SLAMF7 expression is strongly related to sepsis .
To evaluate the inflammatory mediators involved in TLR-signaling and the induction of endotoxin shock, we profiled the expression of cytokines, chemokines, growth factors, and receptors in peripheral blood mononuclear cells (PBMCs) from sepsis and healthy subjects by quantitative PCR. As previously shown8, 21, our results revealed considerable induction of proinflammatory genes, including interleukin-6 (IL6) and interleukin-beta(IL1B). In addition, multiple inflammatory-related receptors, such as TLR4 and integrin subunit alpha M (ITGAM, also called CD11B) were induced in PBMCs from sepsis patients. Among the gene screen, we analyzed the expression of SFRs, which were homotypic activated, but have not been studied in sepsis before. Interestingly, expression of SLAMF7 and SLAMF9 were significantly elevated, whereas SLAMF3, SLAMF4, SLAMF5, SLAMF6 and SLAMF8 were down-regulated (Fig.1a). Since cytokines from monocytes and macrophages are the major sources of inflammatory response involved in endotoxin shock or tolerance22, 23, we isolated human monocytes from healthy donors, and thereafter stimulated with different doses of endotoxin (lipopolysaccharide, LPS). Focused on SFRs, we found that expression of SLAMF2, SLAMF7, SLAMF8 and SLAMF9 were all elevated, but SLAMF7 was more endotoxin- dose-dependent (Fig. 1b). Data suggested that SLAMF7 expression was endotoxin inducible, consistent with the report that SLAMF7 expression elevated on monocytes after TLR ligands stimulation13.
Furthermore, we confirmed that the percentage of SLAMF7 in CD11b+ myeloid cells but not CD3+T cells was significantly increased in sepsis subjects (n=83, Supplementary Table 1) compared to healthy donors (n=81, Supplementary Table 1) by flow cytometry (Fig. 1c,d). Then we investigated whether expression of SLAMF7 in sepsis exhibited disease specificity. As a result, we found that SLAMF7 expression was not related to gender, age, infectious microbe and different disease types (Supplementary Fig. 1a-d). Interestingly, data showed that SLAMF7+CD11b+cells percentage was strongly correlated with the C-reactive protein (CRP) concentration (r= 0.56) (Fig. 1e). Considering that CRP is a marker of disease prognosis24, we evaluated the expression of SLAMF7 during sepsis therapy. We found that levels of CRP (Fig. 1f) and percentage of SLAMF7+CD11b+cells (Fig. 1g) were both gradually decreased from pre-treatment to day 7 post-therapy. Overall, our results showed that expression of SLAMF7 was associated with sepsis process, suggesting a possible diagnosis and prognosis marker of SLAMF7 for sepsis.
SLAMF7 expression is induced by TLR/NF-κB signaling in macrophages.
Upon recognition of LPS or bacteria, TLR4 initiates MyD88-dependent pathway for NF-κB activation, which has been linked to an excessive inflammation and endotoxin shock25. Therefore, we sought to examine whether TLR-MyD88-NF-κB signaling activation contributes to SLAMF7 expression. Using different TLR ligands to stimulate RAW264.7 cells, we observed that LPS induced the most Slamf7 expression than other ligands (Pam3Csk4 for TLR1/2, LPS for TLR4, R848 for TLR7/8, poly(I:C) for TLR3) (Fig. 2a), suggesting that TLR4 engagement was indispensable. To avoid discrimination, we confirmed the expression of Slamf7 in BMDM and huaman monocytes in response to LPS stimulation (Supplementary Figure 2a). Data showed that both TLR4 ligand stimulation and PA infection induced Slamf7 expression of RAW264.7 cells in a time and dose dependent manner (Fig. 2b,c and Supplementary Figure 2b,c). As MyD88 is known to signal downstream of TLR426, we investigate whether it regulates the transcription of Slamf7. After knockdown of MyD88, we found that Slamf7 expression was inhibited in response to LPS stimulation for 24h and 48h (Supplementary Fig. 2d,e). Sequence analysis with the JASPAR program predicted several potential NF-κB and AP-1 binding sites in the promoter region of Slamf7 (Supplementary Fig. 2f). The transcriptional level of Slamf7 was almost diminished by IκB/IKK pharmacological inhibitor (BMS345541) followed LPS stimulation at 24 h or 48 h,while MAPKs inhibitors (led to AP-1 activation) had minimal influence (Supplementary Fig. 2g). Consistant with that, another chemical inhibitor JSH-23 targeting translocation of p65 subunit of NF-κB, also decreased Slamf7 transcriptional levels both in BMDM and RAW264.7 cells (Supplementary Fig. 2h,i). As previous study reported13, these data suggested that TLR-MyD88-NF-κB signaling promoted SLAMF7 expression during endotoxin process.
