Cancer Genetics - (2013) -

MicroRNAs as diagnostic markers for pancreatic

ductal adenocarcinoma and its precursor,
pancreatic intraepithelial neoplasm
Yue Xue a,1, Ahmad N. Abou Tayoun a,1, Kristine M. Abo a, J. Marc Pipas b,
Stuart R. Gordon c, Timothy B. Gardner c, Richard J. Barth Jr d,
Arief
a
A. Suriawinata a, Gregory J. Tsongalis a,*
Department of Pathology, Geisel School of Medicine at Dartmouth, Hanover, NH -and- Dartmouth Hitchcock Medical Center
and Norris Cotton Cancer Center, Lebanon, NH; b Department of Medicine, Section of Hematology/Oncology, Geisel School of
Medicine at Dartmouth, Hanover, NH -and- Dartmouth Hitchcock Medical Center and Norris Cotton Cancer Center, Lebanon,
NH; c Department of Medicine, Section of Gastroenterology, Geisel School of Medicine at Dartmouth, Hanover, NH; Dartmouth
Hitchcock Medical Center and Norris Cotton Cancer Center, Hanover, NH; d Department of Surgery, Geisel School of Medicine
at Dartmouth, Hanover, NH -and- Dartmouth Hitchcock Medical Center and Norris Cotton Cancer Center, Lebanon, NH, USA

Since the discovery of small non-coding RNAs, the analysis of microRNA (miRNA) expression
patterns in human cancer have provided new insights into cancer biology. Evidence suggests
that deregulated miRNA expression is associated with pancreatic cancer development. In this
study, we analyzed the expression of several miRNAs in different types of pancreatic disease
to determine if miRNA expression could aid in the diagnosis of pancreatic ductal adenocarcinoma
(PDAC) and its precursor, pancreatic intraepithelial neoplasm (PanIN). Pancreatic resection
specimens were selected, which included PDAC (n Z 16), benign pancreatic parenchyma from
corresponding carcinoma cases (n Z 16), chronic pancreatitis (n Z 4), normal pancreatic paren-
chyma (n Z 5), and PanIN (n Z 5). The expression levels of five miRNA (miR-148a, miR-217,
miR-21, miR-196a, and miR-10b) were assessed by quantitative real-time reverse transcription-
polymerase chain reaction (qRT-PCR) assays. Our data demonstrate that compared to the
normal pancreatic parenchyma, miR-148a and miR-217 expression levels were down-
regulated in PanIN, particularly in PanIN II-III and PDAC, whereas the level of miR-196 was
significantly up-regulated in PDAC and its precursor, PanIN II-III. In addition, we observed that
miR-21 was significantly overexpressed in PDAC, and miR-10b was highly expressed in PanIN
II-III. Our study demonstrates that certain miRNAs, especially miR-148a, miR-217, and
miR-196a, are significantly deregulated in PDAC, including in the early stage of PDAC. These
markers can potentially be used as diagnostic markers to distinguish PDAC and its precursor
from benign lesions.
Keywords MicroRNA, pancreatic ductal adenocarcinoma, pancreatic intraepithelial neoplasm
ª 2013 Elsevier Inc. All rights reserved.

Pancreatic ductal adenocarcinoma (PDAC) is the fourth with tumors of advanced stage that are inoperable and
leading cause of cancer-related deaths in the United States associated with a poor prognosis. Preoperative evaluation of
(1). Although surgical resection represents the best hope for pancreatic tumors includes radiographic imaging to deter-
cure of early-stage pancreatic cancer, most patients present mine potential resectability and diagnostic biopsy, which is
essential to guiding appropriate surgical and chemothera-
Received February 11, 2013; received in revised form May 24,
peutic management; however, biopsy interpretation may be
2013; accepted May 28, 2013. hindered by limited sampling and inflammatory changes that
* Corresponding author. mask or mimic neoplasms. Therefore, many investigators
E-mail address: [email protected] have tried to identify biomarkers that reliably distinguish
1
Both authors contributed equally to this work. pancreatic neoplasm from benign conditions.

2210-7762/$ - see front matter ª 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.cancergen.2013.05.020
2 Y. Xue et al.

