NITR12⁺ NK Cells Release Perforin to Mediate IgM^hi B Cell Killing in Turbot (Scophthalmus maximus)

The adaptive immune system has emerged and evolved to provide specific protection against pathogenic infections, characterized by specific Ag recognition and immunological memory development (1). Specifically, the B lymphocytes express membrane-binding Igs, which can recognize specific Ags, whereas the secretory Igs are referred to as Abs (2). In recent years, the molecular regulation of the generation, function, and maintenance of humoral immune responses induced by immunization has been revealed, and the protective effects of B cell populations have been identified during specific microbial infections (3, 4). Thus, highlighting the potential regulatory role of B lymphocyte–targeted treatments to improve immune responses in the context of vaccination or infection remains critical for rationale-based vaccine design strategies.

NK cells are innate immune effector lymphocytes that kill intracellular pathogen-infected cells and represent critical roles in immune responses (5, 6). Recently, evidence has highlighted that NK cells could contribute to immune defense through the regulation of adaptive immunity (7, 8). Specifically, NK cell–mediated target cell killing and IFN-γ have been revealed to augment isotype class switching by B cells (8). Moreover, NK cells can inhibit adaptive immunity through the production of immunosuppressive cytokines during viral infection (9, 10). Furthermore, previous reports have demonstrated that NK cells could inhibit the development of B cell responses, resulting in fewer Ag-specific plasma cells and reduced levels of neutralizing Abs (11, 12). However, whether NK cells influence the development of Abs against pathogenic infection remains largely unknown.

Novel immune-type receptors (NITRs) have been proposed to be functional orthologs of mammalian NK cell receptors, which can inhibit or activate NK cells to mediate cytotoxicity or cytokine release processes (13). As identified in mammalian NK cell receptors, the NITRs can be divided into two categories (13, 14). Among them, the inhibitory NITRs are identified on the basis of cytoplasmic immunoreceptor tyrosine-based inhibitory motif (ITIM), whereas the activating NITRs are identified on the basis of the presence of positively charged residues in the transmembrane domains (13, 14). In zebrafish, the recombinant expression of inhibitory NITR3a in human NK cells could inhibit the phosphorylation of MAPK signaling cascades, and the intracellular ITIM is crucial for NITR3a (15). Furthermore, the activating NITR9 in zebrafish was revealed to bind with the adaptor protein DAP12 and activate the MAPK signaling cascades (16, 17). Although the NITR gene sequences have been identified in a variety of teleost fish, including pufferfish, Japanese flounder, rainbow trout, and Atlantic cod, the evolutionary function of these NITRs remains largely unknown (13).

In teleost fish, given the lack of bone marrow, B-cell development occurs in the head kidney, and different B-cell subsets were distinguished according to the expression patterns of ight, ighm, and ighd, including IgM⁺IgD⁺ B cells, IgM⁺IgD⁺ Ab secretory cells, and IgT⁺ B cells (18, 19). Moreover, through single-cell transcriptomes in several fish species, the identification of NK-like cells was well characterized (20). However, the regulatory mechanism between NK cells and B lymphocytes in teleost fish remains largely unknown. Previous studies have established the model of Edwardsiella piscicida infection and found that this intracellular bacterium could provoke cellular immune responses instead of humoral immunity (21, 22); thus, we speculate that E. piscicida infection might result in the suppression of humoral immunity. In the present study, taking advantage of single-cell RNA-sequencing (scRNA-seq)-based technology to comprehensively identify the B cell subsets and NK cells in turbot (23), we demonstrate that NK cells could mediate the IgM^hi B cell impairment through perforin production during E. piscicida infection, resulting in reduced secretion of total IgM Abs. Furthermore, this reaction is tightly activated by inhibiting the expression of NITR12 in NK cells. Thus, our results suggest that the bacterial infection–induced NITR12 inhibitory pathway in NK cells might be a target for the improvement of humoral immunity, which will contribute to better developing efficacious vaccines.

All animal experiments were performed according to the guidelines and approved by the animal experiment committee of East China University of Science and Technology (protocol no. 2006272).

