All experiments with animals in this study were performed in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals published by the Ministry of Science and Technology of China. Animal ethics in all experimental procedures were approved by the Institutional Animal Care and Use Committee of China Agricultural University. All efforts were made to minimize animal suffering.

Murine macrophage ANA1cells (Cox et al., 1989) were cultured in complete DMEM, as described previously (Soloviev et al., 2002). ANA-1 cells were produced from bone marrow macrophages of the C57BW6 mice by infection with the J2 recombinant retrovirus expressing v-myc/v-raf. The cells express markers of the differentiated macrophage, including Ly-5, Mac-I, and FcyR, but do not express either B or T lymphocyte cell markers. T. gondii of Pru strain was maintained in vitro by serial passage on confluent HFF monolayers in DMEM containing 25 mM glucose and 4 mM glutamine supplemented with 10% newborn calf serum (NCS, Hyclone, USA) (Wang et al., 2014). The cells were incubated at 37°C with 5% CO₂ in a humidified incubator. The medium was changed 6 h after inoculation, and thereafter every 24 h.

Inbred strains of 6- to 8-week-old Balb/c female mice (center of Beijing experimental animals) were used for Pru strain T. gondii infection. Briefly, Pru strain tachyzoites were harvested and purified using a 5 μm filter, and infected at 2 × 10⁶/mouse dose through intraperitoneal injection. 5 mL PBS was injected into the abdominal cavity for acquisition of mouse peritoneal exudate cells and exudate at 6, 12, 24, 48, 72, and 96 h after infection, lavage fluid were centrifuged at 600 g for 10 min at room temperature, peritoneal exudate cells in the pellet and supernatant were collected, respectively. Supernatants were then precipitated and concentrated by the precipitation reagent (equal volumes of acidified acetone and methanol) for HMGB1 assay. Cells were harvested for the examination of transcription and expression of HMGB1.

The mice PMΦs were isolated as described previously (Chakraborty et al., 2013). Pru strain tachyzoites were harvested and purified using a 5 μm filter. The parasites were inoculated on to the macrophage monolayers at a parasite to host cell ratio of 5:1. Supernatants were collected at the indicated time points, and then precipitated and concentrated as described above. Cells were harvested for the examination of transcription and expression of HMGB1.

Mouse PMΦs and ANA1 monolayers were cultured in complete RPMI 1640 medium in 24 well culture plates (1 × 106 cells/well) with 10 ng/mL IFN-γ (Sigma–Aldrich, USA) or 2 μM Z-VAD-FMK (Bioversion, USA) or ac-YVAD-cmk (Sigma–Aldrich, USA) for 6 h. All chemicals were used at non-cytoxic concentrations, which means macrophages with >90% viability were examined by Trypan blue staining after treatment. Thereafter, the medium was refreshed and T. gondii tachyzoites were inoculated at a parasite to host cell ratio of 5:1. Further incubation was carried out for different time periods as indicated. The supernatant and cell pellet were collected separately for HMGB1 assay as described above.

Total RNA was isolated from ANA1 cells and mouse PMΦs that were uninfected or infected by pru parasites using the RNA isolation kit version 2.11 (Takara Biotechnology, Dalian, Co., Ltd) according to the manufacturer’s directions, and treated with DNase I (Takara) to remove any residual genomic DNA. cDNAs were synthesized using Oligo (dT)18 and random 6-mers using the cDNA Synthesis Kit (Takara). Specific primers were designed using Primer Premier 5.0, including primers for Mus musculus HMGB1 (GI: 114326547, forward: 5′-GAGACCAAAAAGAAGTTCAAGGAC-3′ and reverse: 5′-TGCAACATCACCAAT GGATAA G-3′), and GAPDH (forward: 5′-GGTGAAGGTCGGTGTGAACG-3′ and reverse: 5′-CTCGCTCCTGGAAGATGGTG-3′). Specificity of these primers was evaluated using conventional real-time (RT)-PCR. Quantitative RT-PCR (ΔΔCt method) was performed using the ABI Prism 7500 System (Applied Biosystems Inc., USA) with Tli RNaseH Plus kit (Takara). The RNA concentration was normalized to GAPDH.

