ORIGINAL ARTICLE

miRNA Signature of Hepatocellular Carcinoma Vascularization:How the Controls Can Influence the Signature

Silvia Fittipaldi1 • Francesco Vasuri1 • Sonia Bonora1 • Alessio Degiovanni1 •

Giacomo Santandrea1 • Alessandro Cucchetti2 • Laura Gramantieri3 •

Luigi Bolondi3 • Antonia D’Errico1

Received: 29 March 2017 / Accepted: 13 June 2017 / Published online: 21 June 2017

� Springer Science+Business Media, LLC 2017

Abstract

Background miRNA deregulation and vascular modifica-

tions constitute promising predictors in the study of hepa-

tocellular carcinoma (HCC). In the literature, the relative

miRNA abundance in HCC is usually determined using as

control non-matched tumoral tissue, healthy liver, or cir-

rhotic liver. However, a common standard RNA control for

the normalization toward the tissue gene expression was

not settled yet.

Aim To assess the differences existing in the quantitative

miRNA gene expression in HCC on tissue according to two

different liver controls.

Methods A wide array of miRNAs was analyzed on 22

HCCs arisen in cirrhotic and non-cirrhotic livers by means

of microfluidic cards. Control samples included total RNA

extracted from healthy and cirrhotic livers.

Immunohistochemistry for CD34 and Nestin was per-

formed to assess the pattern of intratumoral vascular

modifications.

Results Six miRNAs were deregulated in HCCs using

either controls: miR-532, miR-34a, miR-93, miR-149#,

miR-7f-2#, and miR-30a-5p. Notably, the miRNA expres-

sion changed significantly between HCCs arisen in cir-

rhotic and non-cirrhotic livers, according to the control

used for normalization. Different miRNA profiles were

found also in HCCs with different vascular patterns,

according to the control used for normalization.

Conclusions Our data confirm that the choice of the

methodology, and particularly the control used for nor-

malization, represents the main concern in miRNA evalu-

ation, particularly in a heterogeneous model such as liver

pathology. Still we observed the deregulation of some

common miRNAs as promising in HCC cancerogenesis

and progression. A standardized control will be a crucial

achievement to compare miRNA expression among dif-

ferent laboratories.

Electronic supplementary material The online version of thisarticle (doi:10.1007/s10620-017-4654-3) contains supplementarymaterial, which is available to authorized users.

& Francesco Vasuri

vasurifrancesco@libero.it

Silvia Fittipaldi

sv.fittipaldi@gmail.com

Sonia Bonora

soniabonora@hotmail.it

Alessio Degiovanni

alessiodegiovanni@gmail.com

Giacomo Santandrea

giacomo.santandrea@studio.unibo.it

Alessandro Cucchetti

alessandro.cucchetti@aosp.bo.it

Laura Gramantieri

laura.gramantieri@aosp.bo.it

Luigi Bolondi

luigi.bolondi@unibo.it

Antonia D’Errico

derrico@aosp.bo.it

1 ‘‘F. Addarii’’ Institute of Oncology and Transplant Pathology,

Department of Specialty, Diagnostic and Experimental

Medicine (DIMES), S. Orsola-Malpighi University Hospital,

V.le Ercolani 4/2, 40138 Bologna, Italy

2 Department of Medical and Surgical Sciences (DIMEC),

S. Orsola-Malpighi University Hospital, Bologna, Italy

3 Division of Internal Medicine, Department of Medical and

Surgical Sciences, University of Bologna, S. Orsola-Malpighi

University Hospital, Bologna, Italy

123

Dig Dis Sci (2017) 62:2397–2407

DOI 10.1007/s10620-017-4654-3

Keywords Cirrhosis � Hepatocellular carcinoma �Microfluidic card � miRNA � Neoangiogenesis

Introduction

Liver resection represents the therapeutic choice for hep-

atocellular carcinomas (HCCs) without concomitant liver

failure. Although post-resection outcome shows substantial

differences, best results are seen in Child-Pugh class A

patients with small lesions without microvascular invasion

(MVI) or metastases and with free surgical margins [1].

The predictors of HCC recurrence are mainly based on

clinical or morphological features, and new research lines

are strongly needed in order to find and validate new

prognostic markers. The study of miRNAs (small RNAs

characterized by regulatory functions on gene expression

that act as gene promoters or suppressors) represents a

promising approach toward further understanding of

human carcinogenesis and tumor progression [2]. Another

important approach is the analysis of neoangiogenesis and

vascular modifications in HCCs. Villa et al. [3] found five

genes involved in neoangiogenesis and/or endothelial

activation to be up-regulated in fast-growing HCCs. Our

group performed immunohistochemical and gene expres-

sion analysis on HCC vascular modifications. We found

four different vascular patterns in HCCs, each of them

characterized by different tumor architecture, vascular

morphology and phenotype, as well as by different

expression of Nestin, IGF1R and TGF-b1 [4].

In HCCs, the biosynthesis pathways of different miR-

NAs were found to promote neoangiogenesis, tumor

invasion and metastasization at different levels, from

transcriptional (e.g., down-regulation of miR-122) [5] and

post-transcriptional regulations (e.g., miR-99 and miR-21)

[6, 7], up to epigenetic and genetic changes [8].

