Feline tissue samples with natural SFTSV infection and without SFTSV infection were collected from necropsy cases and sent to our laboratory after death for postmortem examination. In necropsy cases with clinical signs of SFTS, SFTSV infection were confirmed by conventional reverse transcription polymerase chain reaction using two primer pairs targeting the nucleoprotein region (Yoshikawa et al., 2014). The clinical information for these cases is provided in the Table 1 of our previous report (Sakai et al., 2021). The samples of systemic organs and cervical, submandibular, cervical, axillary, and mesenteric lymph nodes were collected, immersed in 10% neutral buffered formalin for fixation, and processed routinely for embedding in paraffin. The samples were subjected to histopathological experiments in agreement with the owners. Ten lymph node tissue samples from seven feline necropsy cases of fatal SFTSV infection were selected and analysed in this study. In addition, five lymph node tissues from five SFTSV-negative feline necropsy cases were analysed.

Lymph nodes diagnosed with necrotizing lymphadenitis in natural SFTS cases were excluded because they were inappropriate for cell population analysis. Other lymph nodes from natural infection cases were graded as low or high viral load lymph nodes, based on the SFTSV N protein positivity rate of the cells. Low viral load lymph nodes contained no or only a few SFTSV-positive cells dispersed in the cortex, whereas high viral load lymph nodes contained a sizeable number of positive cell clusters in the parafollicular cortices or massive infiltration of positive cells in the cortices (Supplementary Figure S1). Cell populations were compared among five lymph nodes from non-SFTS cases and low viral load lymph nodes and high viral load lymph nodes from feline SFTS cases. Five lymph node samples from non-SFTS cases, five low viral load lymph nodes from three feline SFTS cases, and five high viral load lymph nodes from three feline SFTS cases were analysed in this study.

The antibodies used in this study are listed in Table 1. Tissue sections of 4 μm thickness were de-paraffinized and heated at 121°C for 15 min in antigen retrieval buffer, as indicated in Table 1. After washing with phosphate-buffered saline (PBS), endogenous peroxidase was quenched with 3% hydrogen peroxide in PBS. After blocking with 5% skim milk in PBS for 30 min, the tissue sections were incubated at 37°C for 60 min with the primary antibodies listed in Table 1. The dilutions of each primary antibody are listed in Table 1. After washing with PBS, sections were incubated with Histofine Simple Stain MAX PO (M) (Nichirei Bioscience, Cat #424131, RRID: AB_2811178), Histofine Simple Stain MAX PO (R) (Nichirei Bioscience, Cat #424141, RRID: AB_3073750), or ImmPRESS HRP goat anti-rat IgG polymer detection kit (Vector Laboratories, Cat # MP-7404-50, RRID: AB_ 2,336,531). After washing with PBS, positive signals were visualized using peroxidase-diaminobenzidine reaction, and the sections were counterstained with hematoxylin.

Double-labelling immunofluorescence of cleaved-caspase 3 and CD3, Iba1, or FAS was performed on 4 μm thick sections. Heat-mediated antigen retrieval was performed using an antigen retrieval solution pH 9.0 (Nichirei Bioscience). After washing with PBS and blocking with 5% skim milk, the sections were incubated with anti-CD3, anti-Iba1, or anti-FAS antibodies at 37°C for 60 min. After washing with PBS, the sections were incubated with Alexa488-conugated anti-rabbit IgG (1:400; Abcam, Cat # ab150077, RRID: AB_2630356) for 30 min at 37°C. After washing with PBS, the sections were incubated with Alexa555-conjugated rabbit monoclonal anti-cleaved-caspase 3 antibody (1:100, Cell Signaling Technology, Cat #9604, RRID: AB_2797708) overnight at 4°C. Tissue sections were incubated with 4′,6-diamidino-2-phenylindole (DAPI; 1:10,000; Dojindo, Cat #D523) for nuclear labelling. The sections were analysed using an LSM710 confocal microscope (Zeiss, RRID: SCR_018063).

Double-labelling immunofluorescence staining for CD3 and granzyme B was performed as follows. After antigen retrieval, blocking, incubation with an anti-CD3 primary antibody, and incubation with a horseradish peroxidase (HRP)-conjugated secondary antibody, the slides were incubated with tyramide-Opal520 (Akoya Biosciences, Cat #OP-001001) for 10 min to label the Opal520 fluorophore and heated at 95°C for 10 min to remove the antibodies from the slides. The slides were blocked again, incubated with an anti-granzyme B primary antibody and an HRP-conjugated secondary antibody, and then incubated with tyramide-Opal690 (Akoya Biosciences, Cat #OP-001006) for 10 min. The slides were then heated at 95°C for 10 min and stained with DAPI and analysed using PhenoImager fusion (Akoya Biosciences, RRID: SCR_023274) and processed using inForm software (Akoya Biosciences, RRID: SCR_019155).

