The Dual Role of Natural Killer Cells in the Septic Liver

Abstract

Sepsis is a life-threatening organ dysfunction caused by dysregulated host responses to infection. As an important immune organ, the liver plays a vital role in the pathophysiological process of sepsis. When sepsis occurs, the liver exhibits unique immunological characteristics, including abnormal activation of pattern recognition receptors, loss of control of the complement system, adaptive immune regulation, and hepatic Kupffer cell polarization, which together determine the outcome of septic liver. Natural killer (NK) cells are innate immune cells with different biological characteristics in the liver than circulating NK cells. Natural killer (NK) cells assume a dual role in hepatic pathophysiology, functioning as both protectors and effectors of injury. On the protective front, these cells mediate antimicrobial defense and immunomodulation through interferon-gamma (IFN-γ) secretion, maintain vascular endothelial barrier integrity via vascular endothelial growth factor (VEGF) downregulation, and regulate hepatic inflammation through the production of anti-inflammatory cytokines such as interleukin-10 (IL-10). Additionally, their cytotoxic activity effectively suppresses viral replication within the hepatic microenvironment. Conversely, NK cells contribute to hepatocellular damage through distinct deleterious mechanisms. They elaborate tumor necrosis factor-alpha (TNF-α) and IFN-γ, thereby precipitating cytokine storm formation. Through the exogenous pathway (Fas ligand/TNF-related apoptosis-inducing ligand [FasL/TRAIL]) and the endogenous pathway (perforin/granzyme), they directly induce hepatocyte apoptosis. Furthermore, NK cells regulate ferroptosis-related gene expression—including heme oxygenase-1 (Hmox1) and solute carrier family 3 member 2 (Slc3a2)—thus driving the progression of sepsis-associated liver failure. This review describes the immunological characteristics of the liver in sepsis and the mechanisms of NK protection and damage to the liver. This study provides a new perspective for exploring immunotherapy of septic liver injury in the future.

Introduction

Sepsis is a life-threatening organ dysfunction caused by a dysregulated host response to infection, with a high incidence and mortality.¹ An epidemiological study of sepsis in the United States demonstrated that the overall mortality rate of patients with sepsis increased from 5.6% to 34.2% as the severity of sepsis increased.² The pathological process of sepsis is characterized by persistent excessive inflammatory responses and immunosuppression.³ As the largest solid organ in the body, the liver serves not only as a central hub for immunity and metabolism, but also as a critical target organ in sepsis, where pathogen recognition triggers systemic inflammatory responses and immunosuppression. In sepsis, the liver is susceptible to damage by pathogens, toxins, or inflammatory mediators, and eventually develops hepatic insufficiency. The average incidence of liver dysfunction in patients with sepsis is 39.9%, whereas the incidence of respiratory, renal, and neurological dysfunctions is 71.1%, 44.4%, and 70%, respectively.⁴ Therefore, although the incidence of hepatic dysfunction is lower than that of other organ dysfunctions in sepsis, it remains an independent risk factor for patient prognosis.⁵ We observed that, within the multivariate phenotypic spectrum of sepsis and septic shock, patients presenting with hepatic dysfunction and coagulopathy demonstrated prolonged durations of hospitalization and increased requirements for mechanical ventilation.⁶ The development of hepatic dysfunction during sepsis may precipitate hepatic failure and subsequent multiple organ dysfunction through the release of inflammatory mediators and the induction of immunosuppression. Therefore, heightened clinical vigilance and early recognition of sepsis-induced liver injury are imperative.⁷

Patients with sepsis usually die from multiple organ failure caused by persistent immunosuppression and secondary infections. This process is related to the apoptosis of various immune cells, such as neutrophils, B cells, monocytes, macrophages, dendritic cells (DC), and natural killer (NK) cells.⁸ NK cells were first discovered in the early 1970s.⁹ They are a subset of lymphocytes that can kill tumor cells and have antiviral effects. NK cells are mainly found in tissues such as the peripheral blood, liver, spleen, lungs, and bone marrow.¹⁰ However, the role of NK cells in the liver during sepsis remains unclear.

NK cells represent crucial immune organs and cells in sepsis, respectively. We systematically elucidate the immunological characteristics of the septic liver Figure 1, delineate the biological features of hepatic NK cells, deeply explore the dual mechanistic roles of NK cells in sepsis-associated liver injury, and prospect the clinical translation of NK cell-based immunotherapy, thereby providing a theoretical basis for the treatment of sepsis-induced hepatic injury.

Figure 1.

Immunological characteristics of the liver in sepsis. When sepsis occurs, pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs) activate pattern recognition receptors (such as TLR4 and NLRP3), thereby promoting the production of inflammatory factors (TNF-α, IL-6, IL-1β, IL-18) via the MAPK/NF-κB pathway and inducing pyroptosis to damage the liver. Concurrently, excessive activation of the complement system produces large amounts of C5a that binds to C5aR1 on neutrophils, triggering neutrophil entrapment and aggravating liver injury. Aβ protein binds to CD74 on B cells to inhibit B cell function and aggravate liver injury. α-GalCer induces B cells to differentiate into regulatory B cells producing IL-10, thereby reducing the production of pro-inflammatory cytokines such as TNF-α and IL-1β to protect the liver. High expression of PD-L1 in liver sinusoidal endothelial cells binds to PD-1 on CD8+ T cells, leading to adaptive immunosuppression. CD4+ T cells and TIGIT+ cells play an anti-infective role by blocking Tim-3 on CD8+ T cells to protect the liver. In addition, Kupffer cells display M1/M2 polarization imbalance. Kupffer cells release CXCL9, while hepatocytes release chemokines such as CXCL9 and CXCL10, which subsequently recruit natural killer cells to the liver through specific chemokine receptors such as CXCR3.

