SEARCH

SEARCH BY CITATION

Keywords:

  • Gene expression;
  • kidney transplantation;
  • liver transplantation;
  • operational tolerance;
  • tolerance

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients, Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

A proportion of transplant recipients can spontaneously accept their grafts in the absence of immunosuppression (operational tolerance). Previous studies identified blood transcriptional and cell-phenotypic markers characteristic of either liver or kidney tolerant recipients. However, the small number of patients analyzed and the use of different transcriptional platforms hampered data interpretation. In this study we directly compared samples from kidney and liver tolerant recipients in order to identify potential similarities in immune-related parameters. Liver and kidney tolerant recipients differed in blood expression and B-cell immunophenotypic patterns and no significant overlaps were detectable between them. Whereas some recipients coincided in specific NK-related transcripts, this observation was not reproducible in all cohorts analyzed. Our results reveal that certain immune features, but not others, are consistently present across all cohorts of operationally tolerant recipients. This provides a set of reproducible biomarkers that should be explored in future large-scale immunomonitoring trials.


Abbreviations: 
NK

natural killer

NKT

natural killer T cells

Bm

mature B cell

K-Tol

tolerant kidney recipients

K-Sta

kidney recipients under maintenance immunosuppression

K-CR

kidney recipients with immunologically-mediated chronic rejection

L-Tol

tolerant liver recipients

L-NonTOL

nontolerant liver recipients under maintenance immunosuppression

HV

nontransplanted healthy individuals

PBMC

peripheral blood mononuclear cells

GEO

gene expression omnibus

SAM

significant analysis of microarrays

PAM

predictive analysis of microarrays

FDR

false discovery rate

GSEA

gene set enrichment analysis

KEGG

Kyoto Encyclopedia of Genes and Genomes

PANTHER

protein analysis through evolutionary relationships

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients, Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

In contrast to the intentional induction of allograft tolerance, which has been rarely accomplished in clinical transplantation (1,2), “spontaneous” long-term acceptance of transplanted organs following discontinuation of conventional immunosuppression has been much more frequently observed (3,4). This state is commonly referred to as spontaneous operational tolerance and is particularly prevalent in liver transplantation, where a sizable proportion of stable recipients could probably cease all immunosuppression without compromising the graft's viability (3,5,6). In recent years, considerable efforts have been devoted to the identification of noninvasive biomarkers of operational tolerance in kidney and liver transplantation (7–20). Several of these studies have employed blood cell immunophenotyping and gene expression profiling to search for immune parameters associated with tolerance (7–9,18,19). This has resulted in the identification of characteristic expression and cell-phenotypic traits that distinguish tolerant patients from stable recipients under immunosuppression and sometimes from age-matched healthy individuals as well. Thus, blood samples from tolerant liver recipients are enriched in NK and γδ T cells and in innate immunity-related transcripts (8,21). In contrast, blood samples collected from tolerant kidney recipients exhibit an expansion of peripheral blood B cells and related transcripts (13,18–20). The mechanistic interpretation of these findings is however challenging. First, whether the peripheral blood compartment is representative of intragraft immune responses is unknown. Second, the potentially confounding effect of pharmacological immunosuppression cannot be completely excluded given the retrospective cross-sectional design of the studies. Third, the use of heterogeneous biological samples (e.g. whole blood vs. peripheral blood mononuclear cells) and different genomic and immunophenotying platforms hampers the integration of data from different studies and different transplantation settings. Finally, the fact that only few tolerant recipients are available for the study prevents the elucidation of the relative contribution to the tolerance-related signatures of different clinical parameters. In this study we sought to address some of these problems by exhaustively comparing blood samples collected from tolerant liver recipients, nontolerant liver recipients, tolerant kidney recipients, nontolerant kidney recipients and healthy individuals. We hypothesized that the simultaneous analysis on the same platforms of blood samples collected from liver and kidney tolerant recipients would allow us to determine the reproducibility of previously-described tolerance-related biomarkers and to study the potential influence on their development of pharmacological immunosuppression.

In accordance with previously reported studies where kidney and liver recipients were not directly compared (7–9,13,18,19), both liver and kidney tolerant patients exhibited distinct transcriptional and cell-phenotypic patterns. However, no or minimal overlaps in blood cell-phenotype and whole-genome expression patterns were observed between recipients tolerant to kidney and liver grafts. Whereas in some cases tolerant liver and kidney recipients coincided in specific NK-related transcripts, this observation was not reproducible in all cohorts analyzed. Our results help contextualize the data derived from previous studies devoted to the search of tolerance-associated biomarkers and provide a refined and highly reproducible set of markers to be employed in future large-scale immunomonitoring trials.

