|Home | About | Journals | Submit | Contact Us | Français|
Patients who survive severe sepsis often display compromised immune function with impairment in innate and adaptive immune responses. These septic patients are highly susceptible to ‘secondary’ infections with intracellular pathogens that are usually controlled by CD8+ T-cells. It is unknown when and if this observed immunoparalysis of CD8+ T-cell immunity recovers and the long-term consequences of sepsis on the ability of naïve CD8+ T-cells to respond to subsequent infections are poorly understood. Here, using the CLP mouse model of sepsis we show that sepsis induces a rapid loss of naïve CD8+ T-cells. However, IL-15-dependent numerical recovery is observed a month after initial septic insult. Numerical recovery is accompanied by IL-15-dependent phenotypic changes where a substantial proportion of naïve (antigen-inexperienced) CD8+ T-cells display a ‘memory-like’ phenotype (CD44hi/CD11ahi). Importantly, the impairment of naïve CD8+ T-cells to respond to viral and bacterial infection was sustained for month(s) after sepsis induction. Incomplete recovery of naïve CD8+ T-cell precursors was observed in septic mice, suggesting that the availability of naïve precursors contributes to the sustained impairment in primary CD8+ T-cell responses. Thus, sepsis can result in substantial and long-lasting changes in the available CD8+ T-cell repertoire affecting the capacity of the host to respond to new infections.
Sepsis, a systemic inflammatory response to severe infection (1-3), is a major public health problem. It is the leading cause of death in non-coronary intensive care units and is the 11th leading cause of death in the United States (4). The early stages of sepsis are associated with a potentially fatal hyper-inflammatory state mediated by pro-inflammatory cytokines (characterized by interferon-γ (IFNγ), interleukin-12 (IL-12) and IL-6 production) (5, 6). As sepsis progresses, the immunologic response shifts to a hypo-inflammatory response, which results in an immunosuppressive state or ‘immunoparalysis’ (5, 7-9). Septic patients exhibit impaired delayed-type hypersensitivity responses and the inability to control infections that would typically be eradicated by normally functioning CD8+ T-cells (10-14). Several factors can contribute to the immunosuppressive state observed in sepsis, such as increased leukocyte apoptosis, deactivated monocyte function and lymphocyte anergy (5, 15). However, the impact of sepsis on naive CD8+ T-cells and their ability to respond to newly introduced pathogen-derived antigens is currently poorly understood.
CD8+ T-cells play a critical role in the control and eradication of intracellular pathogens (16). Because of the need to ensure the capacity to respond to the enormous diversity in the microbial universe, naïve CD8+ T-cells that can recognize particular pathogen-derived epitopes (antigen (Ag)) are infrequent in the total CD8+ T-cell population (ranging from 10 to 1000 cells in an inbred laboratory mouse) (17-22). Upon recognition of cognate antigen, naïve Ag-specific CD8+ T-cells undergo massive proliferative expansion and differentiate into effector cells able to defend against the invading pathogen. Expansion is followed by a contraction phase whereby the numbers of effector Ag-specific CD8+ T-cell decrease by ~95%. The cells that survive the contraction phase initiate the memory Ag-specific CD8+ T-cell pool (23-26). Importantly, the magnitude of the primary CD8+ T-cell response generally correlates with the size of the naïve CD8+ T-cell precursor pool specific for a particular antigen (21, 27). Thus, alterations in naïve Ag-specific CD8+ T-cell precursor frequencies may seriously compromise the capacity of the host to mount an effective immune response.
Sepsis induces apoptosis of immune cells leading to depletion of critical components of the immune system (5). This results in a significant loss myeloid cells and lymphocytes (including CD4+ and CD8+ T-cells) creating a lymphopenic environment (5). Lymphocyte homeostasis is dependent on gamma chain (γc) cytokines such as, IL-2, IL-7 and IL-15 (28, 29). IL-2 and IL-7 are important for T-cell growth and survival, respectively, (28, 30, 31) and gene expression of both of these cytokines has been shown to be deficient in human sepsis (29). Therapeutic IL-15 administration has been shown to prevent sepsis-induced apoptosis and immunosuppression thus improving survival in sepsis (32). Additionally, IL-15 has shown to play an important role in the basal proliferation of memory CD8+ T-cells as well as the sustained proliferation and accumulation of naïve CD8+ T-cells within a lymphopenic environment (33, 34).
