Polymicrobial sepsis and non-specific immunization induce adaptive immunosuppression to a similar degree

Abstract

Sepsis is frequently complicated by a state of profound immunosuppression, in its extreme form known as immunoparalysis. We have studied the role of the adaptive immune system in the murine acute peritonitis model. To read out adaptive immunosuppression, we primed post-septic and control animals by immunization with the model antigen TNP-ovalbumin in alum, and measured the specific antibody-responses via ELISA and ELISpot assay as well as T-cell responses in a proliferation assay after restimulation. Specific antibody titers, antibody affinity and plasma cell counts in the bone marrow were reduced in post-septic animals. The antigen-induced splenic proliferation was also impaired. The adaptive immunosuppression was positively correlated with an overwhelming general antibody response to the septic insult. Remarkably, antigen "overload" by non-specific immunization induced a similar degree of adaptive immunosuppression in the absence of sepsis. In both settings, depletion of regulatory T cells before priming reversed some parameters of the immunosuppression. In conclusion, our data show that adaptive immunosuppression occurs independent of profound systemic inflammation and life-threatening illness.

Fig 1. Suppression of the adaptive immune response after sepsis induction.

Sepsis was induced in C57BL/6 mice, and on day 7, the animals were immunized intraperitoneally with TNP-OVA/OVA in alum (n = 47). Non-septic, not immunized animals (n = 32) and non-septic immunized animals (n = 25) served as controls. The antigen-specific immune response was determined 14 days after the immunization by measuring the serum concentration of anti-OVA IgG by ELISA. (A) The relative concentrations of OVA-specific IgG antibodies in serum normalized to a standard serum are shown (arbitrary units, AU). (B) The proliferative response to restimulation with OVA was measured by thymidine incorporation ex vivo. (C) The proliferative response to the mitogen concanavalin A (ConA) served as control. (D) Animals that had been immunized post-sepsis were divided into two groups: severely immunosuppressed mice (n = 17) had OVA-specific IgG concentrations below 10% of control values, while mildly immunosuppressed animals (n = 30) had serum concentrations greater than or equal to 10%. (E) Numbers of OVA-specific IgG-secreting cells in the bone marrow were determined in an ELISpot assay. (F) The relative antibody affinity of TNP-specific IgG was measured as binding to slightly haptenized TNP-BSA in comparison to highly haptenized TNP-BSA in an ELISA (adapted from [34]). (G-J) Blood was drawn 24 h after sepsis induction and cytokine concentrations of IL-6 (G), IL-10 (H), MCP-1 (I), and TNF (J) were quantified in the sera. Severely (n = 15) and mildly immunosuppressed (n = 29) animals are compared with non-septic, not immunized controls (n = 12). In the panels A-F means and 95% confidence intervals are shown. In the panels G-J box plots depict medians and quartiles. Group differences were tested for significance with the Kruskal-Wallis test and Dunn’s multiple comparison for selected pairs. Results were pooled from five separate experiments. *** P <0.001; ** P <0.01; * P <0.05.

Fig 2. Treg depletion prior to immunization of post-septic animals.

In DEREG mice (n = 15) and C57BL/6 wild-type controls (n = 9) sepsis was induced, and on days 5 and 6 the animals were given 1 μg of diphtheria toxin intraperitoneally to deplete Tregs. On day 7, the animals were immunized intraperitoneally with TNP-OVA/OVA in alum. Non-septic, not immunized animals (n = 11) and non-septic immunized animals (n = 10) served as controls. (A) 14 days after immunization, the antigen-specific immune response was determined by the concentration of anti-OVA IgG in the serum via ELISA and normalized to a standard serum (arbitrary units, AU). (B) The proliferative response of splenocytes to restimulation with OVA was measured by thymidine incorporation ex vivo. (C) The proliferative response to the mitogen concanavalin A (ConA) served as control. Means and 95% confidence intervals are shown. Group differences were tested for significance with the Kruskal-Wallis test and Dunn’s multiple comparison for selected pairs. Pooled results from three separate experiments with the same tendency are shown. *** P <0.001; ** P <0.01; * P <0.05.

Fig 3. Sepsis elicits a strong serum antibody response.

