|Home | About | Journals | Submit | Contact Us | Français|
Adaptive immunity is critical for disease progression and modulates T cell (TC) and antigen-presenting cell (APC) functions. Three signals were initially proposed for adaptive immune activation: signal 1 antigen recognition, signal 2 co-stimulation or co-inhibition, and signal 3 cytokine stimulation. In this article, we propose to term signal 2 as an immune checkpoint, which describes interactions of paired molecules leading to stimulation (stimulatory immune checkpoint) or inhibition (inhibitory immune checkpoint) of an immune response. We classify immune checkpoint into two categories: one-way immune checkpoint for forward signaling towards TC only, and two-way immune checkpoint for both forward and reverse signaling towards TC and APC, respectively. Recently, we and others provided evidence suggesting that metabolic risk factors (RF) activate innate and adaptive immunity, involving the induction of immune checkpoint molecules. We summarize these findings and suggest a novel theory, metabolism-associated danger signal (MADS) recognition, by which metabolic RF activate innate and adaptive immunity. We emphasize that MADS activates the reverse immune checkpoint which leads to APC inflammation in innate and adaptive immunity. Our recent evidence is shown that metabolic RF, such as uremic toxin or hyperhomocysteinemia, induced immune checkpoint molecule CD40 expression in monocytes (MC) and elevated serum soluble CD40 ligand (sCD40L) resulting in CD40+ MC differentiation. We propose that CD40+ MC is a novel pro-inflammatory MC subset and a reliable biomarker for chronic kidney disease severity. We summarize that CD40:CD40L immune checkpoint can induce TC and APC activation via forward stimulatory, reverse stimulatory, and TC contact-independent immune checkpoints. Finally, we modeled metabolic RF-induced two-way stimulatory immune checkpoint amplification and discussed potential signaling pathways including AP-1, NF-κB, NFAT, STAT, and DNA methylation and their contribution to systemic and tissue inflammation.
The online version of this article (doi:10.1186/s13045-017-0504-1) contains supplementary material, which is available to authorized users.
The immune system consists of innate and adaptive immunity. The classical innate immune system provides immediate and non-specific defense. It is activated by pathogens via a pathogen-associated molecular pattern (PAMP), which is recognized by pattern recognition receptors (PRR) in phagocytes (Fig. 1). Innate immunity may also be activated in response to injury, which releases a danger-associated molecular pattern (DAMP) also recognized by PRR. These two pathways are summarized as PAMP/DAMP+PRR recognition which leads to pathogen elimination, inflammatory responses, and antigen-presenting cell (APC) formation . Evidence also suggests that the innate immune system targets innate T cells (TC) leading to TC activation [2, 3].
Different from innate immunity, adaptive immunity is featured by antigen (Ag) specificity, slow response, immunologic memorization, and low responsive cell ratio (Additional file 1: Table S1) . Adaptive immunity comprises of cell-mediated immunity using TC and B cell (BC) humoral immunity. Each type of adaptive immunity contains three activating signals: (1) Ag recognition, (2) co-stimulation (we termed as immune checkpoint in this article), and (3) cytokine stimulation (Fig. 2). The term of immune checkpoint was initially proposed in 2009 referring to co-inhibitory immune checkpoint for TC suppression [5, 6] and was expanded in 2012 to include co-stimulatory immune checkpoint for TC activation . The concept of immune checkpoint has been extensively studied in recent years and summarized in Table 1. It has become evident that the immune checkpoint plays an important regulatory role in adaptive immunity and determines the fate of the immune cell towards activation or suppression.
Increased knowledge in immune checkpoints established advances in cancer medicine. For example, immune checkpoint molecule cytotoxic T lymphocyte-associated protein 4 (CTLA-4)-immunoglobulin (Ig) competes with CD28 to bind to CD80/CD86 and causes CTLA-4:CD80/CD86-induced TC suppression . Antibodies against immune checkpoints, CTLA-4 (ipilimumab) and programmed cell death protein 1 (PD-1) (pembrolizumab and nivolumab), block CTLA-4:B7 and PD-1:PD-L1-induced TC suppression and thus enhance TC-dependent immune reaction [9–11]. These antibodies resulted in clinical regression of melanoma, non-small cell lung cancer, and other cancers [9–11]. Immune checkpoint therapy has also proven beneficial for inflammatory diseases such as rheumatoid arthritis and psoriasis using strategies to alleviate inflammation by engaging the inhibitory immune checkpoint [12, 13]. Immune checkpoint therapy for metabolic disease has not yet been realized, but it is an important consideration to balance TC responses and modulate immune checkpoints in contemplating therapies for metabolic disease.
The initial definition of immune checkpoints refers to receptor:ligand reaction towards TC suppression, also referred as co-inhibitory immune checkpoint. The immune checkpoint concept gradually evolved to incorporate co-stimulatory immune checkpoint and the identification of a reverse function of immune checkpoint towards APC [7, 14]. Recent evidence also suggests that metabolic risk factors (RF) can activate the stimulatory immune checkpoint leading to APC-related inflammatory responses [15–19].
