Cutting Edge: 2B4-Mediated Coinhibition of CD4⁺ T Cells Underlies Mortality in Experimental Sepsis

Sepsis is the leading cause of death among critically ill patients in the United States (1). Importantly, no approved therapeutics are available for sepsis once antibiotics and supportive therapy fail. Recently, studies assessing the immune phenotypes and functionality of septic patients have shed light on the immune dysregulation that occurs during sepsis, as well as its ability to result in profound and protracted immune suppression (2). Previous studies have identified a role for cell surface inhibitory receptors PD-1 and BTLA in controlling macrophage activation and APC function following sepsis (3, 4); however, the mechanisms underlying sepsis-induced immune suppression remain poorly understood.

It is well known that CD4⁺ T cells are heavily impacted during sepsis; human septic patients and experimental mouse models of sepsis exhibit profound CD4⁺ T cell apoptosis and loss of functionality (2, 5). Importantly, preventing CD4⁺ T cell apoptosis has been shown to reverse sepsis-induced mortality in mouse models. In this study, we find that the coinhibitory molecule 2B4 is significantly upregulated on CD4⁺ T cells in human septic patients and in an experimental model of mouse cecal ligation and puncture (CLP). 2B4 (CD244, SLAMf4) is a member of the CD2 subset of Ig superfamily molecules. Previously best known for its role on NK cells (6), more recent work has shown that, in certain settings, 2B4 can be inducibly expressed on CD4⁺ and CD8⁺ T cells and possesses coinhibitory function on these cell populations (7, 8). Therefore, we sought to interrogate the role of 2B4 in mediating immune dysregulation and mortality during sepsis.

C57BL/6, CD4^−/−, CD8^−/− mice were originally purchased from The Jackson Laboratory; they were maintained at Emory University and used at 8–12 wk old. To generate CD4^2B4−/− chimeric mice, CD45.1⁺ mice received 1200 cGy irradiation and 16 × 10⁶ CD4^−/− Thy1.2⁺ bone marrow (BM) cells mixed with 4 × 10⁶ 2B4^−/−Thy1.1⁺ BM cells. Control chimeric mice received wild-type (WT) BM cells instead of 2B4^−/− cells. CD8^2B4−/− chimeric mice were generated by the same method.

Sepsis was induced by CLP (two punctures with a 25 gauge needle) following the method of Baker et al. (9). Septic animals received s.c. antibiotics (50 mg/kg ceftriaxone and 35 mg/kg metronidazole) at 0, 12, 24, and 36 h after surgery. For anti-2B4 treatment, animals received anti-2B4 mAb (clone 2B4, 250 μg per dose, i.p.; Bio X Cell) on days 0, 2, 4, and 6 after CLP. For NK cell depletion, anti–asialo GM1 (100 μg i.p.) was given 1 d before CLP.

Healthy donor and septic patient PBMCs were isolated under Emory University Institutional Review Board protocol no. 00002503. Blood was collected from septic patients within the first 24 h of meeting the consensus clinical definition of sepsis. Patients’ demographics are provided in Supplemental Table I.

Murine 2B4 expression was assessed using clone eBio244F4 (eBioscience). CountBright Beads (Thermo Fisher) were used to determine absolute cell numbers. For intracellular cytokine staining (ICCS), splenocytes were incubated with 30 ng/ml PMA and 400 ng/ml ionomycin in the presence of GolgiStop for 4 h at 37°C. Apoptotic cells were determined by caspase 3/7 staining (Thermo Fisher).

The Student t test, Mann–Whitney U test, and log-rank test were used. Data are presented as mean ± SEM.