Patients, Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients, Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Patients

The following groups of subjects were enrolled in the study: (1) Kidney recipients with stable long-term graft function in the absence of immunosuppressive therapy (K-Tol; n = 12). K-Tol recipients were patients with stable kidney graft function (serum creatinine levels < 150 μmol/L and proteinuria < 1 g/24 h) in the absence of immunosuppression for at least 1 year (range 2–13 years). These patients were noncompliant and refused a kidney biopsy. (2) Kidney recipients with stable graft function under maintenance immunosuppression (K-Sta; n = 12). Stable recipients received standard immunosuppression and had stable graft function (serum creatinine levels < 150 μmol/L and proteinurina < 1 g/24 h) for at least 3 years. (3) Kidney recipients under standard immunosuppression with deteriorating kidney graft function (K-CR; n = 12). These patients displayed serum creatinine levels > 150 μmol/L and/or proteinurina > 1 g/24 h and exhibited transplant glomerulopathy with an active humoral component according to the updated Banff classification criteria and as demonstrated by the presence of graft C4d deposits and/or circulating antidonor HLA antibodies. (4) Liver recipients maintaining stable graft function in the absence of immunosuppressive therapy (L-Tol; n = 12). L-Tol recipients were all hepatitis C virus negative and were intentionally weaned from immunosuppression under medical supervision at least 1 year prior to the study (range 1–2.5 years). Postweaning protocol liver biopsies were available from 7/12 recipients. (5) Liver recipients > 3 years after transplantation and in whom drug weaning was unsuccessfully attempted and led to an acute rejection episode requiring reinstitution of immunosuppresion (L-NonTol; n = 12). At the time of the study all L-NonTol recipients had normalized liver function tests and were receiving low-dose maintenance immunosuppression. (6) Healthy volunteers (HV; n = 12). These individuals had a normal white blood cell count and no infectious or other concomitant diseases for at least 6 months prior to the study. Patients were enrolled from Nantes (kidney recipients) and Barcelona (liver recipients). The two local Ethics Committees approved all aspects of the study and all patients gave their informed consent. Table 1 summarizes the clinical characteristics of all patient groups included in the study.

Table 1.  Demographic characteristics of patient groups
Clinical diagnosisAgeaTime since transplantationa (months)Time off ISa (months)TreatmentSerum creatininea (umol/L)Proteinuriaa g/24hNumber HLA mismatchesaDonor-specific Abs
  1. aMean (range).

  2. CSA, cyclosporine A; Tac, tacrolimus; MMF, mycophenolate mophetil; STD, steroids; Rapa, sirolimus; NA, not available.

L-Tol6015524109NANANA
(30–73)(72–212)(12–47) (88–126)   
L-NonTol44113 Tac 799NANANA
(30–60)(37–215) CSA 3(53–129)   
   MMF 2    
   Rapa 1    
K-Tol44184794 (35–135)0 (0–1)2 (0–5)3
(20–74)(70–288)(3–12)     
K-Sta4598CSA+MMF 611004 (2–5)0
(16–69)(62–177) CSA+AZA 4(81–146)   
   CSA+STD 2    
   Tac+MMF 1    
K-CR52112Tac+STD 1278 (109–621)3 (0–7)3 (1–6)8
(34–78)(15–456) Tac+STD+MMF 2    
   CSA+MMF5    
   Tac 2    
   STD+AZA 1    
HV62
(42–70)       

Peripheral blood samples

Peripheral blood mononuclear cells (PBMCs) were separated on a Ficoll layer (Eurobio, Les Ulis, France) and frozen in either TRIzol® reagent (Invitrogen, Cergy Pontoise, France) or freezing medium.

Microarray experiments

Total RNA was extracted using the TRIzol method according to the manufacturer's instructions. RNA quality and quantity was determined using an Agilent 2100 BioAnalyzer (Palo Alto, CA, USA). The derived cRNA samples from 12 K-Tol, 12 K-Sta, 12 L-Tol, 12 L-NonTol, 12 HV and 12 K-CR were hybridized onto Affymetrix Human Genome U133 Plus 2.0 arrays containing 54 675 probes for 47 000 transcripts (Affymetrix, Santa Clara, CA, USA).