Lymph node tissue sections were immunostained with anti-Ki67, anti-CD3, and anti-Bcl6 antibodies. Images of five lymphoid follicles per tissue section were captured at ×200 magnification. The area of each lymphoid follicle was measured using the QuPath software (RRID: SCR_018257), and the number of positive cells inside each lymphoid follicle was counted. Density (number of positive cells/mm²) of each tissue section was calculated and averaged. Five lymph node tissue sections were analysed for each group.

Lymph node tissue sections were immunostained with anti-Foxp3, anti-granzyme B, anti-cleaved caspase 3, anti-cleaved caspase 8, and anti-Ki67 antibodies. Five images of the parafollicular area per tissue section were captured at ×400 magnification. The number of positive cells per high-power field (HPF) was counted and averaged for each tissue section. Five tissue sections were analysed for each group.

Feline tissue samples with experimental SFTSV infections were analysed in this study. The experimental protocol, clinical information, anti-SFTSV neutral antibody titres, and pathological findings of these cases have been published previously (Park et al., 2019). Briefly, six cats were intravenously inoculated with 10⁷ TCID50/mL of SFTSV SPL010 strain. Four of the six cats exhibited severe illness and were euthanized at 8- or 10-days post infection (dpi) because they met the humane endpoint. The other two cats showed no obvious clinical signs and were euthanized at 14 dpi. The systemic organs were collected, fixed in 10% buffered formalin, routinely processed, and embedded in paraffin. In this study, tissue samples from the submandibular, cervical, axillary, and mesenteric lymph nodes were analysed. The sections were immunostaind with anti-Bcl6, anti-CD3, and anti-cleaved caspase 3 and positive cell densities in the follicular or parafollicular regions were counted.

The immunostaining-positive cell density and positive cell count per HPF data from the three groups (non-SFTS cases, low viral load, and high viral load) were first examined for normality using the Shapiro–Wilk test. For datasets with confirmed normality in all three groups, the significance between groups was examined using one-way analysis of variance (ANOVA), followed by the Tukey–Kramer test. For other datasets, the significance between groups was examined using the Kruskal-Wallis test, followed by Dunn’s test. Analyses were performed using the GraphPad Prism software (RRID: SCR_002798).

As previously reported (Sakai et al., 2021), lymph nodes of cats with SFTS were histopathologically classified into three patterns: hyperplasia without SFTSV-positive cells, accumulation of SFTSV-positive cells, and necrotizing lymphadenitis. Among these three histological patterns, lymph nodes with necrotizing lymphadenitis were excluded from this study because of the massive loss of constituent cells. Lymph nodes with the other two histological patterns were graded as low or high viral load based on the SFTSV N protein immunopositivity rate of the cells. Low viral load lymph nodes contained no or only a few SFTSV-positive cells dispersed in the cortex, whereas high viral load lymph nodes contained a sizeable number of positive cell clusters in the parafollicular cortices or massive infiltration of positive cells in the cortices (Supplementary Figure S1).

Based on the histopathological analysis of H&E-stained tissue sections of the lymph nodes of feline SFTS cases, we suspected a defective germinal centre response. To confirm this observation, lymph node tissue sections from SFTSV-negative and feline SFTS cases were immunostained for Ki67 and Bcl6 to label germinal centre B cells. Immunohistochemical detection of fascin, a marker of follicular dendritic cells, confirmed the preservation of follicular structures in the lymph nodes (Supplementary Figures S2A–C). Immunohistochemistry revealed a decrease in intrafollicular Ki67-positive cells in feline SFTS cases compared to that in SFTSV-negative cases (Figures 1A–C). The densities of Ki67-positive cells in low viral load and high viral load lymph nodes were approximately 50 and 15% of those in non-SFTS cases, respectively (Figure 1G). The density of Bcl6-positive cells in feline SFTS cases dramatically decreased and almost no Bcl6-positive cells were found in the lymphoid follicles (Figures 1D–F,H). These results revealed defects in germinal centre B cell activation.

