranulomatous encapsulated region also stained with the IL-1α, IL-1β, and TNF-α RNA probes. The second type of staining we saw with the three cytokine probes were individual pyogranulomatous lesions located in different areas of the spleen. The three RNA probes also hybridized to most of the lesion with some in cells in the immediate surrounding area (Fig. 5B, C, panels 1–3). With the IFN-γ RNA probe, however, we saw fewer cells stained than with the other three cytokine probes in the infected spleen (Fig. 5B and C, panels 4), and the few IFN-γ stained cells also appeared to be located near the pyogranulomatous lesions. No other region further outside of the pyogranulomatous lesions showed many positive cells arranged in a similar pattern as the proinflammatory cytokine RNA probes. Overall, we saw intense abundant staining with the IL-1α and IL-1β RNA probes within the pyogranulomatous lesion or within the capsule structure surrounding pyogranulomatous lesions. We observed less intense but still abundant staining with the TNF-α RNA probe in the same areas. We saw much fewer cells stained with the IFN-γ RNA probe that were not arranged in the same pattern as the other cytokine probes. In comparison, in spleens from naïve mice we saw few to moderate number of cells scattered throughout the spleen that were positive with the IL-1α, IL-1β, and TNF-α RNA probes but not any with the IFN-γ RNA probe (Fig. 5A). There was no organized pattern to these signals as seen in spleens with pyogranulomatous lesions. In summary, the presence of both the pathogen and the heightened expression of proinflammatory cytokines were both generally confined within the pyogranulomatous lesion in infected spleens from chronically infected mice.

Since we recovered live microorganisms from spleens from chronically infected BALB/c mice after aerosol exposure to different strains of B. pseudomallei, it suggested to us that spleen cells in the infected spleens were persistently stimulated by the presence of the pathogen, and therefore, they would be actively expressing cytokines/chemokines in response to the pathogen, and we should be able to detect this activity ex vivo. We began these series of studies after we started examining the chronically infected mice, and realized that there was a possibility that splenocytes could be immunologically activated. We were able to examine 8 infected spleens (see below) to support our hypothesis. To investigate this theory, we cultured spleen cells from exposed mice ex vivo under three culture conditions without the presence of a protein transport blocker: 1) with media only (unstimulated), 2) with inactivated B. pseudomallei whole-cells, or 3) with phorbol myristate acetate (PMA) and ionomycin as a positive control. After 2 days of incubation, we measured the presence of cytokines/chemokines in the supernatant. In Table 4 we demonstrated the activation of splenocytes by presenting the results of spleen cells (total of 20 preparations shown, that includes 2 from naïve mice) prepared from BALB/c mice exposed to 9 different strains of B. pseudomallei (1106a, 406e, MSHR668, 1026b, MSHR5858, MSHR5855, MSHR5848, HBPUB10134a, and Bp22) and arranged into the three different groups according to the characteristic of the spleen (see Fig. 2). We also showed the results of spleen cells from two naïve mice (Gp 1) incubated under the same conditions for comparison. With spleen cells from naïve mice we saw very little difference in the amount of IFN-γ expressed when incubated with the whole-cell antigen or without the antigen (media only), while with PMA and ionomycin we saw a substantial amount of IFN-γ produced. With spleen cells from normal appearing spleens (Gp. 2, n = 7) in the media only culture condition little IFN-γ was produced, and when these spleen cells where incubated with the whole-cell antigen there was a small increase of IFN-γ produced but not significant when compared to cells from naïve mice. However, a similar amount of IFN-γ was produced by these spleen cells when incubated with PMA and ionomycin as seen with spleen cells from naïve mice. Spleen cells prepared from enlarged spleens (Gp.3, n = 3) produced little to moderate amounts of IFN-γ in media only conditions, but they produced significant amounts of IFN-γ (P < 0.0100), after restimulated with whole-cell antigen [5297 (1.27) pg/ml] compared to naïve cells or slightly higher amounts of IFN-γ with PMA and ionomycin [6381 (1.26) pg/ml]. In contrast, we saw significant amounts of IFN-γ produced (P < 0.0010) by spleen cells prepared from infected spleens (Gp. 4, n = 8) in the presence of media only [4440 (1.48) pg/ml] compared to spleen cells from naïve mice. IFN-γ was also produced by these spleen cells when they were restimulated with the whole-cell antigen [2727 (1.27) pg/ml, P < 0.0009)] and slightly higher amounts with PMA [6117 (1.17) pg/ml. At the same time, we also measured the expression of two other cytokines as examples that no exogenous antigen was needed for their expression by splenocytes from chronically infected mice. In Table 4 we showed the amount of TNF-α and MIP-1α we found in the same cultures as we examined for IFN-γ. However, with these two inflammatory cytokine/chemokines we did not see consistent expression or stimulation in the same cultural conditions as we did for IFN-γ.

