Quantitative Proteomic Analysis of Cardiac Xenograft Failure in a Pig‐to‐Non‐Human Primate Model Identifies NF‐κB as a Critical Immunomodulatory Target

ABSTRACT

Introduction

Shortage of donor organs is one of the greatest challenges of cardiac transplantation. Xenotransplantation is a potential way to solve the contradiction of imbalance and pigs are considered ideal donor sources. However, xenotransplantation still faces the problem of immune rejection at present. In order to further understand the molecular picture of immune rejection after xenotransplantation, and develop immunosuppressive agents to further overcome rejection, we conducted a proteomic analysis of a heterotopic pig‐to‐non‐human primate (NHP) animal model.

Methods

We constructed a heterotopic NHP animal model using wild‐type (WT) and alpha‐1,3‐galactosyltransferase gene knockout (GTKO) porcine hearts as donors. Based on quantitative proteomics analysis, we investigated the changes of protein after CXTx in three groups: Group I: WT donor heart, Group II: GTKO donor heart without immunosuppression, and Group III: GTKO donor heart with immunosuppression. Finally, we assessed the efficacy of the target using a heterotopic heart transplantation model from SD rats to Balb/c mice.

Results

A total of 2425 proteins were identified in the donor heart tissues and approximately 15% of proteins were significantly changed after CXTx, most of them had increased expression. The results of proteomic analysis demonstrated that chronic hypoxia injury induced by microvascular thrombosis may play an important role during cardiac xenograft failure, confirmed by histopathological results. Remarkably, we showed some novel targets especially increased expression of pentraxin 3, MVP, and HSP90AB1 that cannot be suppressed in the present gene editing and immunosuppressive interventions. Because NF‐κB is a common downstream regulator of these three proteins, we hypothesize that it may be crucial to the occurrence of xenograft failure and considered as a potential therapeutic target. Using the SD Rat‐Balb/C Mouse CXTx model and inhibiting NF‐κB with BAY 11–7082, we found that NF‐κB targeting prolonged graft survival from 5 to 8 days and reduced myocardial inflammation.

Conclusions

In summary, the proteomic analysis could help us to solve the mystery of cardiac xenograft failure, confirm the key pathways, and reveal a clear vision of future interventions. NF‐κB inhibition effectively decreased immune cell infiltration and antibody deposition in myocardial tissue, suggesting its potential as a therapeutic strategy to enhance graft survival and reduce inflammation in cardiac xenotransplantation (CXTx).

Abbreviations

CXTx

cardiac xenotransplantation

DEP

differentially expressed protein

GO

gene ontology

GTKO

α‐1,3‐galactosyltransferase knockout

HC

Hierarchical Clustering Analysis

IS

immunosuppression

NHP

pig‐to‐non‐human primate

PCA

principal component analysis

PPI

protein–protein interaction

WGCNA

weighted protein co‐expression network analysis

1. Introduction

Heart transplantation is the ultimate approach for the treatment of end‐stage heart failure patients [1]. The contradiction of the imbalance between the supplement and demand for the donor's heart is prominent and become the major bottleneck for clinical application [2]. Xenotransplantation may be a practical solution to the problems faced by clinicians. Pig is considered the most appropriate species for cardiac xenograft due to its adequate size, physiological similarity to humans, and low risk of zoonosis [3]. Life‐supporting porcine cardiac xenotransplantation (CXTx) has been reported to survive more than half a year due to advances in gene editing and surgical techniques [4]. Xenotransplants are prone to graft failure because of the species barrier, including a complex series of reasons such as hyperacute rejection, acute rejection, acute and chronic rejection, and vasculopathy [5]. The researchers have recently concentrated on trying to prolong the survival time of cardiac xenograft recipients. However, improved understanding of molecular mechanisms underlying cardiac xenotransplant failure has received little attention, which supports researchers in identifying immunosuppressive targets and promotes xenograft in clinical application.

Proteomics has been successfully used to identify the molecular mechanisms of unrecognized pathophysiological pathways in patients with heart diseases [6]. Recent advances in mass spectrometry techniques allow proteomics to analyze thousands of proteins simultaneously [7]. The introduction of proteomics in clinical research increased the efficiency and accuracy because the results drive discovery. In the field of heart transplantation, this technique has been previously used to select biomarkers in predicting heart transplantation outcomes [8]. The pathological mechanism of graft failure especially for CXTx remained poorly investigated.

The aims of the present study were to evaluate the molecular responses of xenograft failure at the tissue proteome level. A pig‐to‐non‐human primate model (NHP) was constructed to perform CXTx. We successively investigated three groups which were classified by different kinds of donor heart and postoperative management. Group I (WT‐Post): The donor porcine heart was wild‐type (WT) and without postoperative immunosuppression, Group II (GTKO‐Post‐NonIS): Donor's heart was from genetically modified homozygous α‐1,3‐galactosyltransferase knockout pig (GTKO) and without postoperative immunosuppression, Group III (GTKO‐Post‐IS): Donor's heart was from GTKO pig and with postoperative immunosuppression. In addition, WT (WT‐Pre) and GTKO (GTKO‐Pre) porcine hearts that were not transplanted were used as controls, respectively. Microvascular thrombosis was found in all three kinds of cardiac xenograft. The activation of the complement system and coagulation was indicated by proteomic results. Bioinformatic analysis showed that differentially expressed proteins (DEPs) were significantly enriched in several processes related to protein synthesis. Furthermore, we identified several important DEPs such as PTX3, MVP, and HSP90AB1. They were consistently high expression after CXTx in all groups, which may be considered as a potential target for further treatment plan optimization. We provided key information on molecular pathways when using the porcine heart with different genetic backgrounds as donor and differential immunosuppression. It could help clinicians improve their understanding of CXTx and identify better immunosuppressive targets.

2. Methods

2.1. Ethics Compliance

All the animals used in this study were provided by the Laboratory Animal Center of Fuwai Hospital. The experimental procedures were approved by the Ethics Committee of Fuwai Hospital and all animals were used in compliance with the guidelines formulated by the Laboratory Animal Center of Fuwai Hospital and Chinese legislation.