Germinal centre reactions are regulated by various cell types including follicular stromal cells and T cell subpopulations (Stebegg et al., 2018). Follicular helper T (Tfh) cells play essential roles in germinal centre B cell activation by providing co-stimulatory molecules and secreting interleukin (IL)-21 (Vinuesa et al., 2016). Because antibodies against Tfh cell marker molecules for cats are not available, T cell numbers in the lymphoid follicles were analysed. The results demonstrated that follicular T cell numbers in feline SFTS cases were significantly decreased to less than half of those in non-SFTS cases (Figures 2A–C,G). In addition to Tfh cells, T cells that suppress the germinal centre reaction, follicular regulatory T (Tfr) cells, constitute the T cell population in lymphoid follicles (Wing et al., 2018). Thus, Tfr cells in the lymphoid follicles were analysed by immunostaining for FoxP3 (Figures 2D–F). Intrafollicular FoxP3-positive cells were rare, constituting less than 3% of the CD3-positive cells in non-SFTS cases (Figure 2H). Their densities in the SFTS cases significantly decreased and were scarcely observed (Figures 2D–F,H).

Because Tfh and Tfr cells are supplied to lymphoid follicles from the paracortex area, the site of T cell activation and proliferation, defects in paracortical T cells were also suspected. Thus, apoptosis and proliferative activity in the paracortical area were analysed by immunohistochemical detection of cleaved-caspase3 and Ki67. Cleaved-caspase3-positive cells increased by approximately five to six times in feline SFTS cases compared to SFTSV-negative cases (Figures 3A–C,G). In contrast, no obvious changes were observed in the number of Ki67-positive cells (Figures 3D–F,H). Thus, an increase in apoptosis rather than a decrease in proliferation was suggested as the cause of the reduction in the number of Tfh and Tfr cells. Further analysis by cleaved-caspase8 immunohistochemistry was performed to determine whether apoptosis of paracortical lymphocytes was induced by the extrinsic pathway. The number of cleaved-caspase8-positive cells was similar to that of number of cleaved-caspase3-positive cells (Figures 4A–D). These results suggest that the apoptosis of paracortical lymphocytes was mostly induced by the extrinsic pathway. Furthermore, apoptotic cells in the paracortices were characterized by double immunofluorescence staining for cleaved-caspase3 and the T cell marker CD3, or the macrophage marker Iba1. The results showed that most cleaved-caspase3-positive cells were CD3-positive T cells and not Iba1-positive cells (Figure 4E). The cleaved-caspase3-positive rate in CD3-positive cells was approximately 20% in both low viral load and high viral load lymph nodes, while the cleaved-caspase3-positive rate in Iba1-positive cells was higher in high viral load lymph nodes than those in low viral load lymph nodes (Supplementary Figure S3). Furthermore, analysis of granzyme B, a marker of cytotoxic T cells and natural killer cells, was performed to analyse cells with antiviral cell-mediated immune responses, and the results demonstrated an approximately 10-fold increase in SFTSV-infected cases compared to non-SFTS cases (Figures 5A–D). Further characterization of granzyme B-positive cells by double staining with CD3 revealed that most granzyme B-positive cells were CD3-positive T cells (Figure 5E; Supplementary Figure S4).

While our analysis demonstrated an impaired germinal centre response and increased apoptosis of T cells, it remains unclear whether these pathological changes affected the immunological responses in these cats, owing to the lack of information regarding their immunological conditions. To address this, we analysed six cats that were experimentally infected with SFTSV, and their anti-SFTSV antibody titres were examined in a previous study (Park et al., 2019). Among the six cats studied, two (Nos. 1 and 5) showed no obvious clinical signs and had high anti-SFTSV antibody titres (>40). The other three (Nos. 2, 3, and 4) became moribund after SFTSV infection and had only low titres of anti-SFTSV antibodies (<10 or 10), while one cat (No. 6) had a high titre of anti-SFTSV antibody (40) and was moribund (Park et al., 2019). According to these antibody titeres, we grouped Nos. 1, 5, and 6 as high antibody responders, and Nos. 2, 3, and 4 as low antibody responders. Immunohistochemical staining for Bcl6, CD3, and cleaved-caspase3 was performed on submandibular, cervical, axillary, and mesenteric lymph node tissue sections. The viral loads in these lymph nodes are summarized in Supplementary Table S1. The results demonstrated that intrafollicular Bcl6-positive cell densities in high antibody responders were much higher than those in low antibody responders (Figure 6A; Supplementary Figure S6A). Additionally, high antibody responders tended to have higher intrafollicular CD3-positive cell densities and lower paracortical cleaved-caspase3-positive cell densities than those of low antibody responders (Figures 6B,C; Supplementary Figures S6B,C).