To interrogate the mechanisms underlying immune dysregulation during sepsis, WT mice were subjected to polymicrobial sepsis via CLP, and expression of 2B4 was assessed by flow cytometry. As negative controls, 2B4^−/− splenocytes and isotype controls were used. 2B4 expression was confirmed by staining with three distinct anti-2B4 clones [eBio244F4, m2B4(B6)458.1, and 2B4], and all yielded similar results (data not shown). Results indicated that NK cells isolated from the sham surgery and CLP groups expressed high levels of 2B4 and maintained 2B4 expression over 7 d postsurgery (Fig. 1A). Intriguingly, 2B4 expression on CD8⁺ and CD4⁺ T cells was increased at 24 h post-CLP in septic animals. The elevated expression of 2B4 on T cells was maintained for 3 d post-CLP and declined by day 4 (Fig. 1B, 1C). PD-1 and BTLA were significantly higher on 2B4⁺ T cells, potentially indicative of an exhausted phenotype (Fig. 1D). In the murine CLP model, PD-1 and CTLA-4 begin to be upregulated on CD4⁺ and CD8⁺ T cells 48–96 h post-CLP and steadily increase until day 7 (10, 11). In contrast, our study found that 2B4 is expressed with distinct kinetics from PD-1 or CTLA-4, suggesting that 2B4 might have a unique and nonredundant coinhibitory function relative to PD-1 or CTLA-4, acting as an earlier inhibitory molecule during T cell activation.

2B4 upregulation was primarily observed on CD44^hi memory T cells during sepsis (Fig. 1D), which is consistent with previous findings and also suggests that 2B4 functions as a cosignaling receptor on memory cell populations. Analysis of expression of other SLAM family members revealed that, although the mean fluorescence intensity of SLAMf2 (CD48, the ligand of 2B4) decreased significantly on CD4⁺ and CD8⁺ T cells following sepsis, it remained highly expressed on all T cells (Fig. 1E), thus confirming ligand availability for 2B4 signaling on T cells during sepsis. Furthermore, SLAMf1, SLAMf3, and SLAMf6 failed to be upregulated on T cells following CLP (data not shown), a finding that further emphasizes the importance of 2B4 cosignaling in the regulation of T cell responses during sepsis. Congruent with previous findings that memory cell populations are more susceptible to sepsis-induced dysfunction (12), our data indicated that 2B4 expression on memory T cell populations might produce inhibitory signaling and lead to sepsis-induced dysfunction.

Given the above results, we sought to determine the effect of disrupting 2B4 signaling during sepsis. Strikingly, 2B4^−/− mice were significantly protected from death during sepsis following CLP compared with WT controls (82% survival compared with 13%, Fig. 2A). Although no difference was found in the bacterial load in peritoneal fluid or blood at 24 h post-CLP, 2B4^−/− mice possessed increased numbers of CD4⁺ T cells at 24 h relative to WT controls (Fig. 2B, Supplemental Fig. 1A). Analysis of caspase 3/7 activity in CD4⁺ T cells revealed that 2B4^−/− CD4⁺ populations contained a lower frequency of apoptotic cells (Fig. 2C) and increased numbers of live cells relative to WT CD4⁺ cells (Supplemental Fig. 1B, 1C). To further determine the functionality of T cells in 2B4^−/− and WT mice, surface markers and intracellular cytokines were assessed. 2B4^−/− and WT animals exhibited no difference with regard to CD25 and CD69 expression in the CD4⁺ or CD8⁺ T cell compartments prior to CLP. In contrast, CD25 was significantly upregulated in the CD4⁺ and CD8⁺ T cell compartments in 2B4^−/− CLP mice relative to WT CLP animals. Furthermore, CD69 was upregulated in 2B4^−/− CLP mice relative to WT CLP animals in the CD4⁺, but not the CD8⁺, T cell compartment (Fig. 2D). After CLP, increased secretion of IFN-γ was also observed in 2B4^−/− CD4⁺ and CD8⁺ T cell populations relative to WT controls (Fig. 2E). Taken together, our data indicate that 2B4^−/− CD4⁺ T cells exhibit a more activated phenotype and have higher functionality during sepsis, a finding that could underlie the increased survival observed in 2B4-deficient mice.

