Autophagy induces apoptosis and death of T lymphocytes in the spleen of pigs infected with CSFV

Lymphocyte depletion and immunosuppression are typical clinical characteristics of pigs infected with classical swine fever virus (CSFV). The apoptosis of virus-infected and bystander cells plays a role in the immunopathology of classical swine fever (CSF). Here, we offer the first evidence that autophagy is involved in apoptosis and death of T lymphocytes in the spleen of pigs infected with CSFV. Using immunohistochemical assays, we observed that more LC3II-positive cells appear in the T-cell zone of spleens. Spleen cell apoptosis was demonstrated using flow cytometry and TUNEL staining. Confocal immunofluorescence revealed that partial LC3II-positive cells were simultaneously TUNEL-positive. By cultivating spleen cells ex vivo, we demonstrated that the inhibition of autophagy by 3-MA treatment inhibited apoptosis and death of T lymphocytes caused by CSFV infection but did not have this effect on B lymphocytes. Further observations demonstrated that uninfected cells in the spleen were also undergoing autophagy in vivo. In summary, these results linked autophagy with the apoptosis and cell death of splenic T cells, providing a new outlook to understand the mechanism of T lymphocyte depletion and immunosuppression during CSF.

The spleen, in which CSFV appears earlier and in which large numbers of viral particles reside, contains various types of macrophage and lymphocyte populations that are sorted by biological origin and behavior [21][22][23] . To uncover the possible mechanism of lymphocyte depletion during CSF, the association between autophagy and apoptosis in spleen cells of pigs infected with CSFV was investigated. The results showed that autophagy and apoptosis pathways were both activated in the spleen of CSFV-infected pigs using western blotting analysis. More LC3II-positive cells appeared in the T-cell zone of spleen paraffin sections. Confocal images revealed that partial LC3II-positive cells were stained by TUNEL. By cultivating spleen cells ex vivo, we offered direct evidence that apoptosis and death of T lymphocytes, but not B lymphocytes, can be prevented by the inhibition of autophagy. For the first time, we demonstrate that autophagy is involved in the apoptosis and death of T lymphocytes in the spleen of pigs infected with CSFV. This finding provides a new idea for exploring the mechanism of immunosuppression caused by CSFV infection.

Pigs infected with the CSFV Shimen strain exhibited typical clinical symptoms. Pigs inoculated
with the CSFV Shimen strain exhibited clinical symptoms of CSF at 3 to 5 dpi in our study. The body temperature of each animal was recorded every two days (Fig. 1A). All animals were sacrificed in accordance with animal ethics guidelines and approved protocols when severe signs appeared at 9 dpi. Typical spleen infarction was observed in CSFV-infected pigs (data not shown). To detect the distribution of CSFV antigen in the spleen, paraffin sections were stained with WH303 antibody, which reacts with E2 protein of CSFV 24 . The immunohistochemical images revealed that most spleen cells exhibited positive staining in infected groups ( Fig. 1B and C). By contrast, control groups displayed negative staining ( Fig. 1B and C).

Increased apoptosis of spleen cells in CSFV-infected pigs.
In vivo, apoptosis is associated with the immunopathology mechanism of CSF 9,10 . To confirm the involvement of apoptosis in spleen cells of pigs infected with CSFV, further research was carried out. Given that phagocytic removal of apoptotic cells is rapid 25,26 , the detection of apoptosis in vivo is difficult. Therefore, Annexin-V, which binds to phosphatidylserine exposed on the surface of early apoptotic cells, was introduced to evaluate cells that were programmed to die 26 . A representative example of flow cytometry detection of apoptosis in spleen cells is presented in Fig. 2A. Statistical (C) Immunocytochemistry score. 3 sections of each sample were prepared, and 3 random visual fields of each section were snapped. The score was evaluated by two different clinical doctors. Score of stained intensity: characteristics of positive-stained cells. Pale yellow, tan and sepia were scored as 1, 2, and 3, respectively. Score of stained area: the mean ratios of positive cells in each visual field. Scores of 1, 2, 3, and 4 were corresponded to 6~25%, 26~50%, 51~75% and >75%, respectively. analysis indicated that CSFV obviously increased the frequency of the early apoptotic cell population (Annexin V + PI − ) (Fig. 2B). To demonstrate that apoptotic signals were activated in spleen cells, hallmark apoptotic proteins were analyzed by immunoblotting. Cleaved caspase-8 and -9 are typically considered as extrinsic and intrinsic initiators, respectively. However, cleaved caspase-3 and PARP are considered as functional downstream effectors 27 . Our results demonstrated that CSFV-mediated up-regulation of cleaved caspase-3 and PARP levels were increased in spleen cells (Figs 2C and S1). In addition, we evaluated caspase-8 and caspase-9 expression to differentiate extrinsic and intrinsic apoptosis. Both initiators were initiated in the spleen cells of pigs infected with CSFV (Figs 2C and S1). To further confirm the presence of apoptotic spleen cells in situ, we stained DNA fragmentation of paraffin sections using the terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) method. An increased ratio of TUNEL-positive cells appeared in the CSFV-infected group when sections were inspected by immunofluorescence microscopy ( Fig. 2D and E). Overall, our data demonstrated that CSFV infection caused the apoptosis of cells in the spleen.

