An experimental data mining strategy (Figure 1) was used to analyze the expression profiles of mRNA transcripts of genes encoding T cell co-stimulation receptors and co-inhibition receptors in 32 different human and 29mouse tissues. We utilized experimentally verified mRNA expression in the expressed sequence tag(EST) databases of the National Institutes of Health (NIH)/National Center of Biotechnology Information (NCBI) UniGene (http://www.ncbi.nlm.nih.gov/sites/entrez?db=unigene) to determine the transcription profile of T cell co-stimulation receptors and co-inhibition receptors in tissues of interest. Transcripts per million of genes of interest were normalized to that of house-keeping β-actin in each given tissue to calculate the arbitrary units of gene expression. A confidence interval of the expression variation of house-keeping genes was generated by calculating the mean plus two times that of the standard deviation of the arbitrary units of three randomly selected housekeeping genes (PRS27A, GADPH, and ARHGDIA in human; Ldha, Nono, and Rpl32 inmouse) normalized by β-actin in the given tissues. If the expression variation of a given gene in the tissues was larger than the upper limit of the confidence interval (the mean plus two times the standard deviation) in house-keeping genes, the high expression levels of genes in the tissues were considered statistically significant. Gene transcripts where the expression level was lower than one per million were technically considered as no expression.

Figure 1

Flow chart of database mining strategy data organization. Abbreviations: NCBI: National Center of Biotechnology Information; IDs: Identifications; EST: Expressed sequence tag.

Microarray datasets were collected from ArrayExpress of European Bioinformatics Institute, which stores data from high-throughput functional genomics experiments (https://www.ebi.ac.uk/arrayexpress).These data includes the information of the expression of T cell co-stimulation receptors and co-inhibition receptors through experiments submitted directly to ArrayExpress or imported from the NCBIGene Expression Omnibusdatabase (https://www.ncbi.nlm.nih.gov/gds/).

Recent reports showed that co-stimulation receptors (CSRs) and co-inhibition receptors (CIRs) expressed in antigen presenting cell (APC) play important roles in modulating T cell activation status (1). An important question remains whether human tissues contain various types and numbers of antigen presenting cells (APCs), which provide various stimulatory/suppressive environments for all the four CD4+ T cell functional aspects including priming, helper cell differentiation, effector and memory functions. We hypothesize that human tissues have different antigen presenting environments and expression levels of CSRs and CIRs. To examine this hypothesis, by searching updated REFERENCES (1, 41, 42), we collected total of 28 co-signaling receptors expressed on the cell surface of APCs including 14 CSRs, 4 CSRs that become CIRs when encounter activated T cells, and 10 CIRs (Table 1). We then examined the expression patterns of all the 28 co-signaling receptors in 32 human tissues by searching DNA sequencing-based data of mRNA levels at physiological conditions. Based on the co-signaling receptor expression amongst human tissues examined, we classified them into following three groups: highly expressed (++), low expressed (+), and not expressed (-) as summarized in Table 2. Of note, information on BTNL2 gene expression was not found in human expression sequence tag (EST) profile in the NIH-UniGene database, therefore expression levels of 27 human co-signaling receptors were analyzed (Table 2).

We found that co-signaling receptors are differentially expressed in 32 human tissues in physiological condition. We identified CD40, CD48, CD58, SEMA4A, and B7-H2 are the highly expressed CSRs and we could not distinguish any CIRs that are highly expressed in tissues. When considering all the levels of expression of co-signaling receptors in tissues, we found that three CSRs including B7-H2, SEMA4A, CD58 and five CIRs such as galectin 9, CD113, HVEM, B7-H3, and VISTA are expressed in more than 60% of human tissues examined. We also found that five tissues such as lymph node, placenta, thymus, nerve and trachea have high T cell co-stimulation and co-inhibition signaling potential. As shown in Table 3, human tumors have higher expression of CIRs than CSRs. SEMA4A, CD40 and CD48 are the highly expressed CSRs; and Galectin9, SEMA4A, B7-H3, B7H4 and VISTA are highly expressed CIRs in human tumors. Bladder carcinoma and ovarian tumors have high co-inhibition potential, whereas chondrosarcoma has high co-stimulator potential (Table 3). Furthermore, three co-signaling receptors such as CD48 (CSR), two CIRs galectin 9, and CD113 are significantly expressed in 29 mouse tissues. Lymph node, lung, skin, and spleen have high expression of co-signaling receptors among mouse tissues examined (Table 4).