To elucidate which immune compartment contributes to the survival benefit obtained in 2B4^−/− mice, we first targeted NK cells because they express high levels of 2B4. NK cells were depleted in 2B4^−/− mice prior to CLP. No survival difference was observed between intact and NK-depleted 2B4^−/− mice (Supplemental Fig. 1D). This result indicated that NK cells likely did not contribute to the survival benefit in 2B4^−/− animals. Next, to further dissect the function of 2B4 on different T cell compartments, we generated BM chimeric conditional knockout mice in which 2B4 is knocked out exclusively in the CD4⁺ or CD8⁺ T cell compartment. Briefly, the CD4-specific 2B4 conditional knockouts were generated by a BM transplant of 80% CD4^−/−Thy1.2⁺ (2B4-intact) BM mixed with 20% 2B4^−/−Thy1.1⁺ BM (or 20% WT BM as control). After 10 wk of reconstitution, all hematopoietic lineages, with the exception of CD4⁺ T cells, were reconstituted primarily from the CD4^−/− WT BM. In contrast, CD4⁺ T cells were reconstituted from the 2B4^−/− BM. The same procedure was done with CD8^−/− BM, thus generating CD4- and CD8-specific 2B4 conditional knockouts (CD4^2B4−/− and CD8^2B4−/−, Fig. 3A, 3B). There was no difference in the number of B or T cells between control chimera and CD4^2B4−/− chimera animals at baseline (Supplemental Fig. 1E). We confirmed our chimera phenotype by subjecting CD4^2B4−/− chimeras to CLP and found that CD8⁺, but not CD4⁺, T cells upregulated 2B4 expression post-CLP (Fig. 3C). Although CD8^2B4−/− animals did not exhibit significantly altered survival relative to controls (Fig. 3D), CD4^2B4−/− chimeric mice exhibited significantly improved survival following CLP compared with CD4^2B4+/+ control chimeras (Fig. 3E). Functional assays showed that CD4^2B4−/− mice exhibited increased numbers of IFN-γ–secreting CD4⁺ T cells in splenocytes relative to controls (Fig. 3F), although they exhibited comparable bacterial loads (Supplemental Fig. 1F). To further investigate the functional effect of 2B4 specifically on CD4⁺ T cells on sepsis-induced immune dysregulation, we analyzed the innate immune compartments in CD4^2B4−/− mice and found that macrophages expressed elevated levels of CD86, suggesting increased activation of macrophages in the spleens of CD4^2B4−/− mice (Fig. 3G). Indeed, macrophage expression of CD86 has been correlated with reduced immune dysregulation and increased number of intensive care unit–free days in human septic patients (13). In addition, reduced levels of CD86 mRNA were identified in lethal pediatric septic shock (14). Of note, increased activated IFN-γ–secreting macrophages did not correlate with higher serum TNF and IL-1β (Supplemental Fig. 1G). Taken together, our data strongly suggest that 2B4 signaling on CD4⁺ T cells results in distinct functionality relative to 2B4 signaling on CD8⁺ T cells and contributes significantly to septic mortality.

Next, to determine whether 2B4 could be pharmacologically targeted to improve mortality following sepsis, we blocked 2B4 signaling in septic animals with a mAb to 2B4 (clone 2B4) given on days 0, 2, 4, and 6 post-CLP. We first confirmed that this clone results in 2B4 blockade and not depletion of 2B4-expressing cells in vivo using GFP⁺ 2B4–expressing retrogenic T cells transferred into WT animals (Supplemental Fig. 1H). Results showed significantly increased survival of animals treated with anti-2B4 relative to PBS-treated controls (Fig. 3H), suggesting that anti-2B4 might be a potential treatment during sepsis.

