N-Octanoyl-Dopamine inhibits cytokine production in activated T-cells and diminishes MHC-class-II expression as well as adhesion molecules in IFNγ-stimulated endothelial cells

IFNγ enhances allograft immunogenicity and facilitates T-cell mediated rejection. This may cause interstitial fibrosis and tubular atrophy (IFTA), contributing to chronic allograft loss. We assessed if inhibition of T-cell activation by N-octanoyl dopamine (NOD) impairs adherence of activated T-cells to endothelial cells and the ability of activated T-cells to produce IFNγ. We also assessed if NOD affects IFNγ mediated gene expression in endothelial cells. The presence of NOD during T-cell activation significantly blunted their adhesion to unstimulated and cytokine stimulated HUVEC. Supernatants of these T-cells displayed significantly lower concentrations of TNFα and IFNγ and were less capable to facilitate T-cell adhesion. In the presence of NOD VLA-4 (CD49d/CD29) and LFA-1 (CD11a/CD18) expression on T-cells was reduced. NOD treatment of IFNγ stimulated HUVEC reduced the expression of MHC class II transactivator (CIITA), of MHC class II and its associated invariant chain CD74. Since IFTA is associated with T-cell mediated rejection and IFNγ to a large extent regulates immunogenicity of allografts, our current data suggest a potential clinical use of NOD in the treatment of transplant recipients. Further in vivo studies are warranted to confirm these in vitro findings and to assess the benefit of NOD on IFTA in clinically relevant models.


Results noD impairs adhesion of activated t cells to cytokine stimulated endothelial cells. Both TNFα
and IFNγ increase the expression of adhesion molecules on endothelial cells and consequently are expected to increase cell adhesion of activated T-cells to such treated monolayers of endothelial cells. As depicted in Fig. 1, anti-CD3/anti-CD28 activated T-cells indeed adhered significantly better to stimulated -, as compared to unstimulated endothelial cells. When T-cells were activated in the presence of 100 µM of NOD, T-cell adhesion was Hereafter the supernatants were replaced by normal medium and adhesion of activated T-cells was assessed as described above. Note that at high dilutions (1/8 and above), supernatants from T-cells that were stimulated in the presence of NOD did not significantly differ from the medium control. (**p ≤ 0.01/***p ≤ 0.001; supernatants of stimulated T-cells without NOD (+stim/−NOD) vs. supernatants of stimulated T-cells with NOD (+stim/+NOD), ns; not significant/**p ≤ 0.01; medium with NOD (+stim/+NOD) vs. supernatant of stimulated T-cells with NOD (+stim/+NOD)). For (A,B) all conditions were tested in triplicates. The results of at least 3 independent experiments are expressed as mean fluorescence + SD. Data were analysed using two-way ANOVA followed by Sidak's multiple comparison test. A p-value < 0.05 was considered to be significant. strongly diminished irrespective of treatment of the endothelial cells (Medium mean RFU ± SD: 34,387 ± 3,509 vs. 17,766 ± 1,236; p < 0.0001; TNFα mean RFU ± SD: 49,542 ± 1,916 vs. 25,556 ± 3,074; p < 0.0001; IFNγ mean RFU ± SD: 41,785 ± 3,113 vs. 22,506 ± 2,935; p < 0.0001; −NOD vs. +NOD). For NOD treated T-cells, adhesion to TNFα or IFNγ treated endothelial cell monolayers was slightly higher as compared to unstimulated monolayers (mean RFU ± SD: 17,766 ± 1,236 vs. 25,556 ± 3,074 vs. 22,506 ± 2,935, p < 0.01) (Fig. 1A).
We next assessed the propensity of supernatants obtained from anti-CD3/anti-CD28 activated T-cells to facilitate T-cell adhesion to monolayers of endothelial cells. To this end, cultured endothelial cells were stimulated overnight with serial dilution of anti-CD3/anti-CD28 activated T-cell supernatants or supernatants of T-cells activated in presence of 100 µM of NOD. While no differences in T-cell adhesion were observed between both types of supernatants in a 1 to 2 dilution, T-cell adhesion decreased to the level of the control culture medium when higher dilutions of supernatants from T-cells that were activated in the presence of NOD were used (Dilution 1/64 mean RFU ± SD: 40,180 ± 2,031 vs. 28,149 ± 1,462 vs. 26,812 ± 3,009; −NOD vs. +NOD vs. medium control p < 0.01) (Fig. 1B).