The majority of research in sepsis focuses on understanding the factors that control early events after sepsis induction. However, survivors of sepsis have an increase risk of death from non-septic causes years after the initial septic episode (35-37). Little is known about the long-term immune consequences for an individual that has survived sepsis. In particular, the long-term effect(s) of sepsis on the ability of the host to mount primary CD8+ T-cell responses to infections is poorly understood. Here, we used the cecal-ligation and puncture (CLP) mouse model to address both short and long-term effects of sepsis on the CD8+ T-cell response to viral and bacterial infections.
C57BL/6 mice (WT, Thy1.2/1.2) were purchased from the National Cancer Institute and used at 6-10 weeks of age. Thy1.1/1.1 P14 TCR-transgenic (specific for LCMV-derived GP33 epitope) mice were provided by Dr. John T. Harty (Department of Microbiology, University of Iowa) and described previously (27, 38-41). Il-15-/- mice were purchased from Taconic (42).
Armstrong strain of Lymphocytic choriomeningitis virus (LCMV-Arm; 2×105 PFU/mouse; i.p.) and Western Reserve strain of Vaccinia virus (VacV; 1×106 PFU/mouse; i.p.) were previously described (43). Attenuated Listeria monocytogenes expressing ovalbumin (attLM-OVA; 1×106 CFU/mouse; i.v.) was used as described previously (44, 45). Infected mice were housed at the University of Iowa under the appropriate biosafety level.
Thy1.1/1.1 P14 CD8+ T-cells were obtained from spleens or peripheral blood of young naïve P14 mice and injected i.v. at indicated numbers into naïve WT C57BL/6 (Thy1.2/1.2) recipients.
Septic insult was induced by CLP (14, 46). Briefly, mice were anesthetized and the abdomen was shaved and disinfected. A mid-line abdominal incision was made, the cecum was identified and the distal one-third was ligated with 4-0 silk sutures. The ligated portion was punctured once using a 25-gauge needle and a small amount of cecal contents was extruded through the puncture. The cecum was returned into the abdomen and the peritoneum was closed with continuous suture. The skin was glued together with Vetbond tissue adhesive (St. Paul, MN) and 1 ml of saline was injected for resuscitation. This level of injury was used to create a chronic septic state characterized by the loss of appetite and body weight, ruffled hair, shivering, diarrhea, and/or periorbital exudates, and with 5-10% mortality. Sham-treated mice underwent the same procedure excluding cecum ligation and puncture. Bupivacaine was administered at the incision site and flunixin meglumine was administered twice for postoperative analgesia to all Sham and CLP-treated mice.
All peptides were synthesized by Bio-Synthesis Inc (Bio-Synthesis, Louisville, TX). LCMV-specific peptides: NP396-404 (FQPQNGQFI), GP33-41 (KAVYNFATM), GP276-286 (SGVENPGGYCL), NP205-212 (YTVKYPNL), NP235-243 (NISGYNFSL) (47). VacV-specific peptides: B8R (TSYKFESV), K3L (YSLPNAGDVI), A47L (AAFEFINSL), A42R (YAPVSPIVI), A19L (VSLDYINTM) (48). OVA peptide: OVA257-264 (SIINFEKL) (49).
To assess expression of cell-surface proteins, cells were incubated with mAb at 4°C for 30 min, washed with FACS buffer (PBS containing 1% FCS and 0.1% NaN3) then fixed with Cytofix/Cytoperm Solution (BD Biosciences). To assess the presence of intracellular cytokines (IFN-γ, TNF-α, IL-2) and granzyme B (GrzB) cells were first stimulated with the appropriate peptide for 5 h incubation at 37°C in the presence of Brefeldin A (BD Biosciences) (50, 51). The cells were then stained for surface markers, fixed/permeabilized with Cytofix/Cytoperm and labeled with the appropriate mAb. Antibodies were used in an appropriate combination of fluorochromes: CD8 (clone 53-6.7, eBioscience), Thy1.2 (clone 53-2.1, eBioscience), CD44 (clone IM7, eBioscience), CD11a (clone M17/4, eBioscience), Thy1.1 (clone HIS51, eBioscience), IFNγ (clone XMG1.2, Biolegend) TNFα (clone MP6-XT22, Biolegend), IL-2 (clone JES6-5H4, Biolegend), GrzB (clone FGB12, Invitrogen) and appropriate isotype control. To assess degranulation potential, FITC-conjugated CD107a (clone 1D48, eBioscience) or isotype control and monensin A (eBioscience) was added to cells during the stimulation period as previously described (52). FlowJo software (Tree Star) was used for analysis of samples acquired on Canto flow cytometer (BD Biosciences).