2.2. Xenotransplantation and NHP Animal Model

Six male Rhesus macaques (weighted around 10 kg and 8–10 years of age) were used as recipients and six male Bama mini pigs (weighed around 5 kg and 3–6 months of age) were used as donors. According to the study design, the genetic background of donor pigs included two WT pigs (Group I: WT1, WT2) and four GTKO pigs (Group II: GTKO‐Non‐IS1, GTKO‐ Non‐IS 2; Group III: GTKO‐IS1, GTKO‐IS2).

The abdominal heterotopic CXTx surgeries were performed by an experienced surgical team in Fuwai Hospital. In brief, the donor pig's chest was opened by a median sternotomy. Inferior and right superior vena cava are ligated, thereafter, the inferior vena cava, left superior vena cava and left atrium were incised. After clamping the ascending aorta, the cold University of Wisconsin solution, approximately 400 mL, was administered retrogradely via the aorta. Perfusion pressure was monitored and recorded, while the aortic root and left ventricle were palpated manually to ensure effective perfusion and to avoid high tension in the left ventricle. The donor's heart was removed by dividing the left atrium, vena cava, pulmonary artery, and aorta. After the donor's heart was taken out of the chest, the left atrium and valvula of inferioris venae cavae were sewed. The recipients were subjected a midline abdominal incision, abdominal aorta, and inferior vena cava were subsequently isolated. Side‐biting clamps were applied then aortotomy and venotomy were performed. The aorta of the donor porcine heart was anastomosed end‐to‐side with the aorta of the recipient, and the pulmonary artery of the donor's heart was anastomosed with the inferior vena cava of the recipient. After clamp removal, the cardiac xenografts were automatically resuscitated and restored to sinus rhythm. After transplantation, the graft underwent regular and long‐term monitoring, particularly through echocardiography, due to its abdominal location. The transplanted heart in the abdomen was evaluated every 6 h using ultrasound and palpation under stable conditions. In addition to echocardiography, continuous monitoring of vital signs such as electrocardiography (ECG), blood pressure, and blood oxygen levels was also conducted. If any instability was detected in the recipient monkey's vital signs or the pig heart's ECG, an immediate ultrasound and palpation assessment were performed to evaluate cardiac function and determine if rejection had occurred. If the heart was found to have ceased beating, the heart was then flushed retrogradely using UW solution and saline. It was then excised by severing the connections between the heart the abdominal aorta and the inferior vena cava. This comprehensive monitoring approach ensured timely detection and intervention during the observation period.

The immunosuppressive regimen was only applied to the recipients in Group III. In short, anti‐CD20 antibody (Rituximab, Rituxan, Genetech, CA, USA) and anti‐CD25 antibody (Basilixumab, Simulect, Novartis Pharma, NJ, USA) were used for induced therapy. Methylprednisolone (MP), mycophenolate mofetil (MMF), and cyclosporin A were used to maintain long‐term immunosuppression. Ganciclovir (Roche, Nutley, NJ, USA) and nystatin were used to prevent viral and fungal infections. All the details of the immunosuppressive regimen including dose and application were in Table 1. In the GTKO donor heart with immunosuppression group, blood drug levels were monitored 1 day post‐transplantation. The cyclosporine levels in two cases exceeded 200 ng/mL, and MMF levels exceeded 1 µg/mL.

TABLE 1.

Immunosuppressive regimen of GTKO recipients.

Agent	Route	Dose and timing
Induction
Rituximab	i.v. infusion	19 mg/kg; days −3, 0, 7, and 14
MP	i.v. infusion	500 mg; immediately after anesthesia and the close of abdomen
Basilixumab	i.v. infusion	10 mg; 1 h before surgery and 4 days after surgery
Maintenance
MMF	o.s.	0.5 g; twice a day then daily
Cyclosporin A	o.s.	25 mg; twice for the first day 50 mg; twice from the second day
MP	i.v. infusion	120 mg/8h before the removing the endotracheal intubation in the first day, start at 10 mg/kg with the reduction of 10 mg every 3 days until the dose of 10 mg/d
Additive therapy
Ganciclovir	o.s.	0.125 g; daily
Nystatin	o.s.	500 000 U; daily

Abbreviations: i.v., intravenous; MP, Methylprednisolone; MMF, mycophenolate mofetil; o.s., oral administration; U., units.

As a control group, the WT pigs or GTKO pigs were sampled directly without undergoing transplantation, and we refer to this control group as the “pre” group. The hearts in the pre‐group were obtained following the same standardized procedure as the donor hearts in the transplantation group.

2.3. TMT Labeling and Fractionation

Tandem mass tag (TMT) labeling was employed to quantify the DEPs in each sample. In brief, each reaction consisted of a set of five kinds of samples including (1) WT‐Pre, (2) GTKO‐Pre, (3) WT‐Post, (4) GTKO‐Post‐NonIS, and (5) GTKO‐Post‐IS. Each kind sample included two biological replicates and a total of 10 samples were labeled using a TMT‐10plex. As the limitation of biological replicates, we performed an additional TMT‐10plex test as technical replication in order to ensure the statistical analysis reliability. The dried peptide samples were resuspended in 100 µL of 100 mM triethylammonium bicarbonate (TEAB). Each sample containing 20 µg peptides was labeled using a commercial TMT‐10plex reagent kit (TMT, Thermo, Pierce Biotechnology, Rockford, IL, USA). The TMT‐labeled peptides were mixed and desalted using SPE cartridges before further fractionation. The detailed procedure was described in our previous publication [9].