CSFV enhanced the autophagy pathway in spleen cells. Autophagy is involved in the replication
of CSFV in vitro, as demonstrated in our previous study 19 . To clarify the effect of autophagy on splenic cells in pigs infected with CSFV, we first evaluated the expression of LC3II and SQSTM1, which were used as markers for assessing autophagy 28 . In contrast to the increased levels of LC3II, which indicate the formation of the autophagosome 29 , SQSTM1 exhibited reduced levels (Fig. 3A), reflecting the increased protein degradation of autolysosomes 30 . Simultaneously, we assessed changes in ATG5, which associates with membranes of precursor autophagosomes 31 . The results revealed that CSFV increased the expression of ATG5 in spleen cells (Fig. 3A). Moreover, increased levels of BECN1, which plays a role in the initiation of autophagy during CSFV infection in vitro 19 , was detected in CSFV-infected spleen cells (Fig. 3A). To strengthen the occurrence of autophagy and to examine the spatial distribution of autophagic cells in the spleen of piglets infected with CSFV, paraffin sections were stained using a LC3II-specific antibody. As presented in Fig. 3B and C, CSFV increased the assessment score of the immunohistochemical assay. Notably, CSFV caused LC3II-positive cells to be mainly distributed around the splenic central arteriole (Fig. 3B, Zoom-2 and Zoom-3), which is a zone of T-lymphocyte predominance (T-cell zone) called the periarteriolar lymphoid sheath (PALS) 32,33 ( Fig. 3B and D). However, relatively few Cell lysates were prepared as described in Materials and Methods, and antibodies specific for the indicated proteins were used for western blotting. The relative levels of target proteins were estimated by densitometry, and the ratios were calculated relative to GAPDH (mean ± SD; n = 3; *p < 0.05, ***p < 0.001). (B) Immunochemistry assay revealed the distribution of autophagic cells in the spleen. The paraffin sections were stained with LC3II antibody as described in Materials and Methods. Visual fields of each section were snapped under 20× objectives. (C) Immunocytochemistry score. 3 sections of each sample were prepared, and 3 random visual fields of each section were obtained. The score was evaluated by two different clinical doctors. The judging standard was the same as described in

Autophagic cells partially underwent apoptosis in the spleen of pigs infected with CSFV.
Our present data demonstrated that CSFV infection caused autophagy and apoptosis simultaneously in the spleen. Apoptosis is associated with the depletion of lymphocytes 4,9,34 . To clarify the relationship between autophagy and apoptosis of spleen cells in pigs infected with CSFV, frozen spleen sections were double stained with LC3II-antibody and TUNEL. Quantitative analysis of confocal images revealed that more apoptotic cells (~65%) appeared in CSFV-infected spleen compared with autophagic cells (~25%). Notably, partial LC3II-positive cells (~40%) were stained with TUNEL in CSFV-infected spleens (Fig. 4). This finding offered evidence that autophagy may participate in the apoptotic process in spleen cells of CSFV-infected pigs.

CSFV caused autophagy-mediated apoptosis and cell death in spleen T lymphocytes.
Based on the findings that LC3II-positive cells were predominant in the T-cell zone of the spleen ( Fig. 3B and D) and that partial LC3II-positive cells were also TUNEL-positive in CSFV-infected spleens (Fig. 4), we assessed whether autophagy played a role in the apoptosis of T lymphocytes in the spleen of pigs infected with CSFV. To answer this question, we attempted to culture spleen cells ex vivo following the method described previously 35 . The drug 3-methyladenine (3-MA) was used to inhibit autophagy in cultured spleen cells. As shown in Fig. 5, CSFV infection not only increased early apoptosis (Annexin-V + ) of CD79a + and CD3 + cells but also increased cell death (PI + ). However, the early apoptosis and death of CD3 + cells but not CD79a + cells are obviously prevented by 3-MA ( Fig. 5B and C). CD79a and CD3 are the special surface receptors of B and T lymphocytes, respectively 36,37 . According to these data, we hypothesized that autophagy resulted in apoptosis and death of T lymphocytes in the spleen of pigs infected with CSFV.