Finally, to interrogate the expression of 2B4 on distinct immune cell types in human septic patients, PBMCs were collected from patients within the first 24 h of a diagnosis of clinically defined sepsis (15). Our results showed no difference in 2B4 expression between septic patients and healthy donors within NK cells and the CD8⁺ T cell compartment (Fig. 4A–C). However, although 2B4⁺ CD4⁺ T cells are rare in healthy human controls, we observed that 2B4 expression was significantly upregulated within the CD4⁺ T cell compartment of septic patients compared with nonseptic controls (Fig. 4D). Further immunophenotyping on human 2B4⁺ CD4⁺ T cells revealed that these cells possess an exhausted phenotype that is characterized by increased PD-1 expression and decreased expression of the costimulatory molecules ICOS and CD28 (Fig. 4E). Overall, these results, combined with our mouse data, highlight the critical role of 2B4 on CD4⁺ T cell populations during sepsis.

The balance of costimulatory and coinhibitory molecules is critical in determining T cell function during infection. In this article, we show that increased 2B4 expression at early time points during sepsis can dampen T cell functionality. Importantly, blockade of 2B4 signaling improves sepsis mortality, thus identifying this pathway as a potential therapeutic target for sepsis immunotherapy. With increased attention on the immunosuppressive stage of sepsis, reversing sepsis-induced immune dysfunction has become a major direction of sepsis therapies. For example, anti–PD-1 has been shown to improve sepsis survival in mice and has entered clinical trials for septic patients. However, it is increasingly apparent that different coinhibitory molecules play distinct and nonredundant roles in inducing T cell dysfunction (16). Our data show that 2B4 expression kinetics is different from those of PD-1 and CTLA-4, indicating potentially distinct roles for these coinhibitory molecules during sepsis. Recent work has also shown that the ability of anti–PD-1 to rescue antitumor T cells is dependent upon CD28 expression (17). In this study, we found that 2B4⁺ CD4⁺ T cells express little surface CD28, suggesting that anti–PD-1 therapy may not effectively rescue the 2B4⁺ T cell compartment within septic patients, further highlighting the concept that targeting multiple coinhibitory molecules during sepsis might be required to adequately reverse immune dysfunction.

Another important and novel finding of the current study is that 2B4 functions in a cell-intrinsic manner on CD4⁺, but not CD8⁺, T cells during sepsis and contributes to decreased macrophage activation. Previous work has shown that 2B4 functions as a costimulatory or coinhibitory molecule on NK cells and primarily as a coinhibitory molecule on memory CD8⁺ T cells, depending upon external stimulation and the downstream adaptor protein SAP and phosphatase SHP-1. Noticeably, distinct 2B4 expression patterns were observed in specific pathogen–free–housed murine versus healthy human CD8⁺ T cells, likely due to differential frequencies of memory and naive T cell populations. Indeed, our murine data showed that 2B4 is expressed primarily on CD44^hi memory cell populations following CLP (Fig. 1D). Because we have previously reported that memory T cells are more susceptible to sepsis-induced dysfunction compared with naive T cells (12), understanding the impact of 2B4 signaling specifically on memory T cell subsets remains an important goal. However, despite the differences in memory T cell composition between murine and human immune populations, our results show that 2B4 expression was significantly increased after sepsis on human and mouse CD4⁺ T cells. Unlike 2B4 expression on human CD8⁺ T cells, which is high at baseline, our data indicate that sepsis selectively induces 2B4 expression on CD4⁺ T cells. However, further investigation is needed to identify the molecular signaling downstream of 2B4 in CD4⁺ T cells during sepsis. In summary, our results suggest a novel therapeutic target for the treatment of septic individuals, as well as highlight the unique roles for 2B4 within CD4⁺ and CD8⁺ T cell compartments.

We thank the patients and healthy donors who contributed to this study, with special thanks to Leona Wells and Mona Brown for collecting patient samples.

M.L.F. and C.M.C. declare intellectual property related to targeting 2B4 for the treatment of sepsis. The other authors have no financial conflicts of interest.