The peptide mixture was fractioned by an Ultra Performance Liquid Chromatography (Ultimate 3000, Thermo‐Fisher Scientific, Waltham, MA, USA) with an XBridge C18 RP column (250 × 4.6 mm, 5 µm, Waters, Milford, MA, USA). Peptides were separated in a binary buffer system of A (98% H₂O and 2% acetonitrile, pH 10.0) and B (98% acetonitrile and 2% H₂O, pH 10.0). The gradient of buffer B was set as follows: 5% to 8%, 5 min; 8% to 18%, 25 min; 18% to 32%, 32 min; 32% to 95%, 6 min; 95% to 5% B, 5 min. A total of 48 fractions were collected and were combined into 12 fractions. Finally, all peptide fractions were dried in a vacuum concentrator and then were resolved in 0.1% FA for further LC‐MS/MS analysis.

2.4. Liquid Chromatography Mass Spectrometry‐Based Proteomics Analysis

LC‐MS/MS analysis was performed on a Q‐Extractive mass spectrometer (Thermo Scientific) coupled to an Ultimate 3000 HPLC system (Thermo Scientific). The TMT‐labeled peptides were separated on a homemade fused silica capillary column (150 mm × 75 µm, 5 µm, Varian, Lexington, MA, USA) and eluted at a 120 min gradient at 0.25 µL/min flow rate. Mobile phase A consisted of 0.1% FA in H₂O and mobile phase B consisted of 0.1% FA in acetonitrile. The mass spectrometer was operated in data‐dependent acquisition mode, and a full‐scan mass spectrum was acquired in the Orbitrap with a resolution of 70 000 (m/z 200) across a mass range of 300–1800 m/z. After the survey scans, the top 20 precursor ions were selected for MS/MS fragmentation (HCD, 32%) with an isolation width of 2.4 m/z. The automatic gain control was 1.7 × 10⁴, and the dynamic exclusion time was 20 s.

Data were analyzed by Proteome Discoverer software(version 1.4, Thermo Scientific). The Sus scrofa UniProt Reference proteome release‐2021_04 (including isoforms, 49 865 sequences) served as the target sequence database for the pig data [10]. The following criteria were used for database searching: Full‐tryptic, up to two missed cleavages, 20 ppm precursor mass tolerance, 0.02 Da MS² fragment ion tolerance, the oxidation of methionine was set as a variable modification, carbamidomethylation of cysteine and TMT‐10plex modification of N‐terminal and lysine were set as fixed modifications. A high level of peptide confidence (q value < 0.01) was considered as positive identification and the false discovery rate (FDR) threshold of 0.01 was used to validate peptide spectrum matches (PSMs). The candidate proteins were identified by at least one unique peptide. Relative protein quantification was performed by TMT reporter ions in an MS² scan.

2.5. Proteomics Data Analysis

DEPs were selected with a cut‐off of fold change (FC) ≥ 1.5 and an FDR‐adjusted p value (adj p value) < 0.05. The gene ontology (GO) analyses of DEPs were performed using the DAVID database (https://david.ncifcrf.gov/). Protein–protein interaction (PPI) analyses of DEPs were performed using the STRING database (https://www.string‐db.org/). Circos plots were processed in R (v4.0.5) followed by a construction (http://circos.ca/) [11]. A weighted protein co‐expression network analysis was generated in the R package (WGCNA) [12].

2.6. Construction of a CXTx Model From SD Rats to Balb/c Mice

We transplanted hearts from weight‐matched SD neonatal rats into the neck region of Balb/c mice, following the procedure outlined in ref. [13]. The intervention was conducted using the drug Bay 11–7082, sourced from MCE. The drug was prepared at a concentration of 0.75 mg/mL, with an administration volume estimated at 0.2 mL/20 g, and was dissolved in a solution of 10% (v/v) DMSO in sodium carboxymethyl cellulose. Daily intraperitoneal injections were administered. For the control group in our study, only the solvent was administered. The viability of the transplanted heart was assessed by palpating for pulsation in the cervical region.

2.7. Immunohistochemistry Staining

In the immunohistochemistry protocol, slides were initially baked at 68°C for 45–60 min, followed by deparaffinization and rehydration to water. Antigen retrieval was conducted by submerging the slides in 1 L of diluted EDTA solution in a pressure cooker, heated initially to 2100 W until boiling, then reduced to 1000 W. After 2.5 min from the start of the steam release, the cooker was cooled in a water bath, and slides were left at room temperature for 20 min. Slides were washed three times with PBS for 5 min each, then treated with 3% H₂O₂ at room temperature for 10 min to block endogenous peroxidase activity. Nonspecific binding was blocked by incubating the slides in either goat serum or an immunostaining blocking solution, potentially with an added 0.1% Triton X‐100, for 0.5–1.5 h based on the reagent's datasheet. The primary antibody was applied overnight at 4°C for 18 h, followed by temperature equilibration at room temperature or 37°C for 45 min. We used the following antibodies for our experiments: Anti‐CD3 (Abcam, ab16669), anti‐CD20 (Abcam, ab64088), anti‐MPO (Abcam, ab208670), anti‐CD68 (Abcam, ab125212), anti‐CD42b (Abcam, ab227669), anti‐mouse IgG (Abcam, ab150113), and anti‐mouse IgM (Abcam, ab190369). After this, slides were washed three times with PBS for 5 min each before and after the secondary antibody was applied at room temperature for 30 min.

2.8. Western Blot Analysis

Tissue specimens were subjected to a wash with ice‐cold phosphate‐buffered saline (PBS) and subsequently lysed using RIPA lysis buffer (Beyotime, Catalog #P0013B), which was further enriched with both phosphatase and protease inhibitors (Roche). The resulting lysates were resolved via SDS‐PAGE, utilizing gels with concentrations ranging from 8% to 15% (Beyotime, Catalog #P0012AC), followed by their transfer onto polyvinylidene difluoride (PVDF) membranes with pore sizes between 0.22 and 0.45 µm (Merck Millipore) utilizing the Trans‐Blot System (Bio‐Rad Laboratories). Post‐transfer, the membranes were blocked with 5% skim milk in TBST for 1 h at a temperature of 25°C, after which they were subjected to incubation with primary antibodies at a dilution, overnight at 4°C. This step was succeeded by incubation with horseradish peroxidase (HRP)‐conjugated secondary antibodies for 1 h at 25°C. For the detection of the blotted proteins, the Western HRP substrate (Millipore) along with an imaging system was employed. The antibodies utilized in this investigation included PTX3(Abcam, Cat#ab90806, 1:1000), HSP90(Abcam, Cat#ab59459, 1:1000) GAPDH(Abcam, Cat#ab1791, 1:1000).