To quantitate the number of LCMV Ag-specific naïve CD8+ T-cells within the spleens of CLP or sham mice, a tetramer-based enrichment protocol (19) using GP33-41 (KAVYNFATC) and NP396-404 (FQPQNGQFI) containing Db MHC I tetramers was employed. Briefly, spleens were harvested for each mouse analyzed, a single cell suspension was prepared, and APC-conjugated GP33 and PE-conjugated NP396 tetramers were added at a 1:400 dilution in Fc block solution (FACS Buffer: PBS supplemented with 2% FCS and 0.02% NaN3, 1:50 normal mouse serum and 1:100 anti-CD16/32 mAb). Cells were incubated in the dark at room temperature for 1 h, followed by a wash in 10 ml cold FACS Buffer. The tetramer-stained cells were then resuspended in 200 μl FACS Buffer, mixed with 50 μl of both anti-APC and –PE mAb-conjugated magnetic microbeads (Miltenyi Biotech), and incubated in the dark on ice for 30 min. The cells were washed and resuspended in 3 ml cold FACS Buffer and passed over a MACS separation column (Miltenyi Biotech) to enrich for the tetramer-specific cells. Columns were washed three times with 3 ml cold FACS buffer, before eluting the bound fraction with 5 ml cold FACS buffer. The resulting enriched fractions were then stained with a mAb cocktail: Pacific Blue-conjugated anti-B220, -CD11b and -CD11c, FITC-conjugated anti-CD3, PerCPCy5.5-conjugated anti-CD4, Alexa Fluor 700-conjugage anti-CD44 (Biolegend), and Orange-conjugated anti-CD8 (Invitrogen). Cell numbers for each sample were determined using AccuCheck Counting Beads (Invitrogen). Samples were then analyzed using an LSR II flow cytometer (BD Biosciences) and FlowJo software (TreeStar Inc., Ashland, OR). The percentage of tetramer-positive events was multiplied by the total number of cells in the enriched fraction to calculate the total number of GP33 and NP396 specific CD8+ T-cells in the spleen.
Data was analyzed with Prism4 GraphPad software and specific tests to determine statistical significance are indicated in figure legends (***P < 0.0001; **P < 0.001; *P < 0.05; not significant (n. s.)). Data generated as scatter dot plots are presented as Mean and data generated as bar graphs are presented as Mean + SEM.
All animal studies were approved by the University of Iowa Institutional Animal Care and Use Committee and meet stipulations of the Guide for the Care and Use of Laboratory Animals (NIH).
To induce septic injury we performed cecal-ligation and puncture (CLP) on C57BL/6 mice (14, 46). CLP is considered the “gold standard” in sepsis research since it mimics human polymicrobial sepsis and provides a better representation of the complexity of human sepsis compared to other animal models (2, 53, 54). Using the CLP mouse model, we recently demonstrated increased susceptibility and reduced primary pathogen-specific CD8+ T-cell responses after bacterial infection with Listeria monocytogenes (LM) (14). These data raised the question of whether impairment in primary CD8+ T-cell responses early after sepsis induction (day 2) was specific for LM or can be generalized to other infections such as, Lymphocytic choriomeningitis virus Armstrong strain (LCMV-Arm) and the Western Reserve strain of Vaccinia virus (VacV). These viruses have been routinely used to study adaptive immune responses to viral infection and multiple H-2b restricted CD8+ T-cell epitopes have been identified in C57BL/6 mice (47, 48). Two days post CLP- or Sham-surgery, mice were infected with LCMV-Arm and Ag-specific CD8+ T-cell responses were examined at the peak of the primary expansion (day 8 p. i.) using peptide-stimulated intracellular IFN-γ staining (50) (Fig. 1A). A significant reduction in the percentage and number of IFNγ+ Ag-specific CD8+ T-cells was observed in the spleen of CLP mice compared to Sham-surgery mice to 4 immunodominant LCMV epitopes (NP396, GP33, GP276 and NP205 (47)) (Fig. 1B, C, D). This resulted in a 9-fold decrease in the total number of LCMV-specific effector CD8+ T-cells (based on the sum of all IFNγ+ CD8+ T-cell responses for the peptides utilized for stimulation) in the spleens of CLP mice (Fig. 1E). Additionally, we observed a similar reduction (11-fold decrease) in the total number of VacV-specific CD8+ T-cells in the spleens of CLP mice compared to Sham controls (Fig. 1F, G). Together, these results show that sepsis significantly compromises the capacity of the host to mount optimal effector CD8+ T-cell responses to systemic viral infections.
