SEARCH

SEARCH BY CITATION

Keywords:

  • Gene expression;
  • HCV recurrence;
  • hepatitis C virus;
  • liver transplantation;
  • microRNA;
  • microarray;
  • prediction

Abstract

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

Diagnosis and prediction of the severity of hepatitis C virus recurrence (HCVrec) after liver transplantation (LT) remain a challenge. MicroRNAs have been recently recognized as potential disease biomarkers. Archival liver biopsy samples from 43 HCV+ LT recipients were collected at clinical HCVrec time and at 3 years post-LT. Patients were classified as progressors (P = F0/F1) or nonprogressors (NP = F3/F4) according to the severity of fibrosis on the 3-year biopsy. Training (n = 27) and validation (n = 16) sets were defined. RNA was isolated from all biopsies at clinical HCVrec time, labeled and hybridized to miRNA-arrays. Progressors versus nonprogressors were compared using the two-sample t-test. A p-value ≤0.01 was considered significant. The ingenuity pathway analysis tool was used for microRNA and miRNA:mRNA ontology data integration. Nine microRNAs were differentially expressed between groups. A supervised cluster analysis separated samples in two well-defined groups (progressors vs. nonprogressors). Pathway analysis associated those microRNAs with hepatitis, steatosis, fibrosis, cirrhosis and T cell-related immune response. Data integration identified 17 genes from a previous genomic study as 9-microRNAs signature targets. Seven microRNAs were successfully validated in the validation set using QPCR. We have identified a 9-microRNA signature able to identify early post-LT patients at high risk of severe HCVrec during long-term follow-up.


Abbreviations
ACR

acute cellular rejection

ALT

alanine aminotransferase

FFPE

formalin-fixed paraffin embedded

HCV

Hepatitis C virus

HCVrec

HCV recurrence

IFNγ

interferon-gamma

IPA

Ingenuity pathway analysis

LT

liver transplantation; miRNA, microRNA

NP

nonprogressor P, progressor

QPCR

quantitative real-time PCR

TH1

T helper 1

TH17

T helper 17

Treg

regulatory T cells

Introduction

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

Hepatitis C virus (HCV)-cirrhosis is the leading indication for liver transplantation (LT) among adults in the Western World [1]. Moreover, a twofold increase in the number of HCV+ LT candidates is expected for the next decade despite a predicted decline in the incidence of HCV infection [1-3]. HCV recurrence (HCVrec) post-LT is triggered by the interaction between virus and host immune response as early as at the time of graft reperfusion [4].

Compared to nontransplant patients, hepatitis C infection in immunosuppressed LT recipients is characterized by an accelerated fibrogenesis and faster decompensation of allograft cirrhosis [5, 6]. Not surprisingly, LT for HCV infection is associated with decreased patient and graft survival when compared to other indications [7, 8]. The mechanisms responsible for the accelerated liver damage in HCV-infected LT recipients remain largely unknown. Numerous donor, recipient and post-LT factors such as donor age [9] and type/intensity of immunosuppression [10-12], have been associated with the aggressiveness of HCVrec. However, no single variable appears accurate enough to predict early after LT long-term histological outcome.

The high variability in the disease progression among LT recipients with HCVrec indicates the urgent need of biomarkers to identify early posttransplant patients at high risk of developing severe fibrosis. Therefore, outcome prediction for those patients will contribute to the development and implementation of interventional strategies to prevent graft failure as well as patient morbidity and mortality. Currently, histological examination of the liver is the gold standard of care to assess severity of HCVrec in the posttransplant setting [13-15]. As disadvantage, the use of liver biopsy is limited by its invasiveness, risk of complications [16], and its sampling variability [16, 17]. Therefore, increasing attention is being drawn to the use of noninvasive tests of liver fibrosis such as the AST-to-platelet ratio index (APRI) [18, 19], the FIB-4-index [20], transient elastography (FibroScanR) [21] and Fibrotest [22]. Development of accurate noninvasive markers of fibrosis is even more relevant after LT considering that allograft dysfunction with fibrosis due to HCVrec is the one of the main causes of graft loss and need of retransplantation among HCV+ LT recipients [1-4]. Unfortunately, none noninvasive biomarkers have been further established as potential predictors for severe HCVrec neither validated for clinical use [23]. Recently we reported that gene expression profiles in liver tissue obtained at the time of clinical HCVrec may predict the severity of fibrosis progression after LT [24]. As consequence of these findings, the assessment of early histopathological parameters and specific genomic profiles justify a liver protocol biopsy in LT patients with HCVrec disease.