2.9. Detection of NF‐κB Signal Intensity

The Beyotime kit SN371 was used, following the manufacturer's instructions.

2.10. Statistical Analysis

A two‐tailed Student's t‐test was implemented in order to determine significance, false discovery rates were used to adjust the p value and the adjusted p value < 0.05 was considered significant. Statistical analysis was performed with SPSS 20.0 (IBM Corp, Armonk, NY, USA).

3. Results

3.1. General Description and Proteome Profiling in Cardiac Xenograft

The experimental design of CXTx is described in Figure 1A. We investigated three different CXTx based on the gene type of the donor's heart and the immunosuppression: Group I, WT‐Pre vs. WT‐Post, Group II, GTKO‐Pre vs. GTKO‐Post‐NonIS, Group III, GTKO‐Pre vs. GTKO‐Post‐IS. The time of cardiac cessation post‐transplantation for each group is shown in Table 2, and the appearance of the hearts at the time of cessation is shown in Figure S1. The relative abundance of proteins spanned almost ten orders of magnitude, and high‐abundant proteins were the constitutive proteins of sarcomeres such as MYL2, TNNT2, and TNNI3 (Figure 1B).

FIGURE 1.

Workflow of this study. (A) Workflow of MS‐based quantitative proteomics, from sample collection to the DEPs analysis. (B) Multiple filtering criteria of protein identified. (C) Dynamic ranges of the CXTx proteomes.

TABLE 2.

Survival time of xenografts.

Group	Number	Survival time
WT‐Post	#1	1.5 h
	#2	2 h
GTKO‐Post‐NonIS	#3	22 h
	#4	45 h
GTKO‐Post‐IS	#5	182 h
	#6	252 h

Note: Group I (WT‐Post): donor porcine heart was wild‐type (WT) and without postoperative immunosuppression, Group II (GTKO‐Post‐NonIS): donor heart was from genetically modified homozygous α‐1,3‐galactosyltransferase knockout pig (GTKO) and without postoperative immunosuppression, Group III (GTKO‐Post‐IS): donor heart was from GTKO pig and with postoperative immunosuppression.

A TMT‐labeled proteomic analysis was performed to investigate the molecular mechanism of xenograft failure. A total of 3892 proteins were identified in the cardiac xenograft. To further increase reliability, we screened proteins by performing the following strategies: (1) Missing data <30%; (2) Unique peptide ≥1; (3) Peptide spectrum matching score >4; (4) Mascot ion score >5. Finally, a list of 2425 proteins was involved in the further bioinformatic analyses after filling in the missing data by using a k‐nearest neighbor algorithm (Figure 1C).

3.2. Alterations of Proteome in WT Donor Hearts

As shown by previous studies, preformed antibodies can rapidly lead to vascular injury in the xenograft. Histopathological results showed a diffuse formation of thrombosis and interfascicular edema, capillary necrosis, and infiltration of neutrophils (Figures 2A, S2). The volcano plot summarized the expression profile of proteins in xenograft while compared to WT‐Pre, the expression of 310 DEPs was significantly upregulated and 65 DEPs was significantly downregulated (Figure 2B). Both Principal Component Analysis (PCA) and HC analysis demonstrated that the WT‐Post were clustered with complete separation from the WT‐Pre, indicating the proteome of donor hearts was significantly changed (Figures 2C and D, S3A).

FIGURE 2.

The different proteome in group I (WT‐Pre vs. WT‐Post). (A) Histology of cardiac xenografts used WT donor hearts. Thrombosis is marked with black arrows, interstitial edema with green arrows, capillary necrosis with red arrows, and neutrophil infiltration with yellow arrows. The heart was harvested on the second hour post‐transplantation (B) The DEPs in WT‐Pre vs. WT‐Post (volcano plot, FDR < 0.05, FC > 1.5). (C) PCA of all proteomes from tissue subsets. (D) Heatmap of protein expression between WT‐Pre and WT‐Post. (E) Gene ontology analysis of the DEPs in WT‐Pre vs. WT‐Post. (F) Protein–protein interaction of the DEPs in WT‐Pre vs. WT‐Post. (G) The WGCNA cluster dendogram of proteins in WT‐Pre vs. WT‐Post. I. (H) The eigenprotein adjacency heatmap of WGCNA clusters in WT‐Pre vs. WT‐Post. (I) The relative expression of eigenprotein of WGCNA clusters in WT‐Pre vs. WT‐Post.

To determine the xenograft failure‐related molecular mechanism, we performed GO analysis using DEPs. The typical significantly enriched GO terms were shown in Figure 2E that involved in protein synthesis and degradation, cell adhesion and junction, and RNA binding and ribosome. Notably, the DEPs in most enriched GO terms were primarily upregulated, except for three GO terms: Negative regulation of coagulation, blood microparticle, and anatomical structure homeostasis (Figure 2E). Thereafter, we constructed a PPI network that found the most cross‐linked DEPs can be clustered in three types based on a k‐mean algorithm, and several proteins associated with cytoskeleton and metabolism, inflammation, and RNA splicing and protein synthesis exhibited tightly internal interactions (Figure 2F). WGCNA classified proteins into eight modules of strongly associated proteins (Figure 2G). Each module was represented as an eigenprotein which was the most representative abstraction component of the proteins. Modules were identified by different colors and eigenprotein adjacency heatmap indicated that these eight modules could be divided into four subgroups and MEturquoise and MEblue showed a correlated protein expression pattern (Figure 2H). Furthermore, the eigenproteins of two modules: MEturquoise and MEblue, showed significantly differential expression between WT‐Pre and WT‐Post (Figure 2I).

