During the acute phase of sepsis, both in animal models and in patients, a significant elevation is observed in white blood cell counts and levels of inflammatory cytokines, such as tumor necrosis factor (TNF)-α, interferon (IFN)-γ, and IL-6 (38–41). When the host is exposed to exogenous and/or endogenous pathogens’ pathogen associated molecular patterns (PAMPs), such as lipopolysaccharide (LPS) (42, 43) and mannose-binding lectin (MBL) (44), or to damage-associated molecular patterns (DAMPs) released from its own damaged tissue cells, such as histones (45), high-mobility group box-1 (HMGB-1) (46, 47), and heat shock proteins (HSPs) (48), these molecules are recognized and interact with pattern recognition receptors (PRRs) on the surface of antigen-presenting cells. This interaction marks the initiation of the acute inflammatory response in sepsis. Current research has categorized PRRs into five subfamilies: the Toll-like receptors (TLRs) (49, 50), the nucleotide oligomerization domain (NOD)-like receptors (NLRs) (51, 52), the retinoic acid-inducible gene-I (RIG-I)-like receptors (RLRs) (53), the C-type lectin receptors (CLRs) (54, 55), and the Absent in melanoma-2-like receptors (ALRs) (56). The effector domains of PRRs mediate the activation of downstream inflammatory signaling pathways by recognizing specific ligands. This activation leads to the release of pro-inflammatory cytokines, recruitment of innate immune cells, and induction of inflammatory responses. Specifically, to address intracellular infections caused by pathogens, certain cytoplasmic PRRs (typically from the NLRs or ALRs families) often serve as receptor proteins that participate in the assembly of inflammasomes. Inflammasomes recruit pro-caspase-1 and activate it to caspase-1, which in turn cleaves pro-interleukin-1β, pro-interleukin-18, and gasdermin (GDSMD) into their mature forms, triggering an inflammatory response and cell pyroptosis. Subsequently, these cytokines induce further proliferation and activation of immune cells, skewing them toward an inflammatory phenotype and elevating levels of various inflammatory mediators such as interleukins IL-1, IL-6, IL-17, TNF-α, IFN-γ, and chemotactic factors such as prostaglandins, histamine. For instance, neutrophils combat pathogens through direct actions such as phagocytosis, degranulation, and the release of neutrophil extracellular traps (NETs), as well as indirectly through the release of reactive oxygen species (ROS), reactive nitrogen species (RNS), and proteolytic enzymes during cell proliferation and migration (57–60). Macrophages polarize toward an M1 inflammatory phenotype and release large amounts of IL-1β, TNF-α, and IL-6 (61–63), while DCs mediate the activation of CD4⁺T (64). This syndrome is also accompanied by an imbalance of redox reactions (65, 66), neutrophil-endothelial cell adhesion (67), activation of the complement system (68, 69), and the coagulation cascade (70, 71). Additionally, intense complement activation (especially C3 and C5a) enhances vascular permeability, and increases the adhesiveness between leukocytes and vascular endothelial cells, further promoting the inflammatory response and damaging self-tissue organs (72, 73). A recent single-cell transcriptomic analysis revealed the inflammatory profile of peripheral blood mononuclear cells (PBMCs) in COVID-19 and sepsis patients, identifying ten highly inflammatory cell subtypes and their characteristics (74). The adaptive immune response typically lags behind the innate immune response, with antigen-presenting cells activating T lymphocytes through the dual signaling system of the major histocompatibility complex (MHC)/antigen peptide-T cell receptor (TCR) and CD80-CD28 (75, 76). Under the influence of various cytokines released by innate immune cells, tissue cells, and activated lymphocytes (77), T lymphocytes differentiate into multiple subgroups, including cytotoxic T cells (CTLs), helper T cells (Th), and regulatory T cells (Tregs), playing a crucial role in the acute phase inflammatory response in sepsis hosts. Additionally, B lymphocytes counteract pathogen invasion by secreting antigen-specific antibodies and inflammatory cytokines (78).

Unfortunately, during the acute phase of sepsis, the excessive activation of immune cells and pro-inflammatory cytokines is not adequately restrained by anti-inflammatory responses. Consequently, this uncontrolled and excessive inflammatory response not only fails to efficiently eradicate pathogens within the host but also leads to severe cellular death, tissue damage, and organ dysfunction, potentially resulting in the early death of the host (79).