Microarray data normalization and analysis

Microarray data from all samples hybridized were normalized using the robust multiarray (RMA) algorithm. Next we employed a probe-filtering step excluding probes not reaching a log2 expression value of 5 in at least 1 sample, which resulted in the selection of a total of 27 317 probes out of the original 54 675 set. To identify genes differentially expressed we employed significant analysis of microarray (SAM)(22) with FDR < 10% and 1000 permutations in the following comparisons: L-Tol versus L-Non-TOL, K-Tol versus K-Non-Tol, L-Tol versus HV and K-Tol versus HV. To study the set of genes differentially expressed between tolerant and nontolerant kidney and liver recipients pooled together, we first employed SAM to identify the genes differentially expressed at FDR < 10%. Then, in order to determine if the differences noted between tolerant and nontolerant recipients were likely to be due to the presence/absence of pharmacological immunosuppression, we used predictive analysis of microarrays (PAM)(23) to tentatively classify HVs as either tolerant or nontolerant. To construct the PAM-derived gene classifier only genes differentially expressed at FDR<5% and with a log2 ratio>0.25 were considered. The data discussed in this publication have been deposited in NCBI's gene expression omnibus (GEO) and are accessible through GEO series accession number GSE22707. Expression data from Sagoo et al. and Newell et al. (18,19) were directly accessed through GEO series accesion nombers GSE14655 and GSE22229, respectively.

Functional analysis of microarray gene expression data

To assess the deregulation of sets of genes associated with specific functional pathways we employed the gene set enrichment analysis (GSEA) method (24,25). The gene list analyzed comprised the filtered 27317-probe set ranked according to the SAM score. The databases of gene sets used in this paper were obtained from the Molecular Signatures Database (MSigDB; http://www.broadinstitute.org/gesa/msigdb/index.jsp), Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway Database (http://www.genome.jp/kegg/pathway.html) and the Protein Analysis Through Evolutionary Relationships (PANTHER) classification system (http://www.pantherdb.org/pathway/). Furthermore, we employed blood cell lineage-specific transcripts identified in the recently reported Haematology Expression Atlas (26) to generate additional gene sets containing transcripts that are preferentially expressed by each of the following PBMC subsets: CD4+ T and CD8+ T lymphocytes, CD14+ monocytes, CD19+ B lymphocytes, CD56+ NK cells and CD66b+ granulocytes. The contribution of cell lineage-specific transcripts was explored by deriving a Z-score from the SAM scores corresponding to the filtered 27 317-gene list and to each of the cell lineage-specific gene sets. SAM scores were then graphically displayed in a density plot. Z-scores > 2.5 (this corresponds to a p-value < 0.012) were considered as statistically significant.

Peripheral blood B-cell immunophenotyping

We used the Bm1–Bm5 classification system to identify B-cell developmental stages in the blood of the patients. Bm3/Bm4 were not included in the analysis because they were absent in the blood. The mature Bm-cell subpopulations were studied according to their expression of CD38, IgD and CD27 (27,28,30). Flow cytometry was performed on a BD LSRII analyzer with FlowJo software (TriStar, Inc., Ashland, OR, USA) using monoclonal antibodies: CD19 (IgG1/J4.119), CD38 (IgG1/HIT2), IgD (IgG2a/IA6), CD80 (IgG1/L307.4), CD86 (IgG1/IT2.2), CD27 (IgG1/M-T271), CD5 (IgG1/UCHT2), CD40 (IgG1/5C3), CD62L (IgG1/Dreg56), IgM (IgG1/G20–127), CD138 (IgG1/BP100), CD1d (IgG1/M-T101) (BD PharMingen, San Diego, CA, USA) and CD20 (IgG1/B9E9) (Immunotech, Luminy, France).

Real-time PCR (qPCR) gene expression experiments

Total RNA samples were reverse transcribed using polydT oligonucleotide and Maloney leukemia virus reverse transcription (Invitrogen, Cergy Pontoise, France), and qPCR was performed using commercially available primer/probe sets for HPRT1: Hs99999909_m1, CD19: Hs00174333_m1, BANK1: Hs01009378_m1 and CD20: Hs00544818_m1 from Applied Biosystems on an ABI Prism 7900 HT (Foster City, CA, USA). HPRT or CD19 for B lymphocytes were used as endogenous controls to normalize RNA amounts and relative expression between a sample and a reference was calculated according to the 2−ΔΔCq method.

Statistical analysis

The nonparametric Mann–Whitney test was used for comparison of the flow cytometry data from two groups of recipients using GraphPad Prism software v.4. The nonparametric Kruskal–Wallis test with Dunn's post-test was used for comparison of more than two groups using GraphPad Prism software v.4. Differences were defined as statistically significant when p < 0.05 (*), p < 0.01 (**) and p < 0.001 (***).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients, Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Liver and kidney tolerant recipients differ from both nontolerant patients and healthy individuals in their transcriptional profile

We employed SAM to conduct a pairwise group analysis comparing peripheral blood Affymetrix microarray expression patterns from K-Tol, K-Sta, K-CR and HV individuals. The number of genes differentially expressed at FDR<10% between the study groups analyzed is depicted in Table 2. K-Tol recipients displayed significant differences in gene expression with all study groups. A similar analysis was conducted on the liver dataset by comparing blood samples from L-Tol, L-NonTol and HV individuals. Again, significant differences in gene expression were noted when comparing L-Tol with both L-NonTol and HV individuals (Table 2). A complete list of all differentially expressed genes is provided as Supporting Information.