MicroRNAs (miRNAs) have received considerable atten- ID 002337). Real-time PCR reactions were subsequently
tion as potential biomarkers of human malignancies (2). performed on the diluted cDNA products with TaqMan
These molecules consist of 18 25 nucleotides and regulate Universal Master Mix II (Applied Biosystems) and specific
gene expression at the post-transcriptional level (3). PCR primers/probes that targeted each of the above miR-
Expression profiles of human miRNAs have demonstrated NAs (TaqMan MicroRNA Assays, Applied Biosystems). PCR
that many of these molecules are deregulated in cancer and reactions were performed on the 7500 Fast Real-Time PCR
could, therefore, potentially assist in further establishing System (Applied Biosystems). Each PCR reaction was run in
molecular diagnosis, prognosis, and therapy. Previous duplicate, and the average cycle threshold (Cq) value was
studies from our laboratory and others have shown that miR- calculated for further analysis.
21, miR-10b, and miR-196a expression levels were signifi- The expression level of each miRNA was calculated using
cantly higher in PDAC compared to normal pancreatic the formula log2 (2 DCq), where DCq is defined as the
tissues (4e8), whereas miR-148a and miR-217 exhibited difference between the Cq value of the target miRNA and that
a consistent decrease or absence of expression only in of the internal control (Cq miRNAeCq RNU6 B). For relative
PDAC (6,9,10). expression between different disease groups, the formula
In this study, we attempted to further validate miRNA log2 (2 DDCq) was used, where DDCq is the difference
expression in different types of pancreatic disease to deter- between the DCq of the neoplastic sample and that of the NP
mine if their expression could aid in the diagnosis of PDAC (DCT neoplastic sampleeDCq NP).
and its precursor, pancreatic intraepithelial neoplasm
(PanIN), in pancreatic resection specimens. Statistical analysis

Materials and methods All reactions were pooled and analyzed by the same exper-
imenter. Statistical significance of the difference between
Case selection different groups was assessed using a two-sample t-test. A
paired t-test was used to calculate statistical significance
between PDAC and its paired control with negative margin.
Analyses of miRNA expression from archived formalin-fixed
paraffin embedded (FFPE) pancreatic resection specimens
were approved by the Dartmouth Committee for the Protec- Results
tion of Human Subjects. FFPE-pancreatic tissue blocks were
selected from our pathology department archive between Clinical and pathologic features of resection
2004 and 2011. These cases included PDAC (16 cases), specimens
benign pancreatic parenchyma from corresponding carci-
noma cases (16 cases), PanIN (5 cases), chronic pancrea- The clinical and pathologic features of the resection speci-
titis (CP) (4 cases), and normal pancreatic parenchyma (NP) mens are summarized in Table 1. Of the 16 PDAC cases,
(5 cases). After hematoxylin and eosin (H & E) staining of two were well-differentiated (Figure 1A), three were moder-
each tissue case, the slides were reviewed and lesions were ately differentiated (Figure 1B), four were poorly differenti-
marked by a gastrointestinal-subspecialized pathologist. ated (Figure 1C), and seven cases were not given a degree
of differentiation due to their status after chemotherapeutic
RNA extraction treatment.
Five cases were associated with lymph node metastases.
Unstained FFPE tissue sections (5 mm) mounted on glass Two cases were associated with metastasis to other organs.
slides were obtained for RNA extraction. The tissue intended At the time this study was performed, nine patients had died
for miRNA analysis was first marked on the unstained slide
using the matched H & E stained slide, after which total RNA
Table 1 Clinical and pathologic features of resected pancre-
was extracted using the Qiagen miRNeasy FFPE Kit (Hilden,
atic specimens
Germany) from the macrodissected tissue. For each case,
4e6 unstained slides were used for RNA extraction. PDAC paired NP
Extracted RNA samples were normalized to 10 ng/mL in with negative PanIN CP parenchyma
RNase-free water and stored at 80 C. margin (n Z 16) (n Z 5) (n Z 4) (n Z 5)
Mean age (y) 68 65 56 66
Quantitative real-time reverse-transcriptase- Male:Female 7:9 2:3 3:1 3:2
polymerase chain reaction (qRT-PCR) ratio
Lesion location in pancreas
Six RT reactions were performed on each sample (50 ng of Head 8 1
RNA for each reaction) using the TaqMan MicroRNA Body 4 0
Reverse Transcription Kit (Applied Biosystems, Foster City, Tail 4 3
CA) and RT primers specific for each of the following six Pathologic tumor stage
miRNAs (TaqMan MicroRNA Assays, Applied Biosystems): I 1
internal control RNU6 B (Assay ID 001093), miR-10b (Assay II 12
ID 002218), miR-21 (Assay ID 000397), miR-148a (Assay ID III 1
000470), miR-196a (Assay ID 241070), or miR-217 (Assay IV 2
MicroRNA alterations of PDAC and its precursor 3

