Article (PDF Available)

Single nucleotide polymorphismin RECQL and survival in resectable pancreatic adenocarcinoma

Department of Molecular and Human Genetics, Human Genome Sequencing Center, Baylor College of Medicine, Houston, TX 77030, USA.
HPB (Impact Factor: 2.68). 08/2009; 11(5):435-44. DOI: 10.1111/j.1477-2574.2009.00089.x
Source: PubMed

ABSTRACT

RECQL is a DNA helicase involved in DNA mismatch repair. The RECQL polymorphism, 3' untranslated region (UTR) A159C, was previously associated with overall survival of patients with resectable pancreatic adenocarcinoma treated with neoadjuvant chemoradiation. In the present study, we examined RECQL for somatic mutations and other polymorphisms and compared these findings with the outcome in patients who received adjuvant or neoadjuvant chemoradiation. We hypothesized that RECQL (i) would be mutated in cancer, (ii) would have polymorphisms linked to the 3'UTR A159C and that either or both events would affect function. We also hypothesized that (iii) these changes would be associated with survival in both cohorts of patients.
We sequenced RECQL's 15 exons and surrounding sequences in paired blood and tumour DNA of 39 patients. The 3'UTR A159C genotype was determined in blood DNA samples from 176 patients with resectable pancreatic adenocarcinoma treated with adjuvant (53) or neoadjuvant (123) chemoradiation. Survival was calculated using the Kaplan-Meier method, with log rank comparisons between groups. The relative impact of genotype on time to overall survival was performed using the Cox proportional hazards model.
Somatic mutations were found in UTRs and intronic regions but not in exonic coding regions of the RECQL gene. Two single nucleotide polymorphisms (SNPs), located in introns 2 and 11, were found to be part of the same haplotype block as the RECQL A159C SNP and showed a similar association with overall survival. No short-term difference in survival between treatment strategies was found. We identified a subgroup of patients responsive to neoadjuvant therapy in which the 159 A allele conferred strikingly improved long-term survival.
The RECQL 3'UTR A159C SNP is not linked with other functional SNPs within RECQL but may function as a site for regulatory molecules. The mechanism of action needs to be clarified further.

Full-text (PDF)

Available from: Sally Hodges
ORIGINAL ARTICLE
Single nucleotide polymorphism in RECQL and survival in
resectable pancreatic adenocarcinoma
Ronald T. Cotton
1,2
, Donghui Li
3
, Steven E. Scherer
1
, Donna M. Muzny
1
, Sally E. Hodges
2
, Robbi L. Catania
1,2
,
Agnieszka K. Witkiewicz
4,5
, Jonathan R. Brody
4
, Eugene P. Kennedy
4
, Charles J. Yeo
4
, F. Charles Brunicardi
2
,
Richard A. Gibbs
1
, Marie-Claude Gingras
1,2
& William E. Fisher
2
1
Department of Molecular and Human Genetics, Human Genome Sequencing Center,
2
Michael DeBakey Department of Surgery and Elkins Pancreas Center,
Baylor College of Medicine,
3
Department of Gastrointestinal Medical Oncology, The University of Texas M.D. Anderson Cancer Center, Houston, TX,
4
Department of Surgery, Jefferson Center for Pancreatic, Biliary and Related Cancers, and
5
Department of Pathology, Kimmel Cancer Center, Thomas
Jefferson University, Philadelphia, PA, USA
Abstract
Background: RECQL is a DNA helicase involved in DNA mismatch repair. The RECQL polymorphism,
3` untranslated region (UTR) A159C, was previously associated with overall survival of patients with
resectable pancreatic adenocarcinoma treated with neoadjuvant chemoradiation. In the present study, we
examined RECQL for somatic mutations and other polymorphisms and compared these findings with the
outcome in patients who received adjuvant or neoadjuvant chemoradiation. We hypothesized that RECQL
(i) would be mutated in cancer, (ii) would have polymorphisms linked to the 3`UTR A159C and that either
or both events would affect function. We also hypothesized that (iii) these changes would be associated
with survival in both cohorts of patients.
Material and methods: We sequenced RECQL's 15 exons and surrounding sequences in paired blood
and tumour DNA of 39 patients. The 3`UTR A159C genotype was determined in blood DNA samples from
176 patients with resectable pancreatic adenocarcinoma treated with adjuvant (53) or neoadjuvant (123)
chemoradiation. Survival was calculated using the Kaplan–Meier method, with log rank comparisons
between groups. The relative impact of genotype on time to overall survival was performed using the Cox
proportional hazards model.
Results: Somatic mutations were found in UTRs and intronic regions but not in exonic coding regions
of the RECQL gene. Two single nucleotide polymorphisms (SNPs), located in introns 2 and 11, were found
to be part of the same haplotype block as the RECQL A159C SNP and showed a similar association with
overall survival. No short-term difference in survival between treatment strategies was found. We iden-
tified a subgroup of patients responsive to neoadjuvant therapy in which the 159 A allele conferred
strikingly improved long-term survival.
Discussion: The RECQL 3`UTR A159C SNP is not linked with other functional SNPs within RECQL but
may function as a site for regulatory molecules. The mechanism of action needs to be clarified further.
Keywords
RECQL, RECQ1, polymorphism, resectable pancreatic adenocarcinoma, neoadjuvant and adjuvant therapy
Received 20 March 2009; accepted 7 May 2009
Correspondence
William E. Fisher, Elkins Pancreas Center, Michael E. DeBakey Department of Surgery, Baylor College
of Medicine, 1709 Dryden, Suite 1500, Houston, TX 77030, USA. Tel: +1 713 798 8695; Fax:
+1 713 798 4530; E-mail: wfisher@bcm.edu
Marie-Claude Gingras, Human Genome Sequencing Center, Baylor College of Medicine, One Baylor Plaza,
N1519, Houston, TX 77030, USA. Tel: +1 713 798 1286; Fax: +1 713 798 3741; E-mail: mgingras@bcm.edu
Presented at the 9th Annual Meeting of the American Hepato-Pancreato-Biliary Association, 12–15 March 2009, Miami, FL, USA.