Autophagy and apoptosis also occurred in bystander spleen cells. Previous reports demonstrated
that the depleted leukocyte populations are mainly bystanders during CSFV infection in vivo 5,7 , so we further examined whether autophagy and apoptosis in spleen cells were directly caused by CSFV infection. By staining E2 protein with WH303 antibody, confocal immunofluorescence of LC3II or TUNEL staining merged with the CSFV antigen were investigated. Most apoptotic cells were not infected with CSFV in vivo (Fig. 6A). However, it is worth noting that a small proportion of autophagic cells were not infected by CSFV, and that not all CSFV-infected cells were undergoing autophagy (Fig. 6B), in contrast to the observation in vitro 19 .

Discussion
Leukopenia is a typical feature of piglets infected with CSFV, which may ultimately result in immunosuppression 4,38,39 . Although previous reports have linked apoptosis with the depletion of lymphocytes 9,34 , whether other cell death signals play a role in the process remains unknown. As the type II programmed cell death signal, autophagy can exert an independent or cooperative function with apoptosis to determine the fate of cells 40 . Continuous with our previous findings that CSFV induces autophagy to inhibit replication and apoptosis in host cells in vitro 19,20 , we present evidence for the first time that autophagy participates in the apoptosis and death of T lymphocytes in the spleen of pigs infected with CSFV. In the spleen, apoptosis is associated with pathological lesions caused by CSFV infection 10 . We sought to confirm the apoptosis of splenic cells and to explore the relationship between autophagy and apoptosis in the spleen of CSFV-infected pigs. An increased ratio of Annexin-V positive cells was detected using flow cytometry in the spleen. The results in the previous publication demonstrated that CSFV increased the percentage of cells with low mitochondrial transmembrane potential (ΔΨm) in peripheral blood lymphocytes 9 . Similar to Annexin-V, which binds to phosphatidylserine exposed on the surface of early apoptotic cells 26 , reduction of ΔΨm is also an early marker of apoptotic cells 41 . Previous findings demonstrated that CSFV inhibits apoptosis signals of vascular endothelial cells in vitro 11 . In the present study, extrinsic and intrinsic apoptosis signals were activated using western blotting. The contrasting roles of apoptosis in vitro and in vivo have long been discussed 9,11 . Furthermore, TUNEL staining in situ revealed the frequent emergence of apoptotic cells in the spleen. This finding is consistent with the results demonstrating that TUNEL-positive cells are frequently observed in periarterial lymphatic sheaths (PALS) from 3 dpi 10 .
Previous in vitro studies demonstrated that virus-induced autophagy enhances the replication of particles and reduces apoptosis in host cells 42,43 . However, few reports have focused on autophagy and viruses in vivo. Whether autophagy is helpful for the existence of the virus in the host is unknown. The changes in autophagy markers (LC3II and SQSTM1) and the increased expression of autophagy-related genes, including ATG5 and BECN1, demonstrated that CSFV enhanced the formation of autophagy bulk and autophagy flux in splenic cells. These findings are consistent with those of a previous in vitro study showing that CSFV induced autophagy 19 . Activation of autophagy in CSFV-infected pigs demonstrated that viral immunopathology in vivo is related to autophagy or autophagy-related genes [44][45][46] . Moreover, examination of paraffin sections stained by LC3II-specific antibody demonstrated that autophagic cells were mainly scattered around PALS, which are mainly populated by T lymphocytes and referred to as the T-cell zone based on previous reports 10 . This observation links T cells with autophagy during CSFV infection in vivo. and death (PI + ) ratios of CD79a + cells (mean ± SD; n = 3; **p < 0.01, ***p < 0.001). (C) Statistical analysis of apoptosis (Annexin-V + PI − ) and death (PI + ) ratios of CD3 + cells (mean ± SD; n = 3; *p < 0.05, **p < 0.01).