Under normal circumstances, as infections are cleared and host compensatory anti-inflammatory responses are modulated, a patient’s heightened inflammatory response gradually subsides and returns to physiological levels, ultimately restoring the immune homeostasis of the internal environment. However, in the later stages of sepsis, the host often experiences a dysregulated anti-inflammatory response, which is not conducive to moderating inflammation but instead manifests as immune tolerance or paralysis, showing a low response to pathogens (80). These factors significantly increase the risk of secondary infections and adverse prognoses, including late or posthumous deaths (81, 82).

Lymphocyte apoptosis or reduction in their numbers is a significant factor contributing to immune paralysis. Preclinical studies using sepsis animal models indicate widespread apoptosis of parenchymal and immune cells across multiple organs, including the thymus, spleen, lungs, intestines, and skeletal muscle (83–86). More importantly, in mice undergoing cecal ligation and puncture (CLP), it has been observed that sepsis-induced impairment in T cells’ ability to combat pathogen infections can persist for several months (87). In clinical research, uncontrolled circulating immune cell apoptosis is considered a primary cause of impaired immune function (88–90). Additionally, transcriptomic analysis revealed that circulating lymphocytes in the later stages of the disease exhibit low inflammatory activity and immune suppression (91). There is a notable reduction in the numbers of CD4⁺ and CD8⁺T cells, with an increased proportion of Tregs (90, 92). This impaired proliferative capacity and sustained apoptosis of T lymphocytes may be associated with the upregulation of negative signaling pathways, such as the programmed death receptor-1/programmed death ligand-1 (PD-1/PD-L1) axis (93–96). The role of B lymphocytes in immune suppression remains unclear, yet studies have identified selective depletion or increased apoptosis of memory B cells in sepsis patients (97). These cells exhibit reduced MHC II expression and tend toward a CD21^lowCD95^high exhausted-like phenotype (98), with a significant reduction in antigen-specific antibody release. During the process of immune suppression in the host, innate immune cells are both victims and perpetrators (99–102), including neutrophils (103–105), DCs (106–108), and monocytes/macrophages (109–112). These cells commonly experience abnormal differentiation, functional impairments, and extensive tissue infiltration. Research has shown that an increased proportion of immature neutrophils (CD10^lowCD16^low cells) in the whole blood of sepsis patients is associated with an increased risk of early death within 48 hours after sepsis onset (113). This may be linked to the upregulation of myeloid-derived suppressor cells (MDSCs) subgroups (113) and the integrin Mac-1 (αMβ2) (114), which mediates suppression of T cell proliferation. Recent studies have reported a higher abundance of immature neutrophil subgroups expressing genes related to IL1R2, PADI4, and MPO in sepsis patients, and in vitro experiments suggest that these immature neutrophils can inhibit the proliferation and activation of CD4⁺T cells (115). Significant changes also occur in DCs, characterized by the acute phase’s systemic high-inflammatory microenvironment excessively activating the immature DCs stored in parenchymal tissues and lymphoid organs until they are exhausted. However, newly generated DCs are functionally immature, which includes acquiring the immunogenic phenotype of pathogens, capturing, processing, and/or presenting antigens, as well as the capacity to stimulate T cell activation (116). Recent studies have found that low expression of monocyte human leukocyte antigen DR (_MHLA-DR) of in sepsis patients reduces activation of T lymphocytes (117). Increasing evidence suggests that low expression of _MHLA-DR can serve as a biomarker for predicting immune paralysis or poor prognosis in sepsis (118–121). Concurrently with extensive apoptosis of immune cells, a class of immature myeloid cells collectively known as MDSCs—which include progenitors or precursors of neutrophils, DCs, and monocytes—proliferate abundantly and are released into the bloodstream (122, 123). These MDSCs exhibit significant immunosuppressive properties, including inhibiting the proliferation and activation of effector lymphocytes while activating Tregs, reducing the production of inflammatory cells and promoting the release of anti-inflammatory cytokines, as well as upregulating the expression of immune checkpoint molecules.