To further explore the idea that spleen cells from B. pseudomallei chronically infected BALB/c mice were persistently stimulated, we wanted to examine spleen cells from chronically infected mice for the intracellular expression of TNF-α and IFN-γ, and at the same time identify the type of cells that were expressing the cytokines by flow cytometry. These two cytokines play an important role in protection and pathogenesis in melioidosis [37‐41]. Similar to the previous ex vivo study where these studies were started after we examined chronically infected mice, we were able to demonstrate this phenomenon in 14 chronically infected spleens that were infected by 7/10 different strains of B. pseudomallei (see below), as well as 28 normal appearing spleens and 7 swollen spleens. Results from spleens that were not complete were not included. For this study spleen cells were neither exogenously stimulated before examination nor were inhibitors of intracellular transport/secretion, such as Brefeldin A or Monensin used [42‐44]. Examples of the flow cytometric analysis of intracellular TNF-α or IFN-γ in spleen cells from different types of spleens from aerosol exposed BALB/c mice are shown in Additional file 3: Figure S2AB, respectively.

Figure 6A shows the results of the cytometric analysis of the presence of intracellular TNF-α in spleen cells from 49 BALB/c mice that were exposed to different strains of B. pseudomallei, and compared them with the analysis of spleen cells from naïve mice (n = 10). Spleen cells from naïve BALB/c mice (Gp 1) showed very low levels of intracellular TNF-α in all the cell types that were examined (CD4⁺ T cells, CD8⁺ T cells, B cells, monocyte/macrophages, NK cells, and granulocytes). In spleen cells from normal appearing spleens (Gp. 2, n = 28, exposed to B. pseudomallei strains 406e, K96243, MSHR668, MSHR5855, MSHR5848, and HBPUB10134a), we detected small increases of intracellular TNF-α in CD4⁺ and CD8⁺ T cells and detected a modest increase in the percentage of B cells (P = 0.0349) with intracellular TNF-α compared to that in spleen cells from naïve mice. We saw a further increase in the percentages of monocyte/macrophages (P < 0.0001), and NK cells (P = 0.0186) that contained intracellular TNF-α, but not in granulocytes from normal appearing spleens. In spleen cells from enlarged spleens (Gp. 3, n = 7, exposed to B. pseudomallei strains 406e, 1106a, MSHR305, MSHR668, 1026b, and MSHR5854, we saw small increases in the percentage of CD4⁺ and CD8⁺ T cells with intracellular TNF-α, and a modest increase in the number of B cells with intracellular TNF-α. However, we detected small but significant increases in the number of monocyte/macrophages (P = 0.0021), NK cells (P = 0.0080), and granulocytes (P = 0.0007) that contained TNF-α. In infected spleens (Gp. 4, n = 14, exposed to B. pseudomallei strains 406e, 1026b, 1106a, K96243, MSHR668, MSHR5848, and MSHR5855) we observed a notable increase in the percentage of CD4⁺ T cells (P = 0.0417) but not in CD8⁺ T cells or in B cells that contained TNF-α. In contrast, we detected a significant (P < 0.0001) number of monocyte/macrophages [21.8 (1.16)%], NK cells [19.3 (1.22)%], and granulocytes [11.23 (1.24)%] that contained TNF-α compared to spleen cells from naïve mice.

When we examined the same spleen cells for the presence of intracellular IFN-γ (Fig. 6B), we saw very little IFN-γ expressed in all cell types from naïve BALB/c mice (Gp. 1). In spleen cells from normal appearing spleens (Gp. 2) there was very little IFN-γ detected, except in NK cells (P = 0.0008) compared to that present in naïve spleen cells. We saw an increase in the percentage of cells that contained intracellular IFN-γ in enlarged spleens (Gp. 3), specifically in monocyte/macrophages (P = 0.0189), NK cells (P = 0.0160), and granulocytes (P = 0.0020), but not in CD4⁺ or CD8⁺ T cells or in B cells. When we examined cells from infected spleens (Gp. 4), we saw a significant increase (P < 0.0001) in the percentage of monocyte/macrophages [21.6 (1.16)%], NK cells [17.8 (1.26)%], and granulocytes [14.0 (1.21)%] that contained intracellular IFN-γ compared to that in naïve spleen cells, but not in CD4⁺ and CD8⁺ T cells or in B cells. We have summarized the findings in the present study when comparing different strains of B. pseudomallei in the chronically infected BALB/c mouse model of melioidosis (Table 5).