Recent reports showed that CSRs and CIRs expressed in APCs play important roles in modulating four inter-connected processes such as:1) priming T cells, 2) promoting T helper cell differentiation, 3) facilitating T effector function, and 4) memory function (1). We hypothesized that CSRs and CIRs have differences in modulating these four functional processes. To examine this hypothesis, we performed an extensive literature search. As summarized in Table 5, we found that 28 co-signaling receptors expressed in APC that can bind to 23 counter receptors on T cell surface. These receptors functionally differ in modulating 4 main immunological functions of T cells mentioned above. CD40 and CD70 are the only CSRs that exert all four immunological functions on T cells. Six CSRs including CD48, CD58, TL1A, OX40L, LIGHT, B7-H2 and one CIR Galectin 9 can regulate three functions mentioned above. The other co-signaling receptors have less effect on functional processes of the T cells. Further, our data revealed that most CIRs focus on suppressing T cell effector function.

To determine whether co-signaling receptors have functional overlaps in modulating four T cell functional processes, we performed a Venn analysis. As shown in Figure 2, we found that CSRs have diversified effects on stimulating T cells in priming, differentiating, effector and memory processes (Figure 2A) whereas most CIRs regulate T effector function (Figure 2B). In addition, we found that GITRL and TIM4 are the only two CSRs that have one function. CD113, CD112, and CD155 are the only CIRs that have one function, others have multiple effects on T cells. Priming, differentiation and memory function have no unique co-signaling receptors, which indicate that these three functions are all connected to T cell effector functions.

Figure 2

The Venn diagram analysis indicate that co-stimulation receptors exert diversified functions on T cells. A. CSRs exert diversified immunological regulating functions on T cells. B. Most CIRs exert limited effects on T cells.

Recently, Dr. Pober’s laboratory showed that engagement of the T-cell receptor not only activates T cells but also triggers CD4+T cell trans-endothelial migration (TEM) by a process that is distinct from that induced by activating chemokine receptors on T cells. This distinct process is slower, require microtubule-organization of T cell and engagement of proteins of the endothelial cell (43), suggesting that MHC class I/II-antigen epitopes expressed in tissue APCs and potential other signaling receptors in tissue endothelial cells modulate various T cell functions including trans-endothelial cell migration. We hypothesized that human tissues have significant differences in co-signaling for four major T cell processes. As shown in Figure 3, based on the highly expressed co-signaling receptors in human tissues (Table 2) and various T cell modulation processes of these co-signaling receptors (Table 5), we estimated the human tissue co-signaling potential (Table 6). We found that among 32 human tissues, 10 tissues including blood, bone, heart, kidney, lymph node, mammary gland, muscle, spleen, thymus, and trachea have high cumulative score of co-signaling potentials. In addition, we found the following results: 1) embryonic tissue, esophagus, intestine, liver, prostate, umbilical cord and uterus do not have highly expressed co-signaling receptors, suggesting these tissues have low co-signaling potential, and may have immune privilege status(23); 2) when meet with naïve T cells versus activated T cells, the co-signaling potential scores of blood, spleen and trachea are decreased by“-8”, and the co-signaling potential scores of kidney are decreased by“-3”; and 3) skin, ovary and placenta have high co-inhibition potentials.

Figure 3

Human tissues have significant differences in co-signaling for 4 main immunological functions of T cells in 32 human tissues. A. The effects of CSRs on immunological functions of naïve T cells. B. The effects of CIRs on immunological functions of activated T cells. The results indicate that: lymph node has the highest co-signaling potential, ovary and skin only have co-inhibition potential. * Co-signaling scores of different tissues were estimated on the highly expressed receptors data given in tables 2 and 3.

We further examined whether human tissues have significant differences in co-signaling for four major T cell functions in 32 tissues. As demonstrated in Figure 3, we found that nine human tissues including bone, eye, kidney, lung, lymph node, mammary gland, thymus, trachea and vasculature have dominant CSR signaling for T effector functions (shown in green). In addition, three human tissues such as eye, ovary, placenta and skin have dominant CIR function for T effector process (shown in light blue).

It has been reported that the infiltrated immune cells play a critical role for immunosurveillance in cancers and the prognosis of patients with cancer (44). We hypothesized that decreased expression of CSRs and increased expression of CIRs in tumor cells may be associated with certain human tumors. Based on our analysis of the expression changes of co-signaling receptors in human tumors in comparison to that of corresponding human tissues, we generated human tumor co-signaling potentials as shown in Figure 4. We found that first, five human tumors including uterine tumor, germ cell tumor, leukemia, glioma and chondrosarcoma have high co-stimulation potential and low co-inhibition potential, suggesting that these five tumors may have better response to immunotherapy; second, six human tumors such as esophageal tumor, skin tumor, liver tumor, colorectal tumor pancreatic tumor, and prostate tumor have moderate co-stimulation potential and moderate co-inhibition potential, suggesting that these six tumors may not respond to immunotherapy well; and third, other 11 tumors (50.0%) including retinoblastoma, cervical tumor, lymphoma, adrenal tumor, gastrointestinal tumor, kidney tumor, ovarian tumor, soft tissue muscle tissue tumor, breast tumor, lung tumor and bladder carcinoma have low co-stimulation potentials and high co-inhibition potential, suggesting that these 11 tumors may have poor response to immunotherapy.

