Gene Expression Signatures of Peripheral Blood Mononuclear Cells during the Early Post-Transplant Period in Patients Developing Cardiac Allograft Vasculopathy

Background. Cardiac allograft vasculopathy (CAV) is a major cause of graft loss and death after heart transplantation. Currently, no diagnostic methods are available during the early post-transplant period to accurately identify patients at risk of CAV. We hypothesized that PBMC gene expression profiles (GEP) can identify patients at risk of CAV. Methods. We retrospectively analyzed a limited set of whole-genome PBMC microarrays from 10 post-transplant patients who did (n = 3) or did not (n = 7) develop advanced grade CAV during their long-term follow-up. We used significance analysis of microarrays to identify differentially expressed genes and High-Throughput GoMiner to assess gene ontology (GO) categories. We corroborated our findings by retrospective analysis of PBMC real-time PCR data from 33 patients. Results. Over 300 genes were differentially expressed (FDR < 5%), and 18 GO-categories including “macrophage activation”, “Interleukin-6 pathway”, “NF-KappaB cascade”, and “response to virus” were enriched by these genes (FDR < 5%). Out of 8 transcripts available for RT-PCR analysis, we confirmed 6 transcripts (75.0%) including FPRL1, S100A9, CXCL10, PRO1073, and MMP9 (P < .05). Conclusion. Our pilot data suggest that GEP of PBMC may become a valuable tool in the evaluation of patients at risk of CAV. Larger prospectively designed studies are needed to corroborate our hypothesis.


Introduction
Cardiac allograft vasculopathy (CAV) is a major cause of graft loss and death after heart transplantation (HTx). Identification of surrogate makers for late cardiac allograft survival has been of major interest to improve long-term outcomes of HTx [1]. After HTx, alloantigens (including molecules from donor endothelium) are presented by antigen-presenting cells to the recipient's T-cells, often generating a differentiated inflammatory response. That response includes T-cells, B-cells, and a coordinated pattern of cytokine release. Cells of innate immunity (monocytederived macrophages) are also involved [1]. Non-antigenspecific perioperative events including microvascular insults may play pivotal roles related to subsequent development of CAV, probably related to ischemia reperfusion injury, advanced donor age, hyperlipidemia, depletion of arteriolar tissue plasminogen activator factor and systemic 2 Journal of Transplantation inflammation [1][2][3][4][5][6][7][8]. The inflammatory response culminates in migration of mononuclear cells through the coronary vascular endothelium and phenotypic switching of medial smooth muscle cells mediated by generation of growthpromoting cytokines. Those processes contribute to chronic damage of the coronary arteries of the transplanted heart. The result is the development of a diffuse, obliterative form of vasculopathy characterized by production of a "neointima" rich in vascular smooth muscle cells and extracellular matrix [9,10]. After the first year post-transplantation, 30% to 50% of patients have some evidence of CAV and after 5 years CAV is one of the leading causes of death with <50% 1-year survival rate in those with extended disease [11,12].
Limited interventions have been shown to prevent, delay, or reverse CAV. While no definite well-validated surrogate marker for late cardiac allograft outcome is available, early detection of CAV represents the key strategy as an effective surrogate [1]. Early identification of CAV became possible with the introduction of intravascular ultrasound (IVUS) [13], but the technique is invasive, it is usually not initiated until at least one year post-transplantation, is expensive, and requires the use of nephrotoxic contrast agents. Noninvasive tests, including stress perfusion, dobutamine echocardiography, ultrafast tomography, and MRI have not proven to be sufficiently sensitive or specific to detect early stages of the disease [14,15]. Therefore, there are clear needs to explore and develop new options for the early evaluation of patients at risk of CAV.
Recently, gene expression profiles of peripheral blood mononuclear cells (PBMC) were used to identify patients with [16,17] or without moderate and severe acute cellular cardiac allograft rejection [18] and patients at risk of antibody-mediated rejection [19].
Since the peripheral recirculation of recipient leukocytes after allo-endothelial-cell contact in the allograft may carry information about immune activation conducive to chronic rejection development, we hypothesized that gene expression profiles of PBMC obtained early after HTx carry molecular signatures that correlate with the future development of CAV.  [18]. Columbia University Medical Center (CUMC) contributed 121 patients. Out of the Columbia University cohort, independent whole genome PBMC samples from 10 patients were available. The study protocol was approved by the Institutional Review Board (IRB) of CUMC. PBMC were isolated from eight mL of venous blood using density gradient centrifugation (CPT, Becton-Dickinson, Franklin Lakes, NJ). Samples were frozen in lysis buffer (RLT, Qiagen, Valencia, CA) within 2 h of phlebotomy. Total RNA was isolated from each sample (RNeasy, Qiagen, Valencia, CA).