Turbot (Scophthalmus maximus) weighing 30 ± 5 g obtained from a fish farm (Tianyuan Corp., Yantai, China) were acclimatized for 14 d at 17°C in a recirculating tank system. The E. piscicida strain (CCTCC no. M208068) was grown at 30°C in a trypticase soy broth (TSB) medium supplemented with 16.7 μg/ml colistin. The Escherichia coli strain (American Type Culture Collection; ATCC 25922) was grown at 37°C in TSB medium.

The lymphocytes from the indicated immune organs were obtained as described before (23, 24). Briefly, the turbot was anesthetized with MS-222 (Aladdin). Head kidney and spleen were sampled, incubated in PBS containing 0.04% BSA, and mechanically disaggregated on 70-μm cell strainers. Posterior intestines were incubated in DMEM (Gibco) at 4°C for 30 min, and the tissue homogenates were treated with PBS containing 0.37 mg/ml EDTA and 0.14 mg/ml DTT for 30 min, followed by enzymatic digestion with 0.15 mg/ml collagenase (Yuanye) for 2 h at 28°C. The cell suspension was washed with PBS-BSA and layered over a 51/35% discontinuous Percoll gradient (Yeasen Biotech). After 30 min of centrifugation at 400 × g, lymphocytes lying at the interface of the gradient were collected and washed with RPMI 1640 (Gibco).

For bacterial infection in lymphocytes, E. piscicida and E. coli were cultured in TSB medium at 30°C or 37°C with shaking overnight, respectively. The bacteria were inoculated into fresh RPMI 1640 and cultured at 30°C or 37°C in an incubator until the OD_{600 nm} reached 1. Then, the harvested bacteria suspension in PBS was added to lymphocytes at a multiplicity of infection (MOI) of 50 with or without 5 mM EGTA treatment or 1 μM concanamycin A (CMA) pretreatment for 1 h. The suspension was centrifuged at 600 × g for 10 min and maintained in an incubator for 2 or 4 h at 28°C. Uninfected lymphocytes were used as controls.

For bacterial infection in turbot, E. piscicida was cultured in 5 ml TSB medium at 30°C with shaking overnight and inoculated into 400 ml TSB medium at 30°C until the OD_{600 nm} reached 3.0. Bacteria were collected and centrifuged at 6000 × g for 10 min at 4°C and rinsed twice in PBS. Turbot (30 fish per group) were infected by immersion with 2 × 10⁷ CFU/ml E. piscicida for 10 min at 17°C and reared in the clean tank system. After infection, three fish were sampled from each group at the indicated time points. Uninfected turbot was used as a control.

Lymphocytes isolated from the head kidney were infected with E. piscicida with or without 5 mM EGTA at 28°C for 2 or 4 h, then lysed with 1% (v/v) Triton X-100 for 10 min. The lysates were serially diluted with PBS and enumerated by plate counting.

Lymphocytes isolated from the head kidney were stained with 1 μg/ml mouse anti-turbot IgM (Aquatic Diagnostics) on ice for 30 min and washed twice with PBS-BSA. Samples were incubated with 4 μg/ml Alexa Fluor 647–labeled goat anti-mouse IgG Ab (Thermo Fisher Scientific, Waltham, MA) on ice for 30 min, then washed twice and analyzed by CytoFLEX LX (Beckman Coulter).

After staining with mouse anti-turbot IgM and Alexa Fluor 647–labeled goat anti-mouse IgG Abs, lymphocytes were tested with the FITC-annexin V/propidium iodide (PI) apoptosis detection kit (Yeasen Biotech), according to the manufacturer’s instructions. FITC–annexin V and PI were used for double staining, followed by CytoFLEX LX (Beckman Coulter).

scRNA-seq data for turbot are available at https://github.com/EIB202/Turbot-scRNA. The scRNA-seq dataset used has been deposited in the Gene Expression Omnibus database under accession numbers GSE195628 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE195628) and GSE174019 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE174019). The detailed analysis method is consistent with our previous studies (23, 25).