Stock cultures of human clinical isolates of B. pseudomallei were collected by the Critical Reagents Program at the US Army Medical Research Institute of Infectious Diseases (USAMRIID) under the guidance of the Defense Threat Reduction Agency and Biomedical Advanced Research and Development Authority (Table 1) [29]. The Bp strains were obtained from sources that had the least number of passages after isolation from clinical samples. Single use stock cultures were obtained from the Critical Reagents Program and stored at − 70 °C until used.

Groups of female BALB/c mice 6–8 weeks old were obtained from the National Cancer Institute, Frederick, MD. For each aerosol challenge study with a particular strain of B. pseudomallei, groups of mice were each placed into five groups with 10 mice per group per housing pan for a total of 50 mice in each study. Mice were housed in clean, autoclaved, filtered top pans, with mouse litter, an enrichment item, and given water and food ad lib., and mouse litter and pans were changed twice a week. Frozen stock cultures of B. pseudomallei selected for aerosol challenge were used only once to inoculate 200 ml of 4% glycerol-1% tryptone (BD, Sparks, MD) broth containing 0.5% NaCl (GTB) and incubated overnight for 24 h at 37 °C at 200 rpm. The 24 h culture was diluted to appropriate concentrations in GTB (five concentrations) and used for aerosol challenges by the Center for Aerobiological Science at USAMRIID. Whole-body aerosol exposures of mice were performed as previously described [74] using a 3-jet Collison nebulizer and aerosols monitored with the Automated Bioaerosol Exposure system, after which they are kept in Biosafety Level 3 conditions. Estimation of the amount of aerosol exposure colony forming units (CFU) were determined after each run using the all-glass impinger (AGI), and CFU were determined after an aliquot of each approximately 10-fold dilution of the culture in the AGI was plated in triplicate on sheep blood agar (SBA) plates and incubated at 37^o C for 2–3 days.

All animals after challenge were observed twice daily for changes in their condition for the LD₅₀ study, and survivors were watched up to 60–70 days post-challenge for changes in appearance. Animals that appeared to be moribund according to an intervention score (which considers overall their appearance, natural behavior, and provoked behavior: Scores of 0–2, normal behavior; 3–7, requires increased monitoring, twice daily; ≥8, moribund) during the study or left over after the studies were humanely euthanized by CO2 inhalation or after anesthesia with a ketamine-acepromazine-xylazine solution (0.1–0.2 ml, given intraperitoneal) and euthanized with a pentobarbital based euthanasia solution given intraperitoneal. Animals that were used in the chronic infection study, were deeply anesthetized with the anesthesia solution given intraperitoneal (0.1–0.2 ml, as stated above), before exsanguination for sera and removal of spleens.

All animal use and ethical evaluation had been approved at USAMRIID before starting any proposed animal studies by the Institute IACUC committee, which issued the following statement: Research was conducted under an IACUC approved protocol in compliance with the Animal Welfare Act, PHS Policy, and other federal statutes and regulations relating to animals and experiments involving animals. The facility where this research was conducted is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, International and adheres to principles stated in the 8th Edition of the Guide for the Care and Use of Laboratory Animals, National Research Council, 2011.

Immunoglobulin IgG and subclasses IgG1 and IgG2a levels were determined by ELISA in triplicate and performed at least once as previously described [75]. We did not report the results of samples that did not perform properly, for example we could not obtain an end point to determine the titer. Briefly, irradiated-inactivated B. pseudomallei K96243 cells (10 μg/ml in 0.1 M carbonate buffer, pH 9.5), was used to coat a 96-well Immulon 2HB, round-bottom plate (Thermo-Fisher, Pittsburgh, PA) overnight at 4 °C. After washing and blocking the antigen-coated plate, two-fold dilutions of the mouse sera prepared in blocking solution was made in the 96-well plate in triplicate, and plate incubated for 1 h at 37 °C. The plates were then washed and developed with 50 μl of 1/5000 dilution of anti-Ig-horseradish conjugate (Southern Biotechnology Associates, Inc., Birmingham, AL). After the final color development step, the plates were read at 450 nm with a reference filter (570 nm). The results are reported as the reciprocal of the highest dilution giving a mean OD of at least 0.1, which was at least 2-fold over the background ±1 SD, and the final antibody titer was reported as the geometric mean with geometric standard error of the mean (SEM). The results were combined with other mice belonging to the same spleen group, and geometric mean with the standard error of the geometric mean determined. The antibody titer against B. pseudomallei K96243 in naïve mice (n = 10) was usually below 50, but for statistical purposes it was reported as 50, and it was used to compare with the antibody titer of the mice in the different spleen group for statistical analysis.