We propose a novel metabolism-associated danger signal (MADS) recognition, which promotes reverse stimulatory immune checkpoint leading to APC inflammation in both innate and adaptive immunity systems. MADS refers to intermediates and products of glucose, lipid, amino acid, nucleotide, hormone, and/or chemical metabolism, which can be recognized by the immune system via a metabolic sensor in a receptor-independent fashion.
In this article, we updated the molecular basis regulating innate and adaptive immunity. We proposed two novel nomenclatures, MADS recognition and reverse immune checkpoint, and suggested a new theory that MADS recognition regulates innate and adaptive immune response, via metabolic sensor, leading to immune cell activation and inflammation. Information described in this article should provide systemic knowledge and comprehensive insights into our understanding about immune response and immune checkpoints, especially the reverse stimulatory immune checkpoint in diseases.
The innate immune system is activated by pathogens via PAMP+PRR recognition and by injury-generated molecules via DAMP+PRR recognition (Fig. 1a). PRR are receptors presented on all immune cells and somatic cells, which bind to DAMP and PAMP to initiate inflammation [2, 3, 16, 20, 21]. Phagocytes, including macrophage (MØ), monocyte (MC), dendritic cell (DC), neutrophil, and natural killer (NK) cells, are activated by PAMP/DAMP+PRR recognitions which lead to pathogen elimination and inflammatory responses such as APC formation and cytokine generation . Toll-like receptors (TLR) are a key PRR located on the cell surface and endosomes. Nucleotide binding and oligomerization domain-like receptors (NLR) are another important cytosolic-sensing DAMP receptor. In addition, transmembrane C-type lectin (TmCL), retinoid acid-inducible gene I (RIG-I), absent in melanoma 2 (AIM2), and receptor for advanced glycation end products (RAGE) are also characterized as classical DAMP-sensing receptors .
We and others provided evidence suggesting that metabolic RF activate innate immune systems leading to inflammatory responses. For example, lipid metabolite ox-LDL promoted NLRP3 inflammasome activation in MØ and foam cell formation . Intermediate amino acid homocysteine (Hcy) induced nucleotide-binding oligomerization domain and leucine-rich repeat and pyrin domain containing protein 3 (NLRP3), causing NLRP3-containing inflammasome assembly, caspase-1 activation, and interleukin (IL)-1β cleavage/activation in EC . Glucose, ceramide, islet amyloid polypeptide, and cholesterol crystals can be sensed by TLR or NLRP3-stimulating NLRP3 inflammasome complex assembly [16, 24, 25]. We  and others  demonstrated that MADS, such as Hcy or ox-LDL, induced MC activation in the absence of Ag within 48 h. Our data supported the notion that metabolism sensors mediate metabolic RF-induced inflammatory response in the innate immune system (Fig. 1a). Recently, we identified increased Hcy and a reduced ratio of S-adenosylmethionine (SAM)/S-adenosylhomocysteine (SAH), an indicator of cellular methylation, as the metabolic mediator/sensor for pro-inflammatory MC differentiation caused by uremic toxin in chronic kidney disease (CKD) .
CD4+ or CD8+ TC, including regulatory TC (Treg), express TLR and is directly involved in innate immunity (Fig. 1b). It is reported that PAMP/DAMP-TLR signaling lead to TC proliferation, inflammatory cytokine production, and glycolysis [2, 3]. Some inflammatory cytokines, such as IL-18, IL-12, IL-1β, IL-23, transforming growth factor (TGF)-β, and interferon (IFN)-α, quickly induced TC subset differentiation and proliferation and IFN-γ, IL-17, and IL-4 secretion in an Ag-independent fashion [6, 27, 28]. Super Ag caused non-specific TC activation and cytokine release . In addition, lipid mediators, such as leukotrienes, are important activators for CD4+ and CD8+ TC recruitment to the site of infection and control fungal infection . These evidences support the concept of innate immune response in TC via five mechanisms: PAMP/DAMP+PRR recognition, inflammatory cytokines, super Ag, and MADS recognition (Fig. 1).
The major features of adaptive immunity are Ag specificity and immunologic memory which led to TC and BC activation (Additional file 1: Table S1). It was initially proposed that TC and BC activation involve three signals: signal 1 Ag recognition, signal 2 co-stimulation or co-inhibition, and signal 3 cytokine stimulation (Fig. 2). In this article, we termed signal 2 as the immune checkpoint which is in agreement with Dr. Pardoll’s suggestion in 2012 . We defined immune checkpoint as interactions of paired molecules leading to either stimulatory or inhibitory immune response in TC and BC (other cells as well).
The discovery of TC receptors (TCR) led to defining TC activation signal 1, Ag recognition. Moreover, TC activation signal 2, immune checkpoint, was found to be essential for complete TC activation. For example, CD28 monoclonal antibody administration with simultaneously stimulating TCR leads to complete TC activation . Signal 3, cytokine stimulation, is also involved in TC activation . CD8+ TC’s response to virus was shown to be IFN-α dependent. We proposed a novel signal 4 because metabolic RF,, such as uremic toxin and hyperhomocysteinemia (HHcy), activated CD40:CD40L co-stimulatory immune checkpoint and increased serum soluble CD40L (sCD40L) levels .