In miRNA studies, choice of methodology is of utmost

importance. Several studies focused on the normalization

of gene expression toward reference genes [9–11]. How-

ever, methodological studies using a standard for normal-

ization toward tissue gene expression are needed. Hitherto,

research has focused on the importance of the reference

genes but not on the appropriateness of the control tissues

[9]. Despite the high number of papers about miRNA,

shared criteria for standardizing the methodology for the

quantification have not been established yet. There is also

lack of details on the sample source, the RNA isolation,

and the control tissues treatment and origin [12]. In the last

decade, some miRNA studies have been carried on with

arrays on human HCC and non-neoplastic tissue, and most

of them utilized adjacent liver tissue for the normalization

[6, 13, 14]. With this approach, Budhu et al. [14] found 20

miRNAs significantly associated with metastases, but only

one miRNA overlapped with the results obtained by

Murakami et al. [13]. In a more recent paper by Sato et al.,

healthy liver tissue was used for the normalization in the

analysis of both HCC and non-neoplastic tissue. Authors

found that different miRNAs were deregulated in both

HCCs and non-neoplastic livers and significantly correlated

with the tumor recurrence after transplantation [15].

Another methodology used the matched non-neoplastic

tissue as control for each HCC, this method is useful to

minimize the individual genetic variability, but it is

expensive and prevents the identification of a universal

control RNA with intralaboratory reproducibility. [16–18].

Thus, according to the goals of the studies, different con-

trols were selected to assess differential miRNA

expression.

The choice of the control tissues as reference for nor-

malization in each analysis is the key point, since it was

demonstrated that results can vary widely with different

reference controls, and a single miRNA can turn out to be

down-regulated in an experiment and up-regulated in

another, just by changing the control [19, 20]. The topic is

even more tangled in the study of liver tumors, where the

control tissue is seldom represented by ‘‘healthy’’ liver,

since most HCCs arise in a cirrhosis or in a chronic hep-

atitis. In the perspective of a miRNA-regulatory therapy,

this could be a challenge [20].

We hereby will demonstrate the concern and the dif-

ferences that emerged by using two different control tissues

to normalize miRNA gene expression in HCC. These two

controls represent the opposite biological backgrounds in

HCCs pathogenesis; total RNA was extracted from virus-

free and cancer-free healthy livers, as well as from HCV-

related cirrhotic livers. The different miRNA profiles of

resectable HCCs arisen in cirrhotic (c-HCC) and non-cir-

rhotic (nc-HCC) livers will be determined. Finally, miRNA

expression will be correlated with the patterns of vascular

modifications of HCCs, in order to confirm the existence of

different steps in HCC progression. The identification of

pattern-specific miRNAs possibly correlated with neoan-

giogenesis and tumor architecture could contribute to better

understand HCCs behavior.

Methods

Ethics

The present tissue study was approved in advance by the

Ethical Committee of the S. Orsola-Malpighi University

Hospital (protocol code APHCC-2012, reference number

85/2013/O/Tess). All patients were treated according to the

2398 Dig Dis Sci (2017) 62:2397–2407

123

ethical guidelines of the 1975 Declaration of Helsinki (6th

revision, 2008); informed consent was obtained from each

patient at the time of surgery.

Case Selection, Histopathology

and Immunohistochemistry

We retrospectively selected 22 consecutive patients sub-

mitted to surgical resection for HCC. Patients were 19

(86.4%) males and 3 (13.6%) females, mean age at surgery

62.9 ± 9.9 years (range 48–80 years). The etiology of

cirrhosis or chronic liver disease was HCV in 12 cases,

HBV in 5, metabolic in 2, alcoholic in 1, and cryptogenic

in 2.

Tissue was promptly fixed in formaldehyde solution 4%

directly in surgery room and buffered pH 6.9 before routine

processing. For each tumor, we evaluated the occurrence of

cirrhosis, the mean tumor size, Edmondson’s grade, the

occurrence of infiltrative margins, and MVI.

Immunohistochemistry (IHC) for CD34 (mouse mono-

clonal Q-BEnd-10, Dako A/S, pre-Diluted Copenhagen

Denmark) was automatically performed on FFPE 2-lm-

thick sections with Benchmark ULTRA� immunostainer

(Ventana/Roche, Ventana Medical Systems). Nestin

(mouse monoclonal 10C2, 1:400 dilution, Millipore USA)

immunostaining was manually performed as previously

described [21] with NovoLink Polymer Detection Kit

(Novocastra, Newcastle, UK).

The four main patterns of HCC vascular modifications

were determined according to the qualitative sinusoidal and

arteriolar positivity for CD34 and Nestin, as well as tumor

architecture, as previously described (4). Briefly, ‘‘Pattern

a’’ was defined in HCCs with microtrabecular and acinar

architecture, with CD34?/Nestin- sinusoids; ‘‘pattern b’’

in HCCs with similar architecture, but showing CD34?/

Nestin? sinusoids; ‘‘pattern c’’ was defined in HCCs with

macrotrabeculae surrounded by a CD34?/Nestin?

endothelium; ‘‘pattern d’’ characterized solid HCCs with

new-formed CD34?/Nestin? arteries (Fig. 1) [5]. These

vascular patterns were found only in HCCs, as they reflect

neoplastic modifications of the vascularization, and they

are not already occurred in healthy liver and/or in cirrhosis.

To address the important issue of the choice of tissue

control for normalization in miRNA array, we selected two

different pools:

1. A pool composed of 10 healthy livers obtained from

healthy multiorgan donors (five males and five

females; mean age 64.2 ± 10.2 years, range 48–75).

This tissue was retrieved during the histological graft

evaluation for donor safety in our Institution. Of note,

this selected tissue is virus-free and cancer-free.

2. A pool composed of 10 cirrhotic livers obtained from

practice routine analyses (five males and five females;

mean age 55.2 ± 8.4 years, range 46–72), all affected

by HCV-driven end-stage liver disease.

miRNA Arrays

RNA Extraction

A commercial kit was used for RNA extraction from FFPE

sections (FFPE Recover All, Life Technologies, Carlsbad,

CA, USA). Tissue slides were dewaxed with xylene. Tissue

was recovered from the slides with a scalpel by adding

10 ll of digestion buffer. The protocol was followed as

stated by manufacturer’s instruction. The RNA quality and

concentration were evaluated by using a ND-1000 spec-

trophotometer (NanoDrop, Fisher Thermo, USA); RNA

was considered pure if A260/A280 ratio = 1.9–2.1.