DOI:10.1111/j.1477-2574.2009.00089.x
HPB
HPB 2009, 11, 435–444 © 2009 International Hepato-Pancreato-Biliary Association
Page 1
Introduction
With an estimated 37 680 new diagnoses and 34 290 deaths in
2008, cancer of the pancreas is the eleventh most common
cancer in the United States and fourth in fatalities.
1,2
Most
American centres treat resectable pancreatic cancer with resec-
tion followed by adjuvant chemoradiation, which extends sur-
vival from 12 months with surgery alone to about 20 months
with adjuvant chemoradiation, with a 20% 5 year post-operative
survival. Conversely, the neoadjuvant approach, with assured
administration of chemoradiation to all potential surgical
candidates and potentially improved patient tolerance prior
to surgery, offers some theoretical advantages over immediate
surgery. Pre-operative treatment may decrease tumour size and
increase the chances of an R0 (microscopically negative margin)
resection. However, about 30% of patients treated with pre-
operative chemoradiation will develop progressive disease and
will never come to surgery. Clearly, discovery of a means to
predict the outcome with the adjuvant or neoadjuvant approach
would be extremely useful.
Gemcitabine (GEM) is the standard chemotherapy for
pancreatic cancer. GEM interferes with DNA replication and
prevents the proofreading enzymes from detecting, excising and
repairing the DNA.
3
In vitro studies have shown that cells defi-
cient in DNA mismatch repair (MMR) enzymes are more sen-
sitive to GEM/radiation.
4
The RecQ family is a highly conserved
group of DNA helicases required for the maintenance of genome
stability and integrity. They have an important role in DNA
replication, telomere maintenance, DNA damage signalling and
DNA repair pathways including mismatch repair, nucleotide
excision repair and direct repair.
5,6
The most highly expressed
but smallest member of this family in human cells is the RecQ
protein-like (DNA helicase Q1-like) (RECQL often named
RECQ1). RECQL catalytic activities include the DNA unwinding
of diverse but specific DNA structures, the annealing of comple-
mentary single-stranded DNA and DNA branch migration.
7–11
This protein plays an important role in chromosomal stability
and genome maintenance.
11
It interacts with and binds to mis-
match repair proteins that regulate genetic recombination
12
and
participates in the repair of endogenous or exogenously induced
DNA damage.
13
RECQL is highly expressed in rapidly proli-
ferating cancer cells and transformed cells and provides these
cells with a growth advantage
14,15
suggesting that greater copy
numbers of RECQL may be needed to repair the elevated load
of endogenous DNA damage generated during their accelerated
cell cycle. Accordingly, RECQL acute depletion (silencing by
small interference RNA) in cancer cells induces growth retar-
dation, sensitivity to DNA damaging agents, accumulation of
DNA damage, chromosomal instability and ultimately results
in mitotic catastrophe coupled with mitotic cell death in cancer
cells already compromised in their checkpoint.
12,15
Conversely,
allelic loss at the RECQL locus (chromosome 12p12) is frequent
in different tumour types.
11
Alteration in DNA repair pathways may affect the cytotoxicity
of chemotherapy and radiotherapy, more specifically the resis-
tance to gemcitabine-induced DNA replication arrest and the
repair of DNA double-strand breaks caused by radiation. The
importance of RECQL in damage repair is further supported
by the observation that a single nucleotide polymorphism (SNP)
located in the 3 untranslated region (UTR) (A159C) negatively
affects the overall survival and response to gemcitabine-induced
radiosensitization of patients diagnosed with pancreatic adeno-
carcinoma treated with neoadjuvant chemoradiation.
16,17
This
SNP exerts its effect in a dominant-inheritance mode.
Genetic changes can be considered either germline or somatic.
Germline variation refers to an alteration in DNA sequence inher-
ited from one’s parents. Single nucleotide polymorphisms (SNPs)
account for more than 90% of germline variation in the human
genome, and have been implicated in phenotype, disease predis-
position and response to therapy.
18,19
Somatic mutations, to the
contrary, are found in diseased tissue (i.e. tumour) only, and
are not a part of the inherited genetic complement. By altering
protein function, somatic mutations can have a profound impact
on tumour development and proliferation (e.g. tumour sup-
pressor genes and oncogenes).
In the present study, the RECQL gene was sequenced in patients
with pancreatic cancer to discover the presence of additional SNPs
or mutations in the coding region that may directly affect the
function of the gene. In addition, to evaluate the prognostic value
of the 3UTR A159C SNP in association with the course of treat-
ment, the effect of the SNP on clinical outcome was compared
in two cohorts of patients with resectable pancreatic cancer:
one treated with resection and adjuvant chemoradiation, and the
other treated with neoadjuvant chemoradiation with surgery or
additional chemotherapy as indicated by repeat CT scan.
Material and methods
Sample collection and processing
Informed consent from patients with resectable pancreatic exo-
crine adenocarcinoma was obtained under an institutional review
board approved protocol at the institutions participating in this
study (BCM: H16215 issued 9/22/04; TJU: 06U.76 issued 4/27/06;
MDA: ID 98–155 issued 09/09/98). A total of 53 patients with
exocrine adenocarcinoma located in the head of the pancreas were
treated with pancreaticoduodenectomy (Whipple procedure) and
adjuvant chemoradiation at the Elkins Pancreas Center at Baylor
College of Medicine (BCM), Houston, and the Thomas Jefferson
University Center (TJU) for Pancreatic, Biliary and Related
Cancers, Philadelphia (26 and 27 patients, respectively). One
hundred twenty-three patients received neoadjuvant chemora-
diation at the University of Texas MD Anderson Cancer Center
(MDACC).
The blood from the BCM patients was directly collected in
PAXgene Blood DNA tubes, and the DNA was isolated with the
PAXgene Blood DNA kit (PreAnalytiX; Qiagen, Valencia, CA,
USA) according to the manufacturer’s instructions. The MDACC
436 HPB
HPB 2009, 11, 435–444 © 2009 International Hepato-Pancreato-Biliary Association
Page 2
blood samples were collected in heparinized vacutainers (BD
Biosciences, Franklin Lakes, NJ, USA), and the peripheral mono-
nuclear cells were immediately separated by Ficoll-Histopaque
(Amersham Pharmacia Biotech, Piscataway, NJ, USA) density gra-
dient centrifugation; the DNA was then extracted with the Flexi-
Gene DNA kit (Qiagen). At TJU, after surgical resection, genomic
DNA was isolated from normal and tumour pancreatic tissue.
gDNA was extracted using the DNeasy Blood and Tissue Kit
(Qiagen).