SLAMF7 regulates proinflammatory cytokine production in macrophages.
Early death in sepsis is largely due to a cytokine storm and the associated multiorgan failure27. It is well known that macrophages are the main mediators of TLR-triggered innate inflammatory responses11, then we investigate the role of SLAMF7 in macrophages. Firstly, we constructed cell line steadily expressing SLAMF7 by lentivirus transfection (RAW-SLAMF7) (Supplementary Fig. 3a). Overexpression of SLAMF7 successfully reduced pro-inflammatory cytokines (Tnf, Il1b and Il6) levels (Fig. 2d and Supplementary Fig. 3b). Consistent with that, BMDM treated with rmSLAMF7 also decreased expression levels of cytokines (Fig. 2e). On the other hand, we observed excessive pro-inflammatory cytokine production in SLAMF7 KO BMDM compared with WT BMDM (Fig. 2f). Consistently, IL-1b and IL-6 production were up-regulated after SLAMF7 was knocked out by crispr-cas9 technique (Fig. 2g and Supplementary Fig. 3c). Taken together, these data indicated that SLAMF7 exerted an inhibitory effect on TLR4-triggered inflammatory response in macrophages.
SLAMF7 suppresses TLR4-triggered inflammatory signaling pathway by activating phosphorylation of SHIP1.
As TLR4-triggered phosphoinositide 3-kinase and protein kinase AKT (PI3K/AKT), NF-κB and MAPKs activation could lead to large amount of inflammatory cytokines in sepsis, we next investigated whether SLAMF7 affected these signaling pathways. We observed that overexpression of SLAMF7 inhibited the phosphorylation of Akt (Fig. 3a), ERK, p38 (Fig. 3b) and IKKα/β (Fig. 3c), but had minimal influence on JNK (Fig. 3a-c). On the contrary, NF-κB (IKKα/β) and MAPKs signaling (p-ERK, p-JNK, p-p38) were activated when SLAMF7 was knocked out in BMDM or knocked down in RAW264.7 cells after LPS stimulation (Fig. 3d and Supplementary Fig. 4a). Previous studies demonstrate that IKKα/β (IkB kinase) phosphorylates IkB lead to its dissociation from NF-κB, thereby allowing NF-κB to enter the nucleus31. We found that the amount of p65 translocation from cytoplasma to nuclei decreased in SLAMF7 over-expressing cells (Fig. 3e), but increased in SLAMF7 KO BMDM (Fig. 3f). A previous study reports that AKT signaling pathway functions in inflammatory process by regulating NF-κB Ser phosphorylation, p38 and ERK activation28. We therefore assessed the role of SLAMF7 on cytokine production by inhibiting AKT or NF-κB using Ly294002 (Fig. 3g) or JSH23 (Fig. 3h) respectively. We observed that these two inhibitors abrogated the decrease of pro-inflammatory cytokine expression in RAW-SLAMF7 cells (Fig. 3g,h). These results indicated that SLAMF7 inhibited inflammatory response by disturbing AKT signaling pathway. To further elucidate how SLAMF7 mediated these effects, we next investigated the adaptor for downstream signal. It has been reported that SFRs utilize SLAM-assocated-adaptor-proteins (SAPs) to transduce signaling from membrane into cytoplasma, including EAT-2 and SAP29. Nevertheless, previous study reported that EAT-2 expressed faintly in primary human monocytes13, suggesting other