Signal 1 (Ag recognition) is a vital immune process and determines the specificity of TC response. Ag is presented by major histocompatibility complexes (MHC) on the surface of an APC, then engaging with Ag-specific TCR on naïve TC contributing to TC activation/proliferation.
Signal 2 (immune checkpoint) plays a key role in regulating TC activation, differentiation, effector function, and deletion. Signal 2 was initially defined as co-stimulation and expanded to include co-inhibitory pathways . In this article, we propose to term the co-stimulatory and co-inhibitory pathways collectively as the immune checkpoint. Immune checkpoint initially described co-inhibitory signal 2 in Topalian et al.’s papers  based on the discovery of T cell function restraint in normal physiologic settings and tumors . This terminology was recently used to describe as a regulatory switch towards either stimulatory or inhibitory pathways . Following Ag recognition or metabolic stimulation, an immune checkpoint ligand on APC binds to its receptor on TC determining TC activation or suppression. For example, CD28:B7 co-stimulatory immune checkpoint is essential for TC expansion and differentiation .
Signal 3 (cytokine stimulation) mediates cytokine-induced TC expansion and differentiation. For example, IL-12 and IFN-α/β, along with Ag and immune checkpoint, enhanced CD8+ TC clonal expansion . The combination of IL-1β and IL-6 induced T helper (Th)-17 cell differentiation from human naïve TC (CD4+CD45RA+CCR7+CD25−), in the presence of anti-CD3 (signal 1) and anti-CD28 (signal 2) antibodies . IL-1β enhanced Th1, Th2, and Th17 cell proliferation with Ag stimulation in IL-1R1−/−Rag1−/− mouse .
Signal 4 (MADS recognition) is a novel signal we proposed based on our and other’s recent findings [15, 26]. Metabolic RF stimulates the expression of immune checkpoint molecules via a metabolic sensor, which in turn activates APC or TC and increases inflammatory cytokine production. We reported that uremic toxin, HHcy, and S-adenosylhomocysteine (SAH) increased CD40+ MC and sCD40L levels during a chronic time frame of CKD patients . CD40:sCD40L molecular pair further promoted pro-inflammatory CD40+ MC and intermediate MC differentiation in 3 days. Moreover, studies in human subjects support that signal 4 MADS recognition may be involved in TC-related adaptive immunity in metabolic disorders . The levels of sCD40L were found to be increased in subjects with metabolic syndrome and hypertension and negatively related to insulin sensitivity . In addition, glucose sustains TC growth and proliferation upon TCR-dependent TC activation .
BC immunity involves the same four signals which leads to antibody production and BC activation . Signal 1 (Ag recognition) is the engagement of Ag with Ag-specific BC receptor (BCR). Signal 2 (immune checkpoint) is the ligation of immune checkpoint molecular pairs. Signal 3 (cytokine stimulation) describes Ag- and immune checkpoint-associated inflammatory cytokine regulation in BC activation. We proposed signal 4 (MADS recognition) for BC activation because the CD40:CD40L immune checkpoint is involved in BC activation  and sCD40L is induced in metabolic disease including CKD, HHcy, hypertension, hyperglycemia, and dyslipidemia [15, 39, 43].
Immune checkpoints are molecular pairs (receptor:ligand) interactions regulating immune response towards TC and APC, also termed signal 2 (Fig. 2). We classified the immune checkpoint into two categories: one-way immune checkpoint for forward signaling towards TC only, and two-way immune checkpoint for both forward and reverse signaling towards TC and APC, respectively (Fig. 3). Each category can be further divided into stimulatory and inhibitory immune checkpoints. The stimulatory immune checkpoint turns up the immune system leading to immune cell proliferation or activation, while the inhibitory immune checkpoint turns down the immune system leading to immune cell suppression or death (Fig. 3a).
One-way immune checkpoint refers to forward signaling only towards TC. It is dual-functional as it can modulate cell fate for proliferation or death (Fig. 3b).
Forward stimulatory molecular pairs promote TC proliferation, cytokine production, differentiation, cytotoxic function, memory formation, and survival. A well-described forward stimulatory molecular pair is CD28:B7. Interaction of CD28:B7 results in distinct phosphorylation, transcriptional activation, and cytokine and chemokine production that are essential for TC expansion and differentiation . Metabolic product ceramide is involved in the forward stimulatory immune checkpoint in TCR-dependent TC activation at multiple levels .
The forward inhibitory molecular pair ligation in the one-way immune checkpoint leads to TC tolerance, exhaustion, apoptosis, cell cycle arrest, and effector function inhibition. For example, CD8+ tumor-infiltrating lymphocytes exhibit high proliferation and IL-2/tumor necrosis factor (TNF)-α production in TC immunoreceptor with Ig and ITIM domain (TIGIT)−/− mice , indicating TIGIT inhibited the effector function and proliferation of CD8+ TC.
The two-way immune checkpoint is bi-directional, towards both TC and APC. Similar as the one-way immune checkpoint, it is also dual-functional as it modulates cell fate for proliferation or death (Fig. 3c).