Most experimental research examines the short-term effects of sepsis (i.e., within the first few days after sepsis induction) on the immune system (8, 55, 56). However, the majority of patients survive the early hyper-inflammatory phase of sepsis, but present with an increased risk of death due to severe impairment of their adaptive immune system (9, 57). Septic patients are highly susceptible to ‘secondary’ infections that are typically controlled by CD8+ T-cells when a normal, functioning immune system is present (5, 13). Viral reactivation of critically ill patients has been reported in the later stages of sepsis, which has been associated with prolonged hospitalization or death (7, 58-60). Furthermore, sepsis survivors have an increased risk of death from non-septic causes years after hospital discharge (35). Therefore, gaining a further understanding of the long-term effects of sepsis is important and clinically relevant.
Thus, to study the long-term effects of sepsis on naïve CD8+ T-cells we induced a mild septic state that most mice survive and that permits longitudinal studies lasting longer than 30 days post-surgery. This mild septic insult was characterized by ruffled fur, hunched back, reduced mobility, and diarrhea, but importantly resulted in less than 10% mortality (Fig. 2A). While CLP mice exhibited weight loss within the first few days following surgery, they were able to regain their weight to pre-surgery levels 7 days post-surgery and had weights similar to Sham-surgery mice one month later (Fig. 2B). Loss of immune cells by apoptosis, resulting in an overall lymphopenic state, has also been reported in septic patients (61, 62). Following mild septic insult, we observed a significant decrease in total cell numbers 2 days post-surgery in the peripheral blood, spleen and inguinal LN in CLP mice compared to Sham-surgery mice (Fig. 2C). This was a transient decrease in cellularity as total cell numbers in the CLP mice returned to Sham-surgery mice levels by 30 days post-surgery (Fig. 2C). Specific evaluation of CD8+ T-cell numbers within the peripheral blood, spleen and inguinal LN of both groups of mice revealed a similar pattern (Fig. 2D). Taken together, these data illustrate that sepsis induces a rapid and transient loss of CD8+ T-cells in all organs examined and that CD8+ T-cell numbers return to normal within a month following sepsis induction.
When lymphocyte numbers are reduced below a certain threshold a lymphopenic environment is created (63, 64). To restore T-cell homeostasis and replenish the T-cell compartment, residual naïve CD8+ T-cells present in the periphery will undergo lymphopenia-induced Ag-independent expansion known as homeostatic proliferation (63, 65). Naïve (Ag-inexperienced) CD8+ T-cells that undergo homeostatic proliferation express activation markers such as CD44, CD11a, CD122 and Ly6c thus exhibiting a ‘memory-like’ phenotype and function (66). In order to determine the extent to which naïve (Ag-inexperienced) CD8+ T-cells acquire a ‘memory-like’ phenotype the expression of CD44 and CD11a was determined at various time after sepsis induction (Fig. 3A, B). Two days post-surgery, when cell numbers in CLP mice were significantly lower (Fig. 2D), there was no difference in the percentage of CD44hi/CD11ahi CD8+ T-cells in CLP- and Sham-surgery mice (Fig. 3A). However, 30 days post-surgery, when cell numbers had recovered in CLP mice (Fig. 2D), we observed a significant increase in the percentage of naïve CD8+ T-cells with a CD44hi/CD11ahi phenotype (Fig. 3A, B). We observed a similar increase in the percentage of CD8+ T-cells with a CD122 and Ly6c phenotype 30 days post-surgery (data not shown). These results demonstrate that a substantial percentage of CD8+ T-cells present after CLP-surgery exhibit a ‘memory-like’ phenotype suggesting that numerical recovery may be driven by lymphopenia-induced homeostatic proliferation.