novel adaptors for SFRs molecular. Here, we used rmSLAMF7 to activate SLAMF7 in BMDM to detect candidate adaptors molecule expression. Interestingly, the Eat-2 and Sap expression had no difference after rmSLAMF7 stimulation, but SHIP1 mRNA expression increased about 3 folds (Fig. 3i). To ascertain this observation, we investigated the phosphorylation of SHIP1 (p-SHIP1) in RAW-SLAMF7 cells or SLAMF7 KO BMDM, followed by LPS stimulation for different time. Data showed that RAW-SLAMF7 increased the phosphorylation of SHIP1 (Fig. 3j and Supplementary Fig.4b). Contrarily, SLAMF7 knockout attenuated the level of p-SHIP1 (Fig. 3k). Accordingly, SHIP1 knockdown also disturbed SLAMF7 signaling transduction. Transfection with siRNA targeting SHIP1 in RAW264.7 (Supplementary Fig. 4c) significantly reduced p-SHIP1 expression, and the inhibition of cytokine production by over-expression of SLAMF7 was reversed (Fig. 3l). A recent study by Guo et al showed that SLAMF7 interacts with SHIP1 and CD45 in multiple myeloma cells dependent of Src kinase30. To elucidate how SLAMF7 mediated these effects in macrophages, we analyzed the contribution of various signaling kinases in SLAMF7-dependent cytokine inhibitions. Results showed that pharmacological inhibitors targeting TBK1 abrogated the inhibition function on inflammatory cytokine responses in RAW264.7 cells, but other kinases, namely Src and Syk were not involved in SLAMF7 signaling pathway (Supplementary Fig. 4d,e), indicating that SLAMF7 functioned in macrophages was different from other cell types. Together, these data suggested that SHIP1 activation was involved in SLAMF7 signaling in suppressing TLR4-triggered inflammatory responses.
SLAMF7 suppresses TRAF6 Ubiquitination by Co-operation with SHIP1.
Previous studies reported that SLAMF7 can interact with SHIP1 to mediate inhibitory effects in the absence of activated adaptor EAT-2 in NK cells15, 31. To investigate whether SLAMF7 can interact with SHIP1 directly, we transfected Flag-tagged SLAMF7 and HA-tagged SHIP1 in 293T cells. Consistant with prior studies15, we found that SLAMF7 interacted with SHIP1 directly (Fig. 4a,b). TNF receptor–associated factors (TRAFs) play a central role in the TNF-a, IL-1–β, and LPS-induced signaling pathways32. Binding of LPS to TLR4 (Toll-like receptor 4) triggers the recruitment of MyD88 and IRAK1/4, which then recruits TRAF6 and triggers downstream signaling33. Here, we observed that activating SLAMF7 obviously up-regulated TRAF6 expression in LPS-stimulated BMDM (Supplementary Fig. 5). Endogenous immunoprecipitation ananlysis showed that SLAMF7 interacted with TRAF6 other than SHIP1 , which was dependent on the SLAMF7 expression after LPS stimulation (Fig. 4c). Exogenous coprecipitation showed the direct interaction of SLAMF7 and TRAF6 (Fig. 4d), as well as SHIP1 and TRAF6 (Fig. 4e). Furthermore, we confirmed that SLAMF7 co-localized with SHIP1 and TRAF6 by laser confocal fluorescence microscopy (Fig. 4f). In line with the endogenous immunoprecipitation, data suggested that SLAMF7 interacted with SHIP1 and TRAF6 directly.