Materials and Methods
Whole genome gene expression profiling was performed on two-color Whole Human Genome 60-mer Oligo Microarrays (Agilent Technologies, Santa Clara, CA), which contain 41,000+ unique transcripts. After the hybridization, data were extracted and Lowess normalized expression files were generously provided by the CARGO study sponsor (XDx Inc., Brisbane, CA). Filtering was done against background and only those probes with more than 1.5-fold change were retained. Probes mapping to the same gene transcript were not averaged.

Cardiac Allograft Vasculopathy.
Patients were eligible for the study if they were evaluated with angiography during their post-transplant course. CAV was defined as any evidence of disease in the angiography as evaluated by one of three expert interventional cardiologists who specialized in heart transplantation care at the Center for Interventional Vascular Therapy at Columbia University. Angiograms were classified as normal (no evidence of vasculopathy), mild (any grade of angiographic luminal stenosis less than 50%), moderate (any angiographic luminal stenosis greater than 50% in a main vessel or two secondary branches), and severe (more than 50% in the left main or two main vessels). Patients with moderate or severe CAV were considered "advanced".

Statistical Analysis.
Quantitative and qualitative clinical variables were compared by the Mann Whitney U-test, Chi-Square test, or Fischer-exact when appropriate using SPSS 11.5.1 (SPSS Inc., 2002). A P-value <.05 was regarded as significant. Gene expression of samples obtained from patients with advanced CAV was compared against those with normal or mild disease using independent, unpaired t-test, and Benjamini-Hochberg correction for multiple comparisons to estimate a false discovery rate (FDR) as implemented in significance analysis of microarrays (SAM) [20] after 1.5-fold change filtering. Genes were retained for further analyses if FDR < 5%. Molecular function and biological processes associated with individual significant genes were determined using PANTHER (Protein ANalysis THrough Evolutionary Relationships) Classification System (http://www.pantherdb.org), a unique resource that classifies genes by their functions, using published scientific experimental evidence and evolutionary relationships to predict function even in the absence of direct experimental evidence. Quantitative, real-time PCR cycle thresholds (CT) for each group were compared with the Mann Whitney U-test. Pvalue <.05 was considered significant.

Hierarchical Clustering.
Gene expression data were correlated and visualized in clustered heat maps [21] using the Genesis web interface ((http://carmaweb.genome.tugraz.at/ genesis/index2.html), Institute for Genomics and Bioinformatics, Graz University of Technology, Graz, Austria). Top 100 genes and microarray samples as well as differentially enriched GO categories and genes were clustered on the basis of co-occurrence of the differentially expressed genes using the Pearson metric and average linkage. .898 HTx: heart transplant; LVAD: left ventricular assist device; RT-PCR: real-time polymerase chain reaction.   Sirolimus levels (ng/ml) .59 .235 HTx: heart transplant; LVAD: left ventricular assist device; RT-PCR: real time polymerase chain reaction.