Sequencing data were obtained from 39 macroscopically
dissected pancreatic adenocarcinoma tissue specimens from
operative cases (head, tail and metastasizing tumours) and match-
ing blood from the same patient (BCM). The tissue samples were
stored in a protease inhibitor solution (Roche Applied Science,
Indianapolis, IN, USA) at -80°C.
19,20
The DNA was isolated from
the preserved tissue using the QIAamp DNA Mini kit (Qiagen)
after washing the tissue several times in phosphate-buffered saline
(PBS) to remove any trace of the stabilizing solution.
Gene amplification and sequencing
The exons and their flanking sequences (300 bases) of RECQL
were identified using the UCSC genome bioinformatics website
(http://genome.ucsc.edu; accessed 18 June 2009). Primer sets
(Table 1) were designed using the HGSC primer design pipeline
that links several software in an integrated automatic approach as
previously described.
19
Ten to fifty nanograms (ng) of DNA was whole genome ampli-
fied (WGA) (GenomiPhi DNA amplification kit, Amersham
Biosciences). PCR was performed on 10 ng of WGA DNA in a final
reaction volume of 8 ml in a 384-well plate using the HotStar
polymerase (Qiagen). Cycling parameters consisted of 40 cycles of
a denaturation step at 95°C for 45 s followed by an annealing step at
60°C for 45 s, and an extension step at 72°C for 45 s. The cycling
process was preceded by a denaturation period at 95°C for 15 min
and followed by a final extension period of 7 min at 72°C. Un-
consumed dNTPs were hydrolyzed and remaining primers were
degraded using a cocktail of Shrimp Alkaline Phosphatase and
Table 1 Primer sets and probes used for sequencing, validation and genotyping
Sanger sequencing primer sets
Exon Forward primer Forward primer
1 TAACTTTCCGGTTTCTCCTCCG TTCCTATTGGCGAAACCTGCTT
1 TGGAGGAAACGCCACTGAGATA AGCTTTGAAGGGTCAAGGGTGT
2 CAAACAGAAATAGAACAGAAGGAAGAAGA TGGTTCTTATTTGAAAGGTCACTGC
3 CTGCAAGTTTCCCATTCCACTG CTGTAATTGATATGGCGGGCAA
4 TGACAAAGCACTTCTTCAACTCAAA TTTGCTTTGCTTAGCTAGTGAGT
5 GCAGGTAAAGCTCCTATTTCAGTG GCATCTGATTCTGAGGGTGGTG
6 AGCAGAGATTTCCATCATGCCA TGCTCCTAGAAGAGCCAAAGGC
7 CCCTCTGCGTAATTCTTCACAAA CCTGCCCATCAAAGAGGCTAAC
8 TCTCTGTCTCCAAAGTTGGTTTG TCTGGGAGGAGGAACAGATGAG
9 TCAATTCCCATACCAAATGCAA CAATGTCCTGAATGTGTGTGTTGA
10 CCATATGCAAGTAAGTGTCAAACTGC TGCTTACCATGCCAATTTGGAG
11 GATGTGCTGGTTCCTACCCTCC GGGTACCCATTACAAATTAACTTCCAC
12 CACTAGGCATTATGACTGTATTAGCC ATCCAGTGAAACCCATAGTACACTTC
13 AAACCGTTTATTCTGTTGCAATTT AGGGTGGCTCACATTGATAACC
14 CAATGTGTTAAGAATTACTGCGCATA GAAAAGCATCCCATAGGCTTT
15 AAACTTAAGACGATTGTATGAACTTATTCTC TTTGTAGGCTGAATCGTCTCAAAC
15 AAACTTAAGACGATTGTATGAACTTATTCTC GAAAAGAAAAATCGATGATGCCT
15 CACTCAGTGAACCTCTGTCAGT TTTGAGAATAAGTTCATACAATCGTC
Biotage primer set
SNP Forward primer Reverse primer Sequencing primer
Intron 2, IVS2 - 17 TTTCCGTAAGTTCTTGAATTTG TCAATCTGGAAACCTAAAGTTGTA CTATGGGAGGCAGCG
intron 11, IVS11 + 30 GGGCATCTTTTCTTGAACTAAGG GCAGAAGCTTTATGAGATGGTATC CTTTCATATTTGCTTTAATT
3` UTR, A159C AATAATGGCATATA CATGCATAAA ACAGGAGCTAAGAAAAGAAAAATC TCGATGATGCCTGATA
TaqMan primer and probe set
SNP Forward primer Reverse primer Probe
3` UTR, A159C CAATCTGGTTCTAAGAATACAGGAGCTA GGCATATACATGCATAAACCATCTTT AGTAACAGTCATATCAGG
AGTAACATTCATATCAGGC
HPB 437
HPB 2009, 11, 435–444 © 2009 International Hepato-Pancreato-Biliary Association
Page 3
Exonuclease I (ExoSAP-IT, USB). The purified PCR products were
diluted and sequenced using BigDye Terminator version 3.1
Cycle Sequencing kit onApplied Biosystems (Foster City,CA, USA)
3730XL DNA Sequencers. The sequences were analysed with SNP
Detector version 3 (created by Jinhui Zhang et al.
21
at the NCI)
using the corresponding sequence deposited in GenBank as refer-
ence. Base disparities from the reference sequence identified by the
SNP Detector software were manually verified in Sequencher
version 4.6 (Gene Codes Corp).Somatic variants were identified by
comparing the sequences from each patient’s matched blood and
tumour samples.
Validation and genotyping
We validated the selected candidate variants by Biotage pyrose-
quencing. Briefly, Biotage assay design software was used to select
new PCR primers at each locus of interest, avoiding all known
SNPs and project-specific variants (Table 1). One of each primer
pair was tailed with a 23 base extension corresponding to the
sequence of a biotinylated universal primer. Primers successfully
tested for amplification product on human cell line DNA were
combined with the universal primer and used to amplify patient
DNA samples. Each amplicon was reduced to single strand by
denaturation and binding to streptavidin beads. Sequence-specific
extension primers were annealed within four bases of the target
variant and extended using the standard Biotage pyrosequencing
protocol and reagents. The resulting 8 to 15 bases of sequence
were analysed for quality based on pyrogram peak geometry and
genotype using Biotage software algorithms.