Sepsis was induced in C57BL/6 mice by AP (n = 7) and non-septic mice served as control (n = 3). After seven days, spleens were harvested and numbers of CD19⁺ B220⁺ total B cells (A) and B220⁺ GL7⁺ IgD^- CD95^hi CD73^int germinal center (GC) B cells (B) were determined by flow cytometry. Gating strategies can be found as supporting information (S4 Fig). Seven days after AP mice were immunized intraperitoneally with TNP-OVA/OVA in alum, and the animals were divided into severely immunosuppressed animals (n = 15) and mildly immunosuppressed (n = 29) as depicted in Fig 1D. Non-septic, not immunized animals (n = 23) and non-septic immunized mice (n = 19) served as controls. 14 days after the test immunization, concentrations of total IgG (C) and IgM (D) in serum were measured by Luminex technology. Means and the 95% confidence interval are shown. Group differences were tested for significance with the Mann-Whitney (A, B) or the Kruskal-Wallis test and Dunn’s multiple comparison for selected pairs (C, D). Results from four separate experiments with the same tendency were combined. *** P <0.001; ** P <0.01; * P <0.05.

Fig 4. Serum cytokine concentrations 24 h after NSI and AP.

C57BL/6 mice were non-specifically immunized (n = 29), or sepsis was induced by AP (n = 23). Untreated animals (n = 12) served as controls. 24 h after NSI or AP, respectively, cytokine concentrations of IL-6 (A), IL-10 (B), MCP-1 (C), IFN-γ (D) and TNF (E) and were determined in serum by CBA. Means and the 95% confidence intervals are shown. Group differences were tested for significance with the Kruskal-Wallis test and Dunn’s multiple comparison for selected pairs. Pooled results from two separate experiments are shown. *** P <0.001; ** P <0.01.

Fig 5. Suppression of the adaptive immune response after non-specific immunization.

(A, B) NSI alone did not induce OVA-specific serum antibodies but increased total serum IgG. OVA-specific IgG (A) and total serum IgG (B) were measured 21 days after NSI (n = 3). Animals without NSI and immunization (n = 3 and 6), only immunized animals (n = 5 and 11), and animals immunized 7 days after NSI (n = 6 and 11) served as controls. (C, D) NSI reduced the specific response to antigen priming. C57BL/6 mice were non-specifically immunized (NSI, adapted from [29]). At the indicated time points after NSI, the animals were primed intraperitoneally with TNP-OVA/OVA in alum (n = 5–6). Animals without NSI and immunization (n = 6) and animals immunized without prior NSI (n = 5) served as controls. The antigen-specific immune response was measured as the concentration of anti-OVA IgG in the serum 14 days after priming with OVA by an ELISA. (C) The relative concentrations of OVA-specific IgG antibodies in serum normalized to a standard serum (arbitrary unit, AU) are shown. (D) The relative antibody affinity of TNP-specific IgG was measured as binding to slightly haptenized TNP-BSA in comparison to highly haptenized TNP-BSA in an ELISA (adapted from [34]). (E) The NSI-induced immune suppression was independent of CpG. C57BL/6 mice were treated with NSI containing CpG (NSI, n = 12) or not (NSI w/o CpG, n = 6). Seven days later, they were immunized intraperitoneally with TNP-OVA/OVA in alum (OVA). Mice without NSI and immunization (n = 7) and animals immunized only (n = 9) served as controls. The antigen-specific immune response was determined as described for panel C. In the panels A-D means and 95% confidence intervals are shown. Kruskal-Wallis test and Dunn’s multiple comparison for selected pairs. In panels B and E pooled results from two separate experiments are shown. *** P <0.001; * P <0.05.

Fig 6. Treg depletion prior to immunization of NSI-treated mice.

DEREG mice (n = 12) were non-specifically immunized (NSI, adapted from [29]). On days 5 and 6, the animals were given 1 μg of diphtheria toxin intraperitoneally to deplete Tregs; control mice received the same volume of PBS (n = 12). Seven days after NSI, the animals were primed by intra-peritoneal immunization with TNP-OVA/OVA in alum. Animals without NSI and immunization (n = 6) and animals immunized with TNP-OVA/OVA without prior NSI (n = 9) served as controls. The relative concentration of anti-OVA IgG (A) and the relative antibody affinity (B) were measured as described in the caption of Fig 5. (C) Numbers of OVA-specific IgG-secreting cells in the bone marrow were determined in an ELISpot assay. (D) The proliferative response of splenocytes to restimulation with OVA was measured in a proliferation assay ex vivo. (E) The proliferative response to concanavalin A (ConA) served as control. Means and the 95% confidence interval are shown. Group differences were tested for significance with the Kruskal-Wallis test and Dunn’s multiple comparison for selected pairs. Results from two separate experiments were combined. *** P <0.001; ** P <0.01; * P <0.05.