Recombinant human IL-17, recombinant human TNFα, and recombinant porcine IFN-γ were purchased from R&D Systems (Minneapolis, MN, USA). Anti-actin antibody was purchased from Cell Signaling Technology (Boston, MA, USA), anti-occludin antibody was obtained from Thermo Fisher Scientific (Rockford, IL, USA), anti-FITC-labeled SLA class I antibody (SLA-I) was obtained from Bio-Rad (Hercules, CA, USA), and Cell Counting Kit-8 (CCK8) was purchased from Dojindo Laboratories (Kumamoto, Japan). The CCR2 (CCL2 receptor)-specific inhibitor RS504393 was from MedChemExpress (Shanghai, China).

PAECs were isolated from wild-type or GGTA1/CMAH double-knockout (DKO) Chinese domestic miniature Wuzhishan pig arteries as previously described (17). In brief, freshly harvested porcine arteries were treated with 0.05% collagenase B (Roche Applied Science, Indianapolis, IN, USA). We collected the cells and washed them with washing medium [RPMI 1640 containing 10% heat-inactivated fetal bovine serum (FBS)]. The isolated PAECs were cultured with endothelial cell medium containing 10% (vol/vol) FBS (cat. no. 0025), 1% (vol/vol) penicillin/streptomycin solution (P/S, cat. no. 0503), and 1% (vol/vol) endothelial cell growth supplement (ECGS, cat. no. 1052) at 37°C with 5% CO_2. The culture medium was purchased from Sciencell (San Diego, CA, USA).

PAECs were harvested after washing with ice-cold phosphate-buffered saline (PBS), lysed for 30 min in ice-cold RIPA lysis buffer supplemented with 10 mM NaF, 1 mM Na₃VO₄, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor cocktail (Roche, Indianapolis, IN, USA), and separated via 10% SDS–PAGE. After transfer onto polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA), the proteins on the PVDF membranes were blocked using 5% non-fat dried milk dissolved in TBST (20 mM Tris-HCl, 150 mM NaCl, pH 7.6) buffer supplemented with 0.1% (vol/vol) Tween 20 at room temperature for 1 h. After washing, the PVDF membranes were incubated with primary antibody overnight at 4°C and then washed with TBST. After incubation with secondary antibody for 1 h at room temperature, the blots were visualized with enhanced chemiluminescence detection reagents (Millipore).

The procedure for real-time PCR has been reported previously (18). Briefly, total RNA was extracted from cells or tissues with TRIzol^® Reagent (Invitrogen, Shanghai, China). cDNA samples were synthesized with PrimeScript™ RT Master Mix (Takara Bio, Dalian, LN, China). The levels of the genes of interest were quantified using TB Green^® Premix Ex Taq ^™ (Tli RNaseH Plus) (Takara Bio). The expression level of the genes was calculated using the 2^-ΔΔCt method and normalized to the rpl13a expression level. cDNA amplification was performed using a ViiA 7 Real-Time PCR system (Applied Biosystems, Foster City, CA, USA), and the oligonucleotide primer sequences are shown in Supplementary Table S1 .

SLA-I binding was assessed using flow cytometry as previously reported (12). Porcine aortic endothelial cells were collected and washed once with PBS, and 1 × 10⁶ cells were resuspended in 100 μl PBS containing 1% BSA and then stained with FITC-labeled SLA-I antibody. Isotype-matched antibody was used as a negative control. After incubation for 30 min at 4°C in the dark, the cells were washed once with PBS, resuspended in 100 μl PBS containing 1% BSA and detected with a BD FACS Aria II flow cytometer (Franklin Lakes, NJ, USA). The extent of SLA-I binding to PAECs was evaluated by geometric mean fluorescence intensity (Gmean).