The cDNA sequence of turbot NITR12 (ENSSMAG00000010797) was obtained from the Ensembl database, and the cDNA sequences of zebrafish NITRs were obtained from the National Center for Biotechnology Information database according to a previous study (17). The phylogenetic analysis between turbot NITR12 and zebrafish NITRs was performed by MEGA 11.

The anti-NITR12 or anti-NITR13 polyclonal Abs were generated by targeting their extracellular peptides (CTFLRIKGPEPDVST, VSGEFPEVLGKTFS, respectively) (GenScript), the specificity of these Abs was confirmed by Western blot analysis. Aliquot lymphocytes (500 μl) were added to 24-well plates, then treated with 20 μg/ml anti-NITR12 or anti-NITR13 polyclonal Abs with or without 5 mM EGTA treatment at 28°C and submitted for subsequent analysis.

Total RNA from lymphocytes was extracted by TRIzol (Invitrogen) according to the manufacturer’s instructions. cDNA was synthesized by using All-in-One First-Strand cDNA Synthesis SuperMix (Transgen). The expression profiles of BCR, cytokines, and cytotoxicity-related genes, including β-actin, cd79a, cd79b, mIgM, and prf1, were detected with specific primers (Supplemental Table I). Each reaction contained 10 μl SYBR Green Mix (Monad), 0.5 μl specific primers, 1 μl cDNA, and 8 μl RNase-free water to a total volume of 20 μl and was analyzed by QuantStudio 3 (Thermo Fisher Scientific) following the manufacturer’s instructions. The relative expression of each gene was analyzed by using the 2^−ΔΔCt method.

Lymphocytes were centrifuged at 600 × g for 10 min at 4 °C. Supernatants were added with 5× SDS-PAGE protein loading buffer (Beyotime) and boiled for 10 min. The pellets were lysed in cell extraction buffer (Beyotime) and centrifuged at 12,000 × g for 10 min at 4°C. The supernatants were added with 5× SDS-PAGE protein loading buffer (Beyotime). The precipitates were subjected to SDS-PAGE and subsequently transferred to polyvinylidene difluoride membranes by electroblotting. Images were obtained using the Tanon 5200 system (Tanon).

Statistical analysis was performed by GraphPad Prism 9 (GraphPad Software Inc., La Jolla, CA). All data were representative of at least three independent experiments. Statistical significance was analyzed by the Student t test or one-way ANOVA and defined as *p < 0.05, **p < 0.01, and ***p < 0.001.

To confirm the Ab production during intracellular bacterial infection in turbot, we established the E. piscicida infection model and found that the level of IgM in turbot serum was gradually decreased during infection compared with that in control groups (Fig. 1A). Moreover, by isolating the lymphocytes from the head kidney, we analyzed the expression of BCR-associated genes obtained from scRNA-seq data (23) and found a significantly decreased expression of CD79a, CD79b, and mIgM during E. piscicida infection (Fig. 1B). These data suggest that E. piscicida infection might mediate B cell impairment and result in reduced Ab production in turbot.

Furthermore, we stepped forward to analyze the relative abundance of IgM^hi B cells in head kidney–derived lymphocytes by flow cytometry, and we observed a reduced frequency of IgM^hi B cells following E. piscicida infection compared with that during E. coli infection (Fig. 2A, 2B). This phenotype was consistent with the expression of BCR-associated genes observed above (Fig. 1). Moreover, we also analyzed the IgM^hi B cells using annexin V and PI staining and found an increased percentage of annexin V⁺/PI⁻ IgM^hi B cells, which indicates early apoptotic cells (26), and annexin V⁺/PI⁺ IgM^hi B cells, which indicates late apoptotic cells or dead cells (27), compared with that in the control and E. coli infection groups (Fig. 2C, 2D). Taken together, these results suggest that E. piscicida infection could induce the apoptosis of IgM^hi B cells in turbot.