Endotoxin tolerance constitutes a critical aspect of immunosuppression (124), often resulting from the innate and adaptive immune cells of the host being persistently exposed to low levels of endotoxins or LPS, entering a transient “desensitized state.” This leads to an unresponsive state of the host immune system to sudden, high-dose endotoxin or LPS exposure. Numerous preclinical studies have shown that mouse monocytes continuously exposed to LPS in vitro can undergo ET, characterized by the downregulation of inflammatory cytokines (including TNF-α, IL-6, and IL-8) and the upregulation of anti-inflammatory cytokine (IL-10) expression (125). Macrophages from mice pretreated with LPS also exhibit a diminished response to subsequent LPS stimulation in vitro, with downregulated mRNA expression of genes encoding recombinant granulocyte-macrophage colony-stimulating factor (GM-CSF), IFN-γ-inducible protein-10, JE/monocyte chemoattractant protein-1, and macrophage-inflammatory protein-1β/2 ( 126). This reduced responsiveness and immune tolerance of monocytes/macrophages to LPS stimulation may be associated with the remodeling of NF-κB function, chromatin modifications, and reprogramming of inflammatory genes (127). Besides, tolerance-inducing DCs in mice, which express lower levels of MHC-II and CD86, can induce the proliferation and recruitment of CD4⁺Foxp3⁺ Tregs through the secretion of TGF-β (128). In clinical research, healthy volunteers continuously stimulated with LPS exhibited a downregulation of pro-inflammatory cytokine levels in vivo ( 129). In vitro experiments showed that under high levels of LPS stimulation, the expression of MHC class II, CD86, and MHC II class transactivator (CIITA) in human monocytes was significantly reduced, leading to impaired antigen presentation (130). Shalova et al. treated monocytes from sepsis patients with LPS in vitro, and their findings indicated that the expression of genes associated with pro-inflammatory cytokines (such as TNF-α, IL-1A, IL-1B, IL-6, IL-12A, IL-23A) and chemokines (such as CCL3, CCL4, CCL5, CCL20, CCL23, CXCL2, CXCL11) were not upregulated. Similarly, gene expression related to activation-associated surface molecules (such as CD80, CD44) and transcription factors (such as ATF5, NFKB1, NFKB2, REL, RELA) were also deficient. This suggests that monocytes in sepsis patients are unable to actively respond to LPS stimulation, indicative of an immune functional defect (131). Research has also reported that hypoxia-inducible factor-1α (HIF1-α) is overexpressed in human tolerant monocytes and targets the upregulation of PD-L1-related gene expression, thereby mediating the suppression of T lymphocyte proliferation and activation (132). Although there are currently no universally recognized biomarkers for ET, genomic and transcriptomic analyses can partially elucidate the genetic variations associated with the onset of immune tolerance in sepsis hosts, which is helpful in distinguishing patients with immune dysfunction (133–135).

In addition to the mechanisms mentioned above, various intrinsic mechanisms contribute to secondary immunosuppression, including dysfunction in the pro-inflammatory functions of the central nervous system (136), epigenetic or transcriptional regulation (137, 138), and immune metabolic dysfunction. For instance, PBMCs exhibit reduced cytokine release capacity, and monocytes display metabolic dysfunctions such as impaired glycolysis and lipid oxidation. Additionally, mitochondrial damage within tissue and immune cells is accompanied by decreased ATP and NAD⁺ levels, reduced lactate production, and diminished oxygen consumption (139–141).