As type II programmed cell death, autophagy is closely related to apoptosis. Autophagy and apoptosis can act as partners, antagonists or enablers to induce cell death in a complex manner 18 . Here, we first observed that autophagic cells were partially undergoing apoptosis simultaneously based on confocal images. This finding suggested that apoptotic cells in the spleen during CSFV infection might be partially caused by autophagy. By cultivating spleen cells ex vivo, we demonstrated that inhibition of autophagy by 3-MA treatment prevented apoptosis and death of T lymphocytes, but not B lymphocytes, caused by CSFV infection. Furthermore, we demonstrated that not all cells undergoing autophagy and apoptosis were infected with CSFV. Thus, autophagy and apoptosis signals in bystander cells were activated in vivo. These overall results were somewhat consistent with the finding that autophagy results in apoptosis of bystander CD4 + T lymphocytes during the development of AIDS 47 . By contrast, CSFV only induced autophagy in infected host cells to enhance viral replication 19 and to reduce the apoptosis of infected cells in vitro 20 . We hypothesized that the induction of autophagy in bystander cells in vivo may be attributed to complex stress inducers of autophagy, including starvation, cytokines, or pathogens 42 . For example, Type I interferons (IFN) are important proteins in the antiviral response, but evidence has highlighted their newly discovered function as inducers of autophagy 48 . Interestingly, up-regulation of IFN-stimulated genes is related to cell death in acute CSFV-infected piglets 49 . However, whether autophagy of bystander cells was induced by high levels of IFN remains to be clarified.
In conclusion, the present study offered the first data that apoptosis and death of T lymphocytes in the spleen are dependent on the activation of autophagy in CSFV-infected pigs. We explored a new approach for understanding the depletion of the T cell population during CSFV infection in vivo, which is associated with The average ratios of LC3II-positive cells infected with CSFV were based on at least 50 cells in each field (mean ± SD; n ≧ 50 cells; *p < 0.05). (B) Frozen sections were double stained with LC3II-antibody (green) and WH303-antibody (red). Nuclei was stained with DAPI. In the zoomed images, the arrows indicated apoptotic cells infected with CSFV. The average ratios of TUNEL-positive cells infected with CSFV were based on at least 50 cells in each field (mean ± SD; n ≧ 50 cells; **p < 0.01). immunosuppression in pigs. Our findings may be helpful to clarify the pathogenesis of CSFV infection and to develop novel antiviral strategies. Chemical drugs target to ATG proteins may be potential modulators for treatment of T leukocytes depletion of pigs caused by CSFV infection. However, future studies should focus on how autophagy is involved in apoptosis and death of T cells and why B cells do not undergo autophagy in this context.
Virus. Blood stock of the virulent CSFV strain Shimen was propagated 1 passage in the swine kidney cell line PK-15. Virus titers were determined by end-point titration on PK-15 cells as previously described 19 . Briefly, cells cultivated in 96-well plates were fixed with 80% acetone at −20 °C for 30 min, and infected cells were detected using the immunofluorescence assay. The primary antibody was mouse anti-CSFV E2 antibody (WH303), and the secondary antibody was Dylight488-conjugated goat anti-mouse IgG. Virus titers were calculated according to Kaerber and expressed as 50% tissue culture infectious doses (TCID 50 ) per milliliter.