To better understand the mechanism of IgM^hi B cell killing during this intracellular bacterial infection, we stepped forward to analyze the perforin production in head kidney–derived lymphocytes because previous studies have proved that the gene encoding perforin (prf1) exhibits a nonredundant function required for rapid depletion of B cells and loss of humoral immune function (12, 28). Consistently, we found a comparatively upregulated expression of the prf1 gene transcript during E. piscicida infection (Fig. 3A); moreover, the production of perforin was also increased in E. piscicida–infected lymphocytes (Fig. 3B). Furthermore, the E. piscicida infection–induced prf1 expression was inhibited by pretreating the lymphocytes with the perforin inhibitor CMA (Fig. 3C) (29). Consistently, the E. piscicida infection–induced downexpression of BCR-associated genes, including CD79a, CD79b, and mIgM, in lymphocytes was alleviated following CMA treatment (Fig. 3D), suggesting that perforin production might contribute to IgM^hi B cell killing during E. piscicida infection.

Moreover, through flow cytometric analysis, the E. piscicida infection–induced lower frequency of IgM^hi B cells was comparatively restored through treatment with another perforin inhibitor, EGTA (Fig. 4A, 4B). Consistently, through treatment with EGTA to inhibit the perforin-mediated IgM^hi B cell killing, we found a lower colonization of E. piscicida than that in untreated E. piscicida–infected lymphocytes (Fig. 4C). These results suggest that bacterial infection–induced perforin production is critical for IgM^hi B cell killing in this teleost fish.

To determine which cell subset plays a critical role in mediating IgM^hi B cell killing, we extracted the expression profiles of prf1 from our previous scRNA-seq–based immune cell atlas (23) and found that prf1 was expressed mainly in NK cells and CD8⁺ CTLs of the head kidney and spleen (Fig. 5A). Moreover, taking advantage of the rare abundance of NK cells in the posterior intestine (Fig. 5B), we further analyzed the expression of BCR-associated genes, including CD79a, CD79b, and mIgM, in these three immuno-organs and found that the transcripts of BCR genes were decreased in the head kidney– and spleen-derived lymphocytes (abundance of NK cells) during E. piscicida infection (Fig. 5C). However, the transcripts of BCR genes were comparatively upregulated in posterior intestine-derived lymphocytes (rarely abundant NK cells) (Fig. 5C), suggesting that prf1-expressed NK cells might play a critical role in inhibiting the B cell response.

Through further analysis of the markers of NK cells in the immune cell atlas of turbot, we found that nitr12 was specifically expressed in NK cells, whereas nitr13 was specifically expressed in CD8⁺ CTLs (Fig. 5D). However, through homologous analysis of NITR12 with the identified NITRs in zebrafish (17), we found that the NK cell–specific expressed NITR12 was homologous with inhibitory receptors NITR3a, 3b, 3c, and 3d (Supplemental Fig. 1A), which was speculated to suppress the cytotoxic function of NK cells (13, 15). Furthermore, we observed a comparatively lower frequency of NITR12⁺ cells during E. piscicida infection (Fig. 5E, 5F), suggesting that the NITR12⁺ cells might be the target for E. piscicida.

On the basis of the clues above, we hypothesize that NITR12 might be involved in regulating the cytotoxic function of NK cells to mediate IgM^hi B cell killing. To this end, we first generated the anti-NITR12 Ab and analyzed the specificity of this polyclonal Ab to probe the lysate of turbot lymphocytes (Supplemental Fig. 1B), and the anti-NITR13 Abs were generated as a control Ab (Supplemental Fig. 1B). Furthermore, we also transiently expressed the NITR12 or NITR13 in HEK293T cells, respectively, and confirmed that the anti-NITR12 Ab can recognize NITR12 but not NITR13 (Supplemental Fig. 1C). Interestingly, through treating the turbot lymphocytes with anti-NITR12 Ab, we found a prominent increase in the prf1 gene transcript and perforin production compared with that in anti-NITR13 Ab treatment (Fig. 6A), suggesting that NITR12 might be the negative regulator of perforin production.

Moreover, the transcripts of BCR genes, including CD79a, CD79b, and mIgM, were comparatively downregulated in anti-NITR12 Ab–treated lymphocytes, but not in anti-NITR13 Ab–treated groups (Fig. 6B). Consistently, a decreased frequency of IgM^hi B cells was observed during anti-NITR12 Ab treatment in lymphocytes (Fig. 6C, 6D). Moreover, the lymphocytes treated with the perforin inhibitor EGTA could reduce the killing of IgM^hi B cells, which was mediated by anti-NITR12 Ab treatment (Fig. 6C, 6D). Taken together, these results suggest that NITR12-expressed NK cells are an orchestrator of perforin production, and this process is critical for mediating the IgM^hi B cell killing in turbot (Fig. 7).