Megaplex Reverse Transcription Reaction

The reverse transcription assay was performed using the

TaqMan� MicroRNA Reverse Transcription Kit associated

with the Megaplex RT Primers Pool A and B (Life Tech-

nologies). Samples were loaded in the GeneAmp PCR

System 9700 (Life Technologies) following the manufac-

turer’s thermal conditions.

cDNAs Megaplex Pre-amplification mix

To increase the yield of cDNAs, a second step of pre-

amplification was performed with the TaqMan� PreAmp

Master Mix associated with Megaplex PreAmp Primers

Pool A and B (Life Technologies). Samples were loaded in

the GeneAmp PCR System 9700 following the manufac-

turer’s thermal conditions. Finally, the pre-Amp products

were diluted in 75 lL of H2O.

RT-PCR Arrays

Relative quantitation of targets was performed with Taq-

Man Human microRNA Arrays 384-wells (Gene expres-

sion Micro Fluidic card, Array A v2.1 e Array B v3.0, Life

Technologies). PCR Reaction mix was composed by

450 lL of TaqMan Universal Master Mix II, no Amperase

UNG (Life Technologies), 9 lL of diluted Pre-Amp pro-

duct with primer Pool A or B. (prepare two separate

reaction tubes) and 441 nuclease free water. Dispense

100 lL of the sample-specific PCR reaction mix (Pool A or

B) in each fill port of the 384 plate in Array A or B.

Amplification was performed on the Real Time 7900HT

System (Life Technologies), and data were collected with

Dig Dis Sci (2017) 62:2397–2407 2399

123

the SDS software v2.2 (Life Technologies). Cycling con-

ditions were as follows: 2 min at 50 �C, 10 min at 94.5 �C,30 s at 97 �C (40 cycles), 1 min at 59.7 �C. The tran-

scription levels were normalized using RnU44 as a refer-

ence gene. The expression values for HCC are presented as

fold expression in relation to controls livers; the actual

values were calculated using the 2�DDCT equation, where

DDCT = [CT Target - CT Rnu44](target sample) - [CT

Target - CT GUSB] (control sample).

Statistics and Data Analysis

Variables were expressed as means ± standard deviations,

ranges, and frequencies. MicroRNA Arrays expression was

analyzed with ExpressionSuite Software v1.0 (Life Tech-

nologies). The hypothesis-testing procedure was applied

once the Ct values were normalized with the endogenous

gene and the control groups. Student’s t test was applied to

compare the means of the normalized Ct values between

two groups. Resulting P values were adjusted to control the

type I error rate using the Benjamini–Hochberg method

[22]. miRNAs with a significant differential expression

compared to controls were plotted on a Volcano plot.

Results

Histopathology and Vascular Patterns

Mean tumor size of the 22 HCCs analyzed was cm

6.33 ± 4.74 (range cm 1.60–18.0); Edmondson’s grade

was 2 in 1 (4.5%), 3 in 17 (77.3%) and 4 in 4 (18.2%)

cases. HCCs showed infiltrative margins in 19 (86.4%),

and MVI was observed in 17 (77.3%) cases. Twelve

Fig. 1 Architectural appearance and immunohistochemical positivity

(qualitative assessment) for CD34 (left inset) and Nestin (right inset)

of the four patterns of vascular modifications in HCCs. a HCCs with

pattern ‘‘a’’ and b pattern ‘‘b’’ were generally trabecular and acinar,

with CD34?/Nestin- sinusoids in ‘‘a’’ and CD34?/Nestin?

sinusoids in ‘‘b’’; c pattern ‘‘c’’ HCCs showed macrotrabeculae

surrounded by a CD34?/Nestin? endothelium; (D) HCCs with

pattern ‘‘d’’ showed a vasculature principally made by newly formed

CD34?/Nestin? arteries. Hematoxylin–Eosin stain, magnification

910 and 920 (insets); scale bar 100 lm

2400 Dig Dis Sci (2017) 62:2397–2407

123

Fig. 2 Significant differential miRNA expression in HCC in relation

to the control group, healthy liver (red bars) or cirrhotic liver (blue

bars). The X axis represents the log10 (fold change); down-regulated

miRNAs are on the left part of the Y axis and up-regulated miRNAs

are on the right part of the Y axis

Dig Dis Sci (2017) 62:2397–2407 2401

123

(54.5%) HCCs arose in a cirrhotic liver (c-HCC) and 10

(45.5%) in a non-cirrhotic livers (nc-HCC).

According to HCC architecture and IHC for CD34 and

Nestin, the HCCs of our series were sorted in vascular

‘‘pattern a’’ in 2 (9.1%), ‘‘pattern b’’ in 6 (27.3%), ‘‘pattern

c’’ in 8 (36.3%), and ‘‘pattern d’’ in 6 (27.3%) cases

(Fig. 1) (4).

miRNA Analysis I: Deregulated miRNAs in HCCs

The first step of our study was to assess the miRNAs

deregulated in HCC. After the Benjamini–Hochberg

posttest, we found 14 miRNAs significantly deregulated

compared to the healthy liver control pool (see Figs. 2, 3,

Supplemental table S1). In particular, 5 miRNAs were

down-regulated (mean fold increase 0.12 ± 0.07) and 9

up-regulated (mean fold increase 8.77 ± 7.42). Using the

cirrhotic liver control pool, we noticed that 53 miRNAs

were significantly deregulated. In particular, 44 were

down-regulated (mean fold increase 0.10 ± 0.11), and 9

up-regulated (mean fold increase 103.55 ± 223.29).