Table 2.  Number of genes differentially expressed at FDR<10% between the study groups compared (numbers displayed represent the up- and downregulated genes, respectively)
 L-TolL-NonTolK-TolK-StaK-CRHV
L-Tol 261/03931/14833723/23354018/2211879/281
L-NonTol  1787/3633723/23351779/2387174/310
K-Tol   621/37235/472556/32
K-Sta    1035/11822120/2264
K-CR     2056/2325

Comparative differential gene expression reveals minimal overlap between the liver and kidney tolerance-related datasets

To identify the potential similarities between tolerance-associated expression patterns in kidney and liver transplantation, we computed the degree of overlap between the L-Tol versus L-NonTol and the K-Tol versus K-Sta expression datasets (Figure 1A). We selected K-Sta recipients as the best counterpart for L-NonTol recipients because (a) stable kidney recipients under maintenance immunosuppression have a very low probability of spontaneous operational tolerance (less than 10% as estimated by Brouard et al. (9); and (b) in common with L-NonTol they exhibit stable graft function. Only two genes were simultaneously differentially expressed at FDR<10% in both datasets (TROVE2 and accession number AI885665). The absence of overlap was also manifested when the set of liver-related differentially expressed genes was compared with the K-Tol versus K-CR dataset (only one coincidence, HIPK2, was found; Figure 1B). To further investigate if a common ground between tolerant and nontolerant blood samples could be elicited, we pooled together tolerant samples from liver and kidney tolerant recipients and compared them with those of nontolerant recipients. At FDR<10% there were 106 genes differentially expressed between the tolerant and nontolerant cohorts (Supporting Information). However, when we measured the expression levels of these genes on the group of HVs, the majority of them were found to closely resemble tolerant recipients (a PAM-derived gene classifier containing the 51 most informative genes with FDR<5% classified 8 out of the 12 HVs as potentially tolerant; data not shown). Furthermore, the most highly overexpressed gene in tolerant recipients was EGR2, a well-known target of NFAT transcription factor. Taken together, these results suggest that the expression differences observed after pooling tolerant and nontolerant liver and kidney recipients are likely to be due to the effects of pharmacological immunosuppression.

image

Figure 1. Tolerance associated expression patterns differ in liver and kidney transplantation. Venn's diagrams display the number of upregulated (black) and downregulated (gray) genes differentially expressed with FDR<10% between tolerant and nontolerant liver and kidney recipients. (A) Differential gene expression of L-Tol versus L-NonTol is compared to that of K-Tol versus K-Sta. (B) Differential gene expression of L-Tol versus L-Non-Tol is compared to that of K-Tol versus K-CR.

Download figure to PowerPoint

Liver and kidney tolerance-related expression patterns are enriched in different functional pathways

To further study potential similarities in gene expression between liver and kidney tolerant recipients at a functional level, we then conducted an unbiased detailed functional enrichment analysis utilizing GSEA. Kidney and liver tolerance-related expression profiles were enriched in different functional pathways, and this was apparent when tolerant recipients were compared with both nontolerant recipients (L-NonTol and K-Sta) and HVs (Table 3). The potential contribution of different PBMC subsets to tolerance-related expression profiles in liver and kidney transplantation was then investigated employing the Haematology Expression Atlas (26), which contains gene sets preferentially expressed by specific blood cell subsets. Again kidney and liver recipients clearly differed. Thus, transcripts preferentially expressed by CD56+ lymphocytes were significantly over-represented in L-Tol recipients (compared with both L-NonTol and HVs), while the L-NonTol phenotype was enriched in CD19+ and CD4+ cell-related transcripts (Figure 2A). In contrast, transcripts characteristic of CD19+ cells were significantly associated with K-Tol recipients (compared with either K-Sta and K-CR but not with HVs) while CD14+ and CD56+ lymphocyte associated gene sets were enriched in K-Sta recipients (Figure 2B).