Immunological Characteristics of the Liver in Sepsis

Abnormal Activation of Pattern Recognition Receptors (PRRs)

Upon pathogen invasion, damage-associated molecular patterns (DAMPs) and pathogen-associated molecular patterns (PAMPs), serving as endogenous and exogenous danger signals, respectively, are recognized by pattern recognition receptors (PRRs) expressed on the surface of immune cells. This interaction subsequently activates intracellular signaling cascades, triggering immune system activation and promoting the release of inflammatory mediators, ultimately culminating in a “cytokine storm” that inflicts damage upon host cells.¹¹

Damaged host cells release endogenous molecules called DAMPs, also known as alarmins, such as high mobility group box 1 (HMGB1), extracellular cold-induced RNA binding protein (eCIRP), adenosine triphosphate (ATP), nicotinamide adenine dinucleotide (NAD), heat shock proteins (HSPs), histone proteins, free DNA, and mitochondrial DNA (mtDNA), which are proinflammatory and immunosuppressive when released into the intercellular space.¹² PAMPs are mainly components of pathogens, including lipopolysaccharides (LPS), nucleic acids, peptidoglycans, and lipoproteins. They bind to the PRRs expressed on the surface of immune cells, the cytoplasm, and the membrane of intracellular compartments. There are five significant pattern subfamilies of recognition receptors: Toll-like receptors (TLRs), nucleotide-binding domain-like receptors (NLRs), retinoic acid-inducible gene I (RIG-I)-like receptors (RLRs; also known as RIG-I-like helicases, RLHs), C-type lectin receptors, and absent in melanoma 2 like receptors (ALRs).¹¹

A mouse model of sepsis induced by cecal ligation and puncture (CLP) illustrated that TLR2 and TLR4 mainly recognized gram-positive bacteria and gram-negative bacteria,¹³ TLR3 and TLR7 are related to viral entry into the human body.¹⁴ In murine models of sepsis, hepatic expression of TLR2, 3, 4, and 7 was significantly upregulated at both the transcriptional and translational levels, concomitant with elevated serum aminotransferases ALT/AST and histopathological evidence of hepatic necrosis.¹⁵ Moreover, TLR9 knockdown suppressed activation of the MyD88‑NF-κB signaling axis, thereby attenuating IL-6 and TNF-α production while concomitantly enhancing hepatic bacterial clearance to mitigate sepsis severity.¹⁶ Several studies have reported that TLR4 plays a key role in sepsis-induced liver injury by activating MAP kinase and nuclear binding factor (NF-κB and/or AP-1) signaling pathways, leading to the activation of liver macrophages and endothelial cells and the release of TNFα, interleukin-1β(IL-1β), COX-2, and nitric oxide synthase(NOS-2) inflammatory mediators and cytokines, leading to liver injury.¹⁷ Chunxia et al illustrated that pyruvate dehydrogenase kinase 2 downregulation inhibited the activation of TLR4 mitogen-activated protein kinase signaling pathway, reduced the production of proinflammatory cytokines from macrophages and DC, and alleviated liver pathological damage in septic mice.¹⁸ Cluster of Differentiation 38 (CD38) is a multifunctional type II transmembrane protein widely distributed on various immune cells, which is closely related to the TLR signaling pathway. When CD38 is absent, TLR4 expression is upregulated, inducing the phosphorylation of p65 in the downstream NF-κB pathway, leading to the upregulation of the expression of downstream inflammatory factors and apoptotic genes, thus aggravating the liver injury of sepsis.¹⁹ Additionally, NLR family pyrin domain-containing protein 3 (NLRP3), serving as a critical component of the inflammasome, exhibits significantly elevated expression levels following bacterial infection in CD38-deficient mice, concomitant with marked upregulation of caspase-1, interleukin-1β (IL-1β), and interleukin-18 (IL-18). These inflammasomes can lead to the cleavage of procaspase-1 and the formation of activated caspase-1. Cleaved caspase-1 promotes the conversion of pro-IL-1β and pro-IL-18 to mature IL-1β and IL-18, recruitment of immune cells, and induction of proinflammatory cell death.^20,21 Activated caspase-1 can also cleave gasdermin-D (GSDMD) and induce proinflammatory cell death, known as pyroptosis. Conversely, TLR4 mutant mice exhibited significant reversal of liver pyroptosis and injury, accompanied by reductions in NLRP3 and GSDMD. Consequently. The TLR4-NLRP3-GSDMD pathway plays a crucial role in liver injury caused by bacterial infection in sepsis.¹⁹ TLR4 signaling is regulated by heme oxygenase-1 (HO-1), reducing TLR4-mediated inflammatory responses. Heme increased hepatic HO-1 protein expression in CLP mice, and inhibition of TLR4 reduced serum IL-6 levels and glutamate dehydrogenase activity, which is a key enzyme present in the mitochondrial matrix and a potential biomarker of mitochondrial damage. Ultimately, this restoration of mitochondrial homeostasis enhances mitochondrial quality control (QC) by attenuating excessive mitochondrial fission, thereby preserving mitochondrial biogenesis and autophagic function, which confers hepatoprotection.²²

In patients with sepsis, the trend of HMGB1 is like that of alanine aminotransferase (ALT), aspartate aminotransferase (AST), and total bilirubin, because HMGB1 promotes liver injury that leads to pyroptosis of liver macrophages in sepsis.²³ Similarly, in an animal model of sepsis, we found that eCIRP levels in the heart and liver were increased. It could be released from the nucleus to the cytoplasm, stimulating macrophages to release TNF-α and HMGB1, eventually leading to tissue damage.²⁴ However, when primary hepatocytes were treated with recombinant heat shock protein 72 (HSP-72), it bound to TLR2 and TLR4 of hepatocytes. This caused an increase in macrophage inflammatory protein 2 through the NF-κB pathway in a dose-dependent manner.²⁵ During sepsis, damaged tissues and immune cells release substantial amounts of extracellular ATP, which activates P2Y2 purinergic receptors expressed on hepatocytes, thereby upregulating NLRP3 inflammasome expression and inducing hepatocyte inflammation and death.²⁶ Conversely, NAD+ has been demonstrated to activate SIRT pathways, ameliorate T-cell exhaustion, and attenuate hepatic injury in experimental sepsis models.²⁷ Citrullinated histone H3 (CitH3), released by immune cells during the early phase of sepsis, drives NETosis and pyroptosis, thereby impairing bacterial phagocytosis within the liver.²⁸ In murine models of sepsis, LPS enhances DRP1-dependent mitochondrial fission and mitochondrial ROS generation in Kupffer cells, promoting the release of mitochondrial DNA (mtDNA) into the cytoplasm. This cytosolic mtDNA subsequently activates the cGAS-STING signaling pathway, triggering hepatocyte death and systemic inflammatory responses.²⁹ Collectively, these findings demonstrate that DAMPs serve as critical modulators of hepatic pathophysiology in sepsis.