Fifty-five miRNAs gave results depending on the tissue

used for normalization: 8 miRNAs were limited to the

analysis using the healthy liver control pool, while 47 were

limited to the analysis using the cirrhotic liver control pool.

Conversely, the deregulation of 6 miRNAs was common

between the two analyses, 5 with the same trend (miR-34a,

miR-93, miR532, miR-149#, and miR-7f-2#) and 1 with the

opposite trend (miR-30a-5p) (Fig. 3). Among these 6 sig-

nificant miRNAs, 5 are already known to be associated with

HCC or with other human cancers (Table 1). No correlations

emerged with etiology in our series (data not shown), except

for miR-149# that showed a significant down-regulation in

HCC with HCV- and HBV-driven chronic liver disease

(P\ 0.001). All raw data prior to statistical analysis are

found in the Supplemental table S2 (sheet 1).

miRNA Analysis II: Deregulated miRNAs in c-

HCCs and nc-HCCs

When the healthy liver control pool is used for normal-

ization, 14 miRNAs were significantly down-regulated in

Fig. 3 Volcano plots representing miRNAs deregulated in HCC

according to the normalization. a compared to the non-cirrhotic and

b cirrhotic liver. The X axis represents the log2 (fold change) and the

Y axis the -log10 (P value). miRNAs with statistically significant

differential expression (P\ 0.05 adjusted for multiple testing and

fold change boundary at 2) are located above the horizontal threshold

blue line and outside the pair of vertical threshold black lines. Down-

regulated miRNAs are found in the upper left (green) of the plot and

up-regulated miRNAs in the upper right parts (red). The 6 miRNAs in

common are miR-34a, miR-93, miR532, miR-149#, miR-7f-2# and

miR-30a-5p

2402 Dig Dis Sci (2017) 62:2397–2407

123

HCC tissue (Table 2, Supplemental table S2, sheet 2):

particularly, only miR-1178 was significantly down-regu-

lated in c-HCCs, while the remaining 13 miRNAs were

deregulated in nc-HCCs. Among these 14 miRNAs, 5 were

in common with those resulting in analysis I (all 22 HCCs):

miR-149#, miR-23a#, miR-640, miR-1259, and miR-7f-2#.

Remarkable results were obtained using the cirrhotic

liver control pool for normalization: 120 miRNAs were

found to be deregulated (Table 2, Supplemental table S2,

sheet 2), 37 in c-HCCs (all down-regulated but 2) and 83 in

nc-HCCs (all down-regulated). Forty-three out of 120

miRNAs were common with those observed in general

HCCs (analysis I). Furthermore, 29 down-regulated miR-

NAs were common to c-HCCs, nc-HCCs and the overall

HCCs, in all cases with the same trend, confirming the

results of analysis I. It is crucial to notice that the pool of

cirrhotic control tissues might not be an appropriate model

for the normalization of nc-HCCs.

Three miRNAs (miR-136#, miR-149# and 188-3p) were

significantly down-regulated independently from the ana-

lyzed tissues (c-HCC or nc-HCC) and the pool used to

compare the fold change (Supplemental table S2, sheet 2).

miRNA Analysis III: Deregulated miRNAs in HCCs

with Different Vascular Patterns

The analysis of the miRNAs expressed in HCCs with dif-

ferent vascular modifications showed that each of the four

vascular patterns is characterized by the differential

expression of miRNAs. In particular, compared to the

healthy liver control pool, 16 miRNAs were significantly

deregulated, 1 in ‘‘pattern a’’ HCCs, 4 in ‘‘pattern b,’’ 6 in

‘‘pattern c’’ and 5 in ‘‘pattern d’’ HCCs (Table 3, Supple-

mental table S2, sheet 3). Compared to the cirrhotic liver

control pool, 112 miRNAs were significantly deregulated,

4 in ‘‘pattern a’’ HCCs (not the same as above), 16 in

‘‘pattern b,’’ 56 in ‘‘pattern c’’ and 36 in ‘‘pattern d’’ HCCs

(Table 3, Supplemental table S2, sheet 3).

Interestingly, HCCs with patterns ‘‘a’’ and ‘‘b’’ share no

miRNAs in common using the two different normaliza-

tions. Conversely, both HCCs with ‘‘pattern c’’ and HCCs

with ‘‘pattern d’’ show 3 common miRNAs, all of them

with similar fold changes, using healthy or cirrhotic liver

control pools as reference. ‘‘Pattern a’’ HCCs are the only

biological group showing up-regulation of 2 miRNAs

(miR-551b# and miR-1825), while all other cases were

down-regulated (all data are found in the Supplemental

table S2, sheet 3). Patterns ‘‘a’’ and ‘‘b’’ HCCs do not

conserve the same profile using different control tissues,

while pattern ‘‘c’’ and ‘‘d’’ seem to show peculiar miRNA

profiles (see ‘‘Discussion’’).