MicroRNAs (miRNAs) are involved in several and diverse physiological and pathological processes. Even more, the interest about miRNAs as biomarkers has increased in recent years due to the deregulated expression of these small molecules in association with disease development. Hepatic diseases are not the exception and several miRNAs have been found to be involved in liver disease and malignancy [25]. In this study, miRNA expression profiles were evaluated in archival formalin-fixed paraffin-embedded (FFPE) liver biopsy samples obtained at the time of clinical HCVrec using microarray technology to (1) identify a miRNA signature able to predict severity of fibrosis post-LT; (2) correlate the identified miRNA with our previous reported genomic profiles and (3) validate the miRNA expression in an independent set of patients.

Patients and Methods

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

Liver tissue samples

The Institutional Review Board approved the study and written informed consent was obtained from all patients. The mean follow-up post-LT was 7.2 ± 2.3 years. All patients underwent liver biopsies at clinical HCVrec time, as defined by detectable HCV RNA in serum at the time of first ALT elevation, at 1-year of LT, annually thereafter and when clinically indicated. The median number of biopsies was 5 (range 3–12) per patient. Fibrosis severity was evaluated using the Knodell score system [26, 27]. The fibrosis stage was defined as 0 (no fibrosis, F0), 1 (fibrous portal expansion, F1), 3 (bridging fibrosis with portal-to-portal or portal-to-central septa, F3) and 4 (cirrhosis, F4). For study purposes, HCV+ LT recipients were classified depending on the histological severity of fibrosis at 3 years of follow-up as ‘nonprogressors’ (NP: fibrosis scores F0-F1) and ‘progressors’ (P: fibrosis score F3-F4) (Figure 1). Histologic activity index (HAI) was estimated to assay the necroinflammatory activity of liver biopsies collected at the time of clinical HCVrec using the Ishak scoring system [28]. Steatosis grade was also evaluated as a risk factor associated with severe HCV recurrence progression [29]. For study purposes, samples from patients with histological diagnosis of acute cellular rejection (ACR) and or active hepatotropic viral infections were excluded to preserve the histological and molecular processes strictly associated with HCVrec disease post-LT. Similarly, the study only included patients diagnosed as infected with HCV genotype 1 to represent the prevalent genotype in the United States, with more than 70% of infected individuals [30]. Liver donors serologically positive for HCV were not included in this study.

image

Figure 1. Study design diagram. Schematic representation illustrating training and validation study sets from HCV-infected liver transplanted recipients. DDLT, deceased donor liver transplantation; HCVrec, HCV recurrence.

Download figure to PowerPoint

Sample processing and miRNA isolation

FFPE liver biopsy samples obtained at clinical HCVrec were used for the molecular studies. Total RNA was isolated using a High Pure miRNA Isolation Kit (Roche, IN, USA) following the manufacturer instructions for 1-column protocol. Briefly, three 10 Mm thickness paraffin-embedded tissue sections per sample were deparaffinized with xylene for 5 min at room temperature. Deparaffinized samples were rehydrated twice using 100% ethanol. Hydrated tissues were centrifuged (14 000 g; 2 min at room temperature) and the resulting pellet dried (10 min; 55°C). Tissue pellet disruption was performed overnight at 55°C in paraffin tissue lysis buffer, 10% SDS and Proteinase K. The 1-column protocol was followed and total eluted RNA treated with DNase I endonuclease. DNase withdrawal was performed by additional column passage. Total RNA concentration and purity were assessed using NanoDrop 1000 spectrophotometer. A 260/280 absorbance ratio range of 1.9–2.1 was considered acceptable for miRNA microarray chip hybridization.