Immunofluorescence assays (IFA) for HMGB1 localization and translocation in ANA1 and mouse PMΦs were performed as described previously (Wang et al., 2014). Briefly, cells cultured on coverslips and subjected to different treatments were fixed, permeabilized, and blocked prior to incubation with the rabbit anti-HMGB1 monoclonal antibody (ABcam, UK, diluted 1:200 in 1% BSA-PBS) for 1 hr at 37°C. The coverslips were washed and then incubated for 1 h at 37°C with FITC-conjugated goat anti-rabbit IgG (H+L) diluted to 1:50 in PBS with 1% BSA (Proteintech, USA). Nuclei were stained with Hoechst 33258 (Sigma–Aldrich, USA) and coverslips were subsequently mounted onto slides. Fluorescence images were obtained through the Leica microsystem (Leica TCS SP5 II, Germany) and epifluorescence optics were obtained using an oil immersion lens with 63× magnification. Collected images and MIFs (mean index of fluorescence) were processed and analyzed using the LAS AF lite 2.2.0 software (Leica). Mouse anti-TgHMGB1a antibodies were used to trace the T. gondii (Wang et al., 2014).

To examine HMGB1 protein, concentrated supernatant and cells (T. gondii infected ANA1 cells and mouse PMΦs), peritoneal lavage cells and fluid from the T. gondii infection mice were lysed using RIPA buffer (Beyotime, Beijing) with a cocktail of protease inhibitors. The lysate was loaded on 12% (w/v) SDS-PAGE. After electrophoresis, separated strips were transferred onto polyvinylidene fluoride (PVDF) membranes (Millipore, USA) together with a visible prestained protein marker. Membranes were blocked overnight with 5% (w/v) skim milk in PBS at 4°C, and then incubated with the rabbit anti-HMGB1 monoclonal antibody (diluted 1:600 by 3% BSA-PBS). After washing in PBST (0.5‰ Tween-20) for three times, the membranes were incubated for 1 h with goat anti-rabbit IgG (H+L) horseradish peroxidase (HRP)-labeled secondary antibody (Proteintech, USA) diluted to 1:10,000 in 3% BSA-PBS. Finally, the blot was developed using enhanced chemiluminescence reagents (Co Win Biotech Co., Ltd., Beijing). Antibodies against GAPDH were used as endogenous control.

Statistical test were performed in Statistical Product and Service Solution 22.0 (SPSS 22.0, IBMCo., USA). Two-tailed P-value was calculated, and P < 0.05 was consider significant.

Pru strain tachyzoites were inoculated into ANA1 cells with or without IFN-γ pretreatment. HMGB1 transcription continuously declined in ANA1 cells over the first 24 h after T. gondii infection, which may suggest inhibition of cell growth by T. gondii invasion. HMGB1 transcription rose at 48 h (Figure 1A), possibly related to cell death in the late stage of T. gondii infection in vitro. In parallel tests, HMGB1 transcription increased at 12 h after T. gondii infection, and then declined before a sharp increase at 48 h in the IFN-γ pretreated ANA1 cells (Figure 1A). To verify whether the changes occurred in vivo, peritoneal exudate cells from the T. gondii infected mice were gathered for examination. Quantitative-RT-PCR results indicated that, in the early 12 h after infection, HMGB1 transcription began to rise, which is similar to the test in vitro using IFN-γ treated ANA1 cells (Figure 1A), but then decreased gradually to below the basal level, and there were no upward trends again in the detecting period (Figure 1B). These results probably implied that IFN-γ lead to upregulated HMGB1 transcription at the early stage of pru infection in vivo.

To investigate whether T. gondii infection induces HMGB1 secretion, supernatant and the infected ANA1 cells were collected for western blot. At the early stage (∼24 h), consistent with the transcription level of HMGB1 (Figure 1A), intracellular HMGB1 gradually decreased after Pru infected the ANA1 cells (Figure 2A). Even at 48 h, the specific HMGB1 strip almost disappeared, but the transcription level of HMGB1 was dramatically rising at 48 h (Figure 1A). Simultaneously, Ana1 cells began to secret HMGB1 to the supernatant soon after infection (Figure 2A). Interestingly, there was only one specific band about 25 kD of HMGB1 in the supernatant at 12 h as expected, but afterward two smaller molecular weight bands emerged, with a trend of increasing with time. Based on the high specificity of Anti-HMGB1 monoclonal abs (Supplementary Figure S1), we proposed that the two new bands might be formed by the splicing of HMGB1 by caspase (LeBlanc et al., 2014) or unknown protein kinases, or other shear degradation mechanisms. Our finding showed that T. gondii infection induces HMGB1 secretion in ANA1 cells; however, the function of HMGB1 release and processing in T. gondii infection need further investigations.

