In sepsis, suppression of innate immunity is marked by the functional reprogramming of myeloid cells. Monocytes from sepsis patients produce fewer proinflammatory cytokines (tumor necrosis factor [TNF]α, interleukin [IL]-1α, IL-1β, IL-6, IL-12) after lipopolysaccharide (LPS) stimulation, while anti-inflammatory cytokines like IL-1RA and IL-10 are preserved or increased. This suggests altered intracellular signaling that limits proinflammatory cytokines while promoting anti-inflammatory responses. This anti-inflammatory reprogramming extends beyond circulating monocytes to dendritic cells and tissue macrophages with reduced proinflammatory cytokine production and enhanced IL-10 production. ^, Macrophages are pushed into a tolerant phenotype, secreting IL-10 and TGF-β. Altogether, this altered state of myeloid cells in sepsis patients is also referred to as tolerance. Although tolerance may reduce antimicrobial activity in sepsis, it is also part of a physiologic response to severe stress. Balancing antimicrobial activity with tolerance to limit tissue damage is crucial for favorable outcomes.

In addition to tolerance, myeloid cells in sepsis exhibit reduced antigen-presenting capacity, reflected by decreased surface expression of human leukocyte antigen–DR (HLA-DR), which is essential for antigen presentation to T cells. This reduction is observed in monocytes as well as in peripheral blood dendritic cells from sepsis patients. ^, Decreased monocytic HLA-DR (mHLA-DR), assessed by flow cytometry, has become the most important biomarker of sepsis-induced immunosuppression and has been successfully used in randomized clinical trials (RCTs). Reduced mHLA-DR expression reflects monocyte dysfunction and is associated with impaired cytokine production in response to LPS. Persistently low mHLA-DR expression in sepsis patients correlates with an increased risk of secondary infections and higher mortality. ^,

Beyond inhibition and reprogramming of the myeloid cell populations in blood and organs, increasing evidence supports the presence of immunosuppressive cells termed myeloid-derived suppressor cells (MDSCs) in sepsis. MDSCs are a heterogeneous group of immature myeloid cells including progenitors of monocytes, neutrophils and dendritic cells, that expand during inflammation or infection. During chronic inflammation, dysregulated myelopoiesis drives the accumulation of MDSCs. MDSCs suppress both innate and adaptive immune cells, particularly by inhibiting T-cell responses through l -arginine consumption and IL-10 production. MDSCs also promote the expansion of regulatory T-cells (Tregs). Studies have shown that MDSC numbers remain elevated for up to 28 days post sepsis and are associated with increased secondary infections, prolonged ICU stays, and poor outcomes. ^, In surgical sepsis survivors, MDSCs remained elevated for 6 weeks post sepsis. Furthermore, their suppressive activity developed around 14 days post sepsis, suggesting that MDSCs may acquire immune-suppressive functions over time. It is important to note however that MDSCs may also play a beneficial role by limiting hyperinflammation during early sepsis and protecting against organ dysfunction.

Sepsis-induced immunosuppression also affects adaptive immune cells, particularly T lymphocytes, which undergo anti-inflammatory reprogramming similar to innate immune cells. T cells from peripheral blood, as well as spleen and lung tissue from postmortem sepsis patients, show altered cytokine production. Reduced production of cytokines includes IFN-γ, TNF, and IL-2, while IL-4 and IL-10 remain largely unchanged. ^, B cells show decreased HLA-DR expression and increased IL-10 production. Antigen presentation defects further impair T cell activation, with reduced expression of costimulatory and signaling molecules, like CD28 and CD3.

Sepsis causes lymphopenia, increasing susceptibility to secondary infections and late mortality. This depletion results from widespread apoptosis in B cells, natural killer (NK) cells, and CD4/CD8 T cells in clinical and postmortem studies. ^, CD4 T cell numbers decrease but typically recover in survivors, while failure to restore CD4 counts correlates with poor prognosis. Sepsis also disrupts immunologic memory by depleting memory CD8 T cells, altering their composition and reducing their diversity, while memory B cells are also disproportionately affected by apoptosis.