Total RNA labeling and microarray hybridization

MiRNA profiles were evaluated in the biopsies collected at HCVrec from 27 patients (training set) including fibrosis progressors (n = 13) and nonfibrosis progressors (n = 14). Total RNA was labeled using the FlashTag™ Biotin HSR RNA Labeling Kit (Genisphere Inc., Hatfield, PA, USA) as recommended by the Affimetrix miRNA microarray user manual protocol. Briefly, total RNA (250 ng) from FFPE biopsies was subjected to Poly (A) tailing reaction followed by ligation of a biotinylated signal molecule. Labeled total RNA was used for GeneChip® miRNA v2.0 Array (Affymetrix, Santa Clara, CA, USA) hybridization. This array covers miRNAs from 131 organisms on a single chip and includes specifically 1105 mature human miRNAs from the Sanger miRNA database v15 and additional human snoRNAs and scaRNAs derived from the snoRNABase and Ensembl Archive. After hybridization each chip was scanned on an Affymetrix GeneChip® Scanner 3000 G7 according to the GeneChip® Expression Analysis Technical Manual procedures (Affymetrix). Raw intensities for every probe were stored in electronic files (.DAT and .CEL formats) by the GeneChip® Operating Software (GCOS) (Affymetrix).

MiRNA microarray intensity data analysis

The detection of individual intensity values above background for each probe set from the microarray raw data was performed using the Wilcoxon Test. Data normalization and expression summaries were obtained using the Robust Multiarray Analysis (RMA) algorithm. Quality control for obtained miRNA microarray results was evaluated by assessing the intensity values of five different spike oligonucleotides (spike in-control-2_st, spike in-control-23_st, spike incontrol-29_st, spike in-control-31_st and spike in-control-36_st) (Supporting Figure S1).

Prior to comparative statistical analysis between patient groups, the dataset was restricted to probe sets for 1105 mature hsa-miRNAs. Probe sets having average signals below the background were eliminated from the dataset. Thereafter, the comparison analysis between progressors (P) versus nonprogressors (NP) was fit using the two-sample t-test. A probe sets p-value ≤ 0.01 was considered significant.

Interaction network and functional analysis

Associated network functions and biological analyses were performed by using the ingenuity pathways analysis tool (IPA; http://www.ingenuity.com). Lists were generated containing those differentially expressed miRNAs between P versus NP alone and in combination with target genes identified from a previously reported genomic signature for HCVrec [24]. Both generated lists contained probeset IDs as clone identifiers, Entrez IDs and fold-change values mapped in the ingenuity knowledge base (genes + endogenous chemicals) as the reference set. The biological role and disease relevance were established for the identified miRNAs and miRNA:mRNA interactions through the associated network functions and canonical pathways analysis. The statistical significance value associated with the identified interaction networks and biological processes was calculated by the IPA tool. The p-value associated with a biological process or pathway annotation represents the statistical significance with respect to the total number of functions/pathways/lists eligible molecules for the dataset and molecules in the IPA reference set. p-values below 0.05 were considered statistically significant.

Reverse transcription and real-time PCR miRNA validation

A total of nine differentially expressed miRNAs were validated by Taqman® real-time PCR (QPCR) in an independent set of liver biopsies collected at clinical recurrence time. HCVrec biopsies from 16 HCV-recipients with (n = 7) and without (n = 9) fibrosis progression at 3 years posttransplantation were used as validation set. Reverse transcription reactions were performed using the TaqMan® miRNA assay kit (Applied Biosystems, Foster City, CA, USA) for each specific miRNA according to the manufacturer's protocol. The identified miRNAs and Taqman® miRNA assays (assay ID) v18 list included: hsa-miR-155 (002623), hsa-miR-34a (000426), hsamiR-222 (002276), hsa-miR-23b (000400), hsa-miR-361 (000554), hsa-miR-455–3p (002244), hsa-miR-30b (000602), hsa-miR-30c (000419) and hsa-miR-27b (000409). Amplification signal for each miRNA was normalized independently to the RNU43 expression as internal standard. The internal control was selected based on the Taqman® miRNA assay manufacturer recommendation for human tissues (Applied-Biosystems) and reported data as specific for liver tissue [31]. Gene expression analysis was performed using the delta-Ct value model (average Ct miRNA – average Ct standard). Statistical significances were evaluated using the analysis of variance (ANOVA) test. A p-value < 0.05 was considered to be significant.