The protocol for human antibody-mediated complement-dependent cytotoxicity assessment has been previously reported (11). In brief, PAECs (4 × 10³) were seeded into 96-well plates and treated with medium (as a negative control), hIL-17 (100 ng/ml), hTNF-α (2 ng/ml), or hIL-17 plus hTNF-α for 48 h. The supernatant was removed and replaced with RPMI-1640 medium containing 20% pooled human serum [experimental group—the serum was pooled from several healthy volunteers (n = 20), including all ABO blood types] or 20% heat-inactivated human serum (control group) for 2 h. After 2 h, CCK8 assays were used to assess the viability of PAECs. The supernatant was removed and replaced with RPMI-1640 medium containing 10% CCK8 reagent for 2 h. At 2 h later, the absorbance values of the wells at OD450 were measured using a multiscan GO spectrophotometer (Thermo Fisher, Vantaa, Finland). The percentage of cell death (cytotoxicity) was calculated according to the following formula:

The chemotaxis assay procedure has been reported previously (19). In brief, isolation of human neutrophils using a human neutrophil isolation kit was performed according to the manufacturer’s instructions. The chemotaxis assay was performed in 24-well plates using 6.5 μm (for THP‐1) or 3 μm (for human neutrophils) transwell inserts with 5-mm pore polycarbonate membranes (Corning). The PAECs were treated with hIL-17 (100 ng/ml), hTNF-α (2 ng/ml), or hIL-17 plus hTNF-α for 48 h, the supernatant was collected and then added to the lower transwell chamber, and 25 × 10⁴ THP-1 or human neutrophils were seeded in the upper chamber. The transmigration assay was performed for 2 h at 37°C. For CCR2 inhibitor assay, THP-1 cells were treated with 10 μM RS504393 for 1 h and then seeded in the upper chamber. The membranes containing the migrated cells were carefully excised. Images were obtained with a microscope, and the migrated cells were counted. The migration data are presented as the number of migrating cells/field.

PAECs were treated with rhIL-17 (100 ng/ml), rhTNF-α (2 ng/ml), or rhIL-17 plus rhTNF-α for 0 or 6 h. RNA was extracted from the treated PAECs using TRIzol reagent (Invitrogen) and was utilized to construct the final library. The sequencing library was generated using the VAHTS mRNA-seq v2 Library Prep Kit for Illumina^® (Vazyme, NR601, Nanjing, JS, China) following the manufacturer’s recommendation. Library concentration was measured using a Qubit^® RNA Assay Kit and Qubit^® 3.0 for preliminary quantification. Insert size was assessed using an Agilent Bioanalyzer 2100 system, and after the insert size was consistent with expectations, the qualified insert size was accurately quantified via qPCR using a Step One Plus Real-Time PCR system (Applied Biosystems). Clustering of the index-coded samples was performed on a cBot Cluster Generation System (Illumina, San Diego, CA, USA) according to the manufacturer’s instructions. After cluster generation, the library preparations of a total of 12 samples were sequenced on an Illumina HiSeq X Ten platform using a 150-bp paired-end module.

Whole-transcriptome sequencing data were filtered and mapped to the porcine genome. The differentially expressed gene-seq method was based on a Poisson distribution (fold change >1.5 and adjusted P-value <0.05). Additive genes and synergistic genes (ASGs) were screened according to the definition. The whole-transcriptome sequencing data can be found in NCBI (https://dataview.ncbi.nlm.nih.gov/object/PRJNA779585?reviewer=3areeogpjo6vkvlr0i3hhb1s64) under accession number PRJNA779585.

According to the Kyoto Encyclopedia for Genes and Genomes (KEGG) annotation results and the official classification, we separately classified the functional and biological pathways of the differentially expressed genes and used R software for enrichment analysis. Bubble charts, heat maps, and volcano maps were generated to visualize the differentially expressed genes on the basis of log-normalized expression values of significant genes using v3 R software.