To validate the putative changes of HMGB1 during T. gondii infection in vivo, peritoneal exudate cells of the pru strain tachyzoites infected mice were collected for examination. No significant changes of cellular HMGB1 were observed in the total peritoneal exudate cells lysates during T. gondii infection (Figure 2B). Further peritoneal lavage fluid results showed that HMGB1 was secreted at 6 h after infection and continuous secretion was detected in the following periods (Figure 2B), similarly to the in vitro assay using ANA1 cells (Figure 2A). Although the amounts were less, which possibly as a result of HMGB1 binding to its receptors in vivo (detected in the cell lysates).

To detect possible phenotype changes of ANA1 cell line and to further investigate the cell source of HMGB1 secretion in the peritoneal cavity of T. gondii infected mice, mouse PMΦs were isolated for T. gondii infection. Results of IFA and MIFs statistics showed that the intracellular HMGB1 of IFN-γ treated PMΦs was at least reduced by half compared with normal PMΦs (Figure 3), while T. gondii infection induced the intracellular HMGB1 increasing gradually again. In parallel experiments, HMGB1 in untreated PMΦs slightly decreased at 6 h post T. gondii infection, and then rose to the basal level at 12 h (Figure 3A). We also monitored HMGB1 in supernatant and HMGB1 released after T. gondii infection from 6 to 36 h (up panel in Figure 4B). Western blot elicited only one HMGB1 specific band with the expected molecular weight (about 25 kD), differing from in the ANA1 test (Figure 2A). These results demonstrated that the macrophage was one source of HMGB1 secretion in T. gondii infection, and T. gondii infection can possibly resist IFN-γ mediated cell killing, which leads to HMGB1 expression recovery in the ANA1 cells and PMΦs.

We explored whether T. gondii-induced HMGB1 release is dependent on NLRP1/3-Caspase-1 inflammasomes pathways. Primary mouse PMΦs were pretreated with a cell permeable irreversible pan-caspase inhibitor Z-VAD-FMK and or ac-YVAD-cmk, which potentially inhibits caspase-1 activity (Dostert et al., 2009; Chakraborty et al., 2013). Intracellular HMGB1 had no obvious changes in caspase inhibited mouse PMΦs without inoculation, while after parasitic infection, the nuclear HMGB1 initially decreased (Figures 3C,D). Unlike the IFA visualization and MIFs data, western blot showed that the total HMGB1 of mouse PMΦs with different treatment were not significantly different after T. gondii infection (Figure 4A). The result of the IFA and MIF statistics should accurately demonstrate that nuclear HMGB1 decreased and dispersion after one cell was infected by T. gondii parasites, while the western blot only shows the total cell HMGB1, and cells without invasion in the experiment can’t be excluded.

We further detected HMGB1 in the supernatant of mouse PMΦs culture with IFN-γ and z-VAD-fmk or ac-YVAD-cmk. After pru strain parasites infection stimulated a sustainable time-independent release of HMGB1 from the normal PMΦs (up panel of Figure 4B), and as expected, a large number of HMGB1 were released upon T. gondii infected the IFN-γ primed PMΦs (bottom panel of Figure 4B), both the z-VAD-fmk or ac-YVAD-cmk inhibitor treatment delayed and suppressed HMGB1 release. HMGB1 was detectable in the supernatant only after 24 h (middle panel of Figure 4B) possibly because of the egress of parasites which probably damage cell membrane. To verify whether this phenomenon is T. gondii strain-specific, RH strain tachyzoites was used as infecting parasites, and similar results were obtained (Figure 4C). All these results indicated that T. gondii, independent of strains, induces HMGB1 release in macrophages, and the release is probably dependent on the Caspase1-inflammasome pathway.