Table 3.  Functional pathways associated with tolerance in liver and kidney transplantation (for each comparison a maximum of five pathways from PANTHER and KEGG pathway collections with p-value < 0.05 and q-value <0.25 are shown)
Tolerance-associated differential gene expression in liver recipients
Enriched in L-Tol recipients (L-Tol vs. L-NonTol)p-valueq-valueGenes with highest enrichment scoresClassification system
 Ras pathway0.0000.000RHOF, RAC2, MAPK1, SOS1, RAF1, MAP2K4PANTHER
 Natural killer cell-mediated cytotoxicity0.0000.002NCR1, PRF1, KIR3DL1, KIR3DL2, NFATC2, NCR3, SH2D1B, IL2RB, KLRF1KEGG
 Angiotensin II-stimulated signaling0.0000.003RHOF, GNGT2, MAPK1, RHOC, ITPR2, ARRB1PANTHER
 Pancreatic cancer0.0000.003RAC2, RAF1, SMAD3, TGFBR1, JAK1, CDK4, NFKB2KEGG
 Semaphorin axon guidance0.0010.004RHOF, RAC2, RHOC, FYN, AKAP13PANTHER
Enriched in L-Tol recipients (L-Tol vs. HV)
 Angiotensin II-stimulated signaling0.0000.005RHOF, GNGT2, MAPK1, RHOC, ITPR2, ARRB1PANTHER
 FAS signaling pathway0.0000.007FAS, FAF1, AKT2, CFLAR, DAXX, GSNPANTHER
 PDGF signaling pathway0.0000.015RPS6KA3, MAPK1, SOS1, ITPR2, DLC1, NRASPANTHER
 Natural killer cell-mediated cytotoxicity0.0000.055FAS, NCR1, KLRC3, LCP2, NRAS, KIR3DL2, KIR2DS1KEGG
 Pancreatic cancer0.0000.037RAC2, RAF1, SMAD3, TGFBR1, JAK1, CDK4, NFKB2KEGG
Enriched in L-NonTol recipients (L-Tol vs. L-NonTol)
 Ribosome0.0000.000RPS8, RPS13, RPL10L, RPS25, RPL11KEGG
 Cell communication0.0000.013KRT72, KRT31, KRT15, KRT14, KRT73KEGG
 Oxidative phosphorilation0.0320.24COX4I2, COS4I1, ATP5L, COX7A2, COX6CKEGG
 Maturity onset diabetes of the young0.0150.24SLC2A2, IAPP, PKLR, GCK, HES1KEGG
Tolerance-associated differential gene expression in kidney recipients
Enriched in K-Tol recipients (K-Tol vs. K-Sta)p-valueq-valueGenes with highest enrichment scoresClassification system
 Cell cycle0.0180.24PIP5K1A, CCNB1, CCNB2, PSMD11, E2F3, CCNE1PANTHER
Enriched in K-Tol recipients (K-Tol vs. HV)
 PDGF signaling pathway0.0000.003RPS6KA3, MAPK1, EHF, STAT3, RIT1, KRASPANTHER
 mTOR signaling pathway0.0000.010RPS6KA3, MAPK1, LYK5, EIF4B, RPS6KEGG
 EGF receptor signaling pathway0.0000.018MAPK1, STAT3, MAP2K7, KRAS, PRKD3, GAB2, CBLBPANTHER
 Integrin pathway analysis0.0000.014MAPK1, ARF1, MAP2K7, KRAS, ARF6, ITGAVPANTHER
 B-cell activation0.0000.016LYN, MAPK1, KRAS, MAP3K3, NFKB2, PRKCE, NFKB1PANTHER
 TGF-beta signaling pathway0.0000.025MAPK1, KRAS, RIT1, CREBBP, SMURF1, TGFBR2PANTHER
Enriched in K-NonTol recipients (K-Tol vs. K-NonTol)
 Galactose metabolism0.0140.186PFKL, PGM1, GLA, RDH12, GLB1, HK1, AKR1B10KEGG
 Pathogenic E. coli infection0.0000.134TUBB1, TUBB2C, HCLS1, WAS, ACTB, NCL, TUBB3KEGG
 Glutathione metabolism0.0330.234GCLC, GSTM4, GSTP1, GSS, G6PD, GPX4, GSTM1KEGG
image

Figure 2. The expression of cell lineage-specific transcripts reveals that different PBMC subsets contribute to the liver and kidney tolerance-related expression patterns. (A) In L-Tol recipients CD56-specific transcripts are significantly overexpressed, while CD19-specific transcripts are significantly underexpressed. In contrast, the expression of CD14-specific genes is randomly distributed. (B) In K-Tol recipients CD19-specific transcripts are significantly overexpressed while CD14- and CD56-specific transcripts are significantly underexpressed. Density plots show the distribution of the expression of the filtered 27 317-gene list (solid lines) compared with the distribution of the expression of cell lineage-specific genes with p-value < 0.01 (dotted lines) in liver (L-Tol vs. L-NonTol) and in kidney (K-Tol vs. K-Sta) transplant recipients. Comparisons with Z-scores > 2.5 are considered statistically significant.