VLA-4 (CD49d/CD29) and LFA-1 (CD11a/CD18) expression on T-cells are decreased by NOD. In keeping with the observation that T-cell adhesion to endothelial cells is diminished when T-cells are activated in the presence of NOD, we addressed if this was reflected by a reduced expression of T-cell ligands for adhesion molecules. Because the expression of ICAM-1 and VCAM-1 on endothelial cells is of pivotal importance for leucocyte extravasation, we assessed in particular if NOD influences the expression of the known ligands for VCAM-1 and ICAM-1 on T-cells, i.e. VLA-4 and LFA-1 respectively. Upon stimulation with anti-CD3/anti-CD28 both the α-(CD49d) and ß-chain (CD29) of VLA-4 were increased (

Induction of endothelial VCAM-1 and ICAM-1 expression by supernatants of activated t-cells.
To test if the propensity of anti-CD3/anti-CD28 activated T-cell supernatants to facilitate T-cell adhesion correlates with their ability to increase VCAM-1 and ICAM-1 expression on endothelial cells, we incubated endothelial cells overnight with these supernatants. While anti-CD3/anti-CD28 activated T-cell supernatants clearly increased the expression of both adhesion molecules even when used at high (1/16) dilutions, supernatants obtained from T-cells that were activated in the presence of 100 µM of NOD did not induce VCAM-1 at all and ICAM-1 only to a little extent (Fig. 3A). We also assessed the concentration of TNFα and IFNγ in these supernatants. As depicted in Fig. 3B, activated T-cells produce approximately 6-fold more IFNγ as compared to TNFα (Mean production ng/ml ± SD 206.20 ± 0.44.30 vs. 34.04 ± 2.84, IFNγ vs. TNFα, p < 0,01) (Fig. 3B).
influence of noD on ifnγ mediated gene expression in endothelial cells. Inasmuch as the presence of IFNγ was pronounced in supernatants of activated T-cells, we subsequently tested if NOD affects IFNγ mediated gene expression in endothelial cells. To this end, genome wide gene expression profiling was performed in endothelial cells stimulated with IFNγ either in the presence or absence of 50 µM of NOD 21 . Gene set enrichment analysis (GSEA) was performed to identify pathways in the Kyoto Encyclopedia of Genes and Genomes (KEGG) database that were influenced by NOD. Enlisted in Fig. 4A is a selection of pathways that are down-regulated when endothelial cells were stimulated with IFNγ in the presence of NOD. Based on an adjusted p-value (p adj ) for significance p adj < 0.05, only 7 of these pathways were significantly affected. The strongest inhibitory effects of NOD were found in the Graft-versus-host disease pathway (Normalised enrichment score (NES): −2.39, p adj = 0.007), Allograft rejection pathway (NES: −2.25, p adj = 0.007), Type I diabetes mellitus pathway (NES: −2.14, p adj = 0.007) and Antigen processing and presentation pathway (NES: −1.96, p adj = 0.007). Figure 4B shows a selection of pathways that were up-regulated in the presence of NOD of which only 3 pathways remained significant after adjustment for multiple testing. At the single gene level genes belonging to MHC class II (HLA-DOA,-DPA1,-DPB1,-DRA,-DRB1,-DMA,-DMB), the MHC class II antigen-associated invariant chain (CD 74), and MHC class II transactivator (CIITA) were identified amongst the genes that were at least 2-fold down regulated by NOD. Likewise interferon regulated factor-1 (IRF-1), a number of interferon-induced proteins (IFI30, -IFI35, IFI44 and IFI44 like) and interferon-induced protein with tetratricopeptide repeats (IFIT3, and IFIT2) as well as genes that are known to be up-regulated by IFNγ, e.g. CX3CL1 23 , TAP1 24 and, IDO1 25 were down regulated by NOD. Yet, for a number of these genes these changes did not reach statistical significance after Bonferroni correction (Table 1).