Loss of Control of the Complement System

The complement system is mainly composed of more than 40 proteins, which have three activation pathways: The classical pathway (antigen-antibody complexes), the lectin pathway (recognition of specific sugar structures on the surface of pathogens), and the alternative pathway (Complement C3 direct activation). These complement activation pathways can form C3/C5 convertases to produce anaphylatoxins (C3a/C5a), opsonins (C3b/C4b), and membrane attack complexes (MAC).³⁰

The complement system is overactivated in the early stages of sepsis, and large amounts of C3a and C5a are generated to trigger a proinflammatory response. As the severity of sepsis increases, the host’s blood vessels leak, immune cell function is impaired, the coagulation system is activated, and multiple organ disorders occur.³¹ In a mouse model of sepsis, complement C3a can reduce the early infection of sepsis through macrophages, whereas combining C5a and C5aR of macrophages and neutrophils can aggravate the systemic inflammatory response.³² C5a can be produced by C5 convertase or phagocyte serine protease cleavage of C5, and then acts by binding to G protein-coupled receptors C5a receptor 1 (C5aR1/CD88) and C5a receptor 2 (C5aR2/C5L2). When C5aR1 was knocked out in septic mice, ALT and AST levels were reduced.³³ C5aR1 expression on immune cells and hepatocytes is upregulated in response to inflammatory stimulation. In a low-dose pathogen infection model, the recruitment of liver immune cells in C5aR1- mice at the early stage of sepsis was reduced, thereby alleviating the direct damage of the C5a-C5aR1 axis on the liver. However, an increase in Interferons γ (IFNγ)/IL6 ratio was observed after 24 h, accompanied by excretory-induced liver dysfunction. Consequently, C5aR1 exerts different hepatic regulatory effects at different stages of sepsis.³⁴ Additionally, C5aR1 exacerbates organ damage by regulating neutrophil extracellular traps (NETs) during sepsis. Knockout of C5aR1 can reduce the peptidyl-arginine deiminase 4 expression in CLP rats, reduce the formation of NETs, inhibit the TLR2/4 pathway and inflammatory factors such as TNF-α and IL-6, and reduce ALT and AST levels.³³

Additionally, complement can directly damage liver cells. Because the MAC of the complement system (C5b9) can bind to target cells, resulting in increased cell membrane permeability, cell swelling, NLRP3 inflammasome activation, and ultimately hepatocyte necrosis through the lectin pathway, sublytic levels of C5b9 can disrupt mitochondrial electron transport and affect mitochondrial function. When mitochondria undergo excessive fission, many oxidative stress substances are produced, aggravating liver damage.^35,36 In the pathological process of sepsis, excessive activation of the complement system aggravates coagulation dysfunction. C5b9 can induce tissue factor expression in endothelial cells and directly activate the exogenous coagulation pathway, expose phosphatidylserine on the surface of the platelet membrane, accelerate the assembly of prothrombin complex, and promote the coagulation cascade, leading to microcirculation disorders, extensive formation of microthrombi, and disseminated intravascular coagulation.³⁷ Studies have found that the microcirculation hemodynamics and inflammatory response disturbance jointly trigger the increase of liver vascular permeability and abnormal coagulation function, leading to hepatic sinusoidal blood perfusion insufficiency. Sepsis can activate hepatic sinusoidal endothelial cells (LSECs) to abnormally express adhesionmolecules and promote fibrin deposition, aggravating microvascular obstruction, eventually manifesting as microcirculation failure and bleeding tendency, and encourages the development of sepsis-induced liver injury.³⁸

The relationship between complement proteins and the liver is complex. First, most complement proteins in the blood are mainly produced by the liver.³⁹ While the liver synthesizes complement components, it can also be targeted by complement, especially in pathological states such as alcoholic fatty liver disease, hepatitis, and liver fibrosis.⁴⁰ However, when the liver is damaged, the synthesis of complement proteins is insufficient, and CRIg, a highly expressed complement receptor in the liver-resident Kupffer cells, decreases, which cannot effectively remove pathogens from the circulation and leads to the death of the host.⁴¹ Consequently, the complement system interacts with sepsis-induced liver injury. The complement system is a component of the innate immune system. Excessive activation of the complement system leads to pathological damage to liver tissue, and liver dysfunction, in turn, affects the synthesis of complement factors and complement receptors to further disrupt the complement homeostasis.

Regulation of Adaptive Immunity

T Cells

In sepsis, innate immune cells become overactivated and are regulated by adaptive immunity. In the immunosuppression stage, the apoptosis of immune cells, such as T cells, B cells, and DC, and the impairment of immune function can be observed in both patients with sepsis and septic animal models. At this time, programmed death ligand 1 (PDL1) expression on the surface of antigen-presenting cells (APCs) increases, which can bind to the inhibitory programmed death protein receptor (PD1) on the T cell surface to further inhibit the T cell function.⁴² LSECs are specialized antigen-presenting cells in the liver. When the PD1 of CD8 T cells binds to the PDL1 expressed by LSECs, it leads to the reduction of T lymphocyte receptor-mediated cytokine secretion and immune tolerance.⁴³ When sepsis occurs, PDL1 expression in LSECs increases, and its binding to PD1 on the surface of Kupffer cells reduces normal endothelial function, which is vital for preventing liver injury during sepsis.⁴⁴ Moreover, simultaneous blockade of IL-6 and PD-1 in a mouse model of sepsis reduced T-lymphocyte apoptosis in the spleen and neutrophil infiltration in the liver and lungs. Consequently, PD-1 can be used as a target for sepsis immunotherapy in the future.⁴⁵

