Metabolism-associated danger signal-induced immune response and reverse immune checkpoint-activated CD40+ monocyte differentiation

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. 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.


Background
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 [1]. 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) [4]. 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 Fig. 1 Innate immunity. a Innate immunity and novel MADS recognition. The classical innate immune system provides immediate and non-specific defense against pathogen or injury-generated molecules via PAMP/DAMP+PRR recognition in phagocytes and TC. Super Ag, a subset of pathogen toxins, can also bind to a multitude of TCR leading to TC activation. In addition, we propose a novel MADS recognition pathway, which allows metabolic risk factors to activate the innate immunity via responsive metabolic sensors in phagocytes and TC. The activation of innate immunity leads to pathogen elimination and inflammation (APC formation, cytokine generation, and TC activation). b Evidences of innate immunity in TC. Stimuli such as PAMP/DAMP, inflammatory cytokines and super Ag activate different subsets of TC and stimulate TC proliferation, inflammatory cytokine production, and phagocytosis. Words in red emphasize our newly proposed recognition pattern. Abbreviations: APC antigen present cell; Ag antigen; Ab antibody; BC B cell; BCR B cell receptor; CpG C, a cytosine triphosphate deoxynucleotide; p phosphodiester; G a guanine triphosphate deoxynucleotide; CTL cytotoxic T lymphocytes; DAMP danger-associated molecular patterns; d days; Foxp3 forkhead box P3; h hours; IL interleukin; IFN interferon; LPS lipopolysaccharide; MHC major histocompatibility complex; MADS metabolismassociated danger signal; NLR NOD (nucleotide-binding and oligomerization domain)-like receptors; PAMP pathogen-associated molecular patterns; PRR pattern recognition receptor; Poly(I:C) polyinosinic-polycytidylic acid; Pam 3 CSK 4 tripalmitoyl-S-glycero-Cys-(Lys)4; RF risk factor; R848 Imidazoquinoline Resiquimod; SEB staphylococcal enterotoxin B; TC T cell; TCR T cell receptor; Th17 T helper 17 cell; TLR Toll-like receptors; SPA staphylococcal protein A; TNF tumor necrosis factor; TGF-β transforming growth factor beta checkpoint for TC suppression [5,6] and was expanded in 2012 to include co-stimulatory immune checkpoint for TC activation [7]. 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 lymphocyteassociated protein 4 (CTLA-4)-immunoglobulin (Ig) competes with CD28 to bind to CD80/CD86 and causes CTLA-4:CD80/CD86-induced TC suppression [8]. 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][10][11]. These antibodies resulted in clinical regression of melanoma, nonsmall cell lung cancer, and other cancers [9][10][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 costimulatory 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][16][17][18][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.