3.3. Alterations of Proteome in GTKO‐NonIS Donor Hearts

Although the GTKO porcine donor can largely reduce its immunogenicity, delayed xenograft rejection also caused xenograft failure in 48 h without immunosuppression. The histological change included microvascular thrombosis, coagulative necrosis, and many neutrophils adhesion to vascular endothelial cells (Figures 3A, S2). The expression of 66 DEPs was significantly upregulated and 20 DEPs were significantly downregulated (Figure 3B). Compared to WT, GTKO hearts presented a smaller change in proteome after xenotransplantation. This indicates that WT pigs exhibit a higher number of upregulated proteins after xenotransplantation compared to GTKO pigs. The absence of the genetic modifications present in GTKO pigs likely results in a more extensive or intense immune and inflammatory response in WT pigs following xenotransplantation. In WT pigs, the upregulated proteins are more associated with acute inflammatory responses and immune activation processes, which are typically observed in unmodified pigs undergoing xenotransplantation. The elevated expression levels of various immune‐related proteins suggest a strong reaction to foreign antigens. Unsurprisingly, some proteins have the same variation tendency in both WT and GTKO hearts, such as fibrinogen A‐α‐chain was significantly upregulated and IGHG was significantly downregulated after xenotransplantation (Figure 3B). PCA and HC (Hierarchical Clustering Analysis) analysis demonstrated a complete separation between GTKO‐Pre and GTKO‐Post‐NonIS (Figures 3C and D, S3B).

FIGURE 3.

The different proteome in group II. (A) Histology of cardiac xenografts used GTKO donor hearts.The heart was harvested on the second day post‐transplantation. (B) The DEPs in GTKO‐Pre vs. GTKO‐Post‐NonIS (volcano plot, FDR < 0.05, FC > 1.5). (C) PCA of all proteomes from tissue subsets. (D) Heatmap of protein expression between GTKO‐Pre and GTKO‐Post‐NonIS. (E) Gene ontology analysis of the DEPs in GTKO‐Pre vs. GTKO‐Post‐NonIS. (F) Protein–protein interaction of the DEPs in GTKO‐Pre vs. GTKO‐Post‐NonIS. (G) The WGCNA cluster dendogram of the DEPs in GTKO‐Pre vs. GTKO‐Post‐NonIS. (H) The eigenprotein adjacency heatmap of WGCNA clusters in GTKO‐Pre vs. GTKO‐Post‐NonIS. (I) The relative expression of eigenprotein of WGCNA clusters in GTKO‐Pre vs. GTKO‐Post‐NonIS.

The typical significantly enriched GO terms of GTKO‐NonIS were showed in Figure 3E. The differences between Figures 2E (WT‐Pre vs. WT‐Post) and Figure 3E (GTKO‐Pre vs. GTKO‐Post‐NonIS) highlight both similarities and distinctions in protein distribution patterns. Biological processes in both figures show translation (GO:0006412) as a key process, indicating increased protein synthesis activity. However, Figure 2E (WT) emphasizes the cellular amide metabolic process (GO:0043603), likely reflecting influence of acute inflammation on metabolism, while Figure 3E (GTKO) highlights regulation of endopeptidase activity (GO:0052548) and response to stimulus (GO:0050896), indicating that due to the extended survival in the GTKO group, not only is translation activated, but there is also a shift toward regulating intracellular protein degradation mechanisms to establish a new protein homeostasis and better manage immune responses. Cellular components differ significantly in the WT group, changes are most pronounced in the extracellular region and extracellular matrix, possibly due to stronger rejection responses, while GTKO primarily affects ribosomal and cytosolic components, indicating intracellular adjustments in protein synthesis. Molecular functions show commonality in ribosomal structural components (GO:0003735), but WT focuses on small molecule binding (GO:0036094), whereas GTKO uniquely shows enzyme regulator activity (GO:0030234). These findings demonstrate distinct pathways affected under WT and GTKO conditions, reflecting differences in cellular structure and function regulation. The PPI network of GTKO‐NonIS was associated with less proteins because of less DEPs in GTKO‐NonIS, and the proteins most involved in the process of RNA splicing, inflammation, and metabolism (Figure 3F). WGCNA classified proteins into six modules of strongly associated proteins (Figure 3G). Eigenprotein adjacency heatmap indicated that these six modules could be divided into three subgroups and MEblue, MEyellow, MEbrown, and MEred showed a correlated protein expression pattern (Figure 2H). The eigenprotein of MEblue and MEbrown, showed significantly differential expression between GTKO‐Pre and GTKO‐Post‐NonIS (Figure 3I). The eigenproteins of MEblue and MEbrown were identified through clustering algorithms, grouping proteins with similar expression levels. These clusters showed significantly higher expression in the GTKO‐Post‐NonIS group compared to the GTKO‐Pre group. MEblue, comprising 441 proteins, is associated with pathways like DNA repair, programmed cell death, inflammatory response, chromatin remodeling, and cytokine production, indicating responses to cellular damage, immune activation, and defense against pathogens. MEbrown, containing 292 proteins, is linked to pathways including cell growth regulation, Wnt signaling, and defense responses, suggesting tissue remodeling and immune challenges. The upregulation of these pathways post‐transplantation reflects adaptive and defensive mechanisms by the graft, highlighting the importance of targeted immunosuppressive strategies to mitigate stress responses and prevent immune‐related graft damage, which could enhance graft survival in xenotransplantation.