Results

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

Study design and liver allograft recipient groups

The present longitudinal retrospective study includes 43 HCV+ patients prospectively evaluated after LT. All patients underwent liver transplantation due to HCV-cirrhosis and received liver allografts from deceased donors. The immunosuppressant protocol was similar in the all HCV-recipients patients and consisted of calcineurin inhibitor based protocol with addition of MMF and prednisone (weaned of at 3 months post-LT). None patient received antiviral treatment post-LT. The study design is illustrated in Figure 1. Surveillance for HCVrec disease progression was performed for at least 3 years post-LT.

Clinical and demographic characteristics of patients are shown in Table 1. Compared to patients with F0/F1 fibrosis stages at 3 years post-LT (NP), patients from the P group had significantly earlier clinical HCVrec onset, higher donor age and significantly higher histological necroinflammatory activity at the first HCVrec diagnosis biopsy. Mild sinusoidal or portal fibrosis was detected at the time of clinical HCVrec in most patients of the P group (85%) but the difference was not statistically significant. Similar statistical trend for necroinflammatory activity and with significantly severe fibrosis development in the P group were observed at the 3-year post-LT biopsy. No significant differences were observed in other donor or recipient variables (Table 1).

Table 1. Demographic and clinicopathological characteristics of the training group
Parameter NonprogressorsProgressorsp-Value
  1. Age at transplant time; 2chi square test without Yates’ correction; 3portal nonbridging fibrosis; 4portal bridging fibrosis.

Patients (n) 1413 
Recipient age (years)1Mean ± SD (range)53.8 ± 5.7 (42–65)53 ± 9.9 (30–67)0.984
Recipient gender (%)Male11 (78.6)11 (84.6)0.686
 Female3 (21.4)2 (15.4) 
Recipient race (%)White8 (57.1)10 (76.9)0.203
 Black African/American6 (42.9)2 (15.4) 
 Other0 (0.0)1 (7.7) 
Clinical HCV recurrence time (month)Mean ± SD (range)9.8 ± 3.0 (5–14)7.3 ± 2.6 (5–12)0.030
Histopathology first biopsyFibrosis3 (%)7 (50.0)11 (84.6)0.0572
 Steatosis (%)4 (28.6)6 (46.1)0.584
 Necroinflammatory Activity Score (range)2.4 (0–5)6.1 (2–12)0.001
Histopathology 3-year biopsyFibrosis (%)8 (57.1)313 (100)40.027
 Steatosis (%)6 (42.9)8 (61.5)0.560
 Necroinflammatory Activity Score (range)4.6 (2–7)9.7 (7–13)<0.0001
Donor genderMale8 (57.1)5 (38.5)0.332
 Female6 (42.9)8 (61.5) 
Donor raceWhite12 (85.7)12 (92.3)0.586
 Black African/American2 (14.3)1 (7.7) 

MicroRNA profiles at clinical HCVrec and progression to fibrosis

The miRNA expression profile was evaluated in the liver biopsy sample at the time of HCVrec in 27 patients included in the training set using miRNA microarray technology. After removal of absence calls or probe set with intensities below background a pairwise comparison analysis was performed between P versus NP groups. From the analysis, nine miRNAs (three upregulated and six downregulated) were individually found to be significantly differentially expressed between groups (Table 2).

Table 2. Identified miRNAs as predictors of the severity of HCV recurrence
MoleculeProbe setmiRNA sequenceP vs. NPp-Value
  1. P = progressors; NP = nonprogressor.

miRNA-155hsa-miR-155_stUUAAUGCUAAUCGUGAUAGGGGUupregulated0.0001
miRNA-34ahsa-miR-34a_stUGGCAGUGUCUUAGCUGGUUGUupregulated0.0105
miRNA-222hsa-miR-222_stAGCUACAUCUGGCUACUGGGUupregulated0.0005
miRNA-23bhsa-miR-23b_stAUCACAUUGCCAGGGAUUACCdownregulated0.0116
miRNA-361hsa-miR-361–5p_stUUAUCAGAAUCUCCAGGGGUACdownregulated0.0027
miRNA-455hsa-miR-455–3p_stGCAGUCCAUGGGCAUAUACACdownregulated0.0049
miRNA-30bhsa-miR-30b_stUGUAAACAUCCUACACUCAGCUdownregulated0.0108
miRNA-30chsa-miR-30c_stUGUAAACAUCCUACACUCUCAGCdownregulated0.0002
miRNA-27bhsa-miR-27b_stUUCACAGUGGCUAAGUUCUGCdownregulated0.0008