ifnγ-mediated expression of CD74, CIITA and HLA-DR is inhibited by NOD. Since gene expression profiling data suggested that NOD might affect MHC class II expression in endothelial cells, we first assessed by TaqMan PCR to what extent CD74 and CIITA mRNA expression was influenced when endothelial cells were stimulated for 24 hrs with IFNγ in the presence or absence of NOD. As depicted in Fig. 5A both CD74 and CIITA mRNA were strongly induced by IFNγ. In the presence of NOD the up-regulation of CD74 and CIITA mRNA was significantly blunted (Fig. 5A). The diminished CIITA mRNA expression translated in a reduced expression of CIITA protein, which became more pronounced after 72 hrs (Fig. 5B). In line with this, it was found that IFNγ

Discussion
Even though current state-of-the-art immunosuppressive strategies to a large extent have contributed to the success of contemporary transplantation medicine, they failed to interrupt the insidious process causing chronic allograft loss. While CNI based immunosuppressive therapies are highly efficacious in preventing acute T-cell mediated rejection, they are also associated with significant adverse effects such as hypertension and nephrotoxicity 26 . Attempts to minimize or wean patients off of CNI have shown that improvement in renal function is often obtainable but only at the expense of increased alloimmune reactions 27,28 . The development of a CNI-based, long-term, maintenance-immunosuppressive drug regimen with improved long-term tolerability is therefore a highly desirable endeavour.
We and others have previously reported that N-acyl dopamine derivatives (NADD) harbour biological properties that might be useful for preventing transplantation associated injury [29][30][31] . As such they limit cold inflicted 32 and ischemia reperfusion-injury 33 and modulate inflammation through inhibition of NFκB 21,34 . We also have reported that NOD, a short synthetic NADD, transiently inhibits T-cell activation and demonstrates synergy with CNI in suppressing T-cell activation 22 .
The results presented herein are in good agreement with our previous results that NOD inhibits anti-CD3/ antiCD28 mediated T-cell activation. This is reflected by a reduction in cytokine production and a diminished expression of VLA-4 and LFA-1 on the cell surface of T-cells. Because VLA-4 can mediate both rolling and arrest of T-cells on endothelial cells 35 , this explains why T-cells that are activated in the presence of NOD display decreased adherence to endothelial cells despite cytokine mediated upregulation of VLA-4's major counterreceptor VCAM-1. The reduced LFA-1 -ICAM-1 interaction is likewise contributing to the diminished adhesive properties of such activated T-cells. In addition to VCAM-1 and ICAM-1 also junctional adhesion molecules (JAM-A, -B and -C) have been implicated in regulating leucocyte extravasation by affecting different steps in this process through their interaction with endothelial cells. Endothelial JAM-A and JAM-C interact with the ß2 integrins LFA-1 and Mac-1 (CD11b/CD18) to sustain firm adhesion and transmigration 36,37 . Although in the present study transmigration was not assessed, our data do suggest that extravasation of T-cells would be impaired as a consequence of a reduced CD18 expression if T-cell activation occurs in the presence of NOD.  Supplementary Fig. 1. Densitometry graphs with significance marks are provided in Supplementary Fig. 2. (B) TNFα and IFNγ concentrations were assessed by ELISA in the T-cell supernatants described in A. Values represent mean cytokine concentration + SD from three independent experiments. Data were analyzed using two-way ANOVA followed by Sidak's multiple comparison test. Significance was defined as p-value < 0.05 (**p < 0.01, ***p < 0.001) for stimulated/−NOD compared to stimulated/+NOD.  www.nature.com/scientificreports www.nature.com/scientificreports/ The results of three independent experiments are expressed as the mean fold change (2 −ΔCT ) + SD. Data were analyzed using two-way ANOVA followed by Sidak´s multiple comparison test on the normalized ΔCT values. Significance was defined as p-value < 0.05 (*p < 0.05; ***p < 0.001) for +IFNγ/−NOD compared to +IFNγ/+NOD. (B) CIITA protein expression was assessed by Western-Blotting using cell-lysates obtained from HUVECs cultured for various times as described in A. The results of a representative blot are depicted, a total of 3 independent experiments with different HUVEC cultures were performed. CIITA protein expression was quantitated by means of densitometry of all independent blots. Data were analyzed using two-way ANOVA followed by Sidak´s multiple comparison test. Significance was defined as p-value < 0.05 for +IFNγ/−NOD compared to +IFNγ/+NOD. Equal protein loading was demonstrated by staining for β-actin. Displayed are the cropped blots. The scanned full-length blots are provided in Supplementary Fig. 3 www.nature.com/scientificreports www.nature.com/scientificreports/ In addition to its effect on T-cell activation, gene expression profiling revealed that NOD also affects the transcriptome of IFNγ stimulated endothelial cells. Significant normalised enrichment scores were amongst others found in the Graft-versus-host disease pathway, Allograft rejection pathway, Type I diabetes mellitus pathway and Antigen processing and presentation in all of which MHC class I and II, as well as IFNγ have been implemented. This not necessarily indicates that NOD per se interferes with IFNγ signalling, since the influence of NOD on other pivotal mediators of these pathway have not been studied. However, because both MHC Class II and IFNγ are of pivotal importance in terms of tissue immunogenicity allograft rejection, the influence NOD on IFNγ mediated MHC class II expression was studied in more detail. Although other pathways were also influenced by NOD, e.g. ribosome, p53 mineral absorption, they were not further studied because an apparent link with tissue immunogenicity or allograft rejection was lacking.