3.4. Differences of Proteome Between WT and GTKO‐NonIS Donor Hearts in Response to CXTx

We compared the DEPs between group I and II. Some DEPs have the same variance tendency and most of them showed an increased expression, including PTX3, MVP, HSP90AB1, SLC25A6, ANXA5, SSB, ITGA5, VARS, RBMX, HMGB1, DHX9, PCBP2, TALDO1, DHX15, RPL7A, RPL14, and RPS8. Others showed a decreased expression such as IGHG, APOH, ITIH4 and LOC396684. We also screened out three DEPs that showed more severe changes in group II than group I (the FC of DEPs in group II divided which in group II > 1.5), as shown in Figure 4. We observed that the expression of PTX3 increased more significantly in GTKO pigs post‐transplant compared to WT pigs. PTX3 is an acute phase protein that promotes complement activation and recruits immune cells. The more pronounced increase in GTKO pigs is likely due to their longer survival time, allowing more time for inflammatory cascades to occur IGHG and ITIH4 showed a more pronounced downregulation in GTKO pigs. The genetic modification in GTKO pigs reduces α‐Gal antigens, thereby decreasing xenoreactive antibodies and lowering IGHG levels. Similarly, the reduction in ITIH4, a protein linked to tissue repair and inflammation, reflects the minimized hyperacute rejection and reduced inflammation due to genetic modification, indicating a lower inflammatory state in GTKO pigs compared to WT pigs.

FIGURE 4.

The DEPs that were more severe changed in GTKO‐Post‐NonIS (the FC of DEPs in group II divided which in group I > 1.5).

3.5. Alterations of Proteome in GTKO‐IS Donor Hearts

Similar to Group II, microvascular thrombosis can also be found in GTKO porcine donors even under immunosuppression when xenograft failure as well as accompanied by focal infiltration of T cells (Figures 5A, S2). The expression of 350 DEPs was significantly upregulated and 41 DEPs was significantly downregulated (Figure 5B). The immunosuppression strongly disturbed the expression of proteins and enhanced the DEPs profiling before and after xenotransplantation. Some proteins that play the role of inflammatory inhibitor and tissue repair were significantly upregulated in GTKO‐Post‐IS compared to GTKO‐Pre, such as THBS1 and CD163 (Figure 5B). It is noteworthy that the expression of MVP and PTX3 still presented a considerable increase after xenotransplantation under the immunosuppressive treatment. PCA and HC analysis also indicated a complete separation between GTKO‐Pre and GTKO‐ Post‐IS (Figures 5C and D, S3C).

FIGURE 5.

The different proteome in group III. (A) Histology of cardiac xenografts used GTKO donor hearts with immunosuppression. The heart was harvested on the tenth day post‐transplantation (B) The DEPs in GTKO‐Pre vs. GTKO‐Post‐IS (volcano plot, FDR < 0.05, FC > 1.5). (C) PCA of all proteomes from tissue subsets. (D) Heatmap of protein expression between GTKO‐Pre and GTKO‐Post‐IS. (E) Gene ontology analysis of the DEPs in GTKO‐Pre vs. GTKO‐Post‐IS. (F) Protein–protein interaction of the DEPs in GTKO‐Pre vs. GTKO‐Post‐IS. (G) The WGCNA cluster dendogram of the DEPs in GTKO‐Pre vs. GTKO‐Post‐IS. (H) The eigenprotein adjacency heatmap of WGCNA clusters in GTKO‐Pre vs. GTKO‐Post‐IS. (I) The relative expression of eigenprotein of WGCNA clusters in GTKO‐Pre vs. GTKO‐Post‐IS.

The result showed most GO terms were more significantly enriched in GTKO‐IS than in GTKO‐NonIS (Figure 5E). Regulation of microtubule‐based process and eukaryotic translation initiation factor 3 complex were specifically enriched in GTKO‐IS. The PPI network of GTKO‐IS was associated with the process of complement activation and coagulation, RNA splicing, and protein synthesis and degradation (Figure 5F). WGCNA classified proteins into seven modules of strongly associated proteins (Figure 5G). The Eigenprotein adjacency heatmap indicated that these seven modules could be divided into three subgroups and MEblue, MEred, MEturquoise, and MEbrown showed a correlated protein expression pattern (Figure 2H). The eigenprotein of only MEturquoise showed significantly differential expression between GTKO‐Pre and GTKO‐Post‐IS (Figure 5I).

3.6. Differences of Proteome Among Three Groups in Response to CXTx

We compared the DEPs among groups I–III. The DEPs that have the same variance tendency are shown in Figure 6A. The DEPs demonstrated more severe changes in group III than group II (the FC of DEPs in group III divided which in group II > 1.5) were MVP, F9, RPL24, BUB3, APOC3, and ATL3, in group III than group I (the FC of DEPs in group III divided which in group I > 1.5) were PTX3, MVP, IGHG, EEF1A, EIF4A1, CD163, CD44, THBS1, FKBP5, FLNA, SNRPD3, MTA2, CTRC, RPL13, RPL11, RPS17, RPS20, and PPIA (Figure 6B). A Sankey diagram showed most of the DEPs in group I were not significantly changed in groups II and III (Figure 7A). It indicated that GTKO donor hearts may have a more stable physiological state than WT donor hearts. Finally, we observed that PTX3, MVP, HSP90AB1, and APOH exhibited consistent variance trends across all three groups, indicating that these DEPs are not affected by the gene editing and immunosuppressive interventions employed in this study (Figure 7B).

FIGURE 6.

The differential CXTx groups contain shared and unique protein partners. (A) The shared and unique DEPs in group I∼III. (B) The DEPs that were more severe changed in GTKO‐Post‐IS (the FC of DEPs in group III divided which in group II and/or I > 1.5).

FIGURE 7.

The variation trend of DEPs among differential CXTx groups. (A) Sankey diagram of the DEPs among differential CXTx groups. (B) The potential target DEPs that were screened out.