The stimulatory molecular pairs in the two-way immune checkpoint activate TC and APC. CD40:CD40L is one of the best-described stimulatory pairs in the two-way immune checkpoint. CD40 binds to its ligand CD40L, which is usually transiently expressed on TC  and modulates effector function and differentiation of TC. This is seen in CD40−/−APOE−/− mice as they have lower effector memory CD4+/CD8+ TC in the spleen . Ligation of CD40L on TC with CD40 on BC promoted BC Ig isotype switching, which was associated with X-linked hyper IgM syndrome in humans . Moreover, metabolic RF cholesterol crystal is required for TCR nanoclustering in TC, which enhances the avidity of the TCR-antigen interaction . Reversely, cholesterol crystals trigger pro-inflammatory cytokine secretion from APC MØ .
The inhibitory molecular pairs in the two-way immune checkpoint lead to TC and APC suppression or death. The ligation of PD-1 and PD-L1 results in TC inactivation, IL-12 reduction, antitumor immunity suppression, and tumor progression . Thus, PD-1:PD-L1 immune checkpoint therapy using PD-1 antibodies (pembrolizumab and nivolumab) achieved great success in melanoma, bladder cancer, and gastric cancer therapy [9–11]. Further, PD-1 delivered inhibitory signals through B7-H1 on APC . Again, metabolic RF cholesterol sulfate inhibited TCR signaling  as well as sterologenesis in APC fibroblasts .
Representative paired immune checkpoint molecules (receptor:ligand) are summarized in Table 1 and listed according to immune checkpoint direction (one-way and two-way) and function (stimulatory and inhibitory). The classification of immune checkpoint families is determined by the checkpoint receptor component. Most immune checkpoint receptors are members of immunoglobulin superfamily (IgSF) and tumor necrosis factor receptor superfamily (TNFRSF), which can be further divided into specific subfamilies based on the primary amino acid sequence, protein structure, and function . Notably, the majority of immune checkpoint ligands are expressed on multiple immune cells.
IgSF checkpoint receptor superfamily contains CD28, B7, CD226, TC (or transmembrane) immunoglobulin, mucin domain (TIM), and CD2/signaling lymphocytic activation molecule (SLAM) subfamilies, which participate in forward stimulatory and forward inhibitory immune checkpoints. For example, CD28 subfamily including CD28 and CD278 (inducible TC co-stimulator, ICOS) transduce stimulatory response. Other members in the CD28 subfamily, such as CTLA-4, PD-1, PD-1 homologue (PD-1H), and B and T lymphocyte attenuator (BTLA), transduce inhibitory response.
TNFRSF checkpoint receptor superfamily contains Type-V, Type-L, Type-s, and orphan subfamilies and recognizes TNF superfamily (TNFSF) molecules . The common feature of TNFRSF:TNFSF is bi-directional (both forward and reverse immune checkpoint) .
The Type-L subfamily, also called conventional TNFRSF immune checkpoint receptors, has the most members in TNFRSF, but only CD40, herpes virus entry mediator (HVEM), death receptor 3 (DR3), and lymphotoxin-β receptor (LTBR) have a co-stimulatory function, while CD120a, CD120b, and CD95 have apoptosis function on TC [52, 55]. The Type-V subfamily, also called divergent, is the only family where all members have co-stimulatory function, including 4-1BB (CD137), OX40 (CD134), CD27, CD30, and glucocorticoid-induced TNFR-related protein (GITR) . Among the Type-s subfamily, transmembrane activator and CAML interactor (TACI), B cell-activating factor receptor (BAFFR), and B cell maturation protein (BCMA) have the function of B cell activation, survival, and differentiation [52, 55]. The function of the orphan subfamily remains unclear, except that the receptor expressed in lymphoid tissues (RELT) has some evidences of stimulating TC proliferation .
We list six pairs of TNFRSF:TNFSF molecules in Table 1: CD40:CD40L, 4-1BB (CD137):4-1BBL, OX40 (CD134):OX40L, CD27:CD70, CD357 (GITR):GITRL, and CD30:CD30L, and discuss their characterizations in the following section.
Emerging evidences suggested that the two-way stimulatory immune checkpoint is critical for TC activation and APC inflammation. We summarized recent studies elucidating two-way stimulatory immune checkpoint with immune cell responses in human and mouse disease models (Tables 2 and and33).