The 3UTR A159C SNP was also genotyped with the Taq Man
SNP Genotyping assay (Applied Biosystems) (Table 1). The reac-
tions were prepared using 10 ng of gDNA, Taq Man universal
master mix and SNP genotyping assay mix in a final volume of
5 ml. The PCR was done using the ABI Prism 7900 HT sequence-
detector.
Statistical methods
Estimations of survival were performed via the Kaplan–Meier
method, with comparisons in survival between groups made pair-
wise by log-rank tests. Differences in distribution of variables
between groups were calculated using Pearson’s c
2
-test. Potential
confounders were identified by univariate analysis; individual
variable’s relative effect on survival was estimated via a Cox
regression model. All statistical calculations were performed with
SPSS version 12.0.1 (SPSS Inc., Chicago, Illinois, USA).
Results
Sequencing results: somatic events (specifically
found in tumour DNA only)
We sequenced the 15 exons and surrounding intronic regions
of the RECQL gene in 39 matched blood and tumour samples of
patients diagnosed with pancreatic adenocarcinoma resected
before chemoradiation. A total of 26 events were identified
(Table 2). Out of these events, two consisted of a somatic base shift
mutation located in the 5UTR, and intron 1 in two patient
tumours (Tables 2,3). Six other somatic events located in five
different introns as well as the 3UTR A159C consisted of loss of
heterozygocity (LOH) at the site of known polymorphisms. Most
of these somatic events were found in two patients (Tables 2,3).
Sequencing results: germline polymorphisms (found
in blood as well as tumour DNA)
The remaining events consisted of 24 polymorphisms, of which15
were known to be present in the general population, and had an
assigned ID number (http://genewindow.nci.nih.gov; accessed 18
June 2009). Fourteen SNPs (nine known) were located in the
intronic region outside the splicing site, six SNPs (three known) in
the UTRs of the gene, one known SNP 115 bases upstream of exon
1, and three SNPs in the coding region of exon 13 (unknown
missense SNP), exons 14, and 15 (both silent known SNPs)
(Table 3). The impact of the missense mutation on recql function
was suggested to be benign based on a PolyPhen analysis (http://
genetics.bwh.harvard.edu/pph/; accessed 18 June 2009).
Nine SNPs were part of three haplotype blocks (Table 3): hap-
lotype 1: IVS2 - 17, IVS11 30, Ex15 + 159 (3 UTR A159C);
haplotype 2: IVS7 - 69, IVS8 - 33, IVS10 + 82, and IVS11 + 103;
and haplotype 3: Ex14 + 64 and Ex15 + 102. First we genotyped
eight of the SNPs showing greater heterozygocity in DNA
extracted from the blood of 77 patients receiving neoadjuvant
therapy. Only three SNPs had any correlation with overall sur-
vival: they were part of the first haplotype block described above.
The total number of genotyped patients was then increased to
123 and the overall survival curves by genotype were established
for each SNP. From the three SNPs, the 3UTR A159C was the
most statistically significant (Fig. 1).
Table 2 RECQL germline and somatic events identified in blood and matching tumour of patients with pancreatic exocrine adenocarcinoma
Gene region Total event Germline SNP Somatic
LOH Base shift
Promoter 1 1
Intron 15 14 5 1
Exon: UTR 7 6 1 1
Coding: synonymous 2 2
Coding: nonsynonymous 1 1
438 HPB
HPB 2009, 11, 435–444 © 2009 International Hepato-Pancreato-Biliary Association
Page 4
Table 3 Somatic and germline events detected in the matching blood and tumour samples of patients treated with adjuvant therapy
Chrom location
(hg 18)
Gene location,
function
Official
nomenclature
Base
shift
SNP ID Genotype
Total patient
number
Germline Somatic
Homo
major
Hetero Homo
minor
Amount Type Patient ID
Somatic events: mutations and LOH
21545758 5` UTR, CDS -433 bases Ex1 + 39 G > A 39 38 1 Base shift 58
21545258 Intron 1 IVS1 + 113 G > A 38 37 1 Base shift 2440
21543908 Intron 1 IVS1 - 92 C > T rs12423412 38 22 13 2 1 LOH 6030
21535934 Intron 2 IVS2 - 17 C > T rs10841834 39 10 15 11 3 LOH 57, 58, 6030
21520058 Intron 8 IVS8 - 33 A > G rs3752648 39 13 18 7 1 LOH 58
21519587 Intron 10 IVS10 + 82 G > A rs10841831 38 13 16 7 2 LOH 58,4520
21519012 Intron 11 IVS11 + 30 T > C rs2159943 39 9 20 9 1 LOH 6030
21514389 3` UTR, CDS +6 bases Ex15 + 159 A > C rs13035 39 10 18 10 1 LOH 6030
Germline polymorphisms in exons and their immediate surrounding regions:
21545759 5` UTR, CDS -434 bases Ex1 + 38 G > A393711
21545736 5` UTR, CDS -411 bases Ex1 + 61 A > C rs1061626 39 29 10
21545674 5` UTR, CDS -349 bases Ex1 + 123 T > C rs1061627 39 25 14 1
21545532 5` UTR, CDS -207 bases Ex1 + 265 C > T39363
21515831 Ile489Val, missense Ex13 + 18 A > G39381
21515236 Asn577Asn, silent Ex14 + 64 T >
C rs6500 39 34 5
21514446 Gln633Gln, silent Ex15 + 102 A > G rs17849408 39 34 5
21514194 3` UTR, CDS +210 Ex15 + 360 A > G39381
21545911 Promoter, Exon 1 -115
bases
-2139 T > G rs1860947 39 32 6 1
21535933 Intron 2 IVS2 - 16 G > C39381
21534665 Intron 3 IVS3 - 86 G > A rs4987216 39 36 3
21534630 Intron 3 IVS3 - 51 C > T39381
21521262 Intron 7 IVS7 - 69 - > G rs5796903 39 13 19 7
21518939 Intron 11 IVS11 + 103 G > C rs35159698 38 13 18 7
21515584 Intron 13 IVS13 + 45 T > -37298
21515576 Intron 13 IVS13 + 53 T > A rs11046076 37 8 9 20
21515571 Intron 13 IVS13 + 58 A > -372611
21515155 Intron 14 IVS14 + 15 TTAA > -39381
CDS, coding sequence; UTR, untranslated region; homo, homozygote; hetero, heterozygote.
