Should We Be Moving From Suppression to Stimulation to Deal With Immunoparalysis in Sepsis Patients?

Evangelos J Giamarellos-Bourboulis


Immunotherapy. 2014;6(2):113-115. 

The failure of most clinical trials has led to considerable consideration regarding the future of immunotherapy in sepsis. Most of these trials studied the efficacy of agents that block pro-inflammatory cytokines such as TNF-α and IL-1β. These cytokines are secreted during the initial stages of sepsis by circulating monocytes and tissue macrophages. When circulating monocytes are isolated for at least 24 h after clinical development of sepsis, they fail to produce similar amounts of TNF-α, IL-1β and IL-6 compared with the monocytes of healthy volunteers.[1] In parallel, their transcriptional activity for these mediators is severely downregulated. It appears that a similar situation takes place in tissues. In a recent study, leukocytes were isolated from the spleen of cadavers of 29 patients dying from trauma and of 40 patients dying from sepsis. The results confirmed the severe suppression of both the innate and adaptive immune responses of splenocytes comprising impaired cytokine production, defective antigen presentation and increased lymphocyte apoptosis.[2] This phenomenon of the exhaustion of the immune response after the initial septic insult is known as sepsis-related immunoparalysis or immunosuppression.

In most conducted trials, agents that block proinflammatory mediators are administered later than the peak of proinflammatory responses; at the time of their administration, circulating monocytes are exhausted for cytokine production. As a consequence, treatment increases further exhaustion and hampers the restoration of their function. This is probably, in part, due to the failure of the strategies to overwhelm the proinflammatory response of the host. Despite the disappointment for the outcome of these trials, knowledge for the phenomenon of immunoparalysis has boosted enthusiasm for future trials with agents that stimulate immune responses.

In a clinical trial of our group, patients with ventilator-associated pneumonia (VAP) were randomized into blind intravenous treatment with placebo or with clarithromycin for 3 consecutive days. Blind treatment was given as adjunctive therapy on top of antimicrobials. Clarithromycin treatment was accompanied by earlier resolution of VAP within 10 days instead of 15.5 days of the placebo arm (p = 0.011). However, no survival benefit from treatment with clarithromycin was seen in the overall study population, but was only seen within the subgroup of patients with septic shock and multiple organ dysfunctions (MODs); odds ratio for death in this subgroup decreased from 19.00 in the placebo arm to 3.78 in the clarithromycin arm (p = 0.048).[3] Monocytes were isolated from the bloodstream of the patients for seven consecutive days; serum TNF-α and IL-10 were measured on the same days. The most striking differences between groups were found among patients with septic shock and MODs who experienced survival benefit from clarithromycin treatment; the ratio of serum IL-10/TNF-α, that is an index of immunoparalysis, decreased, the expression of CD86 on monocytes, that is an index of effective antigen presentation, increased; and the ex vivo production of IL-6 by monocytes was increased.[4] Although these results do not necessarily imply the mode of action of clarithromycin, they underscore that survival benefit from septic shock and MODs is linked with reversal of immunoparalysis.

Until now, no well-powered randomized clinical trial (RCT) of immunostimulation has even been conducted in patients with severe sepsis. The major question is which agent may be considered as a promising candidate for that RCT. Immunoglobulins enriched with IgM (IgM preparations), GM-CSF, G-CSF, recombinant IFN-γ (rIFN-γ) and recombinant IL-7 (rIL-7) tend to be the most promising candidates.

Among them, IgM preparations are more close to the bedside. Part of the phenomenon of immunoparalysis involves exhaustion of B lymphocytes for IgM production. IgM is a polyvalent immunoglobulin that is very effective for the opsonization of microrganisms and for the blockade of bacterial endotoxins and cytokines.[5] Two observational studies have shown that circulating levels are decreased in septic shock.[6,7] Data from the analysis of 332 patients enrolled from 27 study sites of the Hellenic Sepsis Study Group revealed that circulating IgM was at normal levels in severe sepsis and then abruptly decreased when septic shock developed. In these cases, the total body distribution of IgM in survivors was greater than nonsurvivors.[8] Conducted RCTs have failed to demonstrate any survival benefit from adjunctive treatment with intravenous immunoglobulins in severe sepsis. In one meta-analysis, patients treated with IgM preparations were analyzed separately; the relative risk of death decreased by 34% among adults and by 50% among neonates.[9] However, the need to conduct large-scale RCTs remains.