In addition to inhibiting T cell proliferation through the PD-1/PDL1 pathway, T cell immunoglobulin ITIM domain (TIGIT) can be used as a new immune target. According to studies, TIGIT is highly expressed in liver, kidney, and lung tissues, and TIGIT expression on T cells and NK cells is also increased in sepsis. When we knock out TIGIT or block TIGIT, CD4+T cells proliferate, so we believe that TIGIT plays an anti-infection role mainly through CD4+T cells.⁴⁶ In the early stages of sepsis, T cell immunoglobulin mucin domain containing 3 (Tim-3) is highly expressed on the surface of CD8+T cells. Blocking Tim-3 enhances the function of helper T lymphocytes, reduces the number of lymphocytes undergoing apoptosis, and reduces excessive inflammatory responses to improve liver injury. However, the effect of blocking the Tim-3 signaling pathway using Tim-3 antibody in the late stage of sepsis is unclear.⁴⁷

Additionally, sepsis can lead to an increase in Th17 subsets and a decrease in regulatory T cell (Treg) subsets in peripheral blood. Differentiation of Th17 and Treg subsets leads to an increase in TNF-α, IL-6, and IL-17 inflammatory cytokine infiltration in liver tissue. MicroRNAs play a vital role in immune regulation. miR-126 can change the Th17/Treg ratio and reduce damage to multiple organs, such as the liver, lung, and kidney.⁴⁸ Similarly, selenbinding protein 1 (SELENBP1) was elevated in the livers of septic mice and in human blood samples. SELENBP1 regulates the immune function of DC through reactive oxygen species (ROS). It can increase the Treg/Th17 ratio to alleviate liver injury.⁴⁹

B Cells

B cells are important in adaptive immunity. When sepsis occurs, the number of B cells is reduced, and the ability to secrete antibodies decreases. In human sepsis experiments, the number of circulating memory and regulatory B cells (Breg cells) is reduced. This is because the decrease in antigen-presenting T cells in secondary lymphoid organs of patients with sepsis impairs B cell maturation, and B cell apoptosis is increased.⁵⁰ Additionally, the function of B cells changes in sepsis. B1a cells are the major subset of immature B1 cells, characterized by the secretion of natural IgM, which is produced to resist pathogens in the event of infection.⁵⁰ However, as the severity of sepsis increases, the expression level of IgM decreases and is negatively correlated with the prognosis.⁵¹

The research mechanism of B cells in septic liver is minimal. It has been reported that the inhibition of B cell function may further aggravate the immune dysregulation of sepsis and promote the occurrence of liver injury. Activation of the CD47-SIRPα signaling pathway in sepsis induces myeloid cells to produce amyloid-β (Aβ), inhibiting B cell function by binding to CD74 on the surface of B cells. This inhibition not only affects adaptive immunity but also may lead to a decrease in the immune defense capacity of organs, such as the liver, making them more susceptible to bacterial invasion and inflammatory damage.⁵² α-Galactosylceramide (α-GalCer) is an agonist of invariant NK T cells, inducing B cells to differentiate into IL-10-producing regulatory B cells. These Breg cells not only secrete the anti-inflammatory cytokine IL-10 to inhibit the excessive inflammatory responses, but also regulate neutrophil function, T, NK, and NKT cells, and reduce the proinflammatory cytokine production, such as TNF-α and IL-1β, thereby alleviating sepsis-induced liver injury.⁵³ Consequently, B cells play a dual role in sepsis-induced liver injury.

Kupffer Cells

Kupffer cells are liver macrophages that play a key role in regulating the immune response to sepsis. They mainly play a phagocytic role in the hepatic sinusoid to remove bacteria from the bloodstream.⁵⁴ Under the LPS stimulation, Kupffer cells secrete TNF-α, upregulate intercellular adhesion molecule-1 (ICAM-1) in LSECs, and increase fenestration (reduction in the number of fenestrations and space formation) of hepatic sinusoidal cells, leading to liver injury.⁵⁵ Shimizu et al illustrated that eCIRP could activate the TLR4 pathway to induce M1 polarization of Kupffer cells in sepsis, and excessive production of inflammatory factors, such as IL-6 and TNFα, resulted in liver tissue damage.⁵⁶ Peroxisome proliferator-activated receptor γ (PPARγ) is an essential member of the PPAR family, participating in various signaling pathways to affect macrophage polarization. One study concluded that PPARγ agonist pioglitazone could alleviate liver injury by promoting Kupffer cells polarization to the M2 phenotype through the PPARγ/IκBα/NF-κB pathway.⁵⁷

Ferroptosis is a type of programmed cell death caused by excessive accumulation of iron-dependent lipid peroxides and ROS, which is closely related to sepsis-induced liver injury.⁵⁸ We found that coptisine reduced liver injury by inhibiting ferroptosis and inflammatory response in Kupffer cells through the STAT1/IRF1/GPX4 pathway in septic mice.⁵⁹ Hepcidin, produced by hepatocytes, is a major regulator of systemic iron metabolism. Hepcidin deficiency alters the morphology of Kupffer cells and impairs their ability to phagocytose bacteria. Transplantation of intestinal flora or supplementation with indole-3-propionic acid, a tryptophan derivative derived from intestinal commensal bacteria, can maintain the immune function of Kupffer cells and prevent bacterial bloodstream infection in mice.⁶⁰ We also found that LPS promoted the activation of STING signaling in Kupffer cells, initiating a proinflammatory response and inducing liver injury. Activation of this signaling pathway is dependent on dynamin-related protein 1 of Kupffer cells to cause QC and release mtDNA.²⁹ This is the immunological feature of the septic liver.