Download figure to PowerPoint

In contrast to kidney recipients, tolerant and nontolerant liver recipients do not exhibit differences in peripheral blood B-cell phenotypic markers

We and others have recently determined that in kidney transplantation operationally tolerant recipients exhibit an expanded number of peripheral blood B cells displaying a unique set of cell surface membrane markers (11,13,18–20). On the other hand, and in contrast to operationally tolerant liver recipients, no differences in the number of NK cells are observed in tolerant kidney recipients (13,18,19). To investigate the B-cell immunophenotype in operational tolerance following liver transplantation, we analyzed PBMC samples from Tol and Non-Tol liver recipients employing identical reagents and flow cytometry equipment than for kidney recipients (20). In contrast to K-Tol recipients, L-Tol and L-NonTol did not differ in the relative or absolute numbers of peripheral blood B lymphocytes (Figure 3A). Furthermore, a detailed immunophenotypic analysis of these cells according to Bm1–Bm5 classification failed to identify in tolerant liver recipients the increase in the number of activated and memory B-cell subpopulations found expanded in K-Tol recipients (Figure 3B and C). Similarly, the increased surface expression of activation and migratory molecules (CD19, CD80, CD86, CD27, CD28, CD40 and CD62L) and inhibitory/regulatory molecules (CD5 and CD1d) noted on circulating B cells of K-Tol recipients was not observed in L-Tol recipients (data not shown).

image

Figure 3. Operationally tolerant liver recipients do not display and increased number of circulating B-cell subsets. In contrast to drug-free operationally tolerant kidney recipients, tolerant liver recipients do not exhibit an increase in total B cells (A), memory B cells (B) or activated B cells (C).

Download figure to PowerPoint

The enrichment in B-cell-related transcripts is stable over time and reproducible across different cohorts of operationally tolerant kidney recipients

To assess if the enrichment in B-cell-related markers was stable over time we quantified in two sequential blood samples (0.8 to 4 years apart) collected from four K-Tol recipients the number of CD19+ cells and the expression of CD19, CD20 and Bank1. No significant differences between the two time points were noted (Supporting Figure S1). Similarly, no differences were observed over time when a more detailed flow cytometric analysis that included CD27, CD32b, CD81, IgD, CD38, CD86, CD80 was performed (data not shown). Next, to confirm the reproducibility of our results on different cohorts of tolerant recipients we reanalyzed two recently published renal transplantation tolerance expression datasets (18,19). In contrast to our own experiments, these studies had been performed employing either a different microarray platform (RISET 2.0 custom Agilent microarray instead of Affymetrix in Sagoo et al.) or a different type of sample (whole blood instead of PBMCs in Newell et al.). Despite the technical divergences, B-cell-related transcripts were significantly enriched in drug-free kidney tolerant patients when compared with stable immunosuppressed recipients (Figure 4A and B and Table 4). Interestingly, the conjoint analysis of all three kidney tolerant cohorts revealed 35 genes that were differentially expressed in all three studies, 24 of which were preferentially expressed on B cells (Figure 4C). In agreement with our own results, the over-representation of B-cell-related transcripts was not seen when kidney tolerant patients were compared with healthy individuals (see Table 4 and Supporting Figure S2). Lineage-specific transcript analysis of Sagoo et al. and Newell et al. datasets also revealed enrichment of NK-related mRNAs in the two groups of tolerant kidney recipients. However, the statistical significance of these genes either disappeared or was substantially decreased when tolerant kidney recipients were compared to healthy individuals (Table 4).

image

Figure 4. The enrichment in B-cell-related transcripts is common to all three cohorts of tolerant kidney recipients analyzed (Newell et al., Sagoo et al. and our own experiments). The expression of cell lineage-specific transcripts reveals enrichment in B- and NK-related mRNAs in the tolerance expression datasets reported by (A) Newell et al. and (B) Sagoo et al. (C) The three cohorts of tolerant recipients overlap in 35 transcripts, 24 of which (asterisk) are B-cell related. Analyses were performed in all cases by employing the list of genes differentially expressed between drug-free tolerant patients and stable recipients under calcineurin inhibitor-based maintenance immunosupression.