Splenocytes were prepared from spleens from surviving mice after the 21 day LD50 study was completed as previously described [75]. Briefly, spleens were removed from exsanguinated mice and disaggregated in RPMI 1640 medium (Gibco, Life Technologies, Grand Island, NY) containing 25 mM HEPES and 2 mM glutamine (wash medium) to make the spleen extract. CFU in the freshly prepared extract were determined on sheep blood agar (SBA) plates (BD Diagnostics, Franklin Lake, NJ) with undiluted extract or 10-fold dilutions in sterile water on triplicate SBA plates. Plates were incubated at 37 °C for 2-3 days before counting colonies. The limit of detection was 20 CFU/ml. A portion of the spleen extract was saved for cytokine/chemokine determination after being stored at -70 °C and irradiated to sterilize the extract. An aliquot of the irradiated extract was spread on sheep blood agar plates to check for sterility after 3 days of incubation at 37 °C before use. With the rest of the spleen homogenate, the red cells were lysed with ACK Lysing Buffer (BioWhittaker, Walkersville, MD), and the extract was diluted with wash medium and cells pelleted by centrifugation at 1200 rpm for 10 min. Splenocytes were then washed once and suspended in PBS or complete medium [wash medium containing 10% heat-inactivated fetal calf serum (Life Technologies), 1 mM sodium pyruvate, 0.1 mM non-essential amino acids, 100 U/ml of penicillin, 100 μg/ml streptomycin, and 50 μM 2-mercaptoethanol] and cells counted.

Cytokines and chemokines in serum, spleen extracts, and splenocyte culture supernatants were done in duplicate at least once and measured by Luminex MagPix (Life Technologies, Grand Island, NY) as per manufacturer directions using the Mouse Cytokine Magnetic 20-Plex Panel kit. We did not report the results of samples that did not perform properly, for example, if the reading was not within the standard curve. The cytokines/chemokines measured were the following: FGFb, GM-CSF, IFNγ, IL-1α, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12, IL-13, IL-17, IP-10, KC, MCP-1, MIG, MIP-1α, TNFα, and VEGF.

Approximately 10 × 10⁶ splenocytes prepared from each mouse spleen was suspended in PBS, washed in FACS staining buffer (FSB) (1XPBS, 3% fetal calf serum, Life Technologies), and fixed in FSB containing 4% formaldehyde (Pierce, Rockford, IL) at 4 °C. When Live/Dead Fixable Aqua Dead Cell stain (Molecular Probes, Life Technologies, Eugene, OR) was used, 5 μl of the Aqua Dead Cell stain was added (1 μl per 2 × 10⁶ cells) to the cells in PBS, and cells incubated for 30 min at room temperature protected from light. After incubation, the cells were centrifuged for 10 min at 1200 rpm, and cells were washed once with 1.0 ml of FSB and suspended in 1.0 ml of FSB containing 4% paraformaldehyde and incubated for 1 h at 4 °C. The cells were then washed twice with FSB and suspended in 1.0 ml of FSB. Sterility of the cell suspension was checked by plating 0.1 ml (1/10 volume) onto SBA plates, incubating the plates at 37 °C for 3 days and checked for growth. Cells were then distributed into a microtiter plate (2-5 × 10⁵ cells/well) in duplicate and nonspecific binding was inhibited by the addition of Fc Block (BD Biosciences, San Jose, CA). Cells were labeled with the following specific antibodies (BD Biosciences): CD4 T cells, CD4-PE/CD44-FITC; CD8 T cells, CD8-PE/CD44-FITC; B cells, B220-PE/CD86-FITC; monocytes/macrophages, CD11b-PE/CD44-FITC; NK cells, CD49b-PE/CD44-FITC; and granulocytes, Ly6G-PE/CD44-FITC. Corresponding isotype controls were used and all were incubated for 60 min on ice. Samples were then washed with FSB and incubated with 1X Perm/Wash buffer (BD Biosciences) for 20 min. Intracellular cytokine expression was detected by incubation with anti-IFN-γ-APC or anti-TNF-α-APC for 60 min on ice. All samples were stored in FSB at 4 °C until analysis (and between repeat readings). Cells were identified with a BD FACSCalibur using CellQuestPro software (BDBiosciences). All samples were read at least once. Data was then analyzed with FlowJo Software v10 (FlowJo, Ashland, OR). Gating was established using both fluorochrome isotype controls and single stain controls for each respective assay or against naïve, uninfected splenocytes from BALB/c mice prepared as described above. The results from the exposed mice were placed into groups according to appearance of the spleen (see Fig. 1) and compared to the geometric mean of naïve mice (n = 10), unless indicated otherwise, for statistical analysis. We did not report on samples that did not give appropriate readings, such as samples that did not appear to have sufficiently labelled the cell types or have internal labeling perform correctly on infected samples.