Discussion

The present study focused on a series of surgically resected

histologically advanced HCCs, with the aim (1) to define

the different miRNA profiles of HCCs arisen in cirrhotic

and non-cirrhotic livers based on two different human liver

controls extracted from healthy and cirrhotic livers; (2) to

Table 1 miRNA significantly deregulated in HCC independently on the control used, and references to the literature

miRNA expression Known interaction with HCC Reference(s)

miR-34a (UP) HCC progression and aggressiveness

(p53 and b-catenin pathways)

Recurrence prediction after radiofrequency ablation in early HCC

Clinical trial miR-RX34 in patients with primary liver cancer

Budhu et al. [14]

Huang et al. [16]

Pineau et al. [30]

Dang et al. [25]

Gougelet et al. [26]

Cui et al. [27]

miR-93 (UP) HCC progression

Drug sensitivity in vitro

Huang et al. [16]

Pineau et al. [30]

Xu et al. [28]

Shi et al. [29]

miR-532 (UP) Unknown …miR-149# (DOWN) HCC tumorigenesis (AKT/mTOR pathway) Zhang et al. [7]

let-7f-2# (DOWN) Let-7f dysregulated in HCC Gramantieri et al. [31]

Huang et al. [16]

Shimizu et al. [33]

miR-30a-5p (UP/DOWN) HCC progression Wojcicka et al. [34]

Dai et al. [35]

Dig Dis Sci (2017) 62:2397–2407 2403

123

correlate miRNA expression with the patterns of tumor

vascular modifications according to the same two controls.

The tissue used for normalization represents a key point in

order to identify miRNAs deregulation in HCCs. Since

HCCs can arise in both cirrhotic and non-cirrhotic livers

(albeit more frequently in the former), we carried out all

analyses using RNA extracted from both tissues, in the

present study. Total RNA extracted from cirrhotic liver

cannot be considered an appropriate control for all cases, in

particular for nc-HCCs. Even if nc-HCCs usually arise in a

context of chronic hepatitis, we decided here to utilize

healthy livers, obtained from healthy multiorgan donors

without virus exposure, chronic liver disease or neoplasia.

Most recent papers on miRNAs have studied both HCC

and cirrhosis using healthy liver tissue as Refs. [6, 13, 14],

with weak overlaps on the identified HCC-related miRNAs,

and often discordance between the up- and down-regulation

of specific miRNAs. This last issue is very important, since

it was postulated that the up- or down-regulation of miR-

NAs could determine the prognosis in HCC patients [17]

and could be used in specific inhibition or restoration

therapies [23]. Beside the factors that can influence the final

results of miRNA analysis, such as the choice of the array

used, the choice of the housekeeping genes, and the statis-

tical approach [24], the selection of the reference control is

a key step that can impact the expression and the fold

Table 2 miRNA deregulated in HCCs arisen in non-cirrhotic (nc-HCC) or cirrhotic (c-HCC) background, compared to the normal or cirrhotic

controls

Control pool Common miRNAs

according to

normalization poolsNormal Cirrhotic

Biological groups

nc-HCC miR-149#, miR-188-3p, miR159a, miR-543,

miR-483-3p, miR-640, miR-1259, miR-

23a#, 7f-2#, miR-1243, miR-136#, miR-

562

miR-30a-3p, miR-1274B, miR-769-3p, miR-617,

miR-720, miR-213, miR-30a-5p, miR-26a-2#,

miR-1260, miR-132#, miR-29b-2#, miR-543,

miR-429, miR-331-5p, miR-635, miR-200c,

miR-542-5p, miR-186#, miR-136, miR-20a#,

miR-571, miR-181c#, miR-24-2#, miR-223#,

miR-376a, miR-99a#, miR-125b-2#, 7f-2#,

miR-144#, miR-590-3P, miR-596, miR-886-5p,

miR-26a-1#, miR-145#, 7a#, miR-650, miR-

380-5p, miR-149#, miR-31#, miR-338-3p, miR-

199b, miR-199a, miR-1208, miR-1288, miR-

214#, miR-126#, miR-483-3p, miR-483-5p,

miR-200a, miR-662, miR-497, miR-1290, miR-

200b, miR-141, miR-644, miR-1247, miR-

130a#, miR-136#, miR-200a#, miR-584, miR-

195#, miR-656, miR-486-3p, miR-604, miR-

646, miR-1255A, miR-380-3p, miR-519a, miR-

609, miR-889, miR-378, miR-639, miR-1305,

miR-188-3p, miR-202#, miR-30c-1#, miR-

377#, miR-452#, miR-520b, miR-520D-3P,

miR-551a, miR-621, miR-875-5p

miR-7f-2#, miR-483-3p,

miR-136#, miR-149#,

miR-188-3p, miR-543

c-HCC miR-1178 miR-130b#, miR-1255B, miR-125b-1#, miR-

99a#, miR-591, miR-125b-2#, miR-20a#, miR-

31#, miR-145#, miR-380-5p, miR-944, miR-

132#, miR-149#-002164, miR-338-3p, miR-

590-3P, miR-101#, miR-26a-1#, miR-126#,

miR-136#, miR-214#, miR-665-002681, miR-

609, miR-200a#, miR-656, 7a#, miR-338-5P,

miR-92b#, miR-378, miR-188-3p, miR-520D-

3P, miR-875-5p, miR-621, miR-551a, miR-

452#, miR-30c-1#, miR-202#, miR-1305

None

Common

miRNAs

between nc-

HCC and

c-HCC

None miR-136#, miR-149#, miR-188-3p, miR-7a#,

miR-125b-2#, miR-126#, miR-1305, miR-132#,

miR-145#, miR-200a miR-200a# miR-202#

miR-20a# miR-214# miR-26a-1# miR-30c-1#,

miR-31#, miR-338-3p, miR-378, miR-380-5p,

miR-452#, miR-520D-3P, miR-551a, miR-590-

3P, miR-656

miR-136#, miR-149#,

miR-188-3p,

2404 Dig Dis Sci (2017) 62:2397–2407

123

increase in miRNAs. The same concept was already

expressed by Visani et al. [19] in brain tumors.