A previous study has demonstrated that SHIP1 reduces TRAF6 autoubiquitination in NK cells34. Here, we found HA-tagged ubiquitination (Ubs-HA) of Flag-tagged TRAF6 was reduced by transfection of SHIP1 exactly in a dose dependent manner (Fig. 4g). Unexpectedly, we found that SLAMF7 also reduced TRAF6-ubiquitination independent of SHIP1, but the inhibitory function was enhanced by co-operation with SHIP1 (Fig. 4h). TRAF6 is a RING domain protein, which functions to catalyze the synthesis of unique polyubiquitin chains linked through lysine-48 (K48) and lysine-63 (K63) of ubiquitin, and K63 is essential for the activation of IKK and AKT signaling pathway35, 36. We therefore sought to investigate the type of ubiquitination of TRAF6 regulated by SLAMF7 and SHIP1. Through lysine mutation of ubiquitin, we observed that SLAMF7 co-operated with SHIP1 to attenuate K63 linked autoubiquitin but not K48 linked (Fig. 4i). Moreover, we found that the total ubiquitination increased after knockdown of SLAMF7, and increase accumulated by using proteasome inhibitor (MG132), indicating that SLAMF7-induced ubiquitination decrease was mediated independently by proteasome degradation mechanism (Fig. 4j). Finally, we found that Tnf ,IL1b and Il6 expression inhibited by SLAMF7 were reversed after transfection with TRAF6 plasmid, indicating that TRAF6 affected SLAMF7 signaling in regulating cytokine production (Fig. 4k). Collectively, the results revealed that SLAMF7 interacted with TRAF6 and restricted its K63 autoubiquitination by co-operation with SHIP1 to regulate inflammatory cytokines production.
The interaction of SLAMF7, TRAF6 and SHIP1 is dependent on phosphatase domain of SHIP1.
Next, we decided to determine which domains of SLAMF7 and SHIP1 were required for the interaction and suppression of TLR4-triggered inflammatory responses. SLAM family molecules contain a N-terminal extracellular domain, a single transmembrane domain and a cytoplasmic tail37. Unlike the other members, in mice, SLAMF7 has a relatively long tail and three tyrosine-phosphorylation sites in the cytoplasmic domain (Y261, Y266 and Y281)38 (Fig. 5a). SHIP1 possesses an amino-terminal Src homology 2 (SH2) domain that binds preferentially to the sequence pY (Y/S/T) L (M/L) and has been shown to bind to the tyrosine phosphorylated forms of Shc (SH2 domain), a centrally located phosphoinositol phosphatase domain that selectively hydrolyzes the 5’-phosphate (EEP domain), and a critical proline rich C-terminus that binds a subset of SH3-containing proteins (P rich domain)39, 40 (Fig. 5b). Firstly, we generated a panel of serial-deletion constructs of SLAMF7 (deletion of intracellular, trans-membrane and extracellular domain) and performed coimmunoprecipitation assays. Results showed that deletion of any domain of SLAMF7 failed to bind SHIP1, which indicated the interaction was dependent on full-length SLAMF7 (Fig. 5c). However, deletion of SLAMF7 intracellular domain was unable to bind TRAF6 compared to other two truncated fragments and full-length SLAMF7, which demonstrated the interaction between SLAMF7 and TRAF6 was dependent on intracellular domain of SLAMF7 (Fig. 5d). In addition, we ascertained that EEP domain of SHIP1 was responsible for the interaction with SLAMF7 (Fig. 5e) or TRAF6 (Fig. 5f). Moreover, EEP domain of SHIP1 assisted SLAMF7 to inhibit autoubiquitination of TRAF6 significantly (Fig. 5g). However, SHIP1 without EEP domain only partially reversed the inhibition, which suggested that SLAMF7 also had impact on TRAF6 autoubiquitination independent of EEP domain (Fig. 5g).
To elucidate whether SLAMF7 tyrosines mediated the signal transduction, we expressed SLAMF7 with three intra-cytoplasmic tyrosines (Y) mutated to phenylalanines (F) (Y to F mutations) (Fig. 5a). From the co-immunoprecipitation of mutated SLAMF7 with SHIP1, we found that SLAMF7 with Y281 mutation failed to bind SHIP1, indicating that Y281 of SLAMF7 played a decisive role on the interaction (Fig. 5h). Furthermore, we showed for the first time that the interaction between SLAMF7 and TRAF6 was independent of tyrosines (Fig. 5i). These results suggested that SHIP1 and TRAF6 showed no competitive binding to SLAMF7. Lastly, we investigated whether the tyrosine mutations in SLAMF7 resulted in impaired cytokines production. Then RAW264.7 cells were transfected with three mutations of SLAMF7 to detect cytokine production respectively. Intriguingly, SLAMF7 with tyrosine mutations failed to down-regulate cytokine expression, including TNF, Il1b and Il6 (Fig. 5j). Among the three tyrosine mutations, Y281 mutation increased the cytokine production most significantly (Fig. 5j). Overall, our results confirmed that the tyrosines of SLAMF7 is crucial for cytokine production.