Blocking the activity of inflammatory cytokines is a fundamental strategy to inhibit acute-phase excessive inflammatory responses. For instance, in vivo experiments with Infliximab treatment (an anti-TNF-α antibody) have significantly reduced serum TNF-α levels in septic rats, markedly improving acute lung, liver, and kidney injuries, and increased the 7-day survival rate of rats from 0% to 37.5% (142). The IL-6 receptor antagonist Tocilizumab can inhibit NF-κB activation, significantly reducing inflammatory responses and oxidative stress levels in CLP rats, and offers protection against sepsis-induced acute lung and kidney injuries (143). Flierl and colleagues confirmed that anti-IL-17 treatment significantly increased the 7-day survival rate of CLP mice (control group: 10% vs. anti-IL-17 antibody group: 60%), and administering the treatment 12 hours later still offered some protection to the host (144). Intracerebral administration of anti-IL-17 or anti-IL-17R antibody also mitigated microglial activation and central nervous system inflammation in CLP mice by blocking the IL-17A/IL-17R pathway, alleviating sepsis-associated encephalopathy (SAE), and improving cognitive dysfunctions (145). Wan and others found that an anti-CCR6 antibody, by blocking the CCR6-CCL20 axis, inhibited the recruitment and migration of γδT cells and the release of IL-17A in CLP mice, reducing the infiltration of inflammatory cells in the liver induced by sepsis (146). In studies of immune suppression, Unsinger and colleagues tested the efficacy of recombinant human IL-7 (rhIL-7) in a peritonitis-induced sepsis model in mice. The results indicated that rhIL-7 mediated a reduction in apoptosis of CD4⁺ and CD8⁺T cells and promoted the production of IFN-γ by upregulating the expression of Bcl-2, thereby improving immune suppression in mice (147). Conversely, Murphey and others, using a combination of IFN-γ and anti-IL-10 antibody in CLP mice experiencing immune suppression, did not observe a significant improvement in bacterial clearance rates or survival rates (148). However, recent studies have shown that IFN-γ treatment in CLP mice can upregulate CD86 expression on DCs and reduce DC apoptosis, reversing the immune suppression caused by sepsis (149). These findings highlight that while supplementing or modulating cytokines—whether anti-inflammatory or pro-inflammatory—is a relatively straightforward immunomodulatory strategy, the application protocols and therapeutic outcomes require further study.

The activated complement system plays a critical role in the transmission of inflammatory signals. For instance, studies have shown that using an anti-C5aR antibody in CLP mice significantly reduces the accumulation of inflammatory cytokines in plasma and decreases mortality (150). Another in vitro experiment observed that after administering an anti-C5a antibody, chemotaxis of neutrophils activated via the complement alternative pathway was significantly reduced in septic mice, helping to regulate excessive accumulation and abnormal infiltration of neutrophils in tissues (151). However, recent research indicates that C5a/C5aR also participates in anti-inflammatory signaling. Sommerfeld and colleagues observed that mice with a C5aR1 gene deficiency exhibited high levels of IFN-γ and low levels of IL-10 post-CLP (152). Beyond C5a/C5aR, other components of the complement system have also garnered attention. Chung and others discovered that hippocampal tissue expression of complement C1q was upregulated in SAE mice, mediating neuronal damage. Intracerebral injection of a specific C1q blocker significantly protected microglial cells, improving neurocognitive function impairments in mice (69).

Immune checkpoints (ICs) inhibit excessive activation and proliferation of immune cells, regulate inflammatory responses, prevent damage to self-tissues and organs, and promote the restoration of immune homeostasis. However, the continuous transmission of negative signaling pathways can also induce uncontrollable cell apoptosis and immune suppression. Common ICs, including PD-1/PD-L1, cytotoxic T-lymphocyte antigen-4 (CTLA-4)/CD80(CD86), B-and T-lymphocyte attenuator (BTLA)/herpes virus entry mediator (HVEM), and T-cell immunoglobulin and mucin-domain containing-3 (TIM-3)/Galectin-9 (Gal-9), are all potential therapeutic targets. Chang and colleagues reported that the individual use of anti-PD-1, anti-PD-L1, or anti-CTLA-4 antibodies could promote the release of IFN-γ by blocking lymphocyte apoptosis and upregulating the expression of MHC II on DCs, modulating the immune suppression state in mice with primary and secondary fungal sepsis (153). Similarly, Zhang and others observed that in CLP mice, the expression of PD-1/PD-L1 was upregulated on T cells, B cells, and monocytes, and that anti-PD-L1 antibodies could inhibit some lymphocyte apoptosis and exhaustion induced by sepsis, promote the release of TNF-α and IL-6, reduce the production of IL-10, and significantly improve the survival rate of the mice (154). Huang and colleagues discovered that the deletion of TIM-3 in CD4⁺T cells in septic mice could alleviate lymphocyte apoptosis and restore their proliferative activity, thus protecting organ function and reducing mortality (155). In the same year, Liu and others observed that treatment with anti-TIM-3 in CLP mice reduced lymphocyte apoptosis and significantly alleviated sepsis-induced acute lung and liver damage (156). However, the timing and dosage of administration can lead to significant differences in treatment effectiveness. For instance, Inoue and colleagues noted an increase in CTLA-4 expression in T cells in CLP mice, and administering anti-CTLA-4 antibodies reduced sepsis-induced lymphocyte apoptosis, but the 7-day survival rate of the mice showed a clear dose dependency; higher doses of anti-CTLA-4 antibodies decreased survival rates while lower doses increased them (157). Immune checkpoint inhibitors (ICIs) play a crucial role in inhibiting persistent apoptosis and exhaustion of lymphocytes. Other novel checkpoint inhibitors such as VISTA (158) and ICOS (159) are also being developed and validated.