Figure 1 PDAC (H & E staining). (A) well-differentiated; (B) moderately differentiated; (C) poorly differentiated.

of PDAC and seven were still alive. Of the five PanIN cases, P < 0.05) (see Materials and methods for DDCq analysis).
one case was diagnosed as PanIN I, one case was PanIN II, Given that miR-148a was dramatically reduced in PDAC, we
and three cases were PanIN III. Lesions with PanIN I and tested whether this miRNA is differentially regulated between
PanIN III were selected from the cases with well- PDAC and benign pancreatic tissue from corresponding
differentiated ductal adenocarcinoma. Lesions with PanIN II carcinoma cases. Indeed, miR-148a levels were significantly
were selected from the case with CP. reduced in PDAC relative to its benign pancreatic tissue,
which predominantly consists of CP (P < 0.0001)
(Figure 2B).
Expression level of miR-148a is significantly
down-regulated in PanIN and PDAC
miR-217 expression is significantly
Our data show that miR-148a levels were reduced in all down-regulated in PDAC
benign and malignant pancreatic lesions compared to NP
(Figure 2A). This reduction was more apparent in the Analysis of miR-217 expression interestingly shows that,
neoplastic samples (Figure 2A). Comparison of miR-148a unlike the benign lesion- CP, where the miR-217 levels were
levels between the neoplastic lesions and CP showed that comparable to normal pancreatic tissue (P > 0.05, Figure 3A),
miR-148a levels were much lower in PDAC (w3 DDCq the neoplastic lesions exhibited a highly significant reduction
difference, P < 0.05) and in PanIN (w2.5 DDCq difference, in miR-217 expression (Figure 3A). Most important,

Figure 2 Expression of miR-148a level in neoplastic and non-neoplastic pancreatic parenchyma. (A) log2 (2 DCq) values for NP, CP,
PanIN, and PDAC were calculated as shown in Materials and methods (*P < 0.05, **P < 0.0001, two-sample t-test for comparison with
NP). Mean  SEM is shown. (B) log2 (2 DCq) values for N and PDAC (**P < 0.0001, paired t-test for comparison with N). Mean  SEM
is shown. Abbreviations: NP, normal pancreatic parenchyma; CP, chronic pancreatitis; PanIN, pancreatic intraepithelial neoplasm;
PDAC, pancreatic ductal adenocarcinoma; N, paired pancreatic tissue with negative margin.
4 Y. Xue et al.

Figure 3 Expression of miR-217 level in neoplastic and non-neoplastic pancreatic parenchyma. (A) log2 (2 DCq) values for NP, CP,
PanIN, and PDAC were calculated as shown in Materials and methods (**P < 0.0001, two-sample t-test for comparison with NP).
Mean  SEM is shown. (B) log2 (2 DCq) values for N and PDAC (**P < 0.0001, paired t-test for comparison with N). Mean  SEM is
shown. Abbreviations: NP, normal pancreatic parenchyma; CP, chronic pancreatitis; PanIN, pancreatic intraepithelial neoplasm;
PDAC, pancreatic ductal adenocarcinoma; N, paired pancreatic tissue with negative margin.