Figure 4

Human tumors have lower co-signaling potential than correlated human tissues, especially the co-stimulation potential. A. Co-stimulation potential of human tumors. B. Co-stimulation potential of corresponding human tissues. *Tumors with changed co-signaling potential.

In addition, by comparing the detailed changes of co-signaling potentials of human tumors with that of corresponding human tissues (Figure 5), we found that: 1) the co-signaling potential of 16 human tumors are significantly different from that of the corresponding human tissues; and 2) human tumors including leukemia, chondrosarcoma, germ cell tumor, lymphoma, breast tumor, muscle tumor, pancreatic tumor, uterine tumor have dominant co-stimulation for T effector function. Taken together, our results suggest that human tumors have lower co-signaling potential compared to healthy human tissues, suggesting that downregulation of co-stimulation potential may be one of the strategies that tumors have developed to escape immunosurveillance.

Figure 5

Five human tumors have T cell co-stimulation potentials higher than correlated healthy tissues (changes ?0), including in uterine tumor, germ cell tumor, leukemia, glioma, and chondrosarcoma. A. Stimulation potential during tumor genesis B. The correlation between tumor prognosis and co-stimulation potential. Prognosis data is based on the effects of the immune infiltrate cell types discussed in a published article (PMID: 28741618).

Moreover, it has been reported that localized co-stimulation is important for tumor-specific T cell responses (45). We hypothesize that tumorgenesis mediated increase in the expression of CSRs would be correlated with the prognosis of patients with tumors. To examine this issue, we analyzed the new data of the effects of immune infiltrate on the prognosis of patients with cancer, which is discussed in Fridman, WH et al Nature review (Figure 5B) (44). This data was a summary of as many as 200 studies. We defined a type of combined prognosis scores (PS) as follows: 1) No effect of immune infiltrate on prognosis was counted as (0); 2) the strong negative prognosis effect of immune infiltrate was counted as (-2);3) less strong negative prognosis effect of immune infiltrate was counted as (-1);4) positive prognosis effect of immune infiltrate was counted as (+1); and 5) strong positive prognosis effect of immune infiltrate was counted as (+2). The correlation analyses were conducted between the increased co-stimulation potential during tumorigenesis in human tumors shown in Figure 5A and the final prognosis values for six reported immune infiltrated cell types including CD8+ T cells, tertiary lymphoid structure, regulatory T cells (Treg), CD68⁺ macrophages, proinflammatory classically activated macrophages (M1 macrophages), and alternatively activated macrophages (M2 macrophages), respectively. As shown in Figure 5B, among six cell types analyzed, we found that the increased co-stimulation potential in nine human tumors including breast tumor, lung tumor, liver tumor, ovarian tumor, bladder carcinoma, gastrointestinal tumor, esophageal tumor, glioma and kidney tumor was significantly correlated, R²= 0.4851 (p 0.371), with improved final prognosis values based on decreased M2 macrophage infiltrates in the human tumors. The results suggest that first, increased co-stimulation potential during tumorigenesis have potential for predicting the prognosis of patients with tumor; and second, increased co-simulation potentials during tumorigenesis in human tumors are associated with decreased infiltration of M2 anti-inflammatory macrophages in tumors, implying a beneficial prognosis of type I pro-inflammatory macrophages infiltrated in tumors.

We hypothesized that tissue expression of co-signaling receptors are either under the regulation of tissue physiological status (staying downstream) or regulating tissue physiological status (staying upstream) as defined by the expression levels of inflammation-related DAMPRs (46), cellular stress-regulators, hypoxia regulators, blood supply-related vascular endothelial growth factors (VEGFs), tissue regeneration-related four stem cell regulators (Yamanaka factors: octamer-binding transcription factor4 (POU5F1, Oct3/4), sex determining region Y)-box 2 (Sox2),Kruppel-like factor 4(Klf4),c-Myc)), epigenetic modulators such as cellular hyper-, hypo- methylation status and cell markers of three professional antigen presenting cell types. To test this hypothesis, based on the numbers of co-signaling receptors expressed in 32 tissues/27co-signaling receptors, we determined co-signaling receptor potentials. To determine the extent to which factor expression and that of co-signaling receptors are related, we conducted correlation studies.