A supervised hierarchical clustering analysis including the nine miRNAs is shown in Figure 2. Results of this analysis showed that patients were clustered into two well-defined groups featured by progression or nonprogression to severe fibrosis. Only one patient with progression of the disease (case 19, Figure 2) was found miss-classified among the NP group. Interestingly, one patient (case 13, Figure 2), classified as NP at 3 years post-LT, developed fibrosis progression at 5 years of follow-up. Despite clustering together with the NP group, the visual inspection of the miRNA profile for this patient mimics the progressors’ profile.

image

Figure 2. Supervised hierarchical clustering analysis. The heat map and agglomerative dendrogram were determined using the Ward method and represent samples clustering when incorporating the nine miRNAs significantly and differentially expressed. The dendrogram correlation distance bar among subgroups is shown. Each patient was represented by a number to refer within the article text. Green = downregulation; red = upregulation; P = progressors; NP = nonprogressors.

Download figure to PowerPoint

Molecular and disease-related role of the miRNA signature

Ontology and molecular pathway analyses using the IPA tool were performed based on the identified 9-miRNA signature to identify its biological role in HCVrec and fibrosis progression. From this analyses, cellular growth and proliferation and cellular development (miRNAs: 222, 155, 34a and 23b); cell cycle (miRNAs: 155, 34a), cell-to-cell signaling and interaction (miRNA-34a) and cellular compromise (miRNA-23b) were identified as top molecular cellular functions.

All miRNAs were included into a unique function network (score: 24) associated with cancer, endocrine system disorders and gastrointestinal disease (Figure 3). From the network analysis, TGFβ1, TNFSF12 and hydrogen peroxide derived from chemicals were the main regulatory molecules. Regulation of miRNA-34a and miRNA-155 was found to be associated to TGFβ1, which also increase hydrogen peroxide synthesis involved in regulation of the miRNA-361–5p, miRNA-23b, miRNA-30c and miRNA34a expression. TNFSF12 was associated with miRNA-455–3p, miRNA-27b and miRNA-23b. However, TNFSF12 was identified as a regulatory factor of the ICAM1 gene expression (associated with stellate cells activation, also regulated by TGFβ1, and T cell-related immune response) and DNA-damage-inducible transcript 4 encoding gene (DDIT4) that regulates the mTOR signaling pathway (Figure 3).

image

Figure 3. MiRNAs associated network function. The ingenuity pathway analysis tool included all nine miRNA into a unique associated network. Genes involved with network function are represented without color, while miRNAs are color in red or green depending on upregulation or downregulation, respectively. Whole arrow: direct interaction; dash arrow: indirect interaction. Molecular interactions are summarized in Supporting Table S3.

Download figure to PowerPoint

The IPA tool algorithm identified hepatocellular carcinoma (HCC; p = 1.1E-08), hepatitis (p = 6.5E-05), cirrhosis (p = 1.4E-04) and liver steatosis (p = 0.0011) as the most significant hepatic system diseases associated with the miRNA predictive profile. MiRNAs: 155, 30c, 222, 34a, 23b and 27b were found to be associated with HCC; miRNA-222, miRNA-27b with hepatitis and cirrhosis and miRNA-222, miRNA-34a with liver steatosis. Analysis of the expression levels of each miRNA indicated a direct correlation with a predictive outcome trend in favor of the development of each liver disease and disorder in patients with HCVrec progression (Supporting Table S1). HCVrec is also featured by accelerated fibrosis development as the hallmark of the disease. From our analysis, miRNA-155, miRNA30c and miRNA-23b were identified as being involved in the development of fibrosis. Moreover, the expression profile of miRNA-155 (upregulated) and miRNA30c (downregulated) were in agreement with fibrosis development prediction in patients with HCVrec at 3 years post-LT.