CD40:CD40L is the first discovered stimulatory molecular pair of TNFRSF:TNFSF. CD40 is not only expressed on immune cells (BC, MC, MØ, DC) but also on a variety of somatic cells such as endothelial cell (EC), smooth muscle cell (SMC), fibroblast, and platelet . CD40 was initially discovered as a surface receptor on BC binding to CD40L on TC causing TC polyclonal activation and BC proliferation/differentiation . CD40L is the sole ligand for CD40 and is also known as CD154. CD40L has two forms, membrane-bound CD40L and sCD40L. Membrane-bound CD40L is expressed on activated TC, MC, MØ, platelet, mast cell, and EC . sCD40L circulates in the blood and is mainly produced by platelets . The CD40:CD40L two-way immune checkpoint promotes atherosclerosis and inhibits tumor progress and has been used as a cancer immunotherapy target [60–62]. sCD40L is significantly elevated in patients with cardiovascular disease (CVD) and CKD  and proposed as an independent predictor and biomarker for cardiovascular events after acute coronary syndrome and plaque vulnerability . CD40:CD40L interactions stimulate the expression of inflammatory cytokines, adhesion molecules, chemokines, matrix degrading enzymes, and platelet tissue factor. CD40−/−ApoE−/− mice exhibited 55% plaque reduction and less lipid-containing, collagen-rich, stable plaque, and improved reendothelialization . Similarly, anti-CD40L antibody induced a stable lesion with lipid-poor, collagen-rich plaque in ApoE−/− mice . CD40-RNAi-lentivirus prevented plaque progression in ApoE−/− mice .
The influence of CD40:CD40L forward immune checkpoint towards TC is well established. TC presents at all stages of atherosclerotic lesion. The major subset of TC in atherosclerotic plaques is Th1 CD4+ TC. CD40−/−ApoE−/− mice have a lower effector memory CD4+/CD8+ TC in the spleen . Anti-CD40L antibody reduced TC content in mouse atheroma . Moreover, the CD40:CD40L immune checkpoint inhibited Treg activation, as CD40L−/− bone marrow reconstitution in LDLR−/− mice led to increased Treg , and agonistic CD40 antibody reduced Treg in Lewis lung cancer mouse model .
Large amount of evidence described the impact of CD40:CD40L reverse stimulatory immune checkpoint towards APC. In the absence of CD40L on TC, BC only secrete IgM and cannot switch to other Igs (IgG, IgE, IgA). CD40L on TC binds to CD40 on MØ and leads to MØ activation and secretion of matrix metalloproteinase (MMP), pro-inflammatory cytokines (Il-12, TNF-α, IL-1β, IL-6, and IL-8), and platelet tissue factor. Similarly, CD40L gene mutation caused X-linked hyper IgM syndrome which is characterized by low or absent levels of IgG, IgE, and IgA but normal or elevated serum levels of IgM . MC-derived DC from patients with coronary artery disease (CAD) expressed higher CD40 which was associated with smoking history, higher C-reactive protein, and lower high-density lipoprotein cholesterol (HDL-C) . We reported that CD40+ MC was increased in patients with CVD and further elevated in patients with CVD+CKD. Anti-CD40L antibody significantly reduced MØ in mice . CD40−/−ApoE−/− mice exhibited lower pro-inflammatory Ly6C+ MC in blood and M1 MØ in the aorta . Moreover, CD40 agonist activated antitumor MØ infiltration and resulted pancreatic cancer regression in mice .
CD137 is mainly expressed on activated CD4+ TC and also on BC, MC, DC, and EC, while CD137L is constitutively expressed on APC and activated TC . Soluble CD137 (sCD137) is elevated in human acute coronary syndrome (ACS) and atherothrombotic stroke [73, 74] and has been suggested as a prognostic biomarker for acute atherosclerotic disease. The CD137:CD137L immune checkpoint promotes vascular inflammation as CD137−/−ApoE−/− and CD137−/−LDLR−/− mice had reduced atherosclerotic lesions and inflammation  and anti-CD137 antibody decreased atherosclerosis lesion in ApoE−/− mice .
The CD137:CD137L forward immune checkpoint promotes TC activation. CD137 is expressed predominantly in CD8+ TC and occasionally in CD4+ TC in human atherosclerotic lesions and associated with pro-inflammatory factor release such as TNF-α, IL-1β, and IFN-γ. CD137 agonist induced CD8+ TC infiltration in mouse atherosclerotic lesions and promoted the progression of atherosclerosis . In peripheral blood mononuclear cells (PBMC), antibody against CD137 decreased TNF-α and IFN-γ production from CD4+CD28null TC which expresses higher levels of CD137 compared with CD4+CD28+ TC . The CD137:CD137L checkpoint also enhances tumor immunity, as CD137 agonist promoted CD4+ and CD8+ TC proliferation in patients with head and neck cancer .
Recent research emphasized the role of the CD137:CD137L reverse stimulatory immune checkpoint on MC and MØ differentiation. Cross-linking of CD137L by CD137 on human PBMC induced IL-6, IL-8, IL-12, TNF-α, and IFN-γ production and inflammatory DC differentiation . Circulating CD137L+CD14+ MC was increased in patients with acute ischemic atherosclerotic stroke . CD137−/−ApoE−/− mice have lower MC and MØ in the aorta . Anti-CD137 monoclonal antibody induced iNOS-positive MØ differentiation in hepatoma tissue in mice .
CD134 is mainly expressed on activated CD4+ TC, CD8+ TC, and memory TC, while CD134L is expressed on mature APC, activated TC, and EC . The levels of sOX40L were significantly increased in patients with ACS . Anti-CD134L antibody significantly reduced atherosclerotic lesion in LDLR−/− mice [84, 85].