According to our data, when we used the cirrhotic liver

control pool as reference, a greater number of miRNAs was

significantly deregulated in each analysis among biological

groups. Moreover, even the expression of an endogenous

housekeeping gene (U6 snRNA) resulted heterogeneous in

one analysis (Table 3): it could be hypothesized that cir-

rhosis is a too heterogeneous model to be used as control,

albeit representing the background of most HCCs, due to a

high percentage—and variability—of inflammatory cells,

ductular reaction, fibrosis, etc. [17]. For the same reason,

we found many differences in miRNAs deregulation

between c-HCC and nc-HCC (Table 2).

At any chance, our results showed 6 miRNAs to be

deregulated in HCCs, despite the control used: miR-532,

miR-34a, miR-93, miR-149#, miR-7f-2#, and miR-30a-5p.

Apart from miR-532, all of these miRNAs have been

already reported in the HCC literature (Table 1).

miR-34a plays a tumor suppressor role through the

regulation of c-Met and b-catenin pathways [25, 26]. miR-

34a deregulation was linked to HCC progression and

aggressiveness both in vivo and in vitro. Interestingly, a

very recent paper correlated the deregulation of miR-34a

with the recurrence rates of early HCC after radiofrequency

[27].

A deregulation of miR-93 significantly correlated with

poor prognosis in HCCs. In vitro miR-93 activity was also

correlated with HCC sensitivity to sorafenib and tivantinib

[18]: the authors concluded that miR-93 was involved in

cell proliferation, migration, and invasion through the

oncogenic c-Met/PI3 K/Akt pathway, and it also inhibited

apoptosis by PTEN and CDKN1A pathways [18]. Thus, the

use of synthetic regulators of miR-93 may prove to be a

promising approach to liver cancer treatment [28]. Other

studies have found a deregulation of miR-93, together with

other miRNAs, during HCC progression [29, 30].

miR-149 modulates the Akt/mTOR pathway in HCC. A

tumor suppressive role for miR-149, and a prognostic role

in human HCCs were hypothesized [7].

miRNA let-7f-2 was found to be down-regulated in

HCCs [31, 32]. It was seen that the let-7f family of miR-

NAs potentiates sorafenib-induced apoptosis in human

HCC [16]. Recently, a study correlated also serum let-7f

levels with HCC size and early recurrence [33].

Finally, miR-30a-5p expression was altered in HCC

compared to adjacent tissue [34]. In another study, the

overexpression of miR-30a-5p inhibited cell proliferation,

induced apoptosis, increased the number of cells in S

phase, and markedly inhibited invasion and migration of

HCC cells in vitro and in vivo [35].

Table 3 miRNA deregulated in each pattern of vascular modification in HCC compared to the normal or cirrhotic controls

Control pool Common miRNAs

according to

normalization poolsNormal Cirrhotic

Biological groups

Pattern

a

miR-551b# miR-922, miR-548d-5p, miR-516-3p, miR-1825 None

Pattern

b

miR-149#*, miR-1178*, miR-

1243, miR-23a#

U6 snRNA, miR-889, miR-770-5p, miR-662, miR-616, miR-520b,

miR-519b-3p, miR-33a, miR-335#, miR-31#, miR-29a#, miR-

200a#, miR-19b-1#, miR-195#, miR-125b-2#, miR-101#

None

Pattern

c

miR-149#*, miR-1178*, let-

7a#, miR-1300, miR-483-3p,

let-7 g#

miR-29c#, miR-99a#, miR-944, miR-943, miR-92b#, miR-758, miR-

720, miR-708, miR-668, miR-665, miR-661, miR-644, miR-628-

3p, miR-609, miR-604, miR-601, miR-596, miR-592, miR-590-3P,

miR-572, miR-571, miR-497, miR-483-3p, miR-429, miR-33a#,

miR-338-3p, miR-30e-3p, miR-30a-5p, miR-30a-3p, miR-29b-2#,

miR-26a-1#, miR-218, miR-214#, miR-20a#, miR-202#, miR-200c,

miR-188-3p, miR-186#, miR-16-1#, miR-149#, miR-145#, miR-

144#, miR-141, miR-132#, miR-1303, miR-1290, miR-1288, miR-

1267, miR-1260, miR-126#, miR-125b-1#, miR-1228#, miR-1208,

miR-1201, let-7i#, let-7a#

let-7a#, miR-483-3p,

miR-149#

Pattern

d

miR159a, miR-541#, miR-875-

5p, miR-149#*, miR-675

miR-491, miR-875-5p, miR-769-3p, miR-675, miR-656, miR-650,

miR-646, miR-639, miR-621, miR-584, miR-567, miR-551a, miR-

548L, miR-548b, miR-541#, miR-520D-3P, miR-512-5p, miR-

452#, miR-378, miR-377#, miR-30c-1#-, miR-223#, miR-200b,

miR-200a, miR-199b, miR-198, miR-143#, miR-140-3p, miR-136#,

miR-130a#, miR-1305, miR-1298, miR-1274A, miR-1255A, miR-

1247, miR-1225-3P

miR-541#, miR-675,

miR-875-5p

Dig Dis Sci (2017) 62:2397–2407 2405

123

In the study of HCC progression, we sorted HCCs on the

base of different vascular modifications, according to our

recent observations [4]. It is noteworthy, since neoangiogen-

esis plays a pivotal role both in HCC aggressiveness and

metastatic diffusion, and it is responsible from specific radio-

logical features. The present study seems to confirm the pro-

gression from ‘‘pattern a’’ HCCs (characterized by a

microtrabecular architecture,withCD34?/Nestin- sinusoids,

more common in cirrhosis) and the other patterns. Indeed,

‘‘pattern a’’ HCCs showed the deregulation of different miR-

NAs compared to the ‘‘histologically-advanced’’ patterns ‘‘b,’’

‘‘c’’ and ‘‘d’’ HCCs [4]: miR-551b# and miR-1825 were up-

regulated in pattern ‘‘a’’ and down-regulated in the others;

miR-149#was deregulated in pattern ‘‘b,’’ ‘‘c’’ and ‘‘d’’ HCCs

but not in pattern ‘‘a’’; miRNA let-7a# and miR-483-3p were

deregulated in pattern ‘‘c’’ and ‘‘d’’ HCCs. A recent study

correlated miR-149# up-regulation with the inhibition of

tumor neoangiogenesis in vitro and in murine model [36], and

miR-149# down-regulation with HCC aggressiveness [7].

miRNA let-7a# is an inhibitor of tumorigenicity via negative

RAS regulation [37, 38] and a suppressor of HCC invasion

in vitro [39]. miRNA 483-3p is known for his pro-apoptotic

function and was associated with HCC recurrence [15, 40].