Furthermore, early changes in the expression levels of those miRNAs suggest an association with progressive and accelerated liver degeneration after HCVrec disease onset.

Association of the miRNA signature with physiological functions

To study the miRNA signature role in accelerated HCVrec and fibrosis development, a detailed analysis was focused on physiological functions using the IPA tool. A number of 15 physiological system development and functions were identified as significantly associated with the described miRNA signature (Supporting Figure S2), including 3 miRNAs (miRNA-155, miRNA-34a and miRNA-222). This analysis associated miRNA-155 and miRNA-34a with modulation of humoral and cellular immune response. In this regard, upregulated miRNA-155 expression associated and correlated with positive modulation of humoral immune response through proliferation and accumulation of polyclonal pre-B lymphocytes. Moreover, miRNA-155 was found to be involved in T cell-related immune response by positive development of TH1, TH17 and regulatory T cells (Treg). MiRNA-34a was found associated with arrest in development of double-negative T cells in addition to its role in T cell development. This molecular profile snapshot suggests an active T cell-related immune response scenario as predictor of future accelerated HCVrec and disease progression.

Target prediction analyses

To better understand the role of the miRNA signature in predicting fibrosis progression in HCVrec, a target analysis was performed by searching the IPA tool target database and comparison with our previous reported HCVrec predictive genomic signature [24].

From the analysis, 17 genes were identified as putative targets for the miRNAs involved in our miRNA signature. The expression of six genes (SEMA3A, FUT8, MGAT4A, LHFPL2, STIM2 and ZNF200) was found to be directly or indirectly correlated with at least two different miRNAs. The gene list and each associated miRNA are detailed in Table 3. Ontology and pathway analyses using the IPA tool was performed to determine the cellular function of the identified genes. Top molecular and cellular functions included carbohydrate metabolism, cell morphology, cell-to-cell signaling interaction and cellular assembly and organization. A total of eight canonical pathways including N-glycan synthesis, semaphoring signaling, keratan sulfate synthesis, neuregulin signaling, NFAT in immune response regulation and cardiac hypertrophy, calcium signaling and axonal guidance signaling were identified to be associated with these genes.

Table 3. Predictive target gene analysis for the miRNA signature over a previous genomic profile
Genes symbolEntrez IDP vs. NP1miRNAs2 miRNA- 155 ([UPWARDS ARROW])miRNA- 34a ([UPWARDS ARROW])miRNA- 222 ([UPWARDS ARROW])miRNA- 27b ([DOWNWARDS ARROW])miRNA- 30c ([DOWNWARDS ARROW])miRNA- 361 ([DOWNWARDS ARROW])miRNA- 23b ([DOWNWARDS ARROW])
  1. Gene expression data from Mas et al. 2010 [24]; 2miRNA-455 is not listed. Data genes and miRNA expression levels are indicated as ([UPWARDS ARROW]): upregulated, and ([DOWNWARDS ARROW]) downregulated referred to progression (P) versus no progression (NP) of the HCV recurrence disease. (X) Associated target.

SEMA3A10371([DOWNWARDS ARROW])X   X  
FUT82530([UPWARDS ARROW]) X X   
MGAT4A11320([UPWARDS ARROW]) XXX X 
RCAN11827([UPWARDS ARROW]) X     
C4orf33132321([UPWARDS ARROW])  X    
LHFPL210184([UPWARDS ARROW])  X X X
STIM257620([DOWNWARDS ARROW])   XX  
ZNF2007752([UPWARDS ARROW])   XX  
CLIP479745([UPWARDS ARROW])    X  
IL28RA163702([UPWARDS ARROW])    X  
KCNMB210242([DOWNWARDS ARROW])    X  
PHTF257157([UPWARDS ARROW])    X  
TMEM16864418([UPWARDS ARROW])    X  
CCDC6284660([DOWNWARDS ARROW])      X
ERBB2IP55914([DOWNWARDS ARROW])      X
SLC9A26549([DOWNWARDS ARROW])      X
SNAPC36619([DOWNWARDS ARROW])      X