Sepsis alters CD4 T cell subsets, shifting the balance from proinflammatory Th1 cells to anti-inflammatory Th2 cells. This shift, marked by increased Th2-associated cytokines IL-4 and IL-10 and reduced Th1 cytokines TNF-α, IFN-γ, and IL-2, is driven in part by impaired IL-12 production and associated with diminished antimicrobial defense. ^, Persistently elevated Th2/Th1 ratios correlate with higher ICU-acquired infections and 28-day mortality. Additionally, Th17 cells, important for defense against fungal pathogens, are increased in sepsis patients but decline after day 7 in nonsurvivors, while survivors maintain higher levels. This may, in part, explain the increased susceptibility to secondary fungal infections in sepsis.

During sepsis, Tregs increase relative to effector T cells, especially in nonsurvivors. This phenomenon is best interpreted not as an absolute expansion of Tregs, but rather as a relative increase resulting from the loss of other T cell populations. One possible explanation is that Tregs have increased resistance to apoptosis, potentially due to elevated BCL-2 expression. Tregs promote monocyte apoptosis in septic shock, and exert broad immunosuppressive effects, including production of inhibitory cytokines. Increases in Treg proportions may also be beneficial, as they help limit excessive inflammation and promote tissue repair, particularly in conditions such as acute lung injury.

Sepsis-induced hyperinflammation paradoxically drives immunosuppression, promoting tolerance of innate and adaptive immune cells. Supporting this concept, patients who developed secondary infections exhibited a more dysregulated host response during the first days after ICU admission, characterized by heightened inflammation, coagulation, endothelial activation, and vascular disruption. These changes are partly driven by epigenetic and metabolic reprogramming. In parallel, overwhelming inflammation induces immune cell loss through dysregulated activation of programmed cell death.

Sepsis-induced apoptosis involves both death-receptor and mitochondrial pathways, initiated by inflammatory and cellular stress signals. ^, There is evidence that checkpoint regulators, such as programmed cell death-1 (PD-1), contribute to apoptosis of T cells. The interaction of PD-1 and its ligands PD-L1 and PD-L2 on antigen-presenting cells results in T cell suppression and increases susceptibility to apoptosis. ^, The expression of PD-1 on T cells and PD-L1 on antigen presenting cells is associated with lymphopenia, T cell anergy, and T cell apoptosis. ^, Increased expression was also associated with occurrence of secondary infections and mortality. Ex vivo PD-1 blockade in CD8 ⁺ T cells from sepsis patients reduces apoptosis and enhances IFN-γ production. Other checkpoint regulators like cytotoxic T-lymphocyte–associated protein 4 (CTLA-4) may contribute to immunosuppression. CTLA-4 is a negative checkpoint regulator that suppresses T cell activity, promotes apoptosis, and its expression on T cells increases during sepsis.

Sepsis induces dendritic cell apoptosis, linked to secondary infections and higher mortality. ^, Apoptosis of circulating monocytes also increases. Macrophages remove apoptotic cells from inflamed tissues, promoting IL-10 and TGF-β release, suppressing HLA-DR expression, and inducing endotoxin tolerance. In contrast to other innate immune cells, apoptosis of neutrophils is delayed in sepsis. Neutrophils during sepsis also undergo functional changes, including reduced chemotaxis, decreased intracellular expression of myeloperoxidase and lactoferrin, and impaired kinase activity, which may contribute to immunosuppression. Furthermore, IFN-γ–stimulated neutrophils in experimental endotoxemia suppress T cells proliferation through expression of PD-L1.

Alternative forms of programmed cell death, including pyroptosis and ferroptosis, also contribute to immune dysregulation in sepsis, with pyroptosis causing macrophage loss. Similarly, ferroptosis drives macrophage death and skews the immune response toward a more tolerant phenotype. Autophagy, a key process for maintaining homeostasis and supporting antigen presentation, is impaired in sepsis. In a murine model, defective T-cell autophagy accelerates apoptosis and increases IL-10 production, promoting immunosuppression.