3.7. NF‐κB Inhibitor BAY 11–7082 Suppresses Xenograft Rejection Immune Response and Prolongs Survival of Xenogeneic Cardiac Grafts

Among the most significantly DEPs, three are notably associated with immune rejection: PTX3, MVP, and HSP90AB1. Previous studies have demonstrated that these three proteins are involved in various physiological and pathological processes and simultaneously regulate NF‐κB activation. Knockout of PTX3 [14] and pharmacological inhibition of HSP90 (encoded by HSP90AB1) [15] reduce inflammation by inhibiting NF‐κB activation (translocation from the cytoplasm to the nucleus). Upregulation of MVP promotes the translocation of NF‐κB to the promoter regions of various inflammatory cytokines, thereby enhancing their expression [16]. This evidence suggests that the three proteins, which were commonly upregulated across the three experimental groups, may promote inflammation via NF‐κB activation. Even with the use of CD40‐CD154 monoclonal antibodies and 10‐gene edited pigs, NF‐kB activation was observed, as shown in a study where pig hearts were transplanted into brain‐dead patients [17] and further confirmed through multi‐omics analysis [18]. This highlights that, despite the latest CD40 monoclonal antibody‐based regimen, NF‐kB activation persists. Our NF‐κB activation staining results for each group also show NF‐κB activation in the WT donor heart, GTKO donor heart without immunosuppression, and GTKO donor heart with immunosuppression (Figure S4A,B). Therefore, we further investigated whether targeting NF‐κB could alleviate the inflammatory phenotype of CXTx and extend graft survival.

We utilized the SD Rat‐Balb/C Mouse CXTx model to preliminarily explore the effects of targeting NF‐κB (via intraperitoneal injection of BAY 11–7082) on CXTx. This model is widely used to study acute vascular rejection mediated by innate immunity [19, 20, 21]. NF‐κB activation, as well as upregulation of PTX and HSP90, was confirmed in heterotopic heart xenotransplantation samples from rat to mouse. This suggests that the SD Rat‐Balb/C Mouse xenotransplantation model can simulate findings observed in the pig‐to‐monkey xenotransplantation experiment (Figure S5). The primary evaluation metrics were the survival period of the grafts and the inflammatory response of the grafts at POD5 (Figure 8A). The results indicated that targeting NF‐κB significantly prolonged recipient survival (median survival time extended from 5 to 8 days) (Figure 8B) and significantly alleviated myocardial inflammation (Figure 8C). Further quantification revealed that targeting NF‐κB significantly reduced the infiltration of various immune cells, including T cells (Figure 9A, G), B cells (Figure 9B, H), NK cells (Figure 9C, I), macrophages (Figure 9D, J), and neutrophils (Figure 9E, K), in the myocardial tissue, as well as reduced antibody deposition (Figure 9F).

FIGURE 8.

The effect of BAY 11–7082 in SD Rat‐Balb/C Mouse CXTx model. (A) Using the SD Rat‐Balb/C Mouse CXTx model, we explored the effects of targeting NF‐κB with BAY 11–7082 on acute vascular rejection. The primary metrics were graft survival and inflammation at POD5. (B) Targeting NF‐κB extended median survival from 5 to 8 days(p < 0.01). (C) Targeting NF‐κBreduced myocardial inflammation.

FIGURE 9.

Quantification of immunohistochemical results. Targeting NF‐κB significantly reduced the infiltration of various immune cells, including T cells (A, G), B cells (B, H), NK cells (C, I), macrophages (D, J), and neutrophils (E, K), in the myocardial tissue, as well as reduced antibody deposition. Hearts from both the SD Rat‐Balb/C Mouse CXTx model group and the SD Rat‐Balb/C Mouse CXTx model with BAY 11–7082 injection group were harvested on the fifth day post‐transplantation. The scale bar represents 50 µm.

4. Discussion

Xenograft failure is a highly complicated process, yet an understanding of the molecular mechanisms involved in occurrence and progression of xenograft failure is serious lack. In this study, we applied quantitative proteomic to describe the landscape of alterative proteome after CXTx in a pig‐to‐NHP model and to understand the molecular mechanisms of xenograft failure. We took the way of step‐by‐step approach to investigate three CXTx groups: WT‐Pre vs. WT‐Post, GTKO‐Pre vs. GTKO‐Post‐NonIS, and GTKO‐Pre vs. GTKO‐Post‐IS. Histopathological examination showed that microvascular thrombosis was found in cardiac xenograft of WT‐Post, GTKO‐Post‐NonIS and GTKO‐Post‐IS. PCA and HC analysis showed significant difference of heart proteome before and after CXTx in all three CXTx groups. GO analysis indicated blood microparticle was significantly enriched in all three CXTx groups, which would also confirm widespread thrombosis in all three groups of cardiac xenografts. It is noteworthy that the regulation at post‐transcriptional level such as ribosome assembly and protein synthesis was activated in cardiac xenograft of GTKO‐Post‐NonIS and GTKO‐Post‐IS. This result was also confirmed by the result of WGCNA and subsequent KEGG enrichment analysis of the differential expressed modules. We identified some proteins that consistently high expression in cardiac xenograft of all three CXTx groups, such as PTX3, MVP, and HSP90AB1. Our data characterized a comprehensive alteration of proteome after CXTx in pig‐to‐NHP model. The results indicated the activation of complement and coagulation was irrepressible in present GTKO background genotype and immunosuppressive treatment. It appeared necessary to use the donor porcine heart with more genetically modified such as GTKO.hCD46.hTBM.

Xenotransplantation has complex pathophysiological features because the species barrier can profoundly affect xenograft function. Comprehensive proteome analysis showed an obvious disturbance of many biological processes in the cardiac xenograft. Up to now, the priorities in a study about CXTx have been to focus on optimizing the program of gene editing and preservation of donor hearts to prolong survival time [4, 22]. However, there was little investigation of molecular mechanism on cardiac xenograft failure [23]. A recent study reported the transcriptomic data of CXTx in GTKO pig‐to‐NHP model [24]. Similar to our results, Park et al. also found the upregulation of thrombospondin 1 in the level of mRNA expression, which suggested the development of microvascular thrombosis cannot be avoided in only GTKO donor heart [24]. The genes coding for several kinds of collagen related to the remodeling of the heart that were subjected to ischemic condition such as COL1A1, COLA6A3, and COL11A2 were upregulated after CXTx in pig‐to‐NHP model [24]. Whereas the upregulation of COL5A1 was only found in GTKO‐IS group in the present study. We speculated that it was due to the longer survival time under immunosuppressive therapy.