Gene Ontology Analysis.
We used High-Throughput GoMiner (HTGM) [22] (http://discover.nci.nih.gov/) to analyze the list of interesting genes in the context of gene ontology (GO) categories [23]. HTGM analyzes data from all microarrays in a study, provides diagnostics for data interpretation and visualization tools in terms of specialized clustered "heat maps" (also called clustered image maps, CIMs) [21]. Normally, the input to HTGM consists of a totalgenes file (representing the entire Microarray or a randomly generated whole genome seed) and a changed-genes file (representing the genes with altered expression) relevant to the study purpose. The output generated by HTGM includes a summary of the results, a matrix whose rows are categories and whose columns are names of changed gene for hierarchical clustering of experiments and categories, and a statistical summary for each category including one-sided Fisher exact P-value and an FDR. Hierarchical clustering of enriched categories and changed genes allows determining which categories achieved statistical significance by virtue of containing essentially the same set of changed genes.

Patient Demographics.
The baseline characteristics of the study population, including mean age, gender, race, diagnosis, mean cold ischemia time, induction strategy, antirejection prophylaxis, median follow-up time at the time that the microarrays were obtained, median followup time till the diagnosis of CAV, simvastatin use, and median number of rejection episodes (i.e., 1R/1B, 2R, and 2R/3A) requiring augmented immunosuppression are depicted in Tables 1(a) and 1(b) for the overall population and by study subgroups and sample method. All baseline clinical characteristics were comparable across groups in both microarray and PCR studies. Although the difference in timing of blood sample post-HTx between advanced CAV and normal/mild CAV groups in the microarray study population approached the defined level of significance (P = .06) but this difference was not significant in the PCR study population.

Microarray Analysis.
Out of 10 patients who were studied with microarrays, 3 (30%) had advanced CAV while 7 (70%) had angiograms consistent with either absence of disease (n = 3, 30%) or mild disease (n = 4, 40%  (Table 2) and a clustered heat map of top 100 up-and down-regulated genes is shown in (Figure 1). A complete list of genes with their respective fold changes and false discovery rates is provided as supplementary material available online at doi:10.1155/2010/719696.

Gene
Ontology Analysis. HTGM detected 18 changed GO categories (FDR < 5%) enriched with differentially expressed genes (Table 3). Processes related to enriched GO categories included among others: macrophage activation, response to wounding, response to virus, response to biotic stimulus, interleukin-6 biosynthetic process, I-kappaB kinase NF-kappaB cascade, innate immune response, inflammatory response, regulation of interleukin-6 biosynthetic process, defense response, positive regulation of interleukin-6 biosynthetic process, activation of NF-kappaB-inducing kinase, and interleukin-6 production. Evaluation of the clustered image map indicates potential cross-talk among GO categories enriched by advanced CAV genes. Instances of cross-talk are of particular importance because they tie together categories that, based on prior knowledge, might be thought of as unrelated. A clustered heat map of differentially expressed genes and enriched gene ontology categories and cross-talk is shown in (Figure 2).