The additive genes and synergistic genes were defined according to a previous report (15). A gene was considered to be synergistically induced by IL-17/TNF-α when the combined effect was greater than the sum of the separate effects. A gene was considered to be additive when the combined effect was greater than the two individual effects but lower than the sum of the separate effects. In brief, we defined (x) to represent the log₂-FCH (fold change) induced in hIL-17-treated PAECs (y) to represent the log₂-FCH induced by hTNF-α and (x + y) to represent the log₂-FCH induced by the combination of hIL-17 and hTNF-α in treated PAECs. Two tests were then used to define whether a gene was synergistic or additive, and the gene was excluded if it was antagonistic and induced by hIL-17 and hTNF-α. The flow chart was shown in Supplementary Figure S1 . First, we used hypothesis testing to test whether the effect regulated by the two cytokines together was different than the sum of the individual effects, i.e.:

Then, we tested whether the combined effect was different than both individual effects, i.e,:

A gene was considered synergistic if it passed test 1 and |(x + y)| > |(x) + (y)|. A synergistic increase is the difference |(x + y)| - |(x) + (y)|. Synergism can be either positive (if the synergist increase is positive) or negative.

A gene was considered additive (if it is not synergistic) if the combined effect was greater than both individual effects, i.e., the gene passes test 2 for both x and y.

The synergistic genes are highlighted in red, and the additive genes are shown in black in Supplementary Tables S2, S3 (Ho: the null hypothesis; Ha: the alternative hypothesis).

Experimental data are presented as mean ± SEM. GraphPad Prism 5 software is used to perform statistical analysis and graph development. Statistical significance between the groups was calculated using two-tailed Student’s t-test using Microsoft Office Excel software. p-values <0.05 were considered significant.

Human IL-17 plus hTNF-α additively or synergistically regulated the expression of 697 genes in PAECs ( Figure 1A , Supplementary Tables S2, S3 ). A total of 315 genes were upregulated in PAECs, 166 of which were synergistically induced, while another 149 genes were additively induced ( Supplementary Table S2 ). A total of 382 genes were downregulated in PAECs, 249 of which were synergistically regulated, while 133 other genes were additively regulated ( Supplementary Table S3 ). These ASGs were subjected to KEGG pathway enrichment analysis and annotation. Many immune-related signaling pathways, such as the TNF-α, IL-17, MAPK, and Toll-like receptor signaling pathways, were obviously enriched ( Figure 1B ). In addition to these pathways, the cytokine–cytokine receptor interaction was also enriched ( Figure 1B ). The KEGG pathway annotation analysis demonstrated that 159 genes were related to signal transduction ( Supplementary Figure S2 ). Among the organismal systems, the immune system represented the top enrichment, involving 86 genes ( Supplementary Figure S2 ), suggesting that hIL-17 plus hTNF-α primarily activated immune-related genes in PAECs. Cytokines and chemokines play critical roles in xenograft rejection, and we found that various proinflammatory cytokines, such as IL1α and IL6, and chemokines, including CCL2, CCL11, CXCL8, and CXCL2, were upregulated ( Figure 1C ). In contrast, the anti-inflammatory gene IL10 was downregulated ( Figure 1C ). The ligand–receptor analysis identified that CCL11, IL1α, IL6, IL11, and their receptors were upregulated, while IL-10-IL10RB and KITLG-KIT were obviously downregulated ( Figure 1D ). We used RT-PCR to validate the expression of several cytokines and chemokines and found that CCL20, CSF3, IL11, and CXCL2 were slightly induced by IL-17 or TNF-α alone but dramatically induced by IL-17 plus TNF-α, suggesting that these genes were synergistically induced ( Figure 2 ). CCL11 and IL1α were additively induced by IL-17 plus TNF-α ( Figure 2 ). These results were consistent with the transcriptome sequencing data ( Supplementary Table S2 ). Collectively, the data suggest that hIL-17 and hTNF-α synergistically induce proinflammatory cytokine and chemokine expression to amplify the inflammatory response.