We and others have shown that upon antigen stimulation, the increased expression of activation markers (CD44, CD11a) on CD8+ T-cells can distinguish naïve (Ag-inexperienced) from pathogen-specific effector and memory (Ag-experienced) CD8+ T-cells (67-71). Thus, the increase in expression of activation markers on CD8+ T-cells may be a consequence of antigen encounter due to the bacteria that are present following CLP-surgery (46, 72). To address this, we adoptively transferred naïve congenically marked (Thy1.1/1.1) TCR transgenic P14 CD8+ T-cells (specific for the LCMV-derived GP33 epitope (27, 38-41)) into naïve (Thy1.2/1.2) mice prior to CLP- or Sham-surgery and the expression of CD44 and CD11a was examined on CD8+ T-cells on the indicated days post-surgery (Fig. 3C). If the acquisition of a ‘memory-like’ phenotype were driven by cognate antigen, CD44 and CD11a would not be upregulated on the P14 CD8+ T-cells, as GP33 is absent from these uninfected mice. However, if the expression of these markers were driven independently of antigen, the naïve P14 CD8+ T-cells that remain following the sepsis-induced lymphopenia will replenish the T-cell compartment and acquire the ‘memory-like’ phenotype. We observed a significant increase in the expression of CD44 and CD11a on the P14 CD8+ T-cell population (Fig. 3D) as well as on the endogenous CD8+ T-cell population (Fig. 3E) suggesting that the acquisition of a ‘memory-like’ phenotype in CLP-treated mice is not driven by sepsis-induced infection. Although antigen cross-reactivity cannot be completely ruled out in this system, similar results were obtained with naïve TCR transgenic OT-I CD8+ T-cells (specific for the ovalbumin-257 epitope (73)) (data not shown). Taken together, the data suggests that following sepsis-induced lymphopenia, residual naïve CD8+ T-cells replenish the CD8+ T-cell compartment via lymphopenia-induced ‘Ag-independent’ homeostatic proliferation, and subsequently acquire a ‘memory-like’ phenotype.
IL-15 is an important cytokine for the basal proliferation of memory CD8+ T-cells, but it also plays a role in the sustained proliferation and accumulation of naïve CD8+ T-cells within a lymphopenic environment (33, 34, 63). Since our results suggested that CD8+ T-cell numerical recovery was driven by lymphopenia-induced homeostatic proliferation we tested the hypothesis that IL-15 drives numerical recovery and the acquisition of a ‘memory-like’ phenotype of naïve CD8+ T-cells after sepsis. CLP- and Sham-surgery was performed on WT or il-15-/- mice and numerical recovery and phenotype of naïve CD8+ T-cells was determined at late-time points post-surgery (Fig. 4A). Strikingly, we observed that numerical recovery was severely blunted in il-15-/- mice following sepsis induction (Fig. 4B). Furthermore, as opposed to WT hosts, the number of CD44hi/CD11ahi CD8+ T-cells was not significantly increased in il-15-/- mice (Fig. 4C). These data show mechanistically that IL-15 is important for the restoration of CD8+ T-cells following sepsis-induced lymphopenia and the acquisition of a ‘memory-like’ phenotype.
Viral reactivation in critically ill patients has been reported in the later stages of sepsis and sepsis survivors have an increased risk of death from non-septic causes years after the initial septic episode (7, 35, 58-60). Thus, it is important to determine the long-term consequences of sepsis on the ability of naïve CD8+ T-cells to respond to pathogenic challenge. Primary CD8+ T-cell responses are dependent on the environment in which naïve CD8+ T-cells recognize pathogen-derived antigens and the magnitude of the response correlates with the size of the naïve Ag-specific CD8+ T-cell precursor pool (21, 23-27, 74). Since we observed a significant reduction in Ag-specific CD8+ T-cell responses to infections early after sepsis (Fig. 1) next, we examined to what extent primary CD8+ T-cell expansion was impaired at later time points after sepsis induction.
In a preliminary experiment, we adoptively transferred physiological numbers (500 cells per mouse (27)) of naïve congenically marked P14 CD8+ T-cells (Thy1.1/1.1) into naïve (Thy1.2/1.2) mice before CLP- or Sham-surgery. Mice were infected with LCMV-Arm greater than 30 days post-surgery and P14 CD8+ T-cell numbers were determined at the peak of CD8+ T-cell expansion (day 8 p. i.) (Fig. 5A). Interestingly, a significant, 3-fold decrease in the magnitude of expansion of P14 CD8+ T-cells was observed in CLP mice compared to Sham-surgery mice (Fig. 5B, C). In a similar experiment, CLP- or Sham-surgery mice were infected with LCMV-Arm greater than 30 days post-surgery and endogenous Ag-specific CD8+ T-cell responses (based on cytokine production (IFN-γ, TNF-α, IL-2) following ex vivo peptide stimulation) were examined on day 8-post infection (Fig. 6A, B). Despite delaying the infection to a time when CD8+ T-cell numbers had recovered from the sepsis-induced lymphopenia, we still observed a significant reduction in the