HPB 439
HPB 2009, 11, 435–444 © 2009 International Hepato-Pancreato-Biliary Association
Page 5
The 3`UTR A159C polymorphism: correlation between
survival and course of treatment
We compared the influence of the 3UTR A159C SNP on overall
survival between patients treated with neoadjuvant chemo-
radiation and adjuvant therapy. Characteristics of the two
cohorts of patients are detailed in Table 4. The two groups
were similar in age, race, pre-operative CA 19-9 and TNM
stage. There was no difference in overall survival between the
two groups (Fig. 2). Although there were differences in gender,
the number who underwent resection, tumour size and nodal
status, only negative nodal involvement offered a survival
advantage using Cox regression (HR 0.420, CI 0.233–0.760,
P = 0.004).
We determined the germline genotype from the blood DNA
of 123 patients receiving neoadjuvant therapy and 53 patients
receiving adjuvant therapy. No significant difference was
observed in the genotype distribution among the two groups:
the frequency of the A allele was 56.6% vs. 58.1%, respectively
(Table 5). No statistically significant difference in short-term sur-
vival was observed when comparing patients treated by the neo-
adjuvant vs. adjuvant approach within each respective genotype
(Table 6). For example, in comparing neoadjuvant and adjuvant
therapy in patients with the AA genotype, survival was 82% vs.
75% in the 1st year and 59% vs. 45% for the 2nd year. This
represented no significant difference between the two treatment
cohorts.
When comparing the effect of genotype within treatment
groups, patients with the AA genotype receiving neoadjuvant
chemoradiation were observed to have improved long-term sur-
vival compared with their AC and CC counterparts (Fig. 1,
Table 6). Five-year survival was 39% for AA patients, compared
with 14% for genotype AC, (P = 0.025) and 19% for genotype CC
(P = 0.003). The AA genotype was an independent predictor of
improved survival by Cox regression (HR 2.51, CI 1.321–4.756,
P = 0.005). No long-term survival data were available for the
patients receiving adjuvant chemoradiation. However, no statis-
tically significant difference in short-term survival by genotype
was seen after 30 months.
We further analysed the population that received neoadjuvant
therapy by dividing these patients into two groups based on resec-
tability after chemoradiation.A survival benefit was again observed
with the AA genotype in those undergoing surgical resection after
chemoradiation, with survival of 97%, 80%, and 55% at 1, 2 and
5 years, respectively vs. 89%, 62%, and 19% for AC patients (P =
0.010), and 93%, 36%, and 29% for CC patients (P = 0.003) (Fig. 3,
Table 7). The AA genotype was again a predictor of survival (HR
2.92, CI 1.531–5.564, P = 0.001). However, none of the genotypes
conferred a survival benefit in patients that could not undergo
resection as a result of progressive disease during neoadjuvant
chemoradiation (Fig. 3). These patients had a 2-year survival of
10% for all genotypes.The differencein survival was highly empha-
sized when the outcomes of each genotype were compared (Fig. 4).
IVS2-17 C < T
1.0
0.8
0.6
0.4
0.2
0.0
0
p =
CC/TT
CC/CT
0.017
0.020
TT/CC
TT/TC
0.021
0.092
AA/CC
AA/AC
0.003
0.025
20 40 60
Cummulative Survival
80
CC
CT
TT
CC-censored
CT-censored
TT-censored
CC
TC
TT
CC-censored
TC-censored
TT-censored
AA
AC
CC
AA-censored
AC-censored
CC-censored
100 120 0 20 40 60
Time to Event (Months)
80 100 120 0 20 40 60 80 100 120
IVS11+30 T < C 3’ UTR, A159C
Figure 1 Comparison of the overall survival by genotype of three single nucleotide polymorphisms (SNPs) from the same haplotype. The
genotypes CC, TT and AA at IVS32 - 17, IVS11 + 30, and at the base 159 of exon 15 in the 3` untranslated region (UTR) confer a similar
overall survival advantage to the 123 patients that received neoadjuvant therapy. The 3`UTR A159C has a greater statistical significance.
440
HPB
HPB 2009, 11, 435–444 © 2009 International Hepato-Pancreato-Biliary Association
Page 6
Discussion
Although this study provided a more thorough sequencing of the
RECQL gene in the pancreatic cancer patient population, it failed
to identify mutations in the coding region of the gene that would
impair its functionality. Only a few patients harboured an LOH
at the gene locus as previously observed in other cancer types.
11
In this study, two SNPs in introns 2 and 11 of the RECQL gene
were identified as being in linkage disequilibrium with 3 UTR
A159C and had a similar effect on clinical outcome. Therefore, the
mechanism of the association is unlikely to be a SNP resulting
in amino acid substitution and subsequent structural and func-
tional defects in the translated RECQL protein. The RECQL poly-
morphism may represent a binding site for regulatory molecules
that affect gene expression. Also, the polymorphism may impact
GEM’s mechanism of action. Both of these hypotheses will
require further investigation.
The 3UTR A159C SNP is present at different frequency among
different ethnicities as reported by the HapMap project and
other studies (http://www.ncbi.nlm.nih.gov/SNP/snp_ref.cgi?rs=
13035; 18 June 2009). The frequency of the A allele is about 55%
in the Asiatic population, 57% in the Caucasian and 90% in the
Yoruban population. The frequency of the AA genotype is as low
as 16% in some Asiatic ethnicities, about 35% in the Caucasian
population, and over 80% in the Yoruban population. We
observed a similar genotype distribution among our pancreatic
cancer patients.
The present study adds further information to the observation
that the well known and validated RECQL polymorphism, 3UTR
A159C, is associated with overall survival of patients with resect-
able pancreatic adenocarcinoma. No significant difference in
short-term survival was observed by genotype between the patients
treated with neoadjuvant and adjuvant therapy. A key weakness of
this study is the small sample size and limited follow-up, particu-
larly in the group treated with adjuvant chemoradiation. Long-
term follow-up of these patients is necessary to see if trends in
survival similar to that seen in the neoadjuvant group develop.