Due to several gene-modified pigs used for xenotransplantation, we asked whether human IL-17 and TNF-α could synergistically or additively induce proinflammatory cytokine and chemokine expression in PAECs from gene-modified pigs. We isolated PAECs from α1,3-galactosyltransferase (GGTA1) and cytidine monophosphate-N-acetylneuraminic acid hydroxylase (CMAH) gene double-deficient pigs. The two genes are responsible for the synthesis of Galα-1,3-Gal (Gal) and N-glycolylneuraminic acid (Neu5Gc), two carbohydrate xenoantigens that are respectively important for xenotransplantation, and the GGTA1/CMAH DKO pigs could largely reduce immune rejection in xenotransplantation. We treated the GGTA1/CMAH DKO PAECs with hIL-17, hTNF-α, or hIL-17 plus hTNF-α and found that IL-17 plus TNF-α synergistically induced the expression of CCL20, CSF3, and CXCL2 and additively induced the expression of CCL11, IL1α, and CXCL8 ( Supplementary Figure S3 ). The data suggest that human IL-17 and TNF-α synergistically or additively induce proinflammatory cytokine or chemokine expression not only in wild-type PAECs but also in GGTA1/CMAH DKO PAECs. Thus, IL-17 and TNF-α likely promote the rejection of GGTA1/CMAH DKO xenograft in xenotransplantation.

We also found that IL-17 and TNF-α additively or synergistically regulate the expression of 65 cell surface proteins ( Supplementary Figure S4 ): 29 genes were upregulated, and 36 genes were downregulated. The regulated genes included several receptors, adhesion molecules, and tight junction genes, such as IL1R1, ICAM1, and occludin (OCLN). These genes might increase proinflammatory signaling, promote immune cell migration, and enhance xenoantibody-mediated complement-dependent cytotoxicity (CDC). Moreover, some of these regulated cell surface genes might be xenoantigens, which will be investigated in the future.

The transcriptome sequencing data and our previous study showed that hIL-17 and hTNF-α synergistically induce CXCL8 expression and additively induce CCL2 expression in PAECs ( Figure 1C and Supplementary Table S2 ) (12). We also validated their expression levels via RT-PCR analysis ( Figure 3A ). The chemokines CXCL8 and CCL2 have neutrophil and monocyte chemotactic activity, respectively; thus, we asked whether the supernatant of hIL-17- or hTNF-α-treated PAECs had monocyte and neutrophil chemotactic activity. We found that the supernatant of hIL-17- or hTNF-α-treated PAECs alone increased human neutrophil migration and that the supernatant of hIL-17 plus hTNF-α-treated PAECs had greater chemotactic activity than the supernatant of IL-17- or TNF-α-treated PAECs alone ( Figures 3B, C ). In a THP-1 (a human leukemia monocytic cell line) cell migration assay, the results were similar to those of human neutrophil migration ( Figures 3B, C ). In order to investigate whether IL-17 or TNF-α directly mediated THP-1 cell migration, we collected the supernatant from PAECs without IL-17 or TNF-α treatment and then added IL-17 plus TNF-α to the supernatant for THP-1 cell migration assay. We found that the supernatant with IL-17 plus TNF-α did not enhance THP-1 cell migration, suggesting that IL-17 or TNF-α cannot directly mediate THP-1 cell migration ( Supplementary Figure S5 ). To confirm whether IL-17 plus TNF-α increased THP-1 cell migration through inducing CCL2 expression, we used RS504393, a specific inhibitor of CCR2 which is the receptor for CCL2, treated THP-1 cells, and found that RS504393 almost completely blocked THP-1 cell migration, which suggest that IL-17 plus TNF-α increases THP-1 cell chemotaxis through enhancing CCL2 expression ( Supplementary Figure S5 ). Collectively, these data suggest that hIL-17 plus hTNF-α increases human neutrophil or monocyte chemotaxis and that the increased chemotaxis is likely due to increased CXCL8 or CCL2 production by PAECs.