endogenous CD8+ T-cell responses to 4 immunodominant LCMV epitopes (NP396, GP33, GP276 and NP205) in CLP mice compared to Sham-surgery mice (Fig. 6C, D). This resulted in a significant decrease in the total number of Ag-specific CD8+ T-cells (based on the sum of all IFNγ+ CD8+ T-cell responses) in the spleens of CLP mice (Fig. 6E). We also observed a similar decrease in the total number of TNFα+ and IL-2+ CD8+ T-cells in the spleen of CLP mice (Fig. 6F, G). Additionally, we examined the cytotoxic potential (based on granzyme B (GrzB) expression following ex vivo peptide stimulation) of CD8+ T-cells from CLP mice compared to Sham-surgery mice (Fig. 6H, I). On a per-cell basis we observed that CD8+ T-cells from either CLP or Sham-surgery mice produced similar levels of GrzB (Fig. 6H). However, the septic event significantly reduced the total number of GrzB+ CD8+ T-cells in the spleens of CLP mice (Fig. 6I). We observed similar results when measuring surface expression of CD107a (to indicate degranulation potential) on CD8+ T-cells after ex vivo peptide stimulation (Fig. 6J, K). We extended these studies and examined whether primary CD8+ T-cell responses were also impaired to bacterial challenge (attLM-OVA infection) at late time points after sepsis (Supplementary Fig. 1A). We observed similar decrease in the total number of Ag-specific CD8+ T-cells (based on the sum of all IFNγ+, TNFα+ or IL-2+ CD8+ T-cell responses following ex vivo stimulation) (Supplementary Fig. 1B, C) and cytotoxic potential (based on GrzB expression following ex vivo peptide stimulation) (Supplementary Fig. 1F, G) in the spleens of CLP mice compared to Sham-surgery mice. Taken together, these data suggest that sepsis induces long-lasting changes in the host that lead to impaired primary CD8+ T-cell responses to new bacterial and viral infections.
A significant reduction in primary virus-specific CD8+ T-cell responses observed at late time points after sepsis induction might be controlled by several factors such as the environment in which CD8+ T-cells recognize antigen and/or the availability of naïve CD8+ T-cell precursors available at the time of infection. To examine if the post-septic environment affects the ability of the host to mount CD8+ T-cell responses to infection, low numbers of naïve congenically marked TCR transgenic P14 CD8+ T-cells (Thy1.1/1.1) were transferred into CLP- or Sham-surgery (Thy1.2/1.2) mice that were greater than 30 days post-surgery. Mice were subsequently infected with LCMV-Arm and P14 CD8+ T-cell numbers were determined at the peak of CD8+ T-cell expansion (day 8 p. i.) (Supplementary Fig. 2A). Interestingly, no significant difference in the magnitude of expansion of P14 CD8+ T-cells in CLP- and Sham-surgery mice was observed after LCMV infection (Supplementary Fig. 2B). Thus, these data suggest that there are no long-lasting changes within the host's post-septic environment that affect the ability of naïve CD8+ T-cells to mount an immune response to infection.
Our data presented so far suggest that CD8 T-cell intrinsic factors (e.g. availability of naïve CD8+ T-cell precursors) might contribute to the impairment of primary CD8+ T-cell responses after sepsis. In a mouse model of influenza virus it has been demonstrated that an age-associated alteration in the naïve CD8+ T-cell pool diversity can be particularly profound for subdominant responses (low naïve precursor frequencies) resulting in potential ‘holes’ in the CD8+ T-cell repertoire and compromised immunity (75). Therefore, to test the extent to which sepsis-induced alterations in the naïve CD8+ T-cell pool might lead to the inability of the host to mount CD8+ T-cell responses directed to subdominant epitopes we used the LCMV infection model where Ag-specificity of the majority of effector CD8+ T-cells has been determined (47).
CLP- or Sham-surgery was performed on mice and greater than 30 days post-surgery mice were infected with LCMV-Arm (Fig. 7A). However, prior to LCMV infection, the phenotype of the naïve (Ag-inexperienced) CD8+ T-cells (based on expression of CD44 and CD11a) was examined. We observed variability in expression of the ‘memory-like’ phenotype amongst the CLP mice and grouped them according to low CD44hi/CD11ahi expression (CLPlow) or high CD44hi/CD11ahi expression (CLPhigh) (Fig. 7B). We wanted to determine if there was a correlation in the reduction of Ag-specific CD8+ T-cell responses with CLPlow or CLPhigh mice, thus providing clues to which individual was most affected by the initial septic insult. CD8+ T-cell responses to the subdominant NP235-epitope was examined via direct ex vivo peptide stimulation on day 8 post-infection (Fig. 7A). Interestingly, we observed no significant difference in the percentage and number of NP235-specific CD8+ T-cells in the spleens of CLPlow mice compared to Sham-surgery mice (Fig. 7C, D). In contrast, there was no detectable NP235-specific CD8+ T-cell response in any of the CLPhigh mice (Fig. 7C, D).