Epigenetic modifications are another mechanism affecting immunosuppression by influencing gene expression. In experimental models of endotoxin tolerance, sepsis-induced immunosuppression has been characterized by loss of activating histone marks, accumulation of repressive marks, and increased histone deacetylation at promoters of proinflammatory genes, leading to sustained transcriptional repression.

Epigenetic changes also affect adaptive immune cells, promoting repressive histone modifications at key cytokine and transcription factor genes in T cells and favoring regulatory T-cell differentiation. Evidence from murine models suggests that these epigenetic changes can originate in bone marrow progenitor cells, resulting in long-lasting impairment of macrophage inflammatory responses after sepsis. Although less studied than chromatin modifications, posttranscriptional regulation by microRNAs also contributes to immune regulation in sepsis, although their role in immunosuppression remains unclear.

Epigenetic reprogramming in sepsis leads to profound changes in blood leukocyte transcriptomes, affecting 70% to 80% of RNA transcripts compared to healthy controls. ^, These changes, likely also influences by shifts in leukocyte composition, encompass proinflammatory and anti-inflammatory programs, reflecting the heterogeneity of immune dysregulation in sepsis. Most of these transcriptional alterations appear independent of the source of infection or pathogen, supporting sepsis as a host-driven syndrome. Hypoxia-inducible factor-1α (HIF-1α) is a key regulator of this host response. HIF-1α is upregulated and transcriptionally active in sepsis monocytes driving reprogramming toward an endotoxin-tolerant, immunosuppressive phenotype. In addition, early repression of the antiapoptotic gene BCL2 in circulating mononuclear cells is linked to increased lymphocyte apoptosis, CD4 ⁺ and CD8 ⁺ T-cell lymphopenia, and poor outcomes in sepsis. The sepsis transcriptomic response also overlaps with severe blunt trauma or noninfectious respiratory failure, indicating an intrinsic response to inflammatory stimuli.

Alterations in the metabolism of immune cells play a key role in immunosuppression. Usually during inflammation, immune cells shift from oxidative phosphorylation to aerobic glycolysis despite sufficient oxygen (Warburg effect) to meet energy demands. However, this metabolic program is not universal: pathogen- and context-specific stimulation can induce more complex metabolic rewiring in myeloid cells, characterized by increased glycolysis with variable, and in some cases enhanced, oxidative phosphorylation. Warburg-like metabolic reprogramming promotes HIF-1α–dependent transcription of proinflammatory genes, whereas this pathway is dysregulated during LPS-induced tolerance. Tolerant monocytes produce less lactate than nontolerant cells, reflecting an impaired ability to induce Warburg-like glycolysis. Beyond defective glycolysis, these cells also display reduced oxidative phosphorylation and impaired fatty acid oxidation, consistent with a state of immunometabolic paralysis. These metabolic defects were associated with impaired mTOR signaling. ^,

Dysregulation of central nervous system–immune interactions might contribute to immunosuppression. Sepsis-associated encephalopathy, driven by neuroinflammation and cerebral ischemia, suppresses peripheral immune responses, including impaired function of monocytes, dendritic cells, neutrophils, and T cells, although most evidence derives from animal models. The vagus nerve’s cholinergic anti-inflammatory reflex can suppress innate immune responses in animal sepsis. Similarly, the sympathetic nervous system may promote immunosuppression through stimulation of β-adrenergic receptors on immune cells. However, the exact role of these neuroimmune mechanisms in human sepsis remains unclear.

The β-adrenergic receptor also plays a role in therapies used in sepsis, such as norepinephrine, which, in addition to its vasopressor function, can exert anti-inflammatory effects by suppressing proinflammatory cytokine production and increasing IL-10 in experimental sepsis models and in humans, mediated largely through β-adrenergic signaling. Glucocorticoids, like hydrocortisone, are used in sepsis after improving outcomes in an RCT. While beneficial during the hyperinflammatory phase, their anti-inflammatory effects can worsen immunosuppression.