Biological Characteristics of NK Cells

NK cells are lymphocytes that can directly kill virus-infected cells.⁶¹ The origin, development, and differentiation of NK cells are complex processes. NK cells initially originate from hematopoietic stem cells, differentiate into NK precursor cells (NKP) through common lymphoid progenitor cells (CLPs), and eventually develop into mature NK cells.⁶² Human NK cells form CD56+CD16-NK cells in secondary lymphoid organs, which mainly secrete cytokines. They can further differentiate into CD56-CD16+ NK cells, accounting for approximately 90% of peripheral blood NK cells, and have a high cytotoxic effect to kill target cells directly.⁶³ In addition to judging the function of NK cells by recognizing different surface markers, NK cells can activate or inhibit signaling pathways through the interaction of their surface receptors with ligands on target cells to exert cytotoxicity, produce cytokines, and regulate the immune response. The activating receptors of NK cells include NCRs (NKp46, NKp44, and NKp30), DNAM-1, and NKG2D. NCRs combine with ligands, such as nuclear protein, galectin-3, and platelet-derived growth factor, to exert the immunogenic anti-tumor effects of NK cells.⁶⁴ NKG2D recognizes the stress ligands MHCI polypeptide-related sequence A and MHCI polypeptide-related sequence B, which are often overexpressed in stressed or malignant tissues. When they bind, they cause NK cells to secrete cytokines, such as IFN-γ, and degranulation of cytotoxic granules containing perforin and granzyme, which are involved in the destruction of target cells, leading to apoptosis.⁶⁵ DNAM-1 binds to CD155 ligands overexpressed on tumor cells to elicit an immune response.⁶⁶ Inhibitory receptors of NK cells include killer immunoglobulin-like receptor (KIR), lymphocyte activation gene 3, TIGIT, Tim-3, and CD94/NKG2A heterodimers.⁶⁷ KIR expressed on the surface of NK cells are highly polymorphic. It regulates the NK cell activity by interacting with primary histocompatibility complex class I molecules expressed on target cells. KIR has been reported to exist in both centromeres and telomeres. The telomere KIR genotype can increase the susceptibility of the host to bacteria, but the centromere KIR gene sequence can play an antiviral role when adaptive immunity is impaired.⁶⁸

Chemokines and their receptors regulate NK cell regulation in the liver. CXCL9 is a chemokine mainly produced by macrophages, endothelial cells, hepatocytes, and tumor cells. It primarily binds to CXCR3 on the surfaces of T and NK cells. In intrahepatic cholangiocarcinoma, endogenous CXCL9 recruits NK cells to infiltrate the intrahepatic bile ducts through the CXCR3 axis.⁶⁹ During hepatitis B virus infection, CXCL10 is secreted by hepatocytes to recruit NK cells expressing CXCR3 to the inflammatory site in the liver, playing a role in killing virus-infected hepatocytes and regulating immune function.⁷⁰ In a mouse model infected with hepatitis virus strain type 3 (MHV-3), the ligands of CC chemokine receptor 5 (CCR5), MIP-1α, MIP-1β, and RANTES, were increased in the liver, thereby attracting CCR5+ NK cells into the liver and aggravating liver injury.⁷¹ CXCL16, the ligand of CXCR6, is expressed on LSECs, hepatocytes, and cholangiocytes, homing CXCR6+ NK cells to the liver.⁷² Additionally, CD58 and CD54 adhesion molecules of hepatic endothelial cells can recruit NK cells to the liver.⁷³

Some NK cells enter the tissues from the peripheral circulation system to reside in tissues, such as the liver, uterus, salivary glands, and skin, and form tissue-specific or tissue-resident NK cells. Their phenotypes and functions differ from those of circulating NK cells (cNKs) in other tissues. Compared with cNK, LrNK has a stronger ability to secrete TNF-α, IL-2, and granulocyte-macrophage colony-stimulating factor. They are more likely to express granzyme C, whereas perforin expression is reduced.⁷⁴ The transcription of hepatic NK cells is regulated by various transcription factors, regulating their activation, maturation, and effector functions. For example, retinoic acid receptor-associated orphan nuclear receptor α (RORα) regulates the activation and proliferation of liver NK cells. When used as an activator of RORα, it can effectively limit liver cancer metastasis.⁷⁵ Promyelocytic leukemia zinc finger protein (PLZF) is a key transcription factor that regulates NK cell development and tissue residence. PLZF was highly expressed in human hepatic CD56+ NK cells and co-expressed with tissue resident markers CXCR6 and CD69, indicating that PLZF may promote the liver and NK cell localization. Notably, there is a small subset of PLZFhiCD56+NK cells in the peripheral blood, which also expresses CXCR6 and CD69 and exhibits functional characteristics similar to LrNK, suggesting that these cells may be a precursor or transitional population of LrNK.⁷⁶ ID2, a basic helix-loop-helix transcription factorthat inhibits NK cell maturation, is also essential in hepatic NK cells. T-cell factor-1 (TCF1) expression is increased in Id2-deficient NK cells, affecting NK cell function by blocking NK cell maturation and reducing IFN-γ production.⁷⁷ Aryl hydrocarbon receptor (AhR) promotes NK cell differentiation in the liver. The use of AhR agonist 6-formyl indole [3,2-B] carbazole increased the cytotoxic activity of NK cells in the liver and promoted the NK cell maturation-related forkhead box O1 expression.⁷⁸ Mature T-box (T-bet) expressed in T cells is a transcription factor essential for the cytotoxic function of mature NK cells. T-bet expression in hepatocytes and the immune environment can influence the NK cell responses during infection and malignancy, thereby enhancing their anti-tumor activity in the liver.⁷⁹ Overall, the liver is a unique immune microenvironment in which different transcription factors, such as RORα, PLZF, ID2, AhR, and T-bet, regulate the complex mechanisms of hepatic NK cells. These transcription factors act synergistically to determine NK cell development, maturation, and function.

NK cells, as innate immune cells, can also regulate adaptive immunity. According to studies, the high PDL1+NK cell expression in sepsis leads to immune dysfunction, which can be used as a risk factor for the prognosis of sepsis.⁸⁰ Liver-resident NK cells (LrNK/trNK/ILC1) highly express immunosuppressive molecules such as PDL1, TRAIL, and CD39 during viral infection, which inhibit the proliferation and function of virus-specific T cells in the liver through the PD-L1/PD-1 signaling pathway, ultimately delaying viral clearance.⁷⁴ NK cells can also secrete IFN-γ to promote the differentiation of CD4+ T cells into helper T cell subsets and regulate the directional migration of effector CD8+ T cells to the site of viral infection. More importantly, NK cells can indirectly regulate T cell immune response by regulating the functional state of LSECs.⁸¹ NK cells in the liver can be activated by cytokines, such as IL-12 and IL-18, produced by Kupffer cells to produce IFN-γ, exerting cytotoxic effects and causing liver cell damage.⁸²

The Role of NK Cells in the Septic Liver

The mechanism of sepsis-induced liver injury is complex, mainly involving an inflammatory cascade reaction, oxidative stress, mitochondrial dysfunction, microcirculation dysfunction, ischemia-reperfusion, pyroptosis, and ferroptosis.^5,83 NK cells, as innate immune cells, are a double-edged sword in the septic liver, as they can either protect or damage the liver tissue Figure 2.