MHC class II proteins bind microbial peptides in an endosomal compartment and present them to T-cells as part of immune surveillance. While in the endoplasmic reticulum the hydrophobic peptide-binding groove in MHC class II is protected by the invariant chain CD74 38 , in the late endosomal compartment CD74 is degraded leaving class II-associated invariant chain peptides (CLIP) in the peptide binding groove 39,40 . HLA-DM catalyzes CLIP dissociation, stabilizes empty MHC class II proteins, and enables rapid binding of microbial peptides 41 . Since at the single gene level, NOD inhibits the transcription factor for MHC class II expression, i.e. CIITA, as well as HLA-DR, HLA-DM and CD74 expression, our data suggest that NOD might impair MHC class II presentation. Yet, endothelial cells are not endowed with a professional antigen presentation function as they fail to express classical co-stimulatory molecules. Nonetheless increased MHC class I and II expression on endothelial cells may render the allograft more susceptible to alloantigen-dependent vascular damage and to the development and progression of transplant vasculopathy. It needs to be assessed if and to what extent MHC expression and antigen presentation on professional antigen presenting cells is affected by NOD.
The present study has some limitations that should be addressed. In paticular, the mechanism by which NOD inhibits IFNγ mediated gene expression is currently unclear. While the catechol structure provides NOD with strong antioxidant properties, the hydrophobic fatty acid of NOD might allow its passage across membranes to access different intra-cellular compartments and change the redox milieu within these compartments. As such NOD may impair oxidative protein folding in the endoplasmic reticulum (ER) and consequently induce the unfolded protein response (UPR) 42,43 , or inhibit redox dependent transcription factors in the nucleus, e.g. NF-κB, AP-1 and NFAT 22 . Cross-talk between IFNγ signalling and the UPR may partly underlie the inhibitory effect of NOD as the UPR is associated with transient inhibition of mRNA translation 44 . Alternatively, changes in the redox milieu of the cell nucleus may impair epigenetic processes required for gene transcription 45 . Indeed, gene expression profiling studies on endothelial cells have shown that histone acetyl transferase expression was affected by NOD. Even though all the above mentioned mechanisms may explain the in vitro effects of NOD, further in vivo experiments are warranted to elucidate which of these putative mechanisms also occur in vivo. The NOD concentrations used in the current study were based on previous in vitro studies 22,43 , yet the a fast blood clearance, predominantly via renal and hepatobiliary elimination 46 does not allow plasma NOD concentrations to reach the in vitro concentrations used in our current study. Despite this short-coming, in vivo studies demonstrated that NOD is highly efficacious in the treatment of brain dead donors on early post-transplant renal and heart function 47,48 , in ameliorating inflammation in ischemic acute kidney injury 33 and in attenuating the development of transplant vasculopathy in rat aortic allografts 49 . Nonetheless, it remains to be elucidated to what extent the in vivo efficacy of NOD is mediated by its ability to inhibit IFNγ mediated gene expression.
In conclusion, this study provides further evidence for the immunosuppressive properties of NOD on T-cell activation. Not only the present study suggests a potential clinical use for NOD to limit the development of IFTA in allograft recipient, but it also warrants further in vivo testing in relevant models for transplantation. Methods cell culture. Human umbilical vein endothelial cells (HUVECs) were isolated from fresh umbilical cords as described previously 50 . Confluent monolayers were passaged by trypsin 0.025%/ EDTA 0.01% and mostly used for experiments after 2-3 passages. Umbilical cords were obtained from the Department of Obstetrics, University Medical Center Mannheim, after written informed consent of the patients. www.nature.com/scientificreports www.nature.com/scientificreports/ (Novaliq GmbH, Heidelberg, Germany). T-cells were harvested by FACS and supernatants were used for assessment of TNF-α and IFNγ and their ability to induce VCAM-1 und ICAM-1 expression on endothelial cells. The propensity to adhere-, or facilitate adhesion to endothelial cells was tested for T-cells and supernatants.