comparison of miR-217 levels between the neoplastic lesions Unlike other studies, herein we included the PDAC
and CP showed that miR-217 levels decreased the most in precursor, PanIN, with different grades of dysplasia (grade I
PDAC (w8.5 DDCq difference, P < 0.0001). miR-217 to III). Our data showed that, compared to the non-neoplastic
expression levels also decreased, although to a lesser parenchyma, miR-148a was significantly underexpressed,
extent, in PanIN compared to CP (w3.5 DDCq difference, whereas miR-196a and miR-10b were highly overexpressed
P Z 0.09). Dramatic reduction of the miR-217 expression in PanIN. miR-217 expression was shown to decrease in
level in PDAC was further confirmed by comparing its level in PanIN compared to that in the non-neoplastic tissue,
PDAC and corresponding paired control with benign pancre- although the degree of decrease does not reach statistical
atic tissue (P < 0.0001, Figure 3B). significance, which is most likely due to the small number of
samples. Our data suggest that these miRNA markers may
miR-196a and miR-21 expression levels are be involved in an early event in pancreatic carcinogenesis,
significantly up-regulated in PDAC which was reported by only two studies (9,10). Interestingly,
the degree of deregulation of all four miRNA markers
appears to correlate with the degree of dysplasia. In other
Besides miR-148a and miR-217, deregulation of miRNA-
words, the degree of deregulation of miRNA expression is
10b, -21, and -196a have been implicated in pancreatic
much higher in PanIN II-III compared to that in PanIN I (data
carcinogenesis (4e8). Therefore, we analyzed the expres-
not shown). More studies need to be conducted to confirm
sion of these miRNAs to discern whether any or all of them
this phenomenon and further investigate how these markers
could also be differentially regulated in PanIN and PDAC,
are involved in the early stage of pancreatic carcinogenesis.
with its corresponding paired control with benign pancreatic
In summary, our study demonstrates that certain
tissue. Our results showed that miR-196a was significantly
miRNAs, especially miR-148a, miR-217, and miR-196a, are
overexpressed in PanIN (P < 0.0001) and PDAC (P <
0.0001) compared to the group with paired benign pancreatic
tissue (Figure 4). We also observed that expression of miR-
10b was significantly up-regulated in PanIN (P < 0.05),
whereas the miR-21 level was significantly elevated in PDAC
(P < 0.05) (Figure 4).

Discussion

In this study, we evaluated the utility of miRNA as diag-

nostic markers of PDAC and its precursor, PanIN. We used
qRT-PCR to assess the expression of miR-217, miR-148a,
miR-196a, miR-10b, and miR-21 in pancreatic resection
specimens that included PDAC, PanIN, as well as non-
neoplastic pancreatic tissues including NP, CP, and benign
pancreatic tissue from corresponding carcinoma cases. Figure 4 Expression of miR -196a, miR-21, and miR-10b in
Consistent with data from other published studies neoplastic and non-neoplastic pancreatic parenchyma. log2
(5e7,11e15), we observed that miR-148a and miR-217 (2 DCq) values for N, PanIN, and PDAC were calculated as
expression levels were dramatically down-regulated in shown in Materials and methods (*P < 0.05, **P < 0.0001, two-
PDAC, whereas miR-196 and miR-21 expression were sample t-test for comparisons between PanIN and N, paired t-
significantly up-regulated in PDAC. We did not observe any test for comparisons between PDAC and N). Mean  SEM is
correlation between microRNA expression levels in different shown. Abbreviations: NP, normal pancreatic parenchyma; CP,
degrees of differentiation in PDAC and patients’ lymph nodes chronic pancreatitis; PanIN, pancreatic intraepithelial neoplasm;
and metastatic statuses (data not shown). More cases will be PDAC, pancreatic ductal adenocarcinoma; N, paired pancreatic
selected to further address this issue. tissue with negative margin.
MicroRNA alterations of PDAC and its precursor 5