Figure 2.

Dual role of natural killer cells in the liver. In sepsis, natural killer (NK) cells exert both protective and detrimental effects on the liver. Protective effects: (1) Overexpression of SRG3 reduces IFN-γ production by NK cells and promotes polarization of Kupffer cells toward the M2 phenotype, leading to increased IL-10 secretion, which suppresses NK-cell activation and confers hepatoprotection; (2) NK cells down-regulate hepatic VEGF expression and increase CD31⁺ endothelial cells, thereby enhancing liver angiogenesis and maintaining vascular integrity; (3) The phenotypic characteristics of NK cells are associated with the control of hepatic pathogens—elevated levels of CD56⁻CD16⁻NK cells correlate with reduced SIV viral load in the liver and plasma, whereas invariant IrNK cells mediate CD2-dependent mechanisms to decrease HCMV viral load in plasma, contributing to liver protection. Detrimental effects: (1) Lipopolysaccharide (LPS) stimulates Kupffer-cell release of IL-12 and IL-18, resulting in excessive NK-cell activation and overproduction of IFN-γ, which amplifies the inflammatory cascade and induces hepatocyte apoptosis; (2) NK cells compromise hepatic vascular endothelial integrity through secretion of TNF-α and IL-6; (3) NK cells also promote ferroptosis, thereby exacerbating liver injury.

Protective Effect of NK Cells on the Liver in Sepsis

Changes in NK cells during sepsis can significantly affect the liver’s response to inflammation, including cytokine production, NK cell number, and interactions with other immune cells. It is well known that NK cells are an essential source of a variety of cytokines, producing IFN-γ, TNF-α, IL-10, and other immunomodulatory factors, and can also secrete many chemokines and growth factors. Among them, NK cells are the main source of IFN-γ and have proinflammatory and antibacterial effects, which are the first line of defense against bacterial infections.⁸⁴ It has been found that the number of NK cells is reduced in patients with sepsis, which mainly reflects the reduction of CD56-NK cells in the peripheral blood, rather than CD56+NK cells. This indicates that the role of NK cells in killing target cells is weakened, contributing to the maintenance of liver homeostasis.⁸⁵ Mating-type switching/sucrose nonfermentation (SWI/SNF) is an ATP-dependent chromatin remodeling complex, and the SWI3-related gene (SRG3) is a core component of various SWI/SNF complexes. Compared with wild-type mice, the number of NK cells in the spleen and liver of septic mice with SRG3 overexpression driven by the β-actin promoter was decreased, and IFN-γ was also reduced, which finally improved the survival rate of septic mice. Additionally, mice can promote the M2 polarization of liver Kupffer cells, produce a large amount of IL-10 to inhibit the activation of liver NK cells, reduce the production of inflammatory factors, and ultimately protect the host and liver.⁸⁶ It has been reported that IL-18, in conjunction with IL-12, can effectively induce IFN-γ production by NK cells and T cells, which is critical for host defense against bacterial infection. In septic mice, the ability of NK cells in the liver and spleen to produce IFN-γ is significantly reduced because of immune dysfunction. IL-10 is closely related to the severity of sepsis, and antagonism of the IL-10 receptor can restore the sepsis-mediated IL-18/IFN-γ-mediated immune system. This is because it can restore IL-18R expression in liver NK cells of septic mice, significantly improve their IFN-γ production ability, and improve the survival rate of septic mice, confirming that IL-10 is a key factor regulating NK cell function.⁸⁷ Contrary to the literature, when IL-12/IL-18 was used to stimulate septic mice, the secretion of IFN-γ by cNKS remained unaffected. Interestingly, we found that the number of ILC1 decreased, but the ability of ILC1 to secrete IFN-γ increased, which activated macrophages and other immune cells to mount immune responses. This proves that ILC1 is beneficial for pathogen clearance during sepsis and enhances the local immune response to protect the liver.⁸⁸

The phenotype and function of NK cells are related to pathogen control. According to Woolley, NK cell depletion can significantly increase the viral load of acute simiovirus immunodeficiency virus (SIV) in the early stage of viremia, indicating that NK cells play a key role in controlling early viral replication.⁸⁹ During chronic SIV infection in rhesus monkeys, the number of CD56-CD16-NK cells is negatively correlated with viral vectors in the liver and plasma, inhibiting viral transmission mainly through the cytotoxicity and immunomodulatory effects of NK cells. Additionally, CXCR3+ and NK2D+ NK cells enriched in the liver and circulatory system could migrate to the site of viral infection or directly kill infected cells. However, despite the presence of memory-related FcεRIγ defects and CD57+ NK cells in rhesus monkeys, these play a non-dominant role in SIV infection.^90,91 Similarly, phenotypic changes in NK cells can be observed in human cytomegalovirus (HCMV)-infected viremia. HCMV-positive liver transplant recipients can induce adaptive phenotypes of circulating and liver-resident NK cells. cNK cells exhibited upregulation of CD57, NKG2C, and CD2, and downregulation of FcεRIγ, Siglec-7, and NKG2A, whereas lrNK cells expressing CD2 and NKG2C were significantly increased. Simultaneously, the expression frequency of inhibitory receptors, such as NKG2A and Siglec-7, decreased. The inhibition of HCMV by lrNK was significantly stronger than that of cNK, and its mechanism of action was mainly mediated by the CD2 molecule of lrNK rather than NKG2C to enhance the antiviral effect, reduce HCMV viral load in serum, and reduce liver damage.⁹²