Cytokine/chemokine expression was examined after antigen stimulation of splenocytes prepared from exposed or naive mice by a modification of a procedure previously described [76]. Briefly, 2X10⁶ splenocytes were incubated with either no antigen (media control), 2X10⁷ irradiated, inactivated homologous B. pseudomallei (challenge strain) bacterial cells, or PMA (100 ng/ml) with ionomycin (500 ng/ml) in duplicate wells of a 48-well, flat bottom, cell culture plate (Costar 3548, Corning, NY) in a final volume of 0.5 ml for 42–46 h at 37 °C + 5% CO₂. After the incubation period, the cell culture plate was centrifuged at 1200 rpm for 10 min, and the culture supernatant was collected and irradiated to sterilize the sample. Sterility was checked with 0.05 ml of sample plated on a SBA plate, and plate incubated for 3 days at 37 °C. Cytokine/chemokine expression was determined as described above. The combined results from each spleen group is shown in Table 4, and they were compared with the results from the naïve mice group (n = 2) for statistical analysis. Not all samples were analyzed in this manner because the assay was not performed until the study was underway.

Spleens from mice were placed into 10% neutral buffered formalin for at least 21 days before they were taken out of biosafety level 3 conditions. The spleens were embedded in paraffin and sectioned for staining with hematoxylin and eosin (H&E). For immunohistochemical analysis, a rabbit polyclonal antibody raised against an extract of B. pseudomallei K96243 (a kind gift from Dr. Robert Ulrich) was used to identify the B. pseudomallei exopolysaccharide antigen in formalin-fixed, paraffin-embedded tissue [30, 35].

To examine the expression of cytokines (IL-1α, IL-1β, TNF-α, and IFN-γ) in formalin-fix, paraffin-embedded tissue, a RNAscope technology (Advance Cell Diagnostics, Inc., Hayward, CA) was used that targets the RNA of the cytokine of interest by in situ hybridization (ISH) [36]. In brief, tissue sections were fixed onto slides and pretreated to unmask target RNA and cells permeabilized with RNAscope Pretreatment solutions. Special designed proprietary paired Z-probes that hybridized to adjacent RNA sequences ensured the specificity of the amplified signal. It was unlikely that adjacent sequences that are targeted by two separate probes would be found on another RNA target. There were up to 20 paired Z-probes for each targeted (cytokine) RNA. After the initial hybridization, pre-amplifiers hybridized to paired Z-probes which were followed by amplifiers that bound to multiple binding sites on each preamplifier. Labeled probes that contain a chromogenic enzyme bound to multiple sites on each amplifier. After incubation with a red substrate-chromogen solution for 10 min at room temperature, sections were then counter stained with hematoxylin, air dried, and mounted. A bright field microscope was used to examine the stained specimens. See the individual Figure Legends for the number of test samples used that was compared to the samples from a naïve mouse.

Log transformation was applied to cytokine/chemokine concentrations (in sera, cell extracts, culture supernatants, and intracellular), antibody titers, and cell distribution prior to analysis, with comparisons between selected group means made by Welch’s t-test. Results were summarized as geometric mean and geometric standard error. No multiple comparison adjustment has been applied to the reported p-values. Where detection limits were encountered in cytokine/chemokine data, the limit of detection was imputed. Analysis was implemented in SAS version 9.4, SAS Institute Inc., Cary, NC, USA. Results were considered significant when P < 0.05.