Similar to previously reported series, 34% of patients treated
neoadjuvantly progressed to unresectable disease. The subset of
unresectable neoadjuvant patients in our population experienced
extremely poor clinical results, with most dying within a year of
diagnosis. Clearly, the genotype and other factors that determine
response to treatment and overall survival are multiple and
complex. Factors, including mutations in genes impacting tumour
behaviour or regulatory molecules, may be involved in worsening
patient survival in this subset. Chen et al. have also reported
polymorphisms in cell-cycle genes that had an impact on overall
survival in patients with resectable adenocarcinoma treated with
neoadjuvant chemotherapy.
22
It is also possible that this subgroup
Table 4 Patient demographics
Neoadjuvant Adjuvant
123 patients 53 patients
Gender
a
Female 38.2% 58.5%
Male 61.8% 41.5%
Age group <50 12.2% 3.8%
51–60 22.8% 15.4%
61–70 38.2% 38.5%
>71 26.8% 42.3%
Race Caucasian 88.6% 81.1%
Hispanic 4.9% 3.8%
Black 4.9% 9.4%
Asian 1.6% 5.7%
CA 19-9 level <47 26% 26.8%
48–499 48% 39%
500–999 11.4% 12.2%
>999 14.6% 22.0%
Surgical resection
a
Resected 65.9% 100%
Non-resectable 34.1% 0%
Tumour size
a
<2 cm 42.2% 14.0%
>2 cm 57.6% 86.0%
TNM stage IA 7.8% 1.9%
IB 2.6% 3.8%
IIA 31.2% 21.2%
IIB 58.40% 73.10%
a
Significant difference between the two groups (P < 0.05).
1.0
0.8
Adjuvant
ChemoXRT Only
Neo-adjuvant
ChemoXRT Only
Adjuvant
ChemoXRT Only-
censored
Neo-adjuvant
ChemoXRT-
censored
0.6
0.4
0.2
0.0
0 5 10 15
Time to Event (Months)
Cummulative Survival
20 25 30
Figure 2 Cumulative overall survival of patients treated with adjuvant
and neoadjuvant therapy. No difference was noticed between the
two groups.
HPB 441
HPB 2009, 11, 435–444 © 2009 International Hepato-Pancreato-Biliary Association
Page 7
of patients had already developed micrometastases that were
undetected at staging prior to the initiation of neoadjuvant
therapy.
Within the specific subgroup of patients responsive to neoad-
juvant therapy who proceeded to resection, the AA genotype
is associated with a relatively promising long-term survival,
making it a potentially good prognostic marker for this group.
However, the AA genotype was in equal prevalence in patients
who proceeded to resection as well as those who failed neoadju-
vant therapy. It does not appear, then, that the 3UTR A159C
genotype can predict which patients will respond to neoadjuvant
chemoradiation. Thus, the clinical utility of this marker is limited
until a means of prospectively identifying this responsive sub-
group is established.
Conclusion
Recent advances in rapid DNA sequencing technology are
decreasing the cost and improving the efficiency of genetic
testing such that selection of therapy based on genetic analyses
of individual patients may soon become standard practice. This
personalized genomic approach to cancer therapy may improve
Table 5 RECQL A159C genotype distribution among the groups studied
Total AA AC CC % A allele
Adjuvant therapy
All patients 53 34.0% 45.3% 20.8% 56.6%
Neoadjuvant therapy
All patients 123 36.6% 43.1% 20.3% 58.1%
Resected 81 37.0% 45.7% 17.3% 59.9%
Unresected 42 35.7% 38.1% 26.2% 54.8%
Table 6 Survival comparison between the patients treated with neoadjuvant and adjuvant therapy
Genotype AA AC CC
Group Neoadjuvant Adjuvant Neoadjuvant Adjuvant Neoadjuvant Adjuvant
1 year 82% 75% 74% 70% 68% 45%
2 years 59% 45% 45% 52% 24% 34%
5 years 39% na 14% na 19% na
na, not available yet.
Figure 3 Comparison of the overall
survival by genotype between two sub-
groups of patients receiving neoadjuvant
therapy based on their resectability after
chemoradiation. Patients with the AA
genotype receiving the entire course of
treatment (chemoradiation and surgery)
had an impressive overall survival of
55% over 5 years. Conversely, no differ-
ence in overall survival was associated
with any genotype for patients unresec-
table after neoadjuvant therapy.
Neoadjuvant + resection Neoadjuvant, no resection
1.0
0.8
0.6
0.4
0.2
0.0
0
p =
AA/CC
AA/AC
0.003
0.010
20 40 60
Cummulative Survival
80 100 120
AA
AC
CC
AA-censored
AC-censored
CC-censored
AA
AC
CC
AA-censored
AC-censored
CC-censored
Time to Event (Months)
0 5 10 15 20 25 30
442 HPB
HPB 2009, 11, 435–444 © 2009 International Hepato-Pancreato-Biliary Association
Page 8
results by selection of more beneficial treatments for each
patient and by avoiding the application of treatments that are
ineffective and toxic.
23
In an era where the cost–benefit ratio of
treatments will be increasingly scrutinized, selection of patients
with pancreatic cancer, who have a universally fatal disease, that
would be likely to benefit from otherwise highly toxic and
expensive treatment modalities may become critical. The present
study retrospectively identified a distinct subset of pancreatic
cancer patients in whom a specific genotype was associated with
a nearly threefold increase in survival over that which is typically
reported in the literature. Current and future studies, including
whole-genome analysis of this population, may yield clinically
relevant data that can influence prospective medical decision
making and choice of therapy.
Conflicts disclosure
The study was supported by grants from The Effie and Wofford Foundation
and The Don and Coletta McMillian Foundation.
Acknowledgement
The authors acknowledge all the people at the Human Genome
Sequencing Center as well as the Elkins Pancreas Center who made this
work possible.
References
1. Ries LAG, Melbert D, Krapcho M, Stinchcomb DG, Howlader N, Horner
MJ et al. (eds) (2008) SEER Cancer Statistics Review, 1975–2005.