Fig. 2
Adaptive immunity with novel signal 4, the metabolic RF recognition. The adaptive immunity is characterized by Ag specificity and immunologic memory leading to TC and BC activation. There are two types of adaptive immunity: TC immunity (cell-mediated immunity) and BC immunity (humoral immunity). Classically, each involves three activating signals. We propose a novel signal 4 (metabolic RF recognition) mediated by metabolic sensor. a TC immunity. TC activation involves four distinct signals. In signal 1 (Ag recognition), the Ag peptide is presented by MHC on the APC to Ag-specific TCR on TC. Signal 2 (immune checkpoints) involves ligand and receptor binding on APC and TC. Signal 3 responds to inflammatory cytokine stimulation. The novel signal 4 describes metabolic RF using a metabolic sensor leading to MC (APC) differentiation, inflammatory cytokine production, and the enhancement of signals 2 and 3. b BC immunity. BC activation involves Ag binding to BCR (signal 1), ligand and receptor binding (signal 2), cytokine stimulation (signal 3), and metabolic RF recognition (signal 4). Words in red emphasize our newly proposed signal. Abbreviations: APC antigen present cell, Ag antigen, BC B cell, BCR B cell receptor, RF risk factor, HHcy hyperhomocysteinemia, MHC major histocompatibility complex, MC monocyte, sCD40L soluble CD40 ligand 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 Table 1 Immune checkpoint families and paired molecules Immune checkpoints are classified as one-way immune checkpoint and two-way immune checkpoint based on signal 2 direction (forward only or both forward and reverse) and are further divided into stimulatory and inhibitory checkpoints. Listed are a few major immune checkpoint families. For example, in the CD28 family, CTLA-4 receptor binds to ligand B7-1(CD80) or B7-2(CD86) which are enriched in TC, BC, DC, MC, and MØ. The black frame emphasizes the two-way immune checkpoint to be focused. Words in the parentheses are aliases of the receptor and ligands Abbreviations: AC apoptotic cell, BC B cell, BTLA B and T lymphocyte attenuator, CTLA-4 cytotoxic T lymphocyte-associated protein 4, CRTAM cytotoxic and regulatory T cell molecule, DC dendritic cell, DNAM-1 DNAX accessory molecule-1, EC endothelial cell, EPC epithelial cell, FIB fibroblast, GITR glucocorticoid-induced TNFR-related protein, GITRL GITR ligand, HVEM herpes virus entry mediator, ICOS inducible T cell co-stimulator, MC monocyte, MØ macrophage, MAC mast cell, MYC myeloid cell, NKC natural killer cell, NE neuronal, NECL2 nectin-like protein 2, PA platelet, PS phosphatidylserine, PD-1 programmed cell death protein 1, PD-L PD ligand, SMC smooth muscle cell, TC T cell, TNFSF tumor necrosis factor superfamily, TIM T cell (or transmembrane) immunoglobulin and mucin domain, TIGIT T cell immunoreceptor with Ig and ITIM domains knowledge and comprehensive insights into our understanding about immune response and immune checkpoints, especially the reverse stimulatory immune checkpoint in diseases.

Innate immunity recognizes novel MADS and regulates TC activation
Innate immunity and novel MADS recognition 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 [1]. 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 [22].
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 [23]. 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 [16]. Glucose, ceramide, islet amyloid polypeptide, and cholesterol crystals can be sensed by TLR or NLRP3-stimulating NLRP3 inflammasome complex assembly [16,24,25]. We [15] and others [26] 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) [15].

Adaptive immunity recognizes MADS and regulates TC/BC activation
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 [7]. 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).
TC immunity (cell-mediated immunity) (Fig. 2a) 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 [5]. Signal 3, cytokine stimulation, is also involved in TC activation [31]. 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 costimulatory immune checkpoint and increased serum soluble CD40L (sCD40L) levels [15].
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 costimulation and expanded to include co-inhibitory pathways [32]. In this article, we propose to term the costimulatory and co-inhibitory pathways collectively as the immune checkpoint. Immune checkpoint initially described co-inhibitory signal 2 in Topalian et al.'s papers [33] based on the discovery of T cell function restraint in normal physiologic settings and tumors [34]. This terminology was recently used to describe as a regulatory switch towards either stimulatory or inhibitory pathways [7]. 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 [35].
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 [15]. 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 [38]. The levels of sCD40L were found to be increased in subjects with metabolic syndrome and hypertension and negatively related to insulin sensitivity [39]. In addition, glucose sustains TC growth and proliferation upon TCR-dependent TC activation [40].
BC immunity (humoral immunity) (Fig. 2b) BC immunity involves the same four signals which leads to antibody production and BC activation [41]. Signal 1 (Ag recognition) is the engagement of Ag with Agspecific 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 [42] and sCD40L is induced in metabolic disease including CKD, HHcy, hypertension, hyperglycemia, and dyslipidemia [15,39,43].

Immune checkpoint regulates TC and APC activation
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
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 welldescribed 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 [35]. Metabolic product ceramide is involved in the forward stimulatory immune checkpoint in TCR-dependent TC activation at multiple levels [44].
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 [45], indicating TIGIT inhibited the effector function and proliferation of CD8 + TC.