Vascular endothelial growth factor (VEGF) is not only a vascular permeability factor⁹³ but also a pro-angiogenic factor. When VEGF binds to its specific receptor, VEGFR2, it activates NADPH oxidase to directly induce ROS generation to promote proliferation, migration, and differentiation of endothelial cells. Finally, the process of promoting angiogenesis is completed.⁹⁴ The stability of blood vessels is mainly to protect the endothelial barrier to maintain the balance of the internal and external environment of blood vessels, whereas VEGF can cause endothelial barrier dysfunction and increase vascular permeability.⁹⁵ VEGF upregulation in the plasma of patients with sepsis is associated with the severity of multiple organ dysfunction and vascular leakage.^96,97 Despite the downregulation of hepatic VEGFA after NK cell infusion in septic mice, the increase in hepatic CD31-positive mature endothelial cells promotes hepatic angiogenesis, which, together with the maintenance of vascular endothelial stability, protects the liver.⁹⁸ Consequently, the attenuation of vascular permeability mediated by NK cells may involve crosstalk between NK cells and the VEGFA signaling pathway, although the precise underlying mechanism remains elusive. This protective effect appears to be mediated through the modulation of sphingosine-1-phosphate (S1P) signaling, resulting in reduced plasma VEGFA levels, given that decreased VEGFA attenuates hepatic and pulmonary vascular leakage and improves prognostic outcomes in murine models of sepsis.⁹⁹ Conversely, endothelial cell-derived VEGF, upregulated via mitogen-activated protein kinase (MAPK)-dependent pathways, exacerbates microvascular permeability and capillary leakage, ultimately precipitating multiple organ dysfunction.¹⁰⁰ NK cells can also reduce endothelial cell activation by reducing NO and jointly protect target organs.⁹⁸

Role of NK Cells in Liver Damage in Sepsis

Although IFN-γ produced by NK cells can play an anti-inflammatory role in sepsis, it can also cooperate with other cytokines to produce an inflammatory factor storm, leading to organ dysfunction.¹⁰¹ NK cells play a key role in the regulation of liver inflammation during sepsis. The number of liver NK cells in the sepsis mouse model increases, promoting the release of IFN-γ and TNF-α to amplify the inflammatory response and induce liver cell apoptosis.⁵ Kupffer cells are activated to secrete cytokines such as IL-12 and IL-18 under the stimulation of bacterial LPS, which in turn activate liver NK cells and NK1.1 Ag+ T cells to produce IFN-γ and enhance their cytotoxicity against tumors and pathogenic microorganisms. It is the first line of defense against bacterial infections and tumor metastasis. However, insufficient activation of this defense system may lead to multi-organ failure, whereas over-activated NK cells and NK1.1 Ag+ T cells may also damage liver cells.⁸² Additionally, the inflammatory system of NK cells can be activated by the TLR stimulation agonist lipopolysaccharide, which secretes a large amount of IL-6 and TNF-α and destroys the integrity of the liver vascular barrier. However, such studies are limited to in vitro models, and in vivo experiments need to be studied in the future.¹⁰² NK cells exert their effects through two main cytotoxic pathways: First is the endogenous cell death pathway, in which perforin secreted by NK cells forms a channel in the target cell membrane, allowing granzyme B to enter the cell and trigger a cascade of reactions such as cytochrome c release, ultimately inducing target cell apoptosis. The second is the exogenous pathway, in which death ligands, such as FasL and TRAIL on the surface of NK cells, bind to their corresponding receptors Fas and TRAIL-R1/R2 on the target cells to activate effector molecules, such as caspases (for example, caspase-3/7/8/9) and ADP-ribose polymerase. It also induces caspase-dependent apoptosis.¹⁰³ Excessive accumulation of cytotoxic NK cells in septic mice leads to an immune system imbalance and eventually liver dysfunction.¹⁰⁴ Many studies have found that ferroptosis can lead to liver injury in patients with sepsis. Qingli et al conducted ferroptosis-related gene screening and immune correlation analysis on septic liver failure. They discovered that Hmox1, Slc3a2, Jun, and Zfp36 ferroptosis-related genes may promote the development of liver failure through NK cells. Among them, NK cells have the strongest correlation with the stress protein-coding gene Hmox1.¹⁰⁵

Immunotherapy of NK in Sepsis

Adoptive NK cell therapy includes autologous NK cell infusion and allogeneic NK cell infusion,¹⁰⁶ Autologous NK cell infusion involves harvesting peripheral blood mononuclear cells (PBMCs) from the patient, followed by ex vivo expansion and activation, with subsequent reinfusion into the same patient. However, clinical evidence supporting autologous NK cell therapy for sepsis in humans remains virtually nonexistent. Instead, preclinical animal studies and isolated case reports have predominantly utilized NK cells derived from healthy donors or umbilical cord blood. This limitation largely stems from the fact that NK cells in septic patients typically exhibit an exhaustion-like phenotype, characterized by attenuated cytotoxic capacity and impaired cytokine secretion, thereby inherently constraining the therapeutic efficacy of autologous approaches.¹⁰⁷ Consequently, this modality currently remains in a nascent, predominantly theoretical stage for the treatment of human sepsis. Allogeneic NK cell infusion derived from umbilical cord blood (UCB)¹⁰⁸ and induced pluripotent stem cells (iPSCs).¹⁰⁹ A study on LPS-induced septic mice reported that infusion of cord blood-derived NK cells reduced the anti-inflammatory factor IL-10 and improved the survival rate of septic mice.⁹⁸ Additionally, a case report of sepsis secondary to lung cancer indicated that allogeneic NK cell infusion inhibited the production of IL-6 in both patients and in vitro experiments, and it helped enhance the function of NK cells in patients to control infection. However, this therapeutic effect is only short-lived because of the rapid loss of NK cell number and function within 48 h.¹¹⁰ Pretreatment with cytokines (such as IL-2 and IL-15) to enhance NK cytotoxicity and survival is usually performed before allogeneic NK cell infusion.¹¹¹ IL-2 is currently the only cytokine approved for medical use, but its specific and highly toxic effects lead to vascular leakage.¹¹² We used IL-2/α-IL-2 monoclonal antibody complex (IL-2c) to preserve the potency of IL-2 while reducing its harmful effects. Among them, IL-2c (IL-2 bound by S4B6) can expand the number of NK cells to control the MCMV infection of septic mice.¹¹³ As a member of the IL-2 family, IL-15 exhibits strong anti-apoptotic activity. It has been reported that IL-15 can prevent lethal cell apoptosis in vivo and prevent lymphocyte apoptosis.¹¹⁴ IL-15 can block NK cell depletion in sepsis and increase IFN-γ-positive NK cells to regulate cell death and immunosuppression.¹¹⁵ Similarly, recombinant IL-15 is a newly discovered immunotherapy drug, produced by genetic engineering of IL-15. When rat recombinant IL-15 was injected into septic rats, the number of NK cells and the IFN-γ expression increased in a dose-dependent manner.¹¹⁶ IFN-γ is mainly produced by T and NK cells. It has been determined that the new non-steroidal small molecule inhibitors pentamidine and ginanofin can improve the survival rate of septic mice by inhibiting the IFN-γ-induced NOS2 expression and catalyzing NO production.¹¹⁷