Similar as CD137, CD134 is highly expressed in CD4+CD28null TC. CD134 also regulates Treg function by suppressing Treg generation from naïve TC and effector TC in mice . CD134L induced INF-γ CD4+ TC proliferation in cultured splenocytes from ApoE−/− mice . Antibody against CD134 decreased TNF-α and IFN-γ production in CD4+CD28null TC derived from PBMC from ACS patients . Anti-CD134L antibody reduced the populations of circulating CD4+CD134+ TC, CD4+ TC and CD8+ TC, and lesion CD3+ TC in LDLR−/− mice . Anti-CD134 antibody combined with autophagosomes (DRibbles) induced memory and effector TC proliferation and differentiation and promoted tumor regression in mice . Elevating CD134+ CD8+ TC infiltration in colorectal cancer prolonged overall survival in humans .
Even though circulating MC expressed the highest level of CD134L in ACS patients , the atherogenic role of CD134:CD134L may not be mediated by MC and MØ. Anti-CD134L antibody had no effect on both M1 MØ and M2 MØ in ApoE−/− mice . CD134:CD134L may participate in BC Ig isotype switch, as blocking the CD134:CD134L immune checkpoint using anti-CD134L antibody increased anti-ox-LDL IgM, a protective IgM, in LDLR−/− mice . Moreover, agonistic CD134 antibody increased M2 MØ in tumor. M2 MØ produced higher IL-10 and chemokine (C-C motif) ligand (CCL)-17 and lower IL-12-b and IL-23 compared to M1 MØ, which limited the efficacy of CD134 agonist therapy in mice [90, 91].
In contrast to CD134 and CD137, CD27 is expressed on naïve TC, BC, and NK cells and upregulated on activated TC, while CD70 is expressed on APC and activated TC . Evidence for the role of CD27:CD70 in atherosclerosis is conflicting as ruptured atherosclerotic plaques expressed higher CD70 than those in stable lesions , and CD70 transgenic mice attenuated atherosclerotic development .
CD27 promotes activated TC proliferation and survival. CD27+ Treg is reduced in myocardial infarction patients, and this subset has high suppressive potential . CD70 deficiency reduced spleen Treg in ApoE−/− mice  and CD27 deficiency reduced Treg in solid tumor in mice , suggesting that CD27:CD70 may have an immunosuppressive role in atherosclerosis and tumor growth.
The CD27:CD70 reverse stimulatory immune checkpoint towards APC may be protective for atherosclerosis. CD70 transgenic mice displayed increased MC apoptosis . CD70 promoted ox-LDL efflux in MØ  while engineered anti-CD70 increased MØ phagocytosis and prolonged the survival in lymphoma mice .
CD30 is expressed on activated TC and BC, while CD30L is expressed on APC and activated TC . CD30 was originally recognized as a cancer-associated surface antigen in TC. The CD30:CD30L two-way immune checkpoint promotes atherosclerosis and tumor and is a therapeutic target for both diseases. The CD30 antibody is used to treat Hodgkin’s lymphoma, anaplastic large cell lymphoma, and other cancers . A few studies demonstrated that the CD30:CD30L blockade delayed the development of atherosclerosis.
CD30 primarily promotes CD4+ TC activation. Anti-CD30L treatment reduced CD4+ TC counts but had no effect on CD8+ TC, Th1, Th2, Th17, and Treg cell numbers in LDLR−/− mice . Recombinant CD30L inhibited CD30+ Th2 lymphoma cell proliferation .
CD30:CD30L may not affect APC function in atherosclerosis, as anti-CD30L treatment did not change BC counts, ox-LDL-specific IgM/IgG, and aortic MC numbers in LDLR−/− mice . Anti-CD30 antibody enhanced MØ phagocytosis in tumor tissue and increased survival in mice .
GITR is expressed on naïve TC, increased on activated TC, and is also present on BC and NK cells, while GITR ligand (GITRL) is expressed on APC . GITR:GITRL may have a protective role in atherosclerosis via regulating Treg. GITRL transgenic BM transplantation reduced atherosclerotic lesion in LDLR−/− mice .
It is known that GITR:GITRL interaction is important for CD4+ TC, CD8+ TC, and Treg differentiation and expansion. Thus, GIRT is used as a Treg marker. GITRL transgenic chimeric LDLR−/− mice displayed an increased effector TC and Treg and reduced atherosclerosis . sGITRL suppressed Treg infiltration in human liver tumor .
GITR and GITRL are mainly expressed in MØ in plaques. However, the protective role of GITR:GITRL in atherosclerosis is controversial. Anti-GITR mAb induced human MC and MØ activation, MMP-9, and pro-inflammatory cytokine expression, which may promote atherosclerosis and plaque instability . Agonistic GITR antibody promoted M1 MØ differentiation in mice liver tumor .
We summarized two molecular signaling pathways previously reported for the CD40:CD40L immune checkpoint: forward stimulatory immune checkpoint towards TC and reverse stimulatory immune checkpoint towards APC (Fig. 4a, b). In addition, we propose a novel pathway, the TC contact-independent immune checkpoint (Fig. 4c) based on our recent discoveries .