We hypothesized that if there is a positive correlation, the expression of PAMP receptors/metabolite-regulated DAMP receptors (PRRs) and inflammasome components), the factors (such as oxygen sensors, genes that regulate angiogenesis pathway, stem cell master genes, and the co-signaling receptors will modulate their expression accordingly (22). As shown in Figure 6, among 14 genes examined highly expressed co-signaling receptor potentials in tissues were significantly correlated with that of NOD3 (r² = 0.3.068; p = 0.0075), NALP9 (r² = 0.2262; p = 0.0253) and CARD8 (r² = 0.2791; p = 0.0115). Similarly, high and low expressed co-signaling receptor potential in tissues was significantly correlated with that of NOD4 (r² = 0.1822; p = 0.0476); and NALP6 (r² = 0.1825; p = 0.0474) (Figure 7).

Figure 6

The expression of NLRs/ inflammasome sensors are correlated in low levels with the expression of a few highly expressed co-signaling receptors in human tissues. (Human tissues analyzed: adipose tissue, bladder, blood, bone, bone marrow, brain, embryonic tissue, eye, heart, kidney, liver, lung, lymph node, muscle, pancreas, skin, spleen, thymus, placenta, umbilical cord, uterus, vascular). Abbreviations: NOD 1/2/3/4: Nucleotide-binding oligomerization domain-like receptors 1/2/3/4; NALP 2/3/6/9/14: Nod like receptor (NLR) family pyrin domain containing 2/3/6/9/14; NAIP: NLR family apoptosis inhibitory protein; NLRC4: NLR family CARD domain containing 4; ASC: PYD and CARD domain containing; IFI16: interferon gamma inducible protein 16; CARD8: caspase recruitment domain family member 8.

Figure 7

The expression of NLRs/ inflammasome sensors are correlated in low levels with the expression of some co-signaling receptors in human tissues. (Tissues analyzed: adipose tissue, bladder, blood, bone, bone marrow, brain, embryonic tissue, eye, heart, kidney, liver, lung, lymph node, muscle, pancreas, skin, spleen, thymus, placenta, umbilical cord, uterus, vascular). Abbreviations: NOD1/2/3/4: Nucleotide-binding oligomerization domain-like receptors1/2/3/4; NALP2/3/6/9/14: NLR family pyrin domain containing 2/3/6/9/14; NAIP: NLR family apoptosis inhibitory protein; NLRC4: NLR family CARD domain containing 4; ASC: PYD and CARD domain containing; IFI16: interferon gamma inducible protein 16; CARD8: caspase recruitment domain family member 8.

To further consolidate our findings, we used microarray data deposited in the NIH/NCBI-Geo datasets and examined the expression of NOD3, NOD4, CARD8, NALP6 and NALP9 in the deficiencies/knock-down of four CSRs CD80, CD48, TIM and CD40 datasets versus wild-type control microarray data. The results shown in Figure 8 demonstrated that these co-signaling receptors regulate the expression of NOD3, NOD4, CARD8, NALP6 and NALP9 (p 0.05). These results suggest that first, T cell reverse signaling via co-signaling receptors may regulate the expression of some inflammasome components NOD3, NALP9, CARD8, NOD4 and NALP6 in the antigen presenting cells to regulate the DAMPs-, and PAMPs- sensing pathways (Figure 9), which correlated with a recent report on the roles of type 1 T helper cell (Th1) and Th17 in regulating innate immunity for bacterial clearance (47, 48); and/or second, these five inflammasome components may initiate signaling in regulating co-signaling receptor expression in human tissues as the parts of inflammation initiation, suggesting that innate immune sensor components regulate T cell activation (49).

Figure 8

Deficiencies of co-signaling receptors modulate the expression of NOD3, CARD8, NALP6, NALP9 and VEGFR3 (FLT4) (p <0.05). This suggests that T cells regulate the inflammasome pathways in antigen presenting cells via reverse signaling of co-signaling receptors

Figure 9

Reverse signaling from T cells regulate the expression of inflammasome components.

We examined the hypothesis that tissue physiological functional status such as hypoxia responsive pathways, vascular growth factor pathways, stem cell master gene-regulated tissue regeneration pathway and mitochondrial activity pathway may also regulate co-signaling receptor expression in human tissues. To test this hypothesis, we performed correlation analysis for the expression of the genes involved in these pathways with co-signaling receptors. These genes included seven vascular endothelial growth factor (VEGF) pathway genes (VEGFA, VEGFB, VEGFC, C-fos-induced growth factor (FIGF), VEGF receptor 1 (VEGFR1, FLT1), VEGFR2 (KDR), and VEGFR3 (FLT4)), six stem cell regulator genes (CD34, Kit, Myc, Klf4, Pou5F1(Oct3/4), and Sox2) (50), and three mitochondrial activity regulators (transcription factor A, mitochondrial (TFAM), nuclear respiratory factor 1 (NRF1)(51), and the transcription factor nuclear factor erythroid 2 p45-related factor 2, NRF2)(52). Co-signaling receptor expression potentials in human tissues were significantly correlated with that of VEGFR3 (FLT4) (r² = 0.3873; p = 0.0020) (Figure 10). To further consolidate the findings, we used the microarray data deposited in the NIH/NCBI-Geo datasets and examined VEGFR3 expression in the presence of deficiencies/ knock-down of co-signaling receptors such as CD80, CD48, TIM and CD40. The results in Figure 11 showed that CD40 knock-down with RNAi upregulates VEGFR3; and that VEGFR3 knock-down with RNAi regulates the expression of seven co-signaling receptors including four CSRs such as SEMA4A, TIM4, B7-H2, CD48 and three CIRs including Galectin 9, B7-DC and HVEM. These results suggest that in response to hypoxia, blood supply and angiogenesis changes, stem cell regulation and mitochondrial activity changes, all the co-signaling receptors process behavior almost the same; and VEGFR3, as a key vascular growth factor receptor, plays an essential function in regulating these co-signaling receptor expression and/or co-signaling receptor functions may regulate the VEGFR3 expression and lymphatic vascular development in human tissues (53).