Two-way immune checkpoint
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 twoway immune checkpoint. CD40 binds to its ligand CD40L, which is usually transiently expressed on TC [46] 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 [47]. 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 [48]. Moreover, metabolic RF cholesterol crystal is required for TCR nanoclustering in TC, which enhances the avidity of the TCR-antigen interaction [49]. Reversely, cholesterol crystals trigger pro-inflammatory cytokine secretion from APC MØ [50].

Immune checkpoint family and paired molecules
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 B One-way immune checkpoint C Two-way immune checkpoint  Fig. 3 Immune checkpoint and its characterizations. a Immune checkpoint classification. Immune checkpoints are classified as one-way immune checkpoint and two-way immune checkpoint based on signal 2 direction and are further divided into stimulatory and inhibitory immune checkpoints. b. One-way immune checkpoint. The one-way immune checkpoint only involves forward signaling in signal 2. The ligation of co-stimulatory pair molecules triggers forward stimulatory immune checkpoint inducing TC proliferation, whereas the inhibitory immune checkpoint induces TC suppression or death. c Two-way immune checkpoint. The two-way immune checkpoint involves both forward and reverse stimulatory signaling. The reverse stimulatory immune checkpoint induces either TC proliferation or MC (APC) differentiation/ inflammation. The inhibitory immune checkpoint leads to TC suppression/death or APC death. Words in red emphasize our newly proposed signal. Abbreviations: APC antigen present cell, MC monocyte, RF risk factor, sCD40L soluble CD40 ligand, TC T cell [52]. 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.
We list six pairs of TNFRSF:TNFSF molecules in Table 1

Two-way stimulatory immune checkpoint induces tissue and systemic inflammation
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 3).
CD40:CD40L two-way immune checkpoint 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 [57]. CD40 was initially discovered as a surface receptor on BC binding to CD40L on TC causing TC polyclonal activation and BC proliferation/differentiation [46]. 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 [58]. sCD40L circulates in the blood and is mainly produced by platelets [59]. The CD40:CD40L two-way immune checkpoint promotes atherosclerosis and inhibits tumor progress and has been used as a cancer immunotherapy target [60][61][62]. sCD40L is significantly elevated in patients with cardiovascular disease (CVD) and CKD [15] and proposed as an independent predictor and biomarker for cardiovascular events after acute coronary syndrome and plaque vulnerability [63]. 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 lipidcontaining, collagen-rich, stable plaque, and improved reendothelialization [64]. Similarly, anti-CD40L antibody induced a stable lesion with lipid-poor, collagen-rich plaque in ApoE −/− mice [65]. CD40-RNAi-lentivirus prevented plaque progression in ApoE −/− mice [66].
CD40:CD40L reverse immune checkpoint 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 [48]. 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) [70]. 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 [67]. CD40 −/− ApoE −/− mice exhibited lower proinflammatory Ly6C + MC in blood and M1 MØ in the aorta [47]. Moreover, CD40 agonist activated antitumor MØ infiltration and resulted pancreatic cancer regression in mice [71].

CD137:CD137L forward immune checkpoint
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 proinflammatory 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 [76]. In peripheral blood mononuclear cells (PBMC), antibody against CD137 decreased TNF-α and IFN-γ production from CD4 + CD28 null TC which expresses higher levels of CD137 compared with Stimulatory immune checkpoint has been described in human and mouse metabolic diseases. For example, CD40:CD40L two-way stimulatory immune checkpoint was activated and associated with increased lesion CD40L + TC in ACS and CD40 + MC in CKD in humans. CD40 KO reduced spleen effector memory CD4 + /CD8 + TC and blood Ly6C + MC and aorta M1 MΦ. Similar functional change was observed for immune checkpoint CD40:CD40L CD137:CD137L and OX40:OX40L. CD4 + CD28 + TC [77]. The CD137:CD137L checkpoint also enhances tumor immunity, as CD137 agonist promoted CD4 + and CD8 + TC proliferation in patients with head and neck cancer [78].