deregulated in PDAC and its precursor, PanIN II-III, 8. du Rieu MC, Torrisani J, Selves J, et al. MicroRNA-21 is induced
compared to non-neoplastic pancreatic tissue. Thus, these early in pancreatic ductal adenocarcinoma precursor lesions.
markers have the potential to be used as a tumor marker Clin Chem 2010;56:603e612.
panel to distinguish PDAC and its precursor from a benign 9. Hanoun N, Delpu Y, Suriawinata AA, et al. The silencing of micro-
RNA 148a production by DNA hypermethylation is an early event
lesion.
in pancreatic carcinogenesis. Clin Chem 2010;56:1107e1118.
It will be highly important to analyze key downstream 10. Bloomston M, Frankel WL, Petrocca F, et al. MicroRNA expression
targets for better understanding of the molecular mecha- patterns to differentiate pancreatic adenocarcinoma from normal
nism(s) underlying PDAC progression. Several groups have pancreas and chronic pancreatitis. JAMA 2007;297:1901e1908.
previously validated immediate targets for miR-148a, miR- 11. Moriyama T, Ohuchida K, Mizumoto K, et al. MicroRNA-21
217, and miR-196a and delineated their role in tumor modulates biological functions of pancreatic cancer cells
progression. In general, miR-148a and miR-217 act like including their proliferation, invasion, and chemoresistance. Mol
tumor suppressors and their down-regulation was shown to Cancer Ther 2009;8:1067e1074.
enhance tumorigenesis by de-repressing oncogenic targets 12. Park JY, Helm J, Coppola D, et al. MicroRNAs in pancreatic ductal
such as ROCK1 (16), WNT10 B (17), Bcl-2 (18), TGIF2 (19), adenocarcinoma. World J Gastroenterol 2011;17:817e827.
13. Steele CW, Oien KA, McKay CJ, et al. Clinical potential of
DNMT3B (20), and KRAS (21). On the other hand, up-
microRNAs in pancreatic ductal adenocarcinoma. Pancreas
regulation of miR-196a has been shown to promote tumori- 2011;40:1165e1171.
genesis through direct repressing of tumor suppressor genes 14. Frampton AE, Krell J, Jacob J, et al. microRNAs as markers of
such as p27 kip1 (22) and HOXA5 (23). Prospective studies survival and chemoresistance in pancreatic ductal adenocarci-
that analyze the expression of such targets in PDAC will noma. Expert Rev Anticancer Ther 2011;11:1837e1842.
shed more light on the underlying molecular mechanisms 15. Panarelli NC, Chen YT, Zhou XK, et al. MicroRNA expression
and further confirm the utility of the described miRNAs as aids the preoperative diagnosis of pancreatic ductal adenocar-
diagnostic tools for the assessment of patients with PDAC. cinoma. Pancreas 2012;41:685e690.
16. Zheng B, Liang L, Wang C, et al. MicroRNA-148a suppresses
tumor cell invasion and metastasis by downregulating ROCK1 in
gastric cancer. Clin Cancer Res 2011;17:7574e7583.
References 17. Aprelikova O, Palla J, Hibler B, et al. Silencing of miR-148a in
cancer-associated fibroblasts results in WNT10B-mediated stim-
1. Edge SB, Byrd DR, Compton CC, et al, editors. AJCC cancer ulation of tumor cell motility. Oncogene 2013;32:3246e3253.
staging manual. New York, NY: Springer, 2012. 18. Zhang H, Li Y, Huang Q, et al. MiR-148a promotes apoptosis by
2. Iorio MV, Croce CM. microRNA involvement in human cancer. targeting Bcl-2 in colorectal cancer. Cell Death Differ 2011;18:
Carcinogenesis 2012;33:1126e1133. 1702e1710.
3. Pritchard CC, Cheng HH, Tewari M. MicroRNA profiling: app- 19. Lujambio A, Calin GA, Villanueva A, et al. A microRNA DNA
roaches and considerations. Nat Rev Genet 2012;13:358e369. methylation signature for human cancer metastasis. Proc Natl
4. Preis M, Gardner TB, Gordon SR, et al. MicroRNA-10b Acad Sci U S A 2008;105:13556e13561.
expression correlates with response to neoadjuvant therapy and 20. Duursma AM, Kedde M, Schrier M, et al. mir-148 targets human
survival in pancreatic ductal adenocarcinoma. Clin Cancer Res DNMT3b protein coding region. RNA 2008;14:872e877.
2011;17:5812e5821. 21. Zhao WG, Yu SN, Lu ZH, et al. The miR-217 microRNA func-
5. Dillhoff M, Liu J, Frankel W, et al. MicroRNA-21 is overexpressed tions as a potential tumor suppressor in pancreatic ductal
in pancreatic cancer and a potential predictor of survival. adenocarcinoma by targeting KRAS. Carcinogenesis 2010;31:
J Gastrointest Surg 2008;12:2171e2176. 1726e1733.
6. Szafranska AE, Doleshal M, Edmunds HS, et al. Analysis of 22. Sun M, Liu XH, Li JH, et al. MiR-196a is upregulated in gastric
microRNAs in pancreatic fine-needle aspirates can classify cancer and promotes cell proliferation by downregulating
benign and malignant tissues. Clin Chem 2008;54:1716e1724. p27(kip1). Mol Cancer Ther 2012;11:842e852.
7. Lee EJ, Gusev Y, Jiang J, et al. Expression profiling identifies 23. Liu XH, Lu KH, Wang KM, et al. MicroRNA-196a promotes non-
microRNA signature in pancreatic cancer. Int J Cancer 2007; small cell lung cancer cell proliferation and invasion through
120:1046e1054. targeting HOXA5. BMC Cancer 2012;9:348.