Infection of pigs.
A total of 6 animals (Tibet swine, 5 months old) bred at the specific pathogen-free (SPF) unit of the institute were assessed using ELISA and PCR (data not shown). 3 animals were randomly selected, and each one was infected with 10 5 TCID 50 CSFV. The other 3 animals were intramuscularly injected with the same dose of saline water and used as controls. Examination of body temperature and clinical symptoms were conducted every 2 days. All animals were sacrificed at 9 dpi for the collection of spleens when severe clinical symptoms appeared in most infected animals. All experimental procedures were performed in accordance with animal ethics guidelines and approved by the Animal Care and Use Committee of South China Agriculture University, China. The issue number is 2014-12.

Culture and infection of spleen cells ex vivo.
Spleen cells were prepared and cultivated following the methods previously described 35 . Briefly, the spleens of anti-CSFV antibody-negative pigs were removed in sterile conditions. The spleen tissue was cut, softly pressed and dispersed by passing through 70-μm cell strainers. Dispersed cells were washed twice in RPMI 1640 supplemented with 2% antibiotics. Approximately 2 × 10 6 cells were cultivated in a volume of 200 μl culture medium in 96-well U-bottom plates. The culture medium contained RPMI 1640 supplemented with 15% FBS (Gibco) and 1% antibiotics (Gibco). After 18 h, cells were pretreated with 5 mM 3-MA for 4 h followed by inoculating with CSFV at 1 multiplicity of infection for 1 h. Uninfected virus was removed by washing cells twice in RPMI 1640 supplemented with 2% antibiotics. Then, cells continued cultivating in complete medium containing 5 mM 3-MA for 3 days at 37 °C with 5% CO 2 .
Immunoblotting. After washing spleens samples with PBS three times, 50 to 60 mg small mass of each sample was incubated on ice with 1 ml RIPA lysis buffer (Beyotime, P0013B) containing 1 mM PMSF (Beyotime, ST506) and homogenized by Dounce Tissue Grinders. The precipitation was removed by centrifugation at 14,000 g for 10 min at 4 °C. Then, the supernatant was collected, and the protein concentration was quantified by the BCA protein assay kit (Thermo, 23227). Equal amounts of total proteins (20 μg) were separated by 12% SDS-PAGE gels and then electrotransferred onto polyvinylidene fluoride (PVDF) membranes (Millipore, IPVH00010). Then, 5% nonfat milk dissolved in PBS containing 0.1% Tween 20 was used to block the membranes for 1 h at 25 °C. Membranes were incubated with the corresponding primary antibodies at 4 °C overnight and secondary antibodies conjugated to HRP at 37 °C for 1 h. The bands were detected using the ECL Plus kit (Beyotime, P0018) and visualized using the chemiluminescence imaging system (Fine-do × 6, Tanon). Protein blots were measured with Image-Pro Plus (version 6.0).
Immunohistochemistry. Samples of spleen were fixed in 10% phosphate-buffered formalin, embedded in paraffin, and sliced at 3 to 4 mm. Tissue sections were deparaffinized by 3 steps in xylol at 5 min per step; two steps of 5 min each in absolute ethanol; and one 4-min step in 95% ethanol, 85% ethanol and 70% ethanol each followed by 2 min in distilled water. Antigen retrieval with pH 8.0 EDTA buffer was conducted in the pressure cooker (5 min from the beginning of boiling). Afterwards, samples were incubated with 3% hydrogen peroxide for 8 min to block endogenous peroxidase activity. To examine autophagy activation, paraffin sections were stained with rabbit polyclonal anti-LC3II antibody (1:400). For the detection of cells infected with CSFV, mouse monoclonal anti-CSFV E2 protein antibody (1:50) was utilized. ChemMate EnVision kit (DAKO) was used to display the staining. After counterstaining with hematoxylin, 3 digital images for each section at 3456 × 2304-pixel resolution at magnifications under 20× objectives were captured (Olympus, BX43).
Evaluation of immunohistochemistry images. The score of immunohistochemistry images was judged by two different clinical doctors according to the criteria reported previously 50 . The staining pattern is not given a serially numerical score but is only assigned a broad category. The score of staining intensity is judged according to characteristics of positively stained cells. Pale yellow, tan and sepia were given scores of 1, 2 and 3, respectively. The score of the stained area was judged according to the ratio of positive cells. Scores of 1, 2, 3, and 4 corresponded to ratios of 6~25%, 26~50%, 51~75% and >75%, respectively.
Semi-quantitative evaluation of IHC staining by Image-Pro Plus (version 6.0) was conducted according to the method introduced in a previous report 51 . Briefly, after B-cell and T-cell zones were defined according to the structure of the splenic central arteriole and the packed density of the cell populations 32,33 , parameters, including area sum and IOD, were measured. The optical density of the area of interest was calibrated and established as follows: hue, 0 to 30; saturation, 0 to 255; intensity, 0 to 255. The mean optical staining density of B-cell and T-cell zones was calculated by IOD/area sum. TUNEL staining. To analyze the distribution of cells undergoing apoptosis, paraffin sections of spleen were stained with TUNEL BrightGreen Apoptosis Detection Kit (Vazyme, A112-01/02/03) according to the manufacturer's instructions. 3 digital images for each section at 3456 × 2304-pixel resolution at magnifications under 20× objectives were captured (Olympus, BX53).

FACS Analysis.
For spleen tissues, partial spleen blocks were minced and passed through a 70-μm cell strainer. Dispersed tissues were washed in phosphate-buffered saline 3 times and then directly stained with FITC-conjugated Annexin-V and propidium iodide (PI) for 10 min to analyze apoptosis using the Annexin V-FITC Apoptosis Detection Kit I (Vazyme, A211-01/02) according to the manufacturer's instructions.
For spleen cells cultivated ex vivo, cells washed in phosphate-buffered saline 3 times were respectively stained with APC-conjugated antibody against CD79a (Abnova, MAB4515) and PE/Cy5-conjugated antibody against CD3 (Abcam, ab25535) at 37 °C for 1 hour. After cells were washed in phosphate-buffered saline 3 times again, the Annexin V-FITC Apoptosis Detection Kit I was used to analyze apoptosis as described above.