Adhesion assay. Confluent HUVEC monolayer were stimulated for 24 hours with either TNFα (10 ng/ml) (PeproTech, Hamburg, Germany) or IFNγ (50 ng/ml) (Sigma-Aldrich, St. Louis, USA) or left in normal culture medium. In some experiments HUVEC monolayers were incubated with a serial dilution (1/2 to 1/16) of T-cell supernatants. T-cells were stimulated as described above, labelled with CytoTell Green (AAT Bioquest, Sunnyvale, USA) as recommended and added to the endothelial monolayers for 1 hour in a concentration of 10 6 T-cells per well. Before addition of labelled T-cells, the HUVEC monolayers were washed with PBS to remove the stimulating cytokines. Non-adherent T-cells were removed by washing and the remaining cells were lysed in 10 mM Tris, 2% SDS. Fluorescence was measured on a Tecan Spark 10 M with the appropriate filters (Tecan Group, Männedorf, Switzerland). All experimental conditions were performed in triplicate. Results are expressed as mean relative fluorescence units (RFU) ± SD.
Western-blotting. Western-Blotting was performed as described before 21 . Samples (20 µg protein extract) were separated on 10% SDS-polyacryamide gels followed by semi-dry blotting onto PVDF membranes (Roche, Basel, Switzerland). Membranes were blocked for 1 hour at room temperature in tris buffered saline (TBS) (Bio-Rad, Hercules, USA) and incubated overnight at 4 °C with one of the following primary polyclonal antibodies: anti-VCAM-1, anti-ICAM-1 (both from R&D Systems, Wiesbaden, Germany) or anti-CIITA (Santa Cruz Biotechnology, Dallas, USA). Hereafter, the membranes were thoroughly washed with TBS 0,1% Tween-20 and incubated with the appropriate horseradish peroxidase conjugated secondary antibody (Jackson ImmunoResearch, Baltimore, USA). Proteins were visualized using enhanced chemo luminescence technology (Pierce, Rockford, USA). To confirm equal protein loading, membranes were stripped and re-probed with monoclonal anti-β-actin (Santa Cruz Biotechnology, Dallas, USA) antibody. and CD74 (ID: Hs00269961_m1). Samples were run under the following conditions: initial denaturation for 10 min at 95 °C followed by 40 cycles of 15 s at 95 °C and 1 min at 60 °C. The levels of gene expression in each sample were determined with the comparative cycle threshold method. PCR efficiency was assessed from the slopes of the standard curves and was found to be between 90% and 100%. Linearity of the assay could be demonstrated by serial dilution of all standards and cDNA. All samples were normalized for an equal expression of β-actin. fAcS analysis. For FACS analysis HUVEC were incubated with anti-HLA-DR PE conjugated antibody (BD-Biosciences, Heidelberg, Germany). Anti-CD3/anti-CD28 activated T-cells four color-staining was performed using the following monoclonal antibodies: anti-CD49d-PE, anti-CD29-APC, anti-CD11a-BV421 and anti-CD18-FITC (all purchased from BD Biosciences, Heidelberg, Germany). For each antibody optimal dilutions were assessed by serial dilutions. Stained samples were measured on a BD FACSlyricTM Flow cytometer using both negative and fluorescence-minus-one controls. Compensation was performed with BD compBeads (BD Biosciences, Heidelberg, Germany). Data analysis was performed using FlowJo software version 5.2 (FlowJo LLC, Ashland, USA).

Statistical analysis.
A Custom CDF Version 21 with ENTREZ based gene definitions was used to annotate the arrays 51 . The raw fluorescence intensity values were normalized applying quantile normalization and RMA background correction. One-way ANOVA was performed to identify differential expressed genes using a commercial software package SAS JMP10 Genomics, version 6, from SAS (SAS Institute, Cary, USA). A false positive rate of a = 0.05 with FDR correction was taken as the level of significance. Gene Set Enrichment Analysis (GSEA) was used to determine whether defined lists (or sets) of genes exhibit a statistically significant bias in their distribution within a ranked gene list using the software R v3.4.0 52 and RStudio: Integrated development environment for R (RStudio Boston, USA). The fgsea package 53 was used to determine whether defined lists (or sets) of genes exhibit a statistically significant bias in their distribution within a ranked gene list. Pathways belonging to various cell functions such as cell cycle or apoptosis were obtained from public external databases (KEGG, http://www. genome.jp/kegg).