IFN-γ stimulates the gene expression of most proinflammatory cytokines and primes phagocytosis by neutrophils. NK cells are the main reservoir during early immune responses. Production of IFN-γ by NK cells is downregulated in sepsis and this provides a rationale for its use as immunostimulation treatment.[10] The efficacy of rIFN-γ was tested in a setting of experimental endotoxemia of healthy human volunteers. Enrolled volunteers became tolerant to bacterial endotoxin (lipopolysaccharide; LPS) by the first injection of LPS. Subcutaneous doses of blind treatment with placebo or rIFN-γ were then given every 48 h; after the third dose, a second dose of LPS was given. Serum TNF-α and expression of HLA-DR on circulating monocytes decreased after the second dose of LPS in the placebo group; in the IFN-γ arm, serum TNF-α increased, as well as expression of HLA-DR on monocytes.[11] Although results provided a clear-cut rationale for partial restoration of immunoparalysis by IFN-γ, the function of monocytes for cytokine production was not studied.

Sparse evidence is available for the clinical efficacy of rIFN-γ in sepsis patients. In a nonrandomized study, IFN-γ was administered intravenously for 7 days in nine patients with sepsis and low expression of HLA-DR on monocytes (<30%). Treatment was accompanied by an increase in the expression of HLA-DR and by restoration of the function of peripheral blood mononucleated cells for the production of TNF-α.[12] IFN-γ was also administered to one patient with bacteremia by Staphylococcus aureus when marked deterioration was shown on the third week of antibiotic treatment. The patient improved after 1 week. Microarray gene expression profiling was performed in whole blood before and 5 days after start of IFN-γ. A total of 782 genes were differentially expressed; they were mainly linked with upregulation of the pathways of cell cycle, metabolism and HLA-DR.[13]

The phenomena of sepsis comprise both upregulation of some pathways and downregulation of other pathways at the same time. This elaborates the need for the development of genomic biomarkers that can help the clinician to understand whether a patient may benefit from treatment with rIFN-γ or not. To this end, the effect of rIFN-γ on gene expression was first studied on healthy monocytes becoming tolerant to LPS in vitro and then verified on cells from septic patients. More precisely, human monocytes became tolerant to LPS after serial exposure. Changes of the transcriptome after treatment with IFN-γ were studied. Bioanalysis identified five genes that are upregulated after treatment and seven genes that are downregulated after treatment. The expression of these genes was further studied in the whole blood of seven patients with septic shock after exposure to bacterial endotoxin and co-treatment with rIFN-γ; only six of these genes were modulated by rIFN-γ (TNFα, HLADRA, TNFAIP-6, CXCL-10 and GBP-1 being upregulated, and FCN-1 being downregulated).[14]

The efficacy of immunostimulation with G-CSF and GM-CSF in severe sepsis was studied in several RCTs. A meta-analysis of 12 RCTs enrolling 2380 patients was recently published; ten of these trials were blinded. In the meta-analysis, both treatments with G-CSF and GM-CSF were encountered together. Treatment with CSFs increased the rate of reversal of infection significantly without any heterogeneity between studies. Despite this benefit, no effect was shown on hospital mortality and on 28-day mortality.[15]

rIL-7 is the less studied molecule. IL-7 is a nonredundant cytokine for the development of lymphocytes. rIL-7 was added in the growth medium of lymphocytes isolated from the peripheral blood of 70 patients with septic shock. After priming the T-cell receptor, addition of rIL-7 increased proliferation of CD4lymphocytes and the subsequent production of IFN-γ.[16]

Management of sepsis remains an unmet medical need. Failure of clinical development of agents that attenuate proinflammatory phenomena has shed light on strategies of immunostimulation. Much of the development of the concept is based on our understanding of immune tolerance after exposure to LPS. In a recent study by the Hellenic Sepsis Study Group, circulating LPS was measured in the serum of 395 patients.[17] Varying endotoxemia was found in relation with the type of underlying infection. LPS was markedly elevated among patients with community-acquired pneumonia and primary Gram-negative bacteremia compared with acute pyelonephritis, VAP and intrabdominal infections.[17] As a consequence, it is expected that the degree of immunoparalysis generated by LPS will vary considerably from one patient to the other. There is an obvious need for large-scale RCTs to prove the efficacy of strategies of immunostimulation. In these studies, one genomic biomarker should be used to select the appropriate patients for inclusion and to monitor treatment.