Figure 10

Human vascular endothelial cell growth factor receptor 4 (FLT4) expression in tumor tissues are correlated with the expression of highly expressed co-signaling receptors in human tissues. A. Correlation between co-stimulation receptor potential and genes that regulate hypoxia, angiogenesis, stem cells and mitochondrial signaling pathways. B. FLT4 gene is significantly correlated with co-signaling receptor potential. (Tissues analyzed: adipose tissue, bladder, blood, bone, bone marrow, brain, embryonic tissue, eye, heart, kidney, liver, lung, lymph node, muscle, pancreas, skin, spleen, thymus, placenta, umbilical cord, uterus and vascular). Abbreviations: PHD2: Prolyl hydroxylase domain-containing protein-2; HIF1B: Hypoxia-inducible factor-1 beta; HIF1/2A: Hypoxia-inducible factor 1/2-alpha; VEGFA/B/C: Vascular Endothelial Growth Factor A/B/C; FIGF: C-fos-induced growth factor; FLT1/4: Fms related tyrosine kinase ¼; KDR: Kinase insert domain receptor; MYC: MYC proto-oncogene; KIT: KIT proto-oncogene receptor tyrosine kinase; KLF4: Kruppel like factor 4; POU5F1: POU class 5 homeobox 1; SOX2: SRY-box 2; TFAM: transcription factor A; NRF1: nuclear respiratory factor 1; NRF2(GABPA): GA binding protein transcription factor alpha subunit.

Figure 11

nock down of VEGFR3 (FLT4) regulates the expression of co-signaling receptors. This suggests that VEGFR3 regulates the expression and functions of co-signaling receptors in VEGFR3+ endothelial cells and other VEGFR3+ cells.

It has been reported that treatment with DNA methylation inhibitors and histone deacetylase inhibitors activate a growth-inhibitingimmuneresponsewhich can be an effective therapeutic approach for malignant disorders (54). We hypothesized that certain co-signaling receptor expression are under epigenetic regulation such as methylation as we reported for enzyme expression of homocysteine-methionine metabolism pathways. Thus, we analyzed the correlation between mouse co-signaling receptor expression and mouse tissue methylation indices determined by the ratios between S-adenosyl methionine (SAM – the universal methyl donor)/ S-adenosyl homocysteine (SAH - a methyltransferase inhibitor) in mouse tissues (55). As shown in Figure 12A and Figure 12B, we found that relative expression levels of 10 co-signaling receptors in mouse tissues including CIR B7-H1 (r² = 0.9668, P = 0.0004), CIR BTNL2 (r² = 0.9242, P = 0.0022), CSR CD40 (r² = 0.932, P = 0.0018), CSR GITRL (r² = 0.9242, P = 0.0022), CSR CD30L(r² = 0.9242, P = 0.0022), CSR TIM4 (r² = 0.9318, P = 0.0018), CSR SLAM(r² = 0.9242, P = 0.0022), CSR B7-H2 (r² = 0.8807, P = 0.0056), dual function B7-2 (r² = 0.884, P = 0.0053), CSR CD48 (r² = 0.7487, P = 0.0457) are correlated with tissue SAM/SAH levels (hyper-methylation index). These results suggest that first, the expression of 35.7% (10/28) co-signaling receptors may regulate tissue hyper-methylation status or tissue hyper-methylation status may regulate the expression of one-third of co-signaling receptors in mouse tissues; and second, during tumorigenesis certain tumors can switch methylation/demethylation regulation modes on these co-signaling receptors and upregulate these co-signaling receptors, after treated with DNA methylation inhibitors and/or histone deacetylase inhibitor(s), then anti-tumor immune responses can be enhanced.