Bethesda, MD: National Cancer Institute. Based on November 2007
SEER data submission, posted to the SEER web site. http://seer.
cancer.gov/csr/1975_2005 [Accessed 19 June 2009].
2. Yeo TP, Hruban RH, Leach SD, Wilentz RE, Sohn TA, Kern SE et al. (2002)
Pancreatic cancer. Curr Probl Cancer 26:176–275.
3. Huang P, Chubb S, Hertel LW, Grindey GB, Plunkett W. (1991) Action
of 2`,2`-difluorodeoxycytidine on DNA synthesis. Cancer Res 51:6110–
6117.
4. Robinson BW, Im MM, Ljungman M, Praz F, Shewach DS. (2003)
Enhanced radiosensitization with gemcitabine in mismatch repair-
deficient HCT116 cells. Cancer Res 63:6935–6941.
5. Sharma S, Doherty KM, Brosh RM. (2006) Mechanisms of RecQ heli-
cases in pathways of DNA metabolism and maintenance of genomic
stability. Biochem J 398:319–337.
6. Bohr VA. (2008) Rising from the RecQ-age: the role of human RecQ
helicases in genome maintenance. Trends Biochem Sci 33:609–620.
7. Sharma S, Sommers JA, Choudhary S, Faulkner JK, Cui S, Andreoli L
et al. (2005) Biochemical analysis of the DNA unwinding and strand
annealing activities catalyzed by human RECQ1. J Biol Chem 280:
28072–228084.
8. Muzzolini L, Beuron F, Patwardhan A, Popuri V, Cui S, Niccolini B et al.
Table 7 Survival comparison between the original group of patients treated with neoadjuvant therapy and two subgroups: patients
proceeding to resection after chemoradiation and patients for which resection was not feasible because of progression
Genotype AA AC CC
Group All patients Resected Unresected All patients Resected Unresected All patients Resected Unresected
1 year 82% 97% 60% 74% 89% 37.5% 68% 93% 36.4%
2 years 59% 80% 16% 45% 62% 6.3% 24% 36% 9.1%
5 years 39% 55% 14% 19% 19% 29%
Mean (month) 56.5 73.8 15.0 35.2 44.6 10.9 24.0 31.4 11.6
AA AC CC
1.0
Not resected
Resected
Not resected
censored
Resected censored
Not resected
Resected
Not resected
censored
Resected censored
Not resected
Resected
Not resected
censored
Resected censored
0.8
0.6
0.4
0.2
0.0
0 204060
Cummulative Survival
80 100 120 0 20 40 60 80 100 120 0 20 40 60 80
p = 0.00001
Time to Event (Months)
0.0120.00001
Figure 4 Comparison of the overall survival by genotype between patients receiving the full course of neoadjuvant therapy with resection and
patients unresectable after chemoradiation. For any genotype, the patients unable to receive a full course of treatment had a significantly
worsened survival.
HPB 443
HPB 2009, 11, 435–444 © 2009 International Hepato-Pancreato-Biliary Association
Page 9
(2007) Different quaternary structures of human RECQ1 are associated
with its dual enzymatic activity. PLoS Biol 5:e20.
9. Popuri V, Bachrati CZ, Muzzolini L, Mosedale G, Costantini S, Giacomini
E et al. (2008) The human RecQ helicases, BLM and RECQ1,
display distinct DNA substrate specificities. J Biol Chem 288:17766–
17776.
10. Bugreev DV, Brosh RM, Mazin AV. (2008) RECQ1 possesses DNA branch
migration activity. J Biol Chem 283:20231–20242.
11. Sharma S, Brosh RM. (2008) Unique and important consequences of
RECQ1 deficiency in mammalian cells. Cell Cycle 7:989–1000.
12. Doherty KM, Sharma S, Uzdilla LA, Wilson TM, Cui S, Vindigni A et al.
(2005) RECQ1 helicase interacts with human mismatch repair factors
that regulate genetic recombination. J Biol Chem 280:28085–28094.
13. Sharma S, Brosh RM. (2007) Human RECQ1 is a DNA damage respon-
sive protein required for genotoxic stress resistance and suppression
of sister chromatid exchanges. PLoS ONE 12:e1297.
14. Kawabe T, Tsuyama N, Kitao S, Nishikawa K, Shimamoto A, Shiratori M
et al. (2000) Differential regulation of human RecQ family helicases in cell
transformation and cell cycle. Oncogene 19:4764–4772.
15. Futami K, Kumagai E, Makino H, Goto H, Takagi M, Shimamoto A et al.
(2008) Induction of mitotic cell death in cancer cells by small interference
RNA suppressing the expression of RecQL1 helicase. Cancer Sci 99:
71–80.
16. Li D, Frazier M, Evans DB, Hess KR, Crane CH, Jiao L et al. (2006)
Single nucleotide polymorphisms of RecQ1, RAD54L, and ATM genes
are associated with reduced survival of pancreatic cancer. J Clin Oncol
24:1720–1728.
17. Li D, Liu H, Jiao L, Chang DZ, Beinart G, Wolff RA et al. (2006) Significant
effect of homologous recombination DNA repair gene polymorphisms on
pancreatic cancer survival. Cancer Res 66:3323–3330.
18. Ramensky V, Bork P, Sunyaev S. (2002) Human non-synonymous SNPs:
server and survey. Nucleic Acids Res 30:3892–3900.
19. Voidonikolas G, Kreml SS, Chen C, Fisher WE, Charles Brunicardi F,
Gibbs RA et al. (2008) Basic principles and technologies for deciphering
the genetic map of cancer. World J Surg 33:615–629.
20. Voidonikolas G, Gingras MC, Hodges S, McGuire AL, Chen C, Gibbs RA
et al. (2009) Developing a tissue resource to characterize the genome of
pancreatic cancer. World J Surg 33:723–731.
21. Zhang J, Wheeler DA, Yakub I, Wei S, Sood R, Rowe W et al. (2005)
SNPdetector: a software tool for sensitive and accurate SNP detection.
PLoS Comput Biol 1:e53.
22. Chen J, Li D, Killary AM, Sen S, Amos CL, Evans DB et al. (2009)
Polymorphisms of p16, p27, p73, and MDM2
modulate response and
survival of pancreatic cancer patients treated with preoperative chemo-
radiation. Ann Surg Oncol 16:431–439.