The present study confirms the crucial role of the choice of

controls in the miRNA analyses: standard RNA controls—

based on hepatotrophic viruses, pre-cancerous lesions,

inflammatory activity, etc—are required to allow the com-

parison among different laboratories for the routine practice.

A short-term application might be the identification of

miRNAs as marker of HCC with higher aggressiveness and

higher recurrence risk, on biopsy specimen or directly on

serum. Tissue miRNA deregulation is very important in the

study of HCC progression, as also confirmed by our data, and

a standardization will be a crucial achievement for the future.

Funding This work has been supported by the Programma di Ricerca

Regione-Universita 2010–2012, Regione Emilia-Romagna, bando

‘‘Ricerca Innovativa’’ (Professor L. Bolondi), project title ‘‘Innovative

approaches to the diagnosis and pharmacogenetic-based therapies of

primary hepatic tumors, peripheral B and T-cell lymphomas and

lymphoblastic leukaemias.’’ The funding body had no role in the

study design, in the collection, analysis, interpretation of data, in the

writing of the manuscript and in the decision to submit the manuscript

for publication.

Compliance with ethical standards

Conflict of interest Authors declare no conflicts of interest.

References

1. Zhou XD, Tang ZY, Yang BH, et al. Experience of 1000 patients

who underwent hepatectomy for small hepatocellular carcinoma.

Cancer. 2001;91:1479–1486.

2. Wiemer EA. The role of microRNAs in cancer: no small matter.

Eur J Cancer. 2007;43:1529–1544.

3. Villa E, Critelli R, Lei B, et al. Neoangiogenesis-related genes all

hallmarks of fast-growing hepatocellular carcinomas and worst

survival. Results from a prospective study. Gut. 2016;65:

861–869.

4. Vasuri F, Fittipaldi S, Giunchi F, et al. Facing the enigma of the

vascular network in hepatocellular carcinomas in cirrhotic and

non-cirrhotic livers. J Clin Pathol. 2015;69:102–108.

5. Wang B, Hsu SH, Wang X, et al. Reciprocal regulation of

microRNA-122 and c-Myc in hepatocellular cancer: role of E2F1

and transcription factor dimerization partner 2. Hepatology.

2014;59:555–566.

6. Meng F, Henson R, Wehbe-Janek H, Ghoshal K, Jacob ST, Patel

T. MicroRNA-21 regulates expression of the PTEN tumor sup-

pressor gene in human hepatocellular cancer. Gastroenterology.

2007;133:647–658.

7. Zhang J, Jin H, Liu H, et al. MiRNA-99a directly regulates AGO2

through translational repression in hepatocellular carcinoma.

Oncogenesis. 2014;3:e97.

8. Anwar SL, Albat C, Krech T, et al. Concordant hypermethylation

of intergenic microRNA genes in human hepatocellular carci-

noma as new diagnostic and prognostic marker. Int J Cancer.

2013;133:660–670.

9. Edmunds RC, McIntyre JK, Luckenbach JA, Baldwin DH,

Incardona JP. Toward enhanced MIQE compliance: reference

residual normalization of qPCR Gene Expression Data. J Biomol

Tech. 2014;25:54–60.

10. Bruckert G, Vivien D, Docagne F, Roussel BD. Normalization of

reverse transcription quantitative PCR data during ageing in

distinct cerebral structures. Mol Neurobiol. 2016;53:1540–1550.

11. Jacob F, Guertler R, Naim S, et al. Careful selection of reference

genes is required for reliable performance of RT-qPCR in human

normal and cancer cell lines. PLoS One. 2013;8:e59180.

12. Kirschner MB, van Zandwijk N, Reid G. Cell-free microRNAs:

potential biomarkers in need of standardized reporting. Front

Genet. 2013;4:56.

13. Murakami Y, Toyoda H, Tanahashi T, et al. Comprehensive

miRNA expression analysis in peripheral blood can diagnose

liver disease. PLoS ONE. 2012;7:e48366.

14. Budhu A, Jia HL, Forgues M, et al. Identification of metastasis-

related microRNAs in hepatocellular carcinoma. Hepatology.

2008;47:897–907.

15. Sato F, Hatano E, Kitamura K, et al. MicroRNA profile predicts

recurrence after resection in patients with hepatocellular carci-

noma within the Milan Criteria. PLoS ONE. 2011;6:e16435.

16. Huang YS, Dai Y, Yu XF, et al. Microarray analysis of micro-

RNA expression in hepatocellular carcinoma and non-tumorous

tissues without viral hepatitis. J Gastroenterol Hepatol. 2008;23:

87–94.

17. Jiang J, Gusev Y, Aderca I, et al. Association of microRNA

expression in hepatocellular carcinomas with hepatitis infection,

cirrhosis, and patient survival. Clin Cancer Res. 2008;14:

419–427.