Figure 12

The tissue hypermethylation indices as evaluated by S-adenosylmethionine (SAM)/ S-adenosylhomocysteine (SAH) ratio are correlated with the expression of co-signaling receptors in six mouse tissues (tissues: liver, brain, heart, kidney, lung and spleen). A. The correlation between relative expression of co-stimulation receptors and SAM/SAH ratio. B. Depending on the R2 value, the correlation was segmented in to three tiers, named high, middle and low. C) Schematic representation: 1) Histones H3 and H4 methylation promote genes transcription; the methylation residues of H3 are R2, 8, 17, 26 and K4, 9, 23, 27, 36, 79; the methylation residues of H4 are R3 and K20 (PMID: 22982198). 2) CpG island methylation promote co-signaling receptor genes transcription. (PMID: 15506922).

Since the main types of professional APCs are dendritic cells (DCs), macrophages, and B cells (56), we hypothesized that highly expressed co-signaling receptors in tissues are correlated with the expression of macrophage marker adhesion G protein-coupled receptor E1 (F4/80), B cell marker CD20, T cell markers CD3G, CD3E and CD3D and DC marker CD11C in human tissues. To test this hypothesis, we performed correlation analysis for the expression of these professional APC markers with co-signaling receptor potential in 22 human tissues. As shown in Figure 13A and Figure 13B, human tissue expression levels of macrophage marker F4/80 (r² = 0.1259, p = 0.0463), B cell marker CD20 (r² = 0.2392, p = 0.0045), and T cell antigen receptor components CD3G (r²= 0.1565, p = 0.0250) and CD3E (r² = 0.3646, p = 0.0003) are significantly correlated with the high expressed co-signaling receptor potential. However, tissue expression levels of DC marker CD11C and T cell marker CD3D are not correlated with the high expressed co-signaling receptor potential. These results suggest that first, tissue expression of high expressed co-signaling receptors are partially correlated with tissue contents of professional APCs; and second, the expression of highly expressed co-signaling receptors in non-professional APCs in tissues contribute significantly to forward signaling in regulating T cell activation and reverse signaling in regulating tissue physiology and pathology as demonstrated by our report on CD40 (11).

Figure 13

Expression levels of macrophage marker F4/80, B cell marker CD20, and T cell antigen receptor components CD3G and CD3E in human tissues are significantly correlated with the high expressed co-signaling receptor potential. A) Correlation between macrophage, dendritic cells, T-cell and B-cell markers with highly expressed co-signaling potential. B) The cell markers that are significantly correlated with the expression of highly expressed co-signaling potential.

We hypothesized that reverse signaling by upregulated co-stimulation receptors in macrophages may contribute to infections and macrophage polarization. To test this hypothesis, we examined six macrophage microarray datasets of infectious disease including streptococcus pneumonia, leptospiral, influenza H1N1, influenza H3N2, Legionella pneumophila, and listeria, respectively. As shown in Table 7, among 26 upregulated co-signaling receptors, 22 receptors were CSRs (84.6%) whereas other 4 receptors were CIRs (15.4%). Also, the co-stimulation receptors including CD40, 4-1BBL, SLAM, CD48, B7-1, B7-2, and CD155 are significantly upregulated in macrophages in response to infections. In addition, 2 CIRs such as B7-DC and B7-H1 are upregulated in macrophages in response to infection.

Furthermore, we examined five macrophage microarray datasets of type 1 macrophage (M1) polarization versus M0 non-polarized macrophage controls and five microarray datasets of M2 polarization versus M0 non-polarized macrophage controls. We found that 1) 22 CSRs (73.3%) and 8 CIRs (26.7%) were upregulated during M1 polarization; 2) 6 CSRs and 1 CIR were upregulated during type 2 macrophage (M2) polarization; 3) CSRs such as CD40, 4-1BBL, TL1A, CD30L, SLAM, CD48, SEMA4A, B7-1, B7-2, and CD155 are significantly upregulated in M1 macrophage polarization. In contrast, only two CSRs including CD40 and SLAM are upregulated in M2 macrophages; 4) 4 co-inhibition receptors such as CD113, B7-DC, B7-H1, and HHLA2 are upregulated in M1 macrophage polarization; and 5) CD70, GITRL, TIM4, CD112, HVEM, B7-H3, B7-H4 and BTNL2 have no significant changes under M1 or M2 polarization. These results suggest that reverse signaling of CSRs than CIRs play an essential role in macrophages during infections. Further, many CSRs are involved in M1 polarization while only few CSRs play a role in M2 polarization.