From the gene list, nine molecules were found directly involved in liver disease (MGAT4A), inflammatory response (SEMA3A, SLC9A2) and disease (FUT8), hematological and immunological disease (KCNMB2, MGAT4A) and function (RCAN1, SEMA3A, PHTF2) and virus infection disease (IL28RA). In this regard, the MGAT4A gene was found to be target of 4 miRNAs and was involved in liver steatosis and peripheral T cell lymphoma. The increased expression of the FUT8 gene regulated by miRNA-34a and miRNA-27b was associated with embryonic fibroblast activation by regulating TGFβ-R1, and involved in inflammatory disease. Overexpression of the RCAN1 gene (miRNA-34a target) was associated with normal CD4+ TH1-cell development and function in agreement with downregulated SEMA3A gene (a miRNA-155 target) involved in cell-to-cell signaling and inhibition of thymocytes migration. In concordance with the miRNA signature, most of identified target genes are involved in modulation of T cell-mediated immune response and liver tissue injury as predictive mechanisms for HCVrec disease progression.

Independent validation analysis

The identified miRNA signature was tested in an independent validation set of patients (n = 16) by QPCR. Similar to the training set, the independent sample set included FFPE samples collected at the time of clinical HCVrec and classified as P (n = 7) and NP (n = 9) based on histological examination of the 3-year post-LT liver biopsy (Figure 1). As for the validation set, no demographic and clinical significant differences were observed between HCV-infected recipients groups regarding age, gender, race, time of clinical HCVrec and ACR episodes. Patients with progression to fibrosis revealed a trend in significance for histological fibrosis at the clinical HCVrec time (first biopsy) (Supporting Table S2). Also, biopsy at 3 years post-LT evidenced for patients in the P group increased but not significant necroinflammatory activity when compared with nonprogressors. As also shown in Supporting Table S2, despite no statistical significance HCVrec progression was also associated with older donors in the validation set.

From the QPCR results analyses, seven out of nine miRNA from the signature were individually found to be significantly and differentially expressed in the validation set (Table 4). More importantly, the differential expression profiles of those miRNAs were in concordance with the expression trend observed in the training set from miRNA microarrays assays.

Table 4. Validation of the miRNA signature in an independent sample set
miRNA name1miRBase ANTaqman assay name (ID)p-Value
  1. 1miRNA expression was normalized to RNU43.

miRNA-155MI0000681hsa-miR-155 (002623)0.012
miRNA-34aMI0000268hsa-miR-34a (000426)0.029
miRNA-222MI0000299hsa-miR-222 (002276)0.046
miRNA-361MI0000760hsa-miR-361 (000554)0.029
miRNA-455MI0003513hsa-miR-455–3p (002244)0.036
miRNA-30bMI0000441hsa-miR-30b (000602)0.036
miRNA-30cMI0000736hsa-miR-30c (000419)0.027

Discussion

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

One of the main challenges in the LT setting today is to identify early enough patients at risk of developing severe HCVrec during long-term follow-up. Outcome prediction in this patient population remains difficult due to the multifactorial nature of HCVrec [32, 33] and the differential diagnosis with other causes of graft dysfunction such as acute rejection [34, 35]. Therefore, strategies for improving the management of patients undergoing LT for HCV-cirrhosis should include not only the availability of more effective antiviral agents but also the identification of robust predictors of severe disease in order to prevent or minimize their impact on allograft function. The pathogenesis of accelerated fibrogenesis in HCV-infected LT recipients remains largely unknown. Recently, we reported and demonstrated the feasibility of performing gene expression analysis using FFPE liver samples and established the potential clinical usefulness of a set of molecular markers to predict early after LT the long-term severity of HCVrec [24].

MiRNAs are a class of small single stranded noncoding RNAs that function through translational repression of specific target mRNAs [36]. Deregulated miRNA expression has been associated with different type of diseases indicating a functional role of these in pathological development. The stability of these molecules, even in archival samples, their disease specificity and the availability of accurate techniques for detection and monitoring has encouraged miRNA biomarker research and applications in the transplant field.