Previously, we found that hIL-17 plus hTNF-α increased the mRNA level of F3, a procoagulation gene, in PAECs (12). In the present study, we found that F3 was induced in PAECs treated with hIL-17 plus hTNF-α ( Figure 4A ). In addition to F3, we found that the mRNA level of another pro-coagulation gene (SERPINB2) was upregulated, while the mRNA levels of three anti-coagulation genes (TFPI, THBS1, and THBD) were downregulated in PAECs treated with hIL-17 plus hTNF-α ( Figures 4A, B ). To validate the data, we measured the mRNA levels of these genes in PAECs treated with hIL-17 or hTNF-α and found that SERPINB2 was increased, while TFPI, THBS1, and THBD were downregulated in PAECs treated with hIL-17 plus hTNF-α ( Figure 4C ). The mRNA levels of THBS1 and THBD were slightly downregulated by IL-17 or TNF-α alone and largely reduced by IL-17 plus TNF-α. Collectively, these data suggest that hIL-17 and hTNF-α amplify the coagulation reaction in response to xenografts through additive or synergistic effects.

Tight junction genes play an important role in xenoantibody-mediated CDC (20, 21). In the present study, we found that OCLN, ESAM, CLDN6, and CDH5 were downregulated and that JAM2, CLMP, BVES, and CLDN1 were upregulated in PAECs treated with hIL-17 plus hTNF-α ( Figure 5A ). Moreover, we found that hTNF-α alone decreased OCLN expression in PAECs and that the mRNA and protein levels of occludin were much lower in PAECs treated with hIL-17 plus hTNF-α than in PAECs treated with TNF-α alone ( Figures 5B, C ). We previously found that hTNF-α promoted human antibody-mediated CDC in porcine ECs by downregulating occludin expression; thus, we asked whether hIL-17 plus hTNF-α had stronger PAEC cytotoxicity in a human-mediated CDC model. We found that TNF-α alone exerted increased PAEC cytotoxicity, while hIL-17 plus hTNF-α led to more PAEC death than TNF-α alone ( Figure 5D ). To exclude the direct cytotoxicity of hIL-17 or hTNF-α to PAECs, we treated PAECs with or without hIL-17, hTNF-α, or hIL-17 plus hTNF-α for 48 h and assess the viability of PAECs with CCK8. We found that the OD450 of the negative control group was almost equal to hIL-17, hTNF-α, or hIL-17 plus the hTNF-α treated group, suggesting that hIL-17 or hTNF-α does not directly affect the viability of PAECs ( Supplementary Figure S6 ). These data suggest that hIL-17 and hTNF-α likely increase xenoantibody-mediated CDC, which leads to xenograft injury.

We previously found that hTNF-α could induce SLA-I expression; therefore, we asked whether hIL-17 plus hTNF-α could additively or synergistically increase SLA-I expression (12). Surprisingly, we found that hTNF-α significantly induced SLA-I expression; however, hIL-17 almost completely blocked hTNF-α-mediated SLA-I upregulation ( Figure 6A ). We also found that hIL-1β increased SLA-I expression and that hIL-17 slightly reduced hIL-1β-mediated SLA-I upregulation ( Figure 6B ). Moreover, porcine IFN-γ (pIFN-γ) increased SLA-I expression, but IL-17 did not decrease IFN-γ-mediated SLA-I expression ( Figure 6C ). These data suggest that IL-17 has a suppressive effect on TNF-α- or IL-1β-mediated SLA-I upregulation.

Human IL-17 and hTNF-α additively or synergistically regulated the expression of various genes in PAECs, and we wondered whether hIL-17, in combination with hIL-1β or pIFN-γ, had additive or synergistic effects in PAECs. We found that hIL-17 and hIL-1β synergistically induced IL6 expression and additively induced E-selectin, ICAM-1, VCAM-1, CXCL8, CCL2, and CXCL2 expression ( Figure 7A ). Human IL-17 and pIFN-γ also synergistically induced IL6 expression and additively induced E-selectin, ICAM-1, VCAM-1, and CCL2 expression ( Figure 7B ). However, human IL-17 and pIFN-γ did not additively or synergistically induce CXCL8 or CXCL2 expression ( Figure 7B ). These data suggest that hIL-17 and hIL-1β or pIFN-γ can additively or synergistically induce the expression of specific proinflammatory cytokines, chemokines, and adhesion genes in PAECs.