Download figure to PowerPoint

Table 4.  B- and NK-cell lineage-specific transcripts in Newell et al. and Sagoo et al. tolerance expression dataset (only the most significant 30 transcripts with FDR < 10% are shown)
Newell et al.Sagoo et al.
NK-cell-related transcripts enriched in tolerant recipients (tolerant vs. stable recipients)NK-cell-related transcripts enriched in tolerant recipients (tolerant vs. stable recipients)
Gene symbolFDRGene symbolFDRGene symbolFDRGene symbolFDR
BNC20CD2470.47SH2D1B0GZMB0
AKR1C30BNC22.25AKR1C30PRSS230.15
SH2D1B0SPON22.25LINGO20NCAM10.27
TKTL10.27KIR3DL13.38KLRF10PRF10.44
IL12RB20.27PTGDR3.38CD2470CD1600.63
CD2470.47PRF13.38CLIC30GNLY0.81
IL2RB0.71FGFBP23.38FASLG0EDG80.81
ADAMTS10.71MYOM23.38BZRAP10KLRC10.81
GZMB0.71CCL43.38TKTL10GK51.10
KLRF10.95GNLY3.38CCL40CX3CR13.74
KIR2DL11.13PTGDS4.70IL12RB20PDGFRB6.9
COL13A11.13PRSS234.70IL2RB0OSBPL59.99
GK51.13FASLG4.70    
CLIC31.63NCAM14.70    
FEZ11.63GNLY4.70    
NK-cell-related transcripts enriched in tolerant recipients (tolerant vs. healthy individuals)NK-cell-related transcripts enriched in tolerant recipients (tolerant vs. healthy individuals)
Gene symbolFDRGene symbolFDRGene symbolFDRGene symbolFDR
PTGDR3.30CX3CR15.76No significant enrichment
CCL43.93ASCL27.,31    
C1orf215.02CMKLR18.02    
PRSS235.55GK59.59    
BNC25.55      
 
B–cell-related transcripts enriched in tolerant recipients (tolerant vs. stable recipients)B-cell-related transcripts enriched in tolerant recipients (tolerant vs. stable recipients)
Gene symbolFDRGene symbolFDRGene symbolFDRGene symbolFDR
 
TCL1A0FCRL10TCL1A0FCRLA0
TSPAN130HLA-DOB0CD2000CD830
PPAPDC1B0AFF30FCRL20HLA-DOA0
CD2000BTLA0BTLA0FCER20
FCER20BLNK0CD79B0CD400
VPREB30TXNDC50BLNK0FCRL50
PNOC0BACH20CD220IRF40
FCER20PTPRK0HLA-DOB0STAG30
CD720C9orf450PNOC0QRSL10
IGJ0BCL11A0MS4A10CD190
FAM129C0P2RX50CD79A0AFF30
TNFRSF170BACH20BLK0CCR60
SPIB0ID30IGKV3–200IGJ0
KLHL140TSPAN30TSPAN30QRSL10
    PLEKHG10TCF40
 
B-cell-related transcripts enriched in tolerant recipients (tolerant vs. healthy individuals)B-cell-related transcripts enriched in tolerant recipients (tolerant vs. healthy individuals)
Gene symbolFDRGene symbolFDRGene symbolFDRGene symbolFDR
 
HLA-DQB15.76SAV18.02UVRAG0  

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients, Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Achieving long-term acceptance of transplanted organs in the absence of sustained pharmacological immunosuppression is considered a highly desirable goal in clinical organ transplantation. Several recent studies both in liver and kidney transplantation have attempted to characterize the immune parameters of drug-free operationally tolerant recipients with the hope of identifying clinically useful diagnostic biomarkers and gaining insight into the mechanisms of allograft tolerance in humans. Thus, we previously reported that tolerant recipients could be distinguished from nontolerant patients on the basis of blood gene expression patterns and cell phenotypic markers and that these biomarkers could be employed to design predictive models of tolerance both in kidney and in liver transplantation (8,8,19). In this study, our aim was to directly compare blood samples obtained from tolerant and nontolerant liver and kidney recipients in order to address some of the methodological limitations of previous studies. Based on the findings from our previous studies, we chose to employ peripheral blood whole-genome transcriptomic analyses and detailed B-cell immunophenotypic studies. Our results confirm previous reports indicating that drug-free operationally tolerant liver and kidney recipients display blood gene expression profiles that are different from those of nontolerant recipients and healthy individuals. However, tolerance-related expression profiles considerably differed between kidney and liver patients and there was minimal overlap between the two datasets. The absence of similarities was apparent when tolerance-related gene lists were directly compared and also when expression-derived molecular pathways were examined employing GSEA. Functional enrichment analyses such as GSEA are generally considered to provide greater sensitivity than conventional fold-change based statistical analyses to unravel subtle relationships between groups of genes and biological phenotypes. In our study, enrichment analyses employing blood cell lineage-specific transcripts revealed in fact opposite results for liver and kidney recipients. Thus, B-cell-specific transcripts were significantly overexpressed in tolerant kidney recipients while the same genes were downregulated in tolerant liver recipients. Conversely, NK-related genes were upregulated in liver but not in kidney tolerant recipients. In agreement with microarray expression results, total B cells were expanded in tolerant kidney recipients but not in tolerant liver patients. Furthermore, the specific blood B-cell phenotype reported in drug-free tolerant kidney recipients (20) could not be found in tolerant liver patients.