We found that metabolic RF, such as uremic toxin and HHcy, induced circulating sCD40L and CD40+ MC in CKD patients. Also, both sCD40L and HHcy promoted inflammatory CD40+ MC and intermediate MC differentiation in cultured human PBMC . Other metabolic RF, such as triazolopyrimidine, inhibited CD40-associated MC activation . A mechanistic study showed that SAH-related DNA hypomethylation is responsible for CD40+ MC differentiation in human PBMC . We were the first to establish a direct mechanistic link between HHcy and increased cellular SAH and to propose that SAH-related hypomethylation is a key biochemical mechanism for HHcy-induced CVD in EC [109–111]. We believe that Hcy and SAH function as metabolic sensors and are responsible for DNA hypomethylation and APC activation.
The CD40:CD40L forward stimulatory immune checkpoint follows signal 1 (Ag recognition) and leads to TC activation. During this process, MHC presents Ag to TCR, which triggers the assembling of TCR, CD3, and TCRζ chain. The subsequent CD40:CD40L immune checkpoint interaction amplifies the activation of three transduction pathways via the recruitment of zeta chain-associated protein kinase of 70 kDa (ZAP-70) and phosphorylation of linker for activation of T cells (LAT), RAS mitogen-activated protein kinase (MAPK) pathway, calcium-calcineurin pathway, and nuclear factor κB (NF-κB) pathway .
In APC, the CD40:CD40L reverse stimulatory immune checkpoint is associated with proliferation of MC, MØ, BC, SMC, and tumor cells, and inflammatory molecular production. CD40 can bind to TNF receptor-associated factor (TRAF1-3/5-6) and activate three TNF signaling, including signal transducers and activator of transcription-3 (STAT3), NF-κB, and activator protein 1 (AP-1) pathways in cell type and TRAF member-dependent manner. For example, STAT3 can be activated by CD40:TRAF2/3 ligation via JAK in BC ; NF-κB can be activated by CD40:TRAF1-3/5-6 interaction via IKK/I-κB in BC and MC; and AP-1 can be activated by CD40:TRAF6 via MAPK in MC and MØ . Moreover, TRAF1/2/3/5 activation is linked to NF-κB, MAPK/p38, and JUN amino-terminal kinase (JNK) pathway, while TRAF6 activates NF-κB, protein kinase B, and STAT3 pathway . CD40:TRAF6 has a critical role in promoting atherosclerosis, as attenuated atherosclerosis and reduced Ly6C+ MC and M1 MØ were observed in CD40−/−TRAF6−/−ApoE−/− but not in CD40−/−TRAF2/3/5−/−ApoE−/− mice .
This is a novel pathway we proposed based on our and others discoveries [15, 115]. We demonstrated that metabolic RF, such as uremic toxin and HHcy, and sCD40L promoted inflammatory CD40+ MC and intermediate MC differentiation in culture human PBMC in the absence of TC . We hypothesize that metabolic RF promote pro-inflammatory MC differentiation via metabolic sensors, such Hcy and SAH and DNA hypomethylation. This is based on evidence from mediation analysis showing increased plasma Hcy and SAH levels and consequential reduction of SAM/SAH ratio, a recognized indicator of methylation status, and from mechanistic studies showing Hcy-suppressed DNA methylation in CD40 promoter and folic acid, a methylation rescue reagent, reversed CD40+ MC differentiation in human PBMC . We were the first to establish a direct mechanistic link between Hcy and increased cellular SAH with hypomethylation and to propose hypomethylation as a key biochemical mechanism for HHcy-induced CVD [109–111]. Our discoveries suggested that the TC contact-independent immune checkpoint is a critical mechanism for systemic and tissue inflammatory response in metabolic disorders.
MC heterogeneity has been widely acknowledged. MC expresses various receptors, which sense the environment stimulation and mediate cell differentiation towards inflammatory or anti-inflammatory subsets. MC is the most invasive immune cells which can transmigrate into tissue causing tissue inflammation and repair. In humans, MC are divided into three functionally distinct subsets according to the surface marker CD14 and CD16 . CD14 is used as a marker for human MC. The common MC subsets by nomenclature is classified as (1) classical MC (CD14++CD16− phagocytic MC), (2) intermediate MC (CD14++CD16+ pro-inflammatory MC), and (3) non-classical MC (CD14+CD16++ patrolling MC) . However, such human MC classification is not in harmony, as further increased expression of CD16, an inflammatory marker, is associated with anti-inflammatory function in non-classical MC subsets.
Our recent findings resolved the above controversy in MC subset classification and presented CD40+ MC as a novel and stronger pro-inflammatory MC subset compared with the nomenclature-defined intermediate MC (Table 4) . By examining the expression of nine inflammatory markers in three nomenclature-defined MC subsets and CD40+ MC , we discovered that CD40+ MC expressed higher levels of TC activation receptor CD86, chemokine receptor CCR2, and expressed similar levers of other inflammatory surface markers than that on nomenclature-defined intermediate MC (Table 4). In contrast, CD40− MC exhibited much lower levels of TC activation receptor HLA-DR, adhesion receptor CD49d, and chemokine receptor CX3CR1 than that on common nomenclature-recognized anti-inflammatory (patrolling) non-classical MC subset.