CD27:CD70 two-way immune checkpoint
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 [92]. Evidence for the role of CD27:CD70 in atherosclerosis is conflicting as ruptured atherosclerotic plaques expressed higher CD70 than those in stable lesions [93], and CD70 transgenic mice attenuated atherosclerotic development [94].
CD27:CD70 forward immune checkpoint CD27 promotes activated TC proliferation and survival. CD27 + Treg is reduced in myocardial infarction patients, and this subset has high suppressive potential [95]. CD70 deficiency reduced spleen Treg in ApoE −/− mice [93] and CD27 deficiency reduced Treg in solid tumor in mice [96], suggesting that CD27:CD70 may have an immunosuppressive role in atherosclerosis and tumor growth.

CD30:CD30L two-way immune checkpoint
CD30 is expressed on activated TC and BC, while CD30L is expressed on APC and activated TC [98]. 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 [99]. A few studies demonstrated that the CD30:CD30L blockade delayed the development of atherosclerosis.

Molecular mechanisms underlying CD40:CD40L two-way immune checkpoint amplification
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 [15]. 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 [15]. Other metabolic RF, such as triazolopyrimidine, inhibited CD40associated MC activation [108]. A mechanistic study showed that SAH-related DNA hypomethylation is responsible for CD40 + MC differentiation in human PBMC [15]. 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][110][111]. We believe that Hcy and SAH function as metabolic sensors and are responsible for DNA hypomethylation and APC activation.
CD40:CD40L TC contact-independent immune checkpoint ( Fig. 4c) 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 [15]. 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 [15]. 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 HHcyinduced CVD [109][110][111]. Our discoveries suggested that the TC contact-independent immune checkpoint is a critical mechanism for systemic and tissue inflammatory response in metabolic disorders.
CD40 + MC is a novel and stronger inflammatory MC subset 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 [116]. CD14 is used as a marker for human MC. The common MC subsets by nomenclature is classified as (1) [116]. 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) [15]. By examining the expression of nine inflammatory markers in three nomenclature-defined MC subsets and CD40 + MC [15], 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 proinflammatory intermediate MC expressed lower levels of inflammatory markers CCR2, HLA-DR, and CD62L compared with classical (phagocytic) and non-classical (patrolling) MC [15]. This is inconsistent with the inflammatory feature of these MC subsets.

CD40 + MC is a reliable biomarker of CKD severity
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 [117]. MC is the key player in the development of atherosclerosis. Intermediate MC was elevated in patients with CVD compared with healthy subjects [15] and in patients with STelevation myocardial infarction. Its population is positively correlated with cardiovascular events, such as cardiovascular death, acute myocardial infarction, and non-hemorrhagic stroke [118]. However, there are a few contradictory dilemmas regarding the molecular marker and biological function of the currently defined three MC subsets [15]. 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 antiinflammatory 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 [15]. 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.

Conclusion
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 RFinduced 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

Additional file
Additional file 1: Table S1. Features of innate and adaptive immunity. Metabolic risk factors, such as HHcy, uremic toxins, and other RF,, stimulate two-way stimulatory immune checkpoint amplification in TC, APC (MC), and possibly in PL via MADS recognition. In response to metabolic RF stimulation, metabolic sensors mediate TC activation via MAPK/NF-κB/calcineurin pathway, APC inflammation via STAT3MAPK/NF-κB pathway, MC differentiation via DNA hypomethylation, and possibly sCD40L production in PL via MAPK/NF-κB activation. TC activation and APC inflammation finally result in inflammatory cytokine production and systemic/tissue inflammation. Words in red emphasize our newly proposed signal pathway. Abbreviation: APC antigen present cell, HHcy hyperhomocysteinemia, MC monocyte, MAPK mitogen-activated protein kinase, MADS metabolism-associated danger signal, NF-κB nuclear factor κB, RF risk factor, PL platelet, STAT3 signal transducers and activator of transcription-3, sCD40L soluble CD40 ligand, TC T cell