In order to assess the reproducibility of our findings, we reanalyzed the results of two recently published multicenter studies aiming at the transcriptional characterization of tolerant kidney recipients. In agreement with our data, the most significant finding in these two studies was the over-representation in tolerant kidney patients of B-cell-related transcripts, some of which coincided in the three cohorts of tolerant recipients (18–20). B-cell-related transcripts, however, were not differentially expressed when tolerant kidney recipients were compared with healthy individuals. In contrast with our experiments, certain NK-related transcripts were also enriched in the two cohorts of tolerant kidney recipients described by Sagoo et al. and Newell et al. and even overlapped with the expression markers identified in our cohort of tolerant liver recipients. In their original reports (18,19), Sagoo et al., but not Newell et al., described an increased number of NK cells in tolerant recipients compared with stable recipients under immunosuppression. These coincidences are somehow surprising, given the fact that there were substantial methodological differences between the two previously published studies (18,19) and our own experiments in liver recipients. The significance of the NK-related transcripts, however, markedly decreased when tolerant kidney recipients were compared with healthy individuals, raising the concern of the potential confounding effect of pharmacological immunosuppression. Despite this, the role of NK-related mRNAs shared between liver and kidney tolerant recipients merits further investigation.

Our study suffers from the same major methodological constraint common to all previous reports in the field, namely the small number of tolerant recipients available for analysis. The sample size limitation precludes the use of appropriate statistical tools to adjust for all potential confounding variables. Hence, it is not possible to be entirely certain of the specificity of the so-called tolerance fingerprints. Despite this, the observation that operationally tolerant kidney and liver recipients exhibit different immune features could be interpreted as suggesting that the absence of therapeutic immunosuppression is unlikely to be the major driving force in the development B- and NK-cell tolerance-related biomarkers. Along this line, the absence of substantial similarities between operationally tolerant kidney and liver recipients could reflect the divergent immunological outcomes of kidney and liver allografts both in the clinic and in experimental animal models (31). In any case, the unambiguous validation of all proposed biomarkers of operational tolerance, as well as the elucidation of the mechanisms involved in the development and maintenance of this phenomenon, will require enrolment of larger populations of tolerant and nontolerant recipients, simultaneous analysis of blood and graft tissue samples, and, ultimately, the performance of prospective immunosuppression weaning clinical trials.

In short, while tolerant kidney recipients are characterized by an expansion of peripheral blood B cells consistently exhibiting specific transcriptional and immunophenotypic traits, this is not apparent in tolerant liver patients. Operationally tolerant liver and kidney recipients, however, appear to overlap in a set of NK-related transcripts, although this observation is not reproducible in all patient cohorts analyzed. Our results reveal that certain tolerance-related immune features, but not others, are consistently present across all cohorts of operational tolerant recipients. This provides a set of reproducible biomarkers that should be explored in future large-scale immunomonitoring trials.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients, Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

This work was supported by a grant from the Ministerio de Educación y Ciencia, Spain (SAF2008-04092 to A.S-F). CIBEREHD is funded by the Instituto de Salud Carlos III. R. This work is also part of a French Transplantation Research Network (RTRS) supported by the ‘Foundation de Coopération Scientifique’ CENTAURE.

Disclosure

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients, Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients, Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients, Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Figure S1: In blood samples from tolerant kidney recipients the enrichment in B-cell-related markers is stable over time. No differences in the frequency of CD19+ blood cells (A) and in the expression of CD20 (B), CD19 (C) and Bank1 (D), were observed when two sequential blood samples (0.8–4 years apart) from four K-Tol recipients were analyzed and flow cytometry and qPCR, respectively.

Figure S2: Blood samples from tolerant kidney recipients are enriched in B-cell-related transcripts when compared with stable kidney recipients but not with healthy individuals. Heatmap display of the 35 genes overlapping between the three cohorts of tolerant kidney recipients analyzed (Newell et al. in the upper panel, Sagoo et al. in the middle panel and our own experiments in the lower panel). For the three cohorts of patients the expression of each gene in tolerant recipients, stable recipients and healthy individuals is shown. Rows represent genes and columns represent samples. The intensity of each color denotes the standardized ratio between each value and the average expression of each gene across all samples. Red-colored pixels correspond to an increased abundance of the transcript in the indicated sample, whereas green pixels indicate decrease transcript levels. The 35 B-cell-related transcripts tend to be overexpressed in both tolerant kidney recipients (TOL) and healthy individuals (HC) and under-expressed in stable immunosuppressed kidney recipients (STA).

FilenameFormatSizeDescription
AJT_3638_sm_Figure1.ppt178KSupporting info item
AJT_3638_sm_Figure2.ppt322KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.