When activated, liver allograft-derived NK cells are infused into liver transplant recipients, the Fcγ receptors (FcγRs) of NK cells trigger an antibody-dependent cytotoxic effect (ADCC) to release cytotoxins and reduce postoperative bloodstream bacterial infection in liver transplant recipients.¹¹⁸ When we performed ex vivo expansion before NK cell infusion, we found that CCR5 upregulation led to the homing of infused NK cells to the liver. Drugs that promote CCR5 can be studied in the future to reduce liver injury caused by sepsis.¹¹⁹ Exosomes are extracellular vesicles secreted by all cells and carry proteins, lipids, and nucleic acids to facilitate intercellular communication.¹²⁰ Exosomes derived from neutrophils, macrophages, lymphocytes, NK cells, endothelial cells, and platelets can enhance the proliferation and migration of T lymphocytes, inhibit inflammatory responses, and reduce lung and liver tissue damage. Exosomes derived from NK are also a promising new direction in the field of immunotherapy.¹²¹

As an S1P receptor and an agonist of S1P, FTY720 can regulate NK cell migration.¹²² Recently, it was found that S1PR5 mainly regulates the distribution of NK cells in vivo and the transport of NK cells to the lesion site, and it accumulates in the liver.¹²³ S1PR5 expression is increased in patients with sepsis and improves their survival rate.¹²⁴ Consequently, we hypothesized that sinimod, stimulating S1PR5, could be used as a treatment for sepsis-associated liver injury, and its main mechanism involves S1PR5 expression, which is promoted to drive NK cell accumulation in the liver and attenuate immunosuppression and inflammatory response.

Conclusion

Sepsis is a life-threatening organ dysfunction caused by a dysregulated host response to infection. The liver, serving as a critical hub for immune and metabolic functions, undergoes hepatic failure during sepsis through various pathophysiological mechanisms, including aberrant activation of pattern recognition receptors, dysregulation of the complement system (C3a/C5a), functional disturbances in adaptive immunity (T/B lymphocytes), and polarization imbalance of Kupffer cells.

NK cells, as essential components of the immune system, exert bidirectional regulatory effects on sepsis-induced liver injury. NK cells protect the liver through multiple mechanisms: First, as a primary source of IFN-γ, NK cells not only possess direct antimicrobial activity but also modulate hepatic immune responses. Second, NK cells maintain vascular endothelial barrier stability and attenuate vascular leakage by downregulating VEGF expression. Third, NK cells secrete anti-inflammatory cytokines such as IL-10 to suppress excessive inflammatory reactions in the liver. Fourth, through cytotoxic effects and phenotypic switching, NK cells effectively inhibit viral replication within the liver. However, NK cells can also exacerbate hepatic injury through distinct pathways: Initially, they release substantial amounts of TNF-α and IFN-γ, synergistically forming a cytokine storm that amplifies hepatic inflammation. Subsequently, NK cells directly induce hepatocyte apoptosis via endogenous pathways mediated by perforin-granzyme and exogenous pathways involving FasL/TRAIL-death receptor interactions. Finally, NK cells promote the development and progression of sepsis-associated liver failure by regulating the expression of ferroptosis-related genes, including Hmox1 and Slc3a2.

Current immunotherapeutic strategies targeting NK cells primarily encompass adoptive cell transfer (autologous or allogeneic NK cells), cytokine preconditioning (like IL-2/IL-15 and their complexes to enhance cytotoxicity and viability), gene-editing technologies (such as CCR5 upregulation to promote hepatic homing), and the application of NK cell-derived exosomes. Nevertheless, existing therapies confront numerous challenges: infused NK cells exhibit rapid numerical and functional decline within 48 hours, and the high toxicity of IL-2 induces vascular leakage, among other issues requiring resolution. Future research must focus on developing strategies for long-term maintenance of NK cell function, exploring S1PR5 agonists (Siponimod) to promote hepatic NK cell accumulation, optimizing cytokine preconditioning protocols, and thoroughly investigating the therapeutic potential of NK cell exosomes.

In conclusion, the core of sepsis-induced liver injury lies in severe dysregulation of the host immune response, with NK cells serving as critical intermediaries connecting innate and adaptive immunity while balancing inflammatory responses and organ protection. A comprehensive understanding of the dual functional mechanisms of NK cells within the hepatic immune microenvironment will not only elucidate the pathophysiological processes underlying sepsis-related liver injury but also provide a theoretical foundation for developing novel NK cell-targeted immunotherapeutic strategies. Through precise modulation of NK cell function, we anticipate improved clinical outcomes for septic patients.

Funding Statement

The author(s) declare financial support was received for the research, authorship, and/or publication of this article. This study was supported by the Jiangxi Provincial Health Commission Science and Technology project (No. 202410342), The Doctoral Startup Project of the First Affiliated Hospital of Gannan Medical University (No. QD202336), and the Science and technology program of Jiangxi Provincial Administration of Traditional Chinese Medicine (No. 2024B0561).

Data Sharing Statement

Data sharing is not applicable to this article as no data were created or analysed in this study.

Author Contributions

Xiaoyu Wu: Writing – original draft, Conceptualization, Data curation, Writing – review and editing. Jiangfeng Cheng: Writing – original draft, Investigation, Software, Writing – review and editing. Han Yu: Data curation, Formal analysis, Writing – review and editing. Junming Ye: Methodology, Supervision, Writing – review and editing. Hongquan Zhu: Data curation, Supervision, Writing – review and editing. Xiaofeng Liu: Methodology, Supervision, Writing – review and editing. All authors gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work. Xiaoyu Wu and Jiangfeng Cheng should be considered co-first authors.

Disclosure

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.