This longitudinal study using prospectively collected liver biopsy samples at the time of clinical HCVrec and by protocol annually thereafter identified a signature of 9 miRNAs associated with progression of fibrosis. Importantly, this miRNA signature was validated in an independent set of HCV+ LT recipients and demonstrated organ specificity (molecular pathways involved in liver injury, malignancy and hepatic disease), and relationship with immune response (T cell lineages development). To the best of our knowledge, this is the first report describing a miRNA signature able to identify at early post-LT stage patients at high risk to develop severe progression to fibrosis/cirrhosis associated with HCVrec.

Specific histological variables early after LT have been proposed to predict severe fibrosis progression among HCV+ patients [29, 37-39]. Guido et al. [38] and Baiocchi et al. [39] described an association between HAI score and a severe fibrosis or cirrhosis. In agreement with these observations, histological assessment of the biopsies at the time of HCVrec in our study demonstrated significantly increased necroinflammatory activity in the P group, and even it became more significant at the 3-year biopsy, when comparing with patients with mild fibrosis. In another series reported by Pelletier et al. [29] steatosis was found to be an independent predictor of fibrosis development. Although in our study there were no significant differences in the prevalence of steatosis between the P and NP groups, two of the upregulated miRNAs identified (miRNA-222 and miRNA-34a) have been previously associated with steatosis [40].

Increased fibrogenesis early post-LT occurs in most patients developing severe HCVrec, most likely as a consequence of hepatic stellate cell activation [37]. Among our patients on the progressor group, early activation of fibrogenic events was reflected by deregulation of miRNAs involved in fibrosis development (miRNA-155, miRNA-23b and miRNA30c; [41]) and/or embryonic fibroblast activator genes (miRNA-27, mi-RNA-34a). In addition, we found that early post-LT miRNAs that promote positive immune response regulation (miRNA-155, miRNA-34a and miRNA-222) were upregulated in HCV+ LT recipients who developed severe fibrosis. As key regulator, miRNA-155 is thought to be involved in modulation of T cell-mediated immune response and inflammation by Treg cell proliferative activity maintenance (enhancing IL-2R signaling) and by stimulating TH1 and TH17 differentiation [42-46]. A recent study reported increased frequency of HCV-specific TH17 and Treg cells, and proinflammatory cytokines (IL-6 and IL-21β, but not IFN-γ) in peripheral blood of patients with progressive HCVrec [47]. This immune profile has been proposed to be triggered by increased degrees of allograft steatosis grades in the donor grafts [48-50]. Therefore, it is possible to consider differential host immune scenarios at early stages of HCVrec between LT recipients identified as progressors and nonprogressors. The use of biomarkers for early detection post-LT of patients at high risk of severe HCVrec and fibrosis development is critical to allow early therapeutic interventions. Even when the evaluation of the proposed miRNAs requires a liver biopsy as an invasive procedure, their implementations in the clinical practice still relevant considering the lack of biomarkers assessing HCVrec post-LT.

To conclude, herein we identified a set of miRNAs capable of identifying early post-LT HCV patients with high risk of developing severe HCVrec disease leading to fibrosis and cirrhosis. Results were successfully validated in an independent group of patients. These biomarkers were associated with liver injury and inflammation and demonstrating organ specificity. Future studies of larger cohorts of chronic HCV-infected LT patients, also including nonprevalent HCV genotypes in the United States, are required to further validate the predictive value of these biomarkers.

Disclosure

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

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

References

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

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Disclosure
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
ajt12047-sup-0001-FigureS1.tif807KFigure S1: Spike oligonucleotide intensity. The graphic represents the intensity values of different spike oligonucleotides obtained from each hybridized miRNA microarray chip. Patient samples are by unique intensity lines for each microarray.
ajt12047-sup-0002-FigureS2.tif488KFigure S2: Physiological pathways associated with miRNAs miR-155, miR-34a and miR-222. The significance of the identified physiological pathways is graphically represented by p-value negative logarithms.
ajt12047-sup-0003-TableS1.docx14KTable S1: Identified miRNA expression predictive outcome in hepatic disorders
ajt12047-sup-0004-TableS2.docx17KTable S2: Demographic and clinicopathological characteristics of the independent validation study set
ajt12047-sup-0005-TableS3.docx14KTable S3: Direct and indirect interactions between identified miRNAs and genes included in the associated network functions illustrated in Figure 3

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.