18. Ohta K, Hoshino H, Wang J, et al. MicroRNA-93 activates

c-Met/PI3K/Akt pathway activity in hepatocellular carcinoma by

directly inhibiting PTEN and CDKN1A. Oncotarget. 2015;6:

3211–3224.

19. Visani M, de Biase D, Marucci G, Taccioli C, Baruzzi A, Pession

A. Definition of miRNAs expression profile in glioblastoma

samples: the relevance of non-neoplastic brain reference. PLoS

ONE. 2013;8:e55314.

20. Mestdagh P, Hartmann N, Baeriswyl L, et al. Evaluation of

quantitative miRNA expression platforms in the microRNA

quality control (miRQC) study. Nat Methods. 2014;11:809–815.

2406 Dig Dis Sci (2017) 62:2397–2407

123

21. Fittipaldi S, Vasuri F, Degiovanni A, et al. Nestin and WT1

expression in atheromathous plaque neovessels: association with

vulnerability. Histol Histopathol. 2014;29:1565–1573.

22. Shesh NR, Ray HE, Yuan X, Pan J, Hamid T, Prabhu SD. Sta-

tistical analysis of repeated microRNA high-throughput data with

application to human heart failure: a review of methodology.

Open Access Med Stat. 2012;2012:21–31.

23. Callegari E, Gramantieri L, Domenicali M, D’Abundo L, Sab-

bioni S, Negrini M. MicroRNAs in liver cancer: a model for

investigating pathogenesis and novel therapeutic approaches. Cell

Death Differ. 2015;22:46–57.

24. Pradervand S,Weber J, Thomas J, et al. Impact of normalization on

miRNA microarray expression profiling. RNA. 2009;15:493–501.

25. Dang Y, Luo D, Rong M, Chen G. Underexpression of miR-34a

in hepatocellular carcinoma and its contribution towards

enhancement of proliferating inhibitory effects of agents target-

ing c-MET. PLoS ONE. 2013;8:e61054.

26. Gougelet A, Sartor C, Bachelot L, et al. Antitumour activity of an

inhibitor of miR-34a in liver cancer with b-catenin-mutations.

Gut. 2015. doi:10.1136/gutjnl-2014-308969.

27. Cui X, Wu Y, Wang Z, Liu X, Wang S, Qin C. MicroRNA-34a

expression is predictive of recurrence after radiofrequency abla-

tion in early hepatocellular carcinoma. Tumour Biol. 2015;36:

3887–3893.

28. Xu D, He XX, Chang Y, Sun SZ, Xu CR, Lin JS. Downregulation

of MiR-93 expression reduces cell proliferation and clonogenicity

of HepG2 cells. Hepatogastroenterology. 2012;59:2367–2373.

29. Shi KQ, Lin Z, Chen XJ, et al. Hepatocellular carcinoma asso-

ciated microRNA expression signature: integrated bioinformatics

analysis, experimental validation and clinical significance. On-

cotarget. 2015;6:25093–25108.

30. Pineau P, Volinia S, McJunkin K, et al. miR-221 overexpression

contributes to liver tumorigenesis. Proc Natl Acad Sci USA.

2010;107:264–269.

31. Gramantieri L, Ferracin M, Fornari F, et al. Cyclin G1 is a target

of miR-122a, a microRNA frequently down-regulated in human

hepatocellular carcinoma. Cancer Res. 2007;67:6092–6099.

32. Ge W, Yu DC, Li QG, Chen X, Zhang CY, Ding YT. Expression

of serum miR-16, let-7f, and miR-21 in patients with hepato-

cellular carcinoma and their clinical significances. Clin Lab.

2014;60:427–434.

33. Shimizu S, Takehara T, Hikita H, et al. The let-7 family of

microRNAs inhibits Bcl-xL expression and potentiates sorafenib-

induced apoptosis in human hepatocellular carcinoma. J Hepatol.

2010;52:698–704.

34. Wojcicka A, Swierniak M, Kornasiewicz O, et al. Next genera-

tion sequencing reveals microRNA isoforms in liver cirrhosis and

hepatocellular carcinoma. Int J Biochem Cell Biol. 2014;53:

208–217.

35. Dai H, Kang B, Zuo D. Effect of miR-30a-5p on the proliferation,

apoptosis, invasion and migration of SMCC-7721 human hepa-

tocellular carcinoma cells. Zhonghua Gan Zang Bing Za Zhi.

2014;22:915–920.

36. Chamorro-Jorganes A, Araldi E, Rotllan N, Cirera-Salinas D,

Suarez Y. Autoregulation of glypican-1 by intronic microRNA-

149 fine tunes the angiogenic response to FGF2 in human

endothelial cells. J Cell Sci. 2014;127:1169–1178.

37. Johnson SM, Grosshans H, Shingara J, et al. RAS is regulated by

the let-7 microRNA family. Cell. 2005;120:635–647.

38. Salvi A, Abeni E, Portolani N, Barlati S, De Petro G. Human

hepatocellular carcinoma cell-specific miRNAs reveal the dif-

ferential expression of miR-24 and miR-27a in cirrhotic/non-

cirrhotic HCC. Int J Oncol. 2013;42:391–402.

39. Drakaki A, Hatziapostolou M, Polytarchou C, et al. Functional

microRNA high throughput screening reveals miR-9 as a central

regulator of liver oncogenesis by affecting the PPARA-CDH1

pathway. BMC Cancer. 2015;15:542.

40. Bertero T, Bourget-Ponzio I, Puissant A, et al. Tumor suppressor

function of miR-483-3p on squamous cell carcinomas due to its

pro-apoptotic properties. Cell Cycle. 2013;12:2183–2193.

Dig Dis Sci (2017) 62:2397–2407 2407

123

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具