Our results demonstrated that GTKO cardiac xenografts prevented the occurrence of hyperacute rejection compared to WT donors. However, it still cannot overcome acute humoral xenograft rejection even in intensive immunosuppression. A previous study showed the microthrombi were accompanied by the deposition of IgM, IgG, and complement in the GTKO cardiac xenografts, and the severity of thrombotic microangiopathy was correlated with the deposition of immunoglobulin and complement [25]. Our proteomic results appeared to strongly indicate the activation of the complement system in cardiac xenografts after transplantation, even though in GTKO‐Post‐IS. This suggested that the use of existing immunosuppressive measures failed to inhibit complement activation during xenograft failure. Another distinctive characteristic of cardiac xenograft rejection was thrombotic microangiopathy that caused myocardial ischemia and necrosis. We found THBS1 had significantly increased expression in GTKO‐Post‐IS compared to GTKO‐Pre which was associated with the result of platelet‐rich fibrin thrombi in the microvasculature. A previous study showed THBS1 could regulate collagen‐mediated platelet activation through a CD36‐dependent pathway [26]. To address the problem of coagulation after xenotransplantation, hTBM was identified as a promising target. Either complement or coagulation activation might trigger xenograft failure and shorten cardiac xenograft survival [27, 28]. Although transgenic technology is a double‐edged sword with many uncertainties on health, our results still underlined the importance of multi‐transgenic pigs from the molecular mechanism level, such as the typical triple‐transgenic (GTKO/hCD46/hTBM).

A conspicuous expressed feature of proteins in our cardiac xenograft was increased PTX3. PTX3 was known as an amplifier of inflammation and innate immunity as well as its levels and activity may be particularly relevant for the injury response [29]. It can bind to several trigger proteins, such as C1q, mannose‐binding lectin, and ficolin in order to activate complement system [30]. The local production of PTX3 in myocardium was associated with ischemia injury during myocardial infarction and might binding of C1q to active the classic complement cascade [31]. Previous study showed the expression level of PTX3 was strongly associated with the severity of LPS‐induced lung injury and was significantly decreased after the anticoagulant therapy [32]. Introna et al. reported that PTX3 positive staining also can be found in both cardiomyocyte and non‐cardiomyocyte in a transverse aortic constriction (TAC) mouse model [33]. Whereafter Suzuki et al. demonstrated that increased production of IL‐6, a typically proinflammatory cytokine, can be induced by PTX3 in the TAC heart [34]. Yet so far, the role of PTX3 in graft failure is still unclear. The increased expression of PTX3 was found in the biopsies of allograft [34]. Our results confirmed upregulation of PTX3 may also play an important role in xenograft. However, the cause‐and‐effect relationship between PTX3 and xenograft failure remains an open question, and the therapeutic effects using PTX3 as a target need further investigation.

Furthermore, several proteins that have the same variation trend after CXTx in all three groups may also be associated with cardiac xenograft failure. MVP is a primary composition of ribonucleoprotein particles that have been verified to mediate bacterial ingestion by epithelial cells in the innate immune response through rapidly recruited lipid rafts [35]. Shults et al. showed MVP was a cell survival factor that was expressed in myocardium and coronary artery smooth muscle [36]. The Hsp90 family proteins can protect and repair damaged tissues from an ischemic environment when microvascular thrombosis occurs. Hsp90AB1 could stabilize low‐density lipoprotein receptor‐related protein‐1 at the cell surface by binding to its cytoplasmic tail inside the cell in response to hypoxia [37]. In addition, previous studies indicated that APOH may show a biphasic modulation on coagulation, and whether it enhanced or inhibited thrombogenesis depended on its conformation [38].

Given that NF‐κB serves as a shared downstream regulator for these three proteins, we propose that it could be an essential focal point for therapeutic intervention. The significant extension in graft survival and reduction in inflammation confirm the efficacy of NF‐κB targeting in suppressing the immune response in CXTx. Detailed analysis further indicated that this approach significantly decreased the infiltration of various immune cells, including T cells, B cells, NK cells, macrophages, and neutrophils, within the myocardial tissue. Additionally, there was a notable reduction in antibody deposition. These results suggest that the targeted inhibition of NF‐κB may serve as a promising therapeutic strategy to mitigate inflammation and enhance graft survival in CXTx settings. The primary immunosuppressants currently used in xenotransplantation, such as CD40 and CD154 monoclonal antibodies and traditional small molecule immunosuppressants like cyclosporine, mainly target lymphocyte activation [4]. However, NF‐κB inhibitors can also broadly suppress the activation of innate immune cells, serving as an effective supplement to existing immunosuppressive regimens.

This study was a preliminary exploration of the molecular mechanism for xenograft failure during CXTx. In consideration of the further clinical application, we only applied the immunosuppressive regimen that has already widespread application in clinical. The broad‐spectrum immunosuppressor such as anti‐CD40 was not used in this study in view of their clinical safety. In addition, this study used GTKO porcine hearts as donors for CXTx, but not more commonly used GTKO.hCD46.hTBM porcine hearts. This may result in a high risk of xenograft failure and a short survival time for the recipient.

5. Conclusion

In conclusion, the present results were the first comprehensive proteomic analysis of CXTx to our knowledge. We reported the perturbations from protein level in well‐established knowledge in CXTx such as activation of coagulation and complement. Some novel DEPs including PTX3, MVP, and Hsp90AB1 were also found in CXTx and may be considered as potential targets for intervention. Although the etiology of xenograft failure is still controversial, the continuous hypoxia injury induced by microvascular thrombosis may play an important role during xenograft failure. It indicated the use of GTKO.hCD46.hTBM porcine hearts should be necessary for further investigation. The results provided a large amount of DEPs in CXTx that could contribute to understanding the molecular mechanisms and pathways related to cardiac xenograft failure. Finally, NF‐κB inhibition effectively decreased immune cell infiltration and antibody deposition in myocardial tissue, suggesting its potential as a therapeutic strategy to enhance graft survival and reduce inflammation in CXTx. We believe such omics technology could help us to solve the mystery of CXTx and translate it into a potential clinical application in the future.

Supporting information

Supporting Information

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.