Collectively, these results suggest that sepsis induces changes in the composition of the naïve CD8+ T-cell compartment, potentially leading to ‘holes’ in the CD8+ T-cell repertoire thus, contributing to the reduction in primary CD8+ T-cell responses to new infections. In addition, the variability observed amongst the CLP mice suggest that each individual mouse is uniquely affected from the initial septic insult and that the extent of upregulation of the ‘memory-like’ CD44hi/CD11ahi phenotype on naïve CD8+ T-cells may be used to predict the severity of impairment.
Given that our results suggested a sepsis-associated loss of Ag-specific naïve CD8+ T-cell precursors we wanted to formally evaluated this by quantitating endogenous naïve Ag-specific CD8+ T-cell precursors after sepsis utilizing p:MHC I tetramer-based enrichment (19). Tetramer-based enrichment has allowed for the direct enumeration of rare naïve Ag-specific T-cell precursors (19-21). For example, approximately 150 and 280 naïve CD8+ T-cell precursors specific for 2 immunodominant LCMV-derived NP396 and GP33 epitopes, respectively, can be found in a C57BL/6 mouse (20, 47). Spleens were harvested on days 2 and 30 post-surgery and the number of NP396- and GP33-specific CD8+ T-cell precursors was determined (Supplementary Fig. 3). Two days post-surgery, we observed a significant reduction in the number of NP396-and GP33-specific naïve CD8+ T-cell precursors in CLP mice compared to Sham-surgery mice (Fig. 8A, B). Importantly, on day 30 post-surgery, when cell numbers had recovered in CLP mice (Fig. 2D), we still observed a significant reduction in the number of naïve CD8+ T-cell precursors specific for LCMV-derived NP396 and GP33 epitopes in CLP mice compared to Sham-surgery mice (Fig. 8A, B). These data demonstrate that naïve Ag-specific CD8+ T-cell precursors are significantly reduced early after sepsis induction. In addition, these results also show that despite numerical recovery there is incomplete recovery of naïve CD8+ T-cell precursor numbers (at least for the 2 epitope specificities examined). Taken together, the results suggest that the reduction in primary CD8+ T-cell responses observed at late-time points after sepsis could be, at least in part, attributed to changes in the naïve Ag-specific CD8+ T-cell precursor pool at the time of infection.
Most epidemiology studies and experimental models of sepsis focus on short-term outcomes, thus providing the view that sepsis is a deadly acute syndrome related to the initial hyper-inflammatory state that develops. However, as sepsis progresses, an anti-inflammatory response predominates. Septic patients exhibit an immunosuppressive state or ‘immunoparalysis’ that is manifested by the inability to clear infections that would otherwise be eradicated in a host with normally functioning CD8+ T-cell immunity (5, 8, 9, 13). Moreover, it is during this later immunosuppressive phase that viral reactivation and ‘secondary’ infections can occur – events that have been associated with prolonged hospitalization or death (7, 59, 60). Sepsis survivors have an increased risk of death from non-septic causes years after the initial septic incident that has been associated with advanced stages of comorbidities, reducing the mean remaining life span from a predicted 7.66 years to 2.5 years for septic patients after hospital discharge (35-37). Therefore, the time it takes for the immune system to recover from sepsis may be a critical contributing factor to the increased morbidity and mortality associated with ‘secondary’ infections. Currently, little is known about the long-term effects of sepsis on the functional capacity of naïve CD8+ T-cells to respond to new infections. In this study, we provide direct evidence that sepsis has profound and sustained detrimental effects on the ability of the host to mount primary pathogen-specific effector CD8+ T-cell responses. Our results show that reduction in primary CD8+ T-cell responses to infection after sepsis are a consequence of a sepsis-associated changes in the CD8+ T-cell compartment and/or loss of naïve CD8+ T-cell precursors.