Recently, NK cells have been reported to have immunosuppressive functions during pathogen infection (10, 30). For instance, NK cells infected with Trypanosoma brucei might kill splenic B2 B cells by perforin, suppressing humoral immunity (12). Moreover, studies in mammalian cells have implicated that certain receptors might promote immunosuppressive killing of NK cells (31), dependent on the balance of information transmitted through activated and inhibitory receptors (28, 31). In teleost fish, the NITRs) were considered as the specific markers for NK cells, with characteristics similar to those of NK receptors in mammals (20, 32, 33). Generally, these receptors are composed of a single V domain or membrane-proximal I domain, with variations in joining (J) motifs, charged transmembrane residues, and a cytoplasmic tail containing an ITIM or an immunoreceptor tyrosine–based switch motif (17, 34). Previous studies have revealed that NITRs in teleosts could engage in allorecognition and influence the function of mammalian NK cells by recombinant expression of these receptors (35). For example, an activating catfish NITR11 expressed by a cytotoxic NK-like cell line can recognize allogeneic B cell targets (35). Although zebrafish NITR9 can induce the phenotype of cytotoxicity in mammalian NK cells (16), the physiological function of NITRs in teleosts remains largely unknown. In our study, on the basis of identification of a haplotype of 17 NITR genes in turbot through scRNA-seq analysis in the immune cell atlas of turbot (23), we characterized an inhibitory NITR12 receptor in turbot NK cells and found it can regulate perforin expression to mediate the IgM^hi B cell killing through both intracellular bacterial infection and Ab blocking models, which suggests that the NITRs might be the target for bacterial evasion (Fig. 7A). However, during E. piscicida infection, turbot B cells might also be the target of this bacterium; if so, these infected B cells would be killed by NK cells in a perforin-dependent manner, which might be an alternative explanation for the reduction of total serum IgM (Fig. 7B). Thus, further development of the pathogen-specific IgM Ab will contribute to better analyzing the pathogenesis of E. piscicida. Furthermore, to better analyze the function of fish immune cells, we generated the polyclonal Ab for NITR12 and NITR13 for functional analysis, and our scRNA-seq data revealed that NITR12 was specifically expressed in turbot NK cells (23). Although our results do confirm that these Abs possess the ability to recognize their own proteins and suggest that NITR12 plays a critical role in regulating the function of turbot NK cells, whether there is a nonspecific binding of these Abs with other proteins, especially for other NITRs, remains to be further analyzed. Consequently, to establish better models to analyze the function of NITR13 in CT8⁺ CTLs, as well as to analyze the function of other NITRs, the specific Abs for fish immune cells should be generated in the future.

According to various pathogen infections, the regulatory function of NK cells might be beneficial or detrimental to the mammalian host (6, 10). On the one hand, the NK cells could contribute to preventing the overactivation of adaptive immune cells during chronic infections (10, 36). On the other hand, inhibition of NK cells could enhance the responses of T lymphocytes, promoting humoral immune responses to combat the persistence of pathogens (37). Interestingly, recent studies have revealed that NK cells could be triggered to suppress the responses of T and B cells in mice administered with vaccines (30, 38), and blocking the immunosuppressive function of NK cells might be an innovative path to enhance the efficacy of vaccines (31). However, the characteristics of NK cells in regulating T/B lymphocytes in teleost fish remain largely unknown. Our present study delineates the suppression of B cell–mediated humoral immune responses during E. piscicida infection and reveals the critical role of NITR12 in regulating perforin production in NK cells, which is detrimental for IgM^hi B cell killing in turbot (Fig. 7). These results not only provide what is, to our knowledge, a novel target for vaccine design but also contribute to better understanding the function of NK cells in suppressing IgM^hi B cell–mediated responses in teleost fish.

The authors have no financial conflicts of interest.