We previously proposed endothelial cells as innate immune cells, which can act as APC in pathological conditions by upregulating MHC class II and co-stimulation receptors (56). We also reported that vascular smooth muscle cells upregulate innate immune sensor caspase-1/inflammasome pathway in chronic kidney disease (CKD) (31, 57). Thus, we hypothesized that vascular cells including endothelial cells (EC) and vascular smooth muscle cells (VSMC) may upregulate co-stimulation receptors and co-inhibition receptors in pathologies. We analyzed microarray datasets conducted on endothelial cells and VSMCs stimulated with pro-inflammatory cytokines. Additionally, our analysis also included microarray datasets conducted on leukocytes extracted during infections, trauma injury and septic shock. As shown in Table 8, we found that 1) VSMCs upregulate co-stimulation receptors B7-H2, CD40, SEMA4A, B7-1, B7-2, and CD155 after stimulation with interleukin-1beta (IL-1beta); 2) In contrast, endothelial cells upregulate four CSRs B7-H2, CD40, SEMA4A and CD112 and four CIRs Galectin 9, HVEM, B7-DC, and B7-H1 (PD-L1) after stimulation with tumor necrosis factor-alpha (TNF-alpha) and interferon-gamma (IFN-gamma) (58); and 3) three CSRs such as LIGHT, SEMA4A, and CD58 were significantly upregulated and SLAM, B7-2, and CD112 were downregulated in leukocytes extracted during severe trauma and septic shock. These results suggest that first, forward and reverse signaling of three out of 18CSRs, B7-H2, CD40, and SEMA4A (16.7%),play significant roles in vascular cells stimulated with pro-inflammatory cytokines such as TNF-alpha and IFN-gamma; second, pro-inflammatory cytokines TNF-alpha and IFN-gamma also upregulate five out of ten CIRs (50%) in endothelial cells, suggesting that endothelial cells play an essential role in immune tolerance, anti-inflammatory responses, and inflammation resolution, presumably via their forward signaling and reverse signaling; and third, three CSRs, LIGHT, SEMA4A and CD58, upregulated in trauma and septic shock injury in leukocytes have specific forward and reverse signaling in the pathologies.

To further consolidate our findings that reverse signaling of CSRs and CIRs in the pathophysiology of vascular cells, we hypothesized that CSRs and CIRs play important roles in EC activation (22) and VSMC phenotypic switch by modulating the expressions of EC adhesion molecules and VSMC phenotypic switching markers. To examine this hypothesis, we conducted a literature search and selected 20 EC adhesion molecules and 20 VSMC phenotypic switching markers. As shown in Figure 14, we found that: 1) CD30L (a CSR) transgenic overexpression in spleen induces the endothelial cell adhesion molecule FN1 and VSMC phenotypic switching marker alpha-SMA upregulation; 2) B7-H4 (a CIR) overexpression in pancreas downregulates CD31 and upregulates VSMC markers S100A4 and GP110; CD40 RNA interference (RNAi) in human umbilical vascular endothelial cells (HUVEC) induces the upregulation of EC adhesion molecules AGT, TGFB1 in HUVEC and upregulation of VSMC markers MMP-1, MMP-3, and SMMHC; 3) CD70 (a CSR) transgenic overexpression in hematopoietic stem cells (HSC) induces the upregulation of EC adhesion molecule VCAM-1, and VSMC marker cyclin; 4) In CD48 (CSR) negative cells, VSMC marker TIMP-2 was upregulated; and finally, 5) CSR CD112 deficiency induces upregulation of EC adhesion molecules COL18A1, and E-selectin in HUVEC, and downregulates VSMC markers TIMP-1, ACLP, smoothelin, and SM22-alpha. These results suggest that reverse signaling of CSRs and CIRs regulates the pathophysiology of vascular cells by modulating the expression of endothelial cell adhesion molecules and VSMC phenotypic switching markers.

Figure 14

Reverse signaling of CSRs and CIRs regulate pathophysiology of vascular cells. A) Modulation of CSRs and CIRs affect endothelial cell activation and vascular smooth muscles phenotypic switch. B) Schematic representation of modulation co-signaling receptors in vascular inflammatory disorders induce endothelial activation by increasing the presence of adhesion molecules. Further, changes in the co-signaling receptors induce the vascular smooth muscles to undergo phenotypic changes.

To further consolidate our findings that reverse signaling of CSRs and CIRs in the pathophysiology of professional APCs, endothelial cells (EC) and other cells, we extensively searched previous literature. As shown in Figure 15, we found that, 1) in addition to co-stimulation for T cell activation, 12 CSRs including B7-1, B7-2, CD70, CD40, 4-1BBL, OX40L, TL1A, GITRL, CD30L, SLAM, CD155, CD112, and 2 CIRs Galectin9, HVEM have reverse signaling pathways in professional APCs, ECs, monocytes, and other cells; 2) The most common pathway involved in reverse signaling in various cells is mitogen activated protein kinase (MAPK) pathway; 3) Other downstream pathways involved are phosphatidylinositol-3 kinases (PI3K)-AKT (a serine/threonine kinase also known as protein kinase B, PKB), indoleamine 2,3-dioxygenase (IDO), tumor necrosis factor receptor associated factors(TRAF)2/3/4/5/6, signaling lymphocytic activation molecule (SLAM)-associated protein (SAP), focal adhesion kinase (FAK) and caspase-1; 4) CSRs have signaling pathways in various innate immune and non-immune cells; and 5) CSRs regulate various immunological and cell biological roles including apoptosis, autoantibody generation, EC activation (adhesion molecule upregulation), immunoglobulin class switching, expression of CCR7, HLA-DR, and CD86, osteroclastogenesis, cell adhesion and migration, cytotoxicity, cell proliferation.