Apoptosis of lymphoid (including CD4+ and CD8+ T-cells) and myeloid cells and their decreased functionality during early stages of sepsis combine to increase susceptibility to subsequent infections (5, 12, 76). The degree of sepsis-induced lymphocyte apoptosis observed in human and animal studies correlates with disease severity contributing to sepsis-associated immunoparalysis (13, 61, 62). As expected, we observed a significant decrease in CD8+ T-cell numbers early after sepsis induction. The decrease in lymphocyte numbers creates a lymphopenic environment and adoptive transfer of naïve CD8+ T-cells into a lymphopenic host results in the homeostatic proliferation of these cells to fill the ‘empty space’ generated (66, 77-79). During the homeostatic proliferation process the naïve CD8+ T-cells acquire activation markers (ex. CD44, CD11a, CD122, Ly6c) and exhibit a ‘memory-like’ phenotype (66). Lymphocyte homeostasis is dependent on cytokines such as, IL-2, IL-7 and IL-15 (28, 29). IL-15, a cytokine needed for basal proliferation of memory CD8+ T-cells, has also been shown to play a role in the sustained accumulation of naïve CD8+ T-cells within a lymphopenic environment (33, 34, 63). Additionally, therapeutic IL-15 administration prevents apoptosis and immunosuppression and improves survival in sepsis (32). Our results demonstrate that endogenous IL-15 also controls numerical recovery of the naïve CD8+ T-cell compartment following sepsis. Furthermore, we show that numerical recovery was accompanied by IL-15-dependent phenotypic changes of naïve CD8+ T-cells that exhibit a ‘memory-like’ phenotype (ex. CD44hi/CD11ahi). Therefore, IL-15 plays an important mechanistic role in CD8+ T-cell homeostasis after sepsis, controlling both accumulation and phenotypic changes of naïve Ag-inexperienced CD8+ T-cells.
Recently, we showed that septic mice had impaired CD8+ T-cell responses to a bacterial infection (14) and here we demonstrated that expansion and accumulation of primary effector CD8+ T-cells after viral infection(s) are also impaired early after sepsis induction. More importantly, we observed that the impairment of naïve CD8+ T-cells responding to viral and bacterial infection was sustained long after the initial septic insult. These results were unexpected since infection was delayed to a time point when CD8+ T-cell numbers had completely recovered from the initial sepsis-induced lymphopenia. Multiple factors could contribute to this impairment after sepsis, such as alterations in the environment in which CD8+ T-cells recognize antigen and/or the number of available naïve Ag-specific CD8+ T-cell precursors at the time of infection. Although we did not observe any long-lasting post-septic environmental changes we did observe a sepsis-associated loss of a subdominant LCMV-specific CD8+ T-cell response. The generation of ‘holes’ in the CD8+ T-cell repertoire following sepsis-induced lymphopenia may be one possible contributing factor to this immunological impairment. Utilizing p:MHC I tetramer-based enrichment we confirmed that after sepsis there is a reduction in naïve LCMV-specific CD8+ T-cell precursors (at least for the 2 epitope specificities examined) demonstrating that the naïve CD8+ T-cell precursor pool may be altered at the time of infection. Our data are in parallel with aging studies, where an age-associated loss of naïve precursors correlated with the loss of specific CD8+ T-cell responses and resulted in compromised immunity to influenza (75). These findings have public health implications suggesting that septic patients may not be able to mount an appropriate immune response to new infections since certain specificity may be lost after sepsis.
Our results indicate that the availability of naïve Ag-specific CD8+ T-cell precursors after sepsis contributes to the sustained impairment in CD8+ T-cell responses and resultant sepsis-induced immunoparalysis. We demonstrate that the loss of Ag-specific CD8+ T-cell responses results from a significant reduction in numbers and incomplete recovery of naïve Ag-specific CD8+ T-cell precursors after sepsis, indicating, that alterations in the naïve Ag-specific CD8+ T-cell precursor frequencies may seriously compromise host immunity. This in turn might suggest that survivors of sepsis that encounter a new pathogen later in life may not be able to elicit a proper immune response based on a low number of available naïve Ag-specific CD8+ T-cell precursors at the time of infection. It appears likely that the sepsis-induced deletion event occurs stochastically suggesting that the severity of the ‘immunoparalysis’ caused by CD8+ T-cell precursor deletion may vary from patient to patient.
In summary, our results provide new insight into the long-term immunological consequences of sepsis. Increasing our understanding of the sustained sepsis-effects has important clinical implications, which may help design immune-based therapies needed to improve post-septic patient outcome.
We thank A. Pagán (Jenkins Lab, Microbiology, Immunology, and Cancer Biology Graduate Program, Center for Immunology, University of Minnesota) for assistance with p:MHC I tetramer enrichment protocol. We also thank members of the Badovinac Lab for discussion and Drs. Harty and Richer for critical comments on the manuscript.
Work was supported by NIH Grant A183286 (to V. P. B.) and U.S. Department of Veterans Affairs Merit Review Award (to T. S. G.).