Mesenchymal stem cells (MSCs) are described as balancers of the inflammatory microenvironment and immune dysregulation due to their ability to modulate the activation, maturation, proliferation, differentiation, and effector functions of various immune cells (170, 171). This modulation occurs through direct contact with target cells, the release of bioactive factors (such as cytokines, growth factors, chemokines), and paracrine pathways involving the secretion of extracellular vesicles and exosomes that contain cytokines, miRNAs, and other soluble factors. Extensive in vivo and in vitro experimental results support that MSCs and their derivatives can regulate programmed cell death (including apoptosis (160–162), autophagy (163–165), pyroptosis (166–168), and ferroptosis (169)) in immune and tissue cells, maintaining homeostasis within the host environment. They can modulate imbalanced immune responses, alleviate tissue and organ damage, improve multi-organ dysfunction, and reduce mortality. Despite satisfactory results in rodent models, the efficacy of MSCs and their derivatives in larger animal models remains unclear. For instance, Horak and colleagues observed that in pigs with peritoneal sepsis treated with MSCs, there was no significant alleviation of hemodynamic abnormalities, the systemic overactivation of inflammatory responses was unmodulated, and organ failure assessment scores continued to increase (172). Thus, further preclinical experiments are needed to determine the appropriate pathways, dosages, indications, and potential adverse reactions for the use of MSCs and their derivatives in sepsis and related diseases.

The acute inflammatory response in sepsis patients, coupled with the encouraging results from preclinical studies of immunotherapy, has provided a preliminary basis for conducting clinical trials on immunosuppressive therapies. Although some studies indicate that administering anti-inflammatory treatments within the first few hours after the onset of sepsis can somewhat mitigate the systemic inflammatory response and protect against organ dysfunction, overall, immunosuppressive strategies have not yielded satisfactory results in clinical trials for sepsis.

To date, clinical trials exploring immunosuppressive therapies for sepsis have indicated that the “theoretical” or “idealized” strategy of immunosuppression to mitigate the excessive inflammatory response in sepsis and restore host immune homeostasis is not always viable. Simultaneously, with advancing research into sepsis, a significant number of preclinical and clinical studies have observed that hosts often exhibit an excessive state of immunosuppression in the later stages of the disease. This severe immunoparalysis mediates secondary infections, subsequent deaths, or severe adverse prognoses (93, 213). In this context, there has been a shift in focus from “immunosuppressive therapy” to “immune enhancement therapy” or “immune stimulation therapy” in an effort to reverse the state of immune paralysis in sepsis hosts. The goal is to reduce apoptosis of immune cells and stimulate their proliferation and anti-inflammatory effects, enhance the release of inflammatory cells, and improve patients’ long-term survival rates.

Numerous preclinical studies have demonstrated that MSCs and their derivatives can protect against organ dysfunction and improve survival rates in septic animals (217). MSCs and their derivatives have unique advantages such as low immunogenicity, multi-directional differentiation potential, and the ability to modulate immune responses, providing a new option for the treatment of sepsis. However, the ethics, safety, effectiveness, and strategies for application and therapeutic mechanisms still need to be confirmed through extensive clinical trials (207, 208).

As of April 19, 2024, our team entered the following keywords into ClinicalTrials.gov: “sepsis,” “Immunotherapy,” “Immunosuppressive therapy,” “Immunomodulatory therapy,” “Immune checkpoint,” “Checkpoint inhibitor,” “immunoregulation,” “antibody,” “Anakinra,” “Complement inhibitor,” “GM-CSF,” “Thymosin,” “MSCs,” “MSCs-Exo.” After screening, we compiled 35 registered clinical trials. These clinical trials use similar or different immunotherapy strategies but are all aimed at exploring the safety and/or effectiveness of immunotherapy in sepsis patients ( Table 3 ).