Figure 15

In addition to co-stimulation for T cell activation, 12 co-stimulation receptors including B7-1, B7-2, CD70, CD40, 4-1BBL, OX40L, TL1A, GITRL, CD30L, SLAM, CD155, CD112, and 2 co-inhibition receptors Galectin9, HVEM have reverse signaling pathways in professional APCs, endothelial cells, monocytes, and other cells. A) The signaling pathways that are involved in reverse signaling. MAPK pathway is the most common pathway involved in reverse signaling. B) Schematic representation of the biological roles played by the signaling pathways that mediate reverse signaling.

For those less-characterized CSRs and CIRs including five CSRs (B7-H2, LIGHT, TIM4, CD48 and CD58) and seven CIRs (CD113, B7-DC, B7-H3, B7-H4, VISTA, HHLA2 and BTNL2), we also searched the structural evidences of reverse signaling by analyzing the candidate phosphorylation sites in their intracellular domains and protein interaction partners. As shown in Table 9, we found that 1) Most co-signaling receptors have their N-terminus localized intracellularly; 2) all 12 CSRs and CIRs have the intracellular domains and potential phosphorylation sites; and 3) B7-H4 and VISTA has two interaction partners, while all the other receptors have more than three interaction partners. These analyses suggest that all the CSRs and CIRs have intracellular domain with phosphorylation site capable for reverse signaling and regulating the pathophysiology of professional APCs, vascular cells and other cells in addition to regulate T cell activation and other T cell functions including priming, differentiation, effector and memory functions.

It has been well accepted that all the CSRs and CIRs are expressed in professional antigen presenting cells (APCs) and “conditional APCs” including endothelial cells as we proposed (56). Recent reports showed that an increased number of cell types that are capable of acting as atypical APCs and present antigens to major histocompatibility complex (MHC) class II molecules to conventional CD4⁺ T cells (37). Of note, CD4⁺ T cells are among atypical APCs. However, it is unknown what master regulators control the development of T cell plasticity into atypical APCs. We hypothesized that when key transcription factors in T cells have mutations, T cells upregulate MHC class II and CSRs and CIRs. To examine this hypothesis, we analyzed MHC class II expression and the expression of CSRs and CIRs in six T cell master gene mutations. As shown in Figure 16, we found that: 1) T-helper-inducing POZ/ Kruppel-like factor (Thpok, cKROX) deficiency upregulated CIR CD113, CSR B7-1 and downregulated CSRs CD40 and SLAM; 2) Myb deficiency upregulated CSR CD30L and CIR VISTA but downregulated CSRs SLAM and B7-2; 3) signal transducer and activator of transcription 4 (STAT4) deficiency downregulated B7-1; 4) transcription factor 3 (TCF3, E2 alpha) deficiency upregulated CSRs LIGHT and SLAM and downregulated CSR CD48; 5) canonical Wnt/β-catenin effector (Tcf1) deficiency upregulated CSRs LIGHT, CD155 and CD112, but downregulated T cell antigen receptor beta (TCR-beta); and finally 6) STAT1 deficiency (59) upregulated four CSRs such as CD40, CD30L, B7-1, and CD112 and three CIRs including CD113, B7-DC, and BTNL2. In addition, STAT1 deficiency upregulated MHC class II expression. Taken together, STAT1 deficiency not only upregulates MHC class II for delivering T cell activation signal 1, but also upregulates four CSRs for delivering T co-stimulation signal as T cell activation signal 2 (2, 4). Therefore, our results have demonstrated for the first time that STAT1 inhibits T cell plasticity into atypical APCs; and other transcription factors also modulate the expression of CSRs and CIRs in T cells.

Figure 16

Transcription factors modulate the expression of CSRs and CIRs in T-cells. Stat1 attenuate T-cell plasticity and prevents its conversion to APCs. A. The deficiencies of master genes such as THPOK, MYB, STAT1, and TCF1 in T cells upregulate co-signaling receptors and major histocompatibility complex class II (MHC class II) (STAT1 deficiency). B. Schematic representation of how the mutations in T-cell master regulators can modulate T-cell plasticity.