On the other hand, classically defined pro-inflammatory intermediate MC expressed lower levels of inflammatory markers CCR2, HLA-DR, and CD62L compared with classical (phagocytic) and non-classical (patrolling) MC . This is inconsistent with the inflammatory feature of these MC subsets.
CKD is considered as a metabolic complication. Patients with CKD have 10 to 30 times higher cardiovascular mortality than the general population, and 50% of deaths in end-stage CKD were due to CVD . MC is the key player in the development of atherosclerosis. Intermediate MC was elevated in patients with CVD compared with healthy subjects  and in patients with ST-elevation myocardial infarction. Its population is positively correlated with cardiovascular events, such as cardiovascular death, acute myocardial infarction, and non-hemorrhagic stroke . However, there are a few contradictory dilemmas regarding the molecular marker and biological function of the currently defined three MC subsets . For example, (1) the intermediate CD14++CD16+ (pro-inflammatory) MC expresses very high levels of anti-inflammatory marker CX3CR1 and (2) high levels and absence of CD16 were presented in CD14+CD16++ (phagocytic) and CD14++CD16− (patrolling) MC, respectively, both displaying an anti-inflammatory function. Therefore, there is no strong consensus to use the intermediate CD14++CD16+ MC as a reliable biomarker for the severity of CVD and metabolic disease.
Currently, CKD severity is determined by estimated glomerular filtration rate (eGFR) which is a prediction parameter calculated using blood creatinine, age, race, gender, and other factors. We believe CD40+ MC is a more accurate and reliable biomarker for CKD and CVD . As shown in Fig. 5, CD40+ MC subset was elevated in patients with CVD and CVD+CKD compared with healthy subjects and increased with the elevation of CKD severity (Fig. 5a). Similarly, sCD40L was also elevated in patients with CVD and CVD+CKD compared to healthy subjects (Fig. 5b). CD40+ intermediated MC subset was elevated in patients with CVD and CVD+CKD compared with healthy subjects and increased with the elevation of CKD severity (Fig. 5c). However, intermediate MC subset was elevated in CVD patients but not further increased in CVD+CKD patients (Fig. 5d). Future studies will further analyze the relationship of CD40+ MC with different subtypes of CKD; such studies should allow us to better define CD40+ MC as a diagnosis and prognosis biomarker for CKD.
A novel metabolic response was incorporated into the immune system framework, providing an extensive overview of current knowledge in immune checkpoint theory (Fig. 6). This metabolic response is a novel MADS recognition pattern, mediating metabolic RF-induced innate and adaptive immune response. We propose the MADS recognition as signal 4 in adaptive immunity. MADS recognition induces immune checkpoint molecule expression via metabolic sensor leading to amplification of signal 2 two-way stimulatory immune checkpoint amplification. The forward immune checkpoint leads to TC activation. The reverse immune checkpoint leads to APC activation. Metabolic RF, such as uremic toxin or HHcy, was demonstrated to induce CD40 expression in MC and to elevate circulating sCD40L resulting in CD40+ MC differentiation via metabolic sensor. We defined CD40+ MC as a novel and stronger pro-inflammatory MC subset, compared with intermediate MC, and a reliable biomarker for CKD severity. Our studies supported the notion that MADS recognition amplify stimulatory immune checkpoint leading to TC activation and APC inflammation, respectively, which results in systemic and tissue inflammation. Furthermore, we propose the CD40:CD40L immune checkpoint as a therapeutic target for metabolic disease, CVD, and cancer.
This work was supported in part by the following National Institutes of Health (NIH) grant numbers: HL82774, HL110764, HL117654, DK104116, HL131460, HL130233, and DK113775 (HW); HL9445, HL108910, and HL116917 (XFY), HL131460 (HW/EC/XFY), by American Heart Association (AHA) grant 17SDG33671051 (PF), by the National Science Foundation of China grant number 81200208 (JD), by the Medical and Health Technology Plan of Zhejiang Province grant number 2017KY508 (JD), and by the National Science Foundation of China grant numbers 81330004 and 91639204 (JY).
Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.
Additional file 1: Table S1.(30K, doc) Features of innate and adaptive immunity. (DOC 29 kb)
Features of innate and adaptive immunity. (DOC 29 kb)
All authors have contributed to revising the manuscripts. All authors have read and approved the final manuscript.
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
The online version of this article (doi:10.1186/s13045-017-0504-1) contains supplementary material, which is available to authorized users.
Jin Dai, Email: moc.361@606sinnad.
Pu Fang, Email: ude.elpmet@86576aut.
Jason Saredy, Email: ude.elpmet@67793gut.
Hang Xi, Email: ude.elpmet@ixgnah.
Cueto Ramon, Email: ude.elpmet@00645eut.
William Yang, Email: moc.liamg@2161gnaygniy.
Eric T. Choi, Email: ude.elpmet.shut@iohC.cirE.
Yong Ji, Email: nc.ude.umjn@ijgnoy.
Wei Mao, Email: moc.361@wliewoam.
Xiaofeng Yang, Email: ude.elpmet@gnayfx.
Hong Wang, Email: ude.elpmet@wgnoh.