23. Brunicardi FC, Gibbs RA, Fisher W, Chen C. (2008) Overview of the
molecular surgeon symposium on personalized genomic medicine and
surgery. World J Surg 33:612–614.
444 HPB
HPB 2009, 11, 435–444 © 2009 International Hepato-Pancreato-Biliary Association
Page 10
  • Source
    • "These metabolites prevent DNA synthesis by incorporating into the C site of the elongated DNA strand, evading detection by DNA repair machinery, and directly binding to the DNA polymerase enzyme [63]. Such an interaction might be interrupted due to germline polymorphisms on genes such as RECQL, a DNA helicase, for which an SNP in the UTR region confers long-term survival to patients who received a full neoadjuvant treatment regimen [64]. To form the active metabolites, gemcitabine must be transported across the cellular membrane and phosphorylated. "
    [Show abstract] [Hide abstract] ABSTRACT: Pancreatic cancer has the worst five-year survival rate of all malignancies due to its aggressive progression and resistance to therapy. Current therapies are limited to gemcitabine-based chemotherapeutics, surgery, and radiation. The current trend toward "personalized genomic medicine" has the potential to improve the treatment options for pancreatic cancer. Gene identification and genetic alterations like single nucleotide polymorphisms and mutations will allow physicians to predict the efficacy and toxicity of drugs, which could help diagnose pancreatic cancer, guide neoadjuvant or adjuvant treatment, and evaluate patients' prognosis. This article reviews the multifaceted roles of genomics and pharmacogenomics in pancreatic cancer. Copyright © 2015. Published by Elsevier Ireland Ltd.
    Full-text · Article · Apr 2015 · Cancer letters
  • Source
    • "Individual variations in DNA repair capacity affects the clinical response to cytotoxic cancer therapy and overall survival of patients. Li et al. (2006a,b) and Cotton et al. (2009) found that polymorphism of A159C SNP at the 3 -untranslated region in RECQL1 genes significantly affect the overall survival of patients with resectable pancreatic cancer who were treated with adjuvant or neoadjuvant chemoradiation using gemcitabine as the radiosensor. Why the polymorphism at the non-coding region of mRNA affects the clinical progonosis of patients is not clear, but A159C SNP may function as a binding site for regulatory molecules that affect either splicing, translation efficiency, or stability of RECQL1 mRNA. "
    [Show abstract] [Hide abstract] ABSTRACT: RECQL1 and WRN helicases in the human RecQ helicase family participate in maintaining genome stability, DNA repair, replication, and recombination pathways in the cell cycle. They are expressed highly in rapidly proliferating cells and tumor cells, suggesting that they have important roles in the replication of a genome. Although mice deficient in these helicases are indistinguishable from wild-type mice, their embryonic fibroblasts are sensitive to DNA damage. In tumor cells, silencing the expression of RECQL1 or WRN helicase by RNA interference induces mitotic catastrophe that eventually kills tumor cells at the mitosis stage of the cell cycle. By contrast, the same gene silencing by cognate small RNA (siRNA) never kills normal cells, although cell growth is slightly delayed. These findings indicate that RECQL1 and WRN helicases are ideal molecular targets for cancer therapy. The molecular mechanisms underlying these events has been studied extensively, which may help development of anticancer drugs free from adverse effects by targeting DNA repair helicases RECQL1 and WRN. As expected, the anticancer activity of conventional genotoxic drugs is significantly augmented by combined treatment with RECQL1- or WRN-siRNAs that prevents DNA repair in cancer cells. In this review, we focus on studies that clarified the mechanisms that lead to the specific killing of cancer cells and introduce efforts to develop anticancer RecQ-siRNA drugs free from adverse effects.
    Full-text · Article · Jan 2014 · Frontiers in Genetics
  • [Show abstract] [Hide abstract] ABSTRACT: Pancreatic ductal adenocarcinoma (PDA) is a devastating disease that killed nearly 38,000 people in the United States this past year. Treatment of PDA typically includes surgery and/or chemotherapy with gemcitabine. No reliable biomarker exists for prognosis or response to chemotherapy. Two previously proposed prognostic markers, cyclooxygenase-2 (COX-2) and vascular endothelial growth factor (VEGF), are regulated by Hu protein antigen R (HuR), an mRNA binding protein that we have previously demonstrated to be a promising predictive marker of gemcitabine response. This study was designed to evaluate the clinical utility of HuR, COX-2, and VEGF as potential prognostic and predictive biomarkers for PDA. A tissue microarray of 53 PDA specimens from patients who underwent potentially curative pancreatic resection was analyzed. HuR, COX-2, and VEGF status were correlated with clinicopathologic and survival data. We also performed ribonucleoprotein immunoprecipitation assays using an HuR antibody to assess VEGF and COX-2 mRNA binding to HuR in pancreatic cancer cells. Roughly 50% (27/53) of patients had high cytoplasmic HuR expression. These patients had worse pathologic features as assessed by T staging (P = 0.005). Only cytoplasmic HuR status correlated with tumor T staging, whereas VEGF (P = 1.0) and COX-2 (P = 0.39) expression did not correlate with T staging. Additionally, HuR status was an unprecedented positive predictive marker for overall survival in patients treated with gemcitabine, pushing median survival over 45 months in the high cytoplasmic HuR expressing patient population compared with less than 23 months in the low cytoplasmic HuR expressing patient group (P = 0.033 for log-rank test and P = 0.04 in a Cox regression model) for the low versus high cytoplasmic HuR expressing group. We also validated that mRNA transcripts for both VEGF and the gemcitabine metabolizing enzyme, deoxycytidine kinase, are specifically bound by HuR in pancreatic cancer cells. HuR is a useful prognostic biomarker for PDA patients as indicated by its association with higher tumor T stage. Additionally, HuR status is a robust predictor of outcome for patients with resected PDA in the setting of adjuvant gemcitabine therapy. Finally, HuR binds to VEGF mRNA implying that HuR, in part, regulates VEGF expression in PDA. This study supports the notion that HuR status should be used by clinicians for the individualized treatment of PDA in the future.
    No preview · Article · Sep 2010 · Annals of surgery
Show more