Several cytokines have been explored to reverse sepsis-induced immunosuppression. IFN-γ increases monocyte HLA-DR expression in experimental endotoxemia and sepsis patients, improving immune function. Patients with invasive Candida and/or Aspergillus infections regained partial immune function after IFN-γ treatment. Recombinant IFN-γ has shown effectiveness in treating sepsis-induces immunoparalysis, defined by reduced HLA-DR expression, in the ImmunoSep phase II RCT. Another interferon studied in immunosuppression is IFN-β. IFN-β restored the LPS-induced immunosuppressed state of monocytes in experimental endotoxemia. In murine sepsis, IFN-β significantly improved survival when administered after the onset of sepsis.

IL-7 promotes lymphocyte survival, proliferation, and function and reverses sepsis-induced lymphopenia in animal models. In an RCT, recombinant IL-7 increased CD4 ⁺ and CD8 ⁺ T cells in septic shock patients with severe lymphopenia, with a favorable safety profile, although it has yet to show an effect on clinical outcomes. IL-15 improves survival in murine sepsis by preventing apoptosis of NK cells, dendritic cells, and CD8 ⁺ T cells and by enhancing IFN-γ production. However, reported toxicities in oncology trials raise concerns regarding its clinical applicability in sepsis. IL-33 has demonstrated protective effects and reduced mortality in murine sepsis. Finally, granulocyte–macrophage colony-stimulating factor (GM-CSF) restores monocyte HLA-DR expression and reduces ICU length of stay in sepsis patients with immune suppression. Although RCTs and meta-analyses report GM-CSF improved leukocyte recovery and infection clearance, these benefits have not translated into reduced mortality.

Immune checkpoint inhibitors, particularly targeting the PD-1/PD-L1 axis, have also been studied in immunosuppression. Monocyte PD-L1 expression predicts 28-day mortality and may serve as a biomarker to identify patients most likely to benefit from checkpoint inhibition. Phase I/II RCTs of anti-PD-L1 therapy in sepsis have been well tolerated, improving lymphocyte counts and monocyte HLA-DR expression. However, effects on major clinical outcomes such as secondary infections and mortality remain unproven, and plans for a phase III RCT were terminated prematurely. Other inhibitory checkpoints like CTLA-4 improve survival in murine models, but human data are lacking.

Thymosin α1 (Tα1) is a thymic peptide proposed to restore immunity in sepsis, showed a trend toward mortality reduction in one RCT but no effect on 28-day mortality in a recent phase III RCT. A Phase II RCT targeting toll-like receptor 4 (TLR4) showed a trend toward lower mortality, while a Phase III RCT failed to reduce 28-day mortality. The use of caspase inhibitors improves survival in murine sepsis by preventing lymphocyte apoptosis, but results in human studies are still lacking. Intravenous immunoglobulin (IVIg) has immunomodulatory properties and may reduce mortality in sepsis, though results are mixed. The Surviving Sepsis Campaign guidelines recommend against intravenous immunoglobulins in patients with sepsis or septic shock due to low quality of evidence. Mesenchymal stem cells (MSCs) have shown potential for immunomodulation and tissue repair in sepsis, with early studies indicating feasibility and tolerance, but further large trials are needed to confirm efficacy.

Innate immunity can develop a form of memory, termed trained immunity, where prior pathogen or vaccine exposure enhances future immune responses. Trained immunity is driven by epigenetic and metabolic reprogramming of monocytes and is epigenetically and functionally the opposite of endotoxin tolerance. β-glucan-induced trained immunity restores monocyte responses, including cytokine production. A fusion protein of apolipoprotein A1 and IL4 reversed endotoxin tolerance in murine models and ex vivo human endotoxemia models by inducing trained immunity. However, induction of trained immunity with gamma-irradiated Bacille Calmette-Guérin (BCG) did not enhance immune responses in human endotoxemia. Future studies are needed to prove clinical effectiveness of trained immunity-based immunotherapies.