Activation of SLAMF7 rescues septic mice by inhibiting inflammatory response
To further investigate the role of SLAMF7 in sepsis, WT C57BL/6 mice were subjected to sepsis by LPS(endotoxemia model) stimulation, PA(bacterial infection model) infection and CLP (polymicrobial sepsis model)41. Consistent with data from peripheral blood of sepsis patients, the percentage of SLAMF7+ F4/80+ macrophages was elevated significantly after LPS stimulation , PA infection or CLP in peritoneal lavage (Fig. 6a). Data indicated that SLAMF7 expression in macrophages was upregulated during endotoxin stimulation and bacterial infection in mice. To confirm the role of SLAMF7 in the pathogenesis of sepsis, we observed the survival and histopathology of mice after sepsis induction. Since SLAMF7 can be activated homotypically42, we utilized recombinant mouse SLAMF7 peptide (rmSLAMF7, R&D) targeting extracellular domain to investigate the its function in vivo. Firstly, WT mice were subjected to LPS , PA and CLP surgery to induce sepsis , and mortality was monitored. We found that rmSLAMF7 treatment obviously improved the survival of septic mice (Fig. 6b-d). Thus, SLAMF7 receptor appeared to play an important role on rescuing septic mice. Moreover, mice treated with rmSLAMF7 presented less tissue injury and leukocytic infiltrationfter with sepsis (Fig. 5e and Supplementary Fig. 6a-c). Together, these data indicated that SLAMF7 activation rescues the survival and pathology of septic mice. To investigate whether SLAMF7 regulated the inflammation induced by sepsis in vivo, enzyme-linked-immunosorbent assay (ELISA) was used to determine the concentrations of key inflammation factors in the supernatant of liver, lung ,PL and serum of mice. We observed significant reductions in the expression levels of the proinflammatory cytokines TNF-a, IL-1βand IL-6 in the rmSLAMF7 treatment group with CLP- and LPS-induced sepsis (Fig. 6f and Supplementary Fig. 7a). These results suggested that SLAMF7 alleviated LPS- or CLP- induced inflammation and protected septic mice against the acute inflammatory process. Furthermore, we found that the percentage of F4/80+ macrophages from the PL had no difference in WT mice treating with rmSLAMF7 compared with control (Fig. 6g), but expression levels of TNF-a, IL-1β and IL-6 were decreased in rmSLAMF7-treated group after CLP for 24h (Fig. 6h). Thus, SLAMF7 attenued inflammatory response by regulating the homeostasis of macrophages in sepsis.
Knockout of SLAMF7 contributes to the development of exacerbated inflammatory response in sepsis
To further confirm the function of SLAMF7 on the development of sepsis, WT and SLAMF7-deficient (SLAMF7 KO) mice were subjected to LPS , PA and CLP surgery to induce sepsis, and mortality was monitored. We found that the mortality of SLAMF7 KO is higher compared with that of WT mice (Fig. 7a-c). In addition, more severe infiltration of polymorphonuclear cells and interstitial pneumonitits were found in SLAMF7 KO mice after CLP compared to WT mice (Fig. 7d). These data suggested that SLAMF7 was vital in improving survival and reducing pathological damage in septic mice. Furthermore, We demonstrated that SLAMF7 KO mice displayed a remarkable increase levels of inflammatory cytokines (TNF-α, IL-1β and IL-6) in comparison with WT mice post-CLP (Fig. 7e) or LPS treatment (Supplementary Fig. 7b). These results indicated that SLAMF7 was required for inhibition of excessive inflammation in sepsis. Consistant with that, SLAMF7 knockout didn’t affect the percentage of macrophages (Fig. 7f), but more secreted cytokines were found in SLAMF7 KO mice after CLP (Fig. 7g). Taken together,our results indicated that SLAMF7 attenuated sepsis-induced mortality and lung injury through suppressing excessive inflammatory response.