Journal of Transplantation
Immune system process, cell cycle, protein metabolic process, and response to stress     of a disease under investigation. These disease specific gene signatures can be used to select patients at a specific state of a disease with an accuracy that facilitates the diagnosis of the disease and selection of patients for different treatment options.
In this pilot study, patients with advanced CAV show peripheral blood gene expression profiles that differ already in the early period after transplantation from those of patients who will not develop advanced CAV. The findings are consistent with previous suggestions that donor-and recipient-related factors in the perioperative period may play major roles in the immune response and development of a sustained chronic response to immune injury, leading to proliferation of the endovascular matrix, with consequent obstruction of the blood flow and impairment of allograft function [2,4,9,10].
In previous studies, gene expression profiles of PBMC have been shown to correlate with presence or absence of concurrent acute cellular rejection [16][17][18]24], antibody mediated rejection [19], and in response to mechanical circulatory support device implantation [25,26]. Based on these gene expression profiles, the genomic classifier developed to rule out acute cellular cardiac allograft rejection has also been shown to correlate with different organ function related parameters of rejection [27,28], and longitudinal changes in clinical profiles [29,30]. In current study, we hypothesized that gene signatures early after HTx correlate with future development of CAV. To the best of our knowledge, this is the first report of using a high throughput genomic screening approach to identify patients at risk of developing advanced CAV on the basis of leukocyte samples obtained during the early period after transplantation. This observation has potentially important implications.
Recently CAV has been reported as one of the major causes of new-onset graft dysfunction and associated outcomes [31]. Early detection of CAV was proposed as the key strategy, to identify surrogate markers for late cardiac allograft outcomes [1]. Currently, the most accurate diagnostic test to define the phenotype of CAV is intravascular ultrasound [9]. Evaluation of the neointimal proliferation rate between months 3 and 12 has been shown to predict survival and major adverse cardiac events at long-term follow-up [11] but intravascular ultrasound is invasive, expensive and complication prone. Therefore, early profiling of an immune response that differentiates a "high CAVrisk phenotype" from a "low CAV-risk phenotype" may become valuable to early identify these patients, model timely therapeutic interventions and monitor response.
This concept in HTx is not new. Earlier observations in animal models of rejection suggested that "unstable tolerance induction" is associated with the persistent immune activation that mediates destruction of graft parenchymal cells, and that the evaluation of biological processes involved might be useful in identifying different types of rejection and the likelihood of progression to chronic forms [32].
In our study, we found differentially expressed genes that significantly enriched GO categories including the innate immunity, macrophage activation, interleukin-6, activation of the NF-KappaB cascade and response to virus. The finding that the innate immune response plays an important role is interesting because it has been already described in the literature. Differentially expressed genes, among others in this category, included Toll-like receptors 1, 4 and 6, and Interleukin-23 alpha. Several authors proposed an important role of innate immunity in the pathogenesis of CAV [10,[33][34][35]. For example, it has been shown that in patients with allograft endothelial dysfunction (an early clinical indicator of transplant vasculopathy), mRNA transcript level and surface expression of TLR-4 on circulating monocyte is significantly higher than controls [34]. According to our results the categories related to these mechanisms were enriched by the down-regulated genes. The interaction between the up-and down-regulated genes collectively determines the functioning of different biological processes as assessed by high through-put microarray analysis. Our observations show some interesting expression patterns of these mechanisms during the early period post-HTx which needs further understanding. Therapy with Simvastatin that inhibits allograft inflammatory activity and attenuates endothelial coronary dysfunction shows reduced trans-cardiac IL-6 and TNF-alpha gradients [33] and polymorphisms within the promoter region of the IL-6 gene has been proposed as risk markers for CAV [36]. In addition, the transcription factor NFKB becomes up-regulated in response to CD40/ CD40 ligand mediating apoptosis of endothelial cells. Inhibition of this pathway has been shown to be protective against the development of CAV [37]. The GO analysis shows that the STAT1 gene involved in the NFKB pathway is also part of the response to virus showing a "cross-talk" between both categories and suggesting that mechanisms of immune damage related to viruses (i.e., CMV) [9,10] may be related to the NFKB pathway.
The findings presented in this paper should encourage the development of strategies to advance evaluation and management of heart transplant recipients in a preventative, preemptive and personalized way, but the conclusion from this pilot study should be interpreted within the important limitations imposed by the small sample size, the use of angiography which is an imperfect standard for detection of CAV, and the known variable reproducibility of gene expression studies [38]. Corroboration of genes was based on retrospective PCR data. Only a small number of genes were available that were present in the Microarray-based candidate gene list.
In conclusion, our data show that peripheral blood leukocyte genes, specifically innate immune response genes, are differentially expressed during the early time after HTx in patients who develop advanced CAV. Larger prospectively designed studies are needed to corroborate our findings.

Disclosure
All authors report no conflict of interests in relationship to the data presented here.