+ model ARTICLE IN PRESS
Cancer Letters xx (2006) 1 3
www.elsevier.com/locate/canlet
Molecular overlap of fly circadian rhythms
and human pancreatic cancer
a, b a,c
*
Kay L. Pogue-Geile , James Lyons-Weiler , David C. Whitcomb
a
Division of Gastroenterology, Hepatology and Nutrition, Department of Medicine, Hillman Cancer Center,
University of Pittsburgh School of Medicine, Pittsburgh, PA 15213, USA
b
Center for Pathology and Oncology Informatics, University of Pittsburgh Cancer Institute and Department of Pathology,
University of Pittsburgh School of Medicine, Pittsburgh, PA 15232, USA
c
Department of Human Genetics, Graduate School of Public Health and Department of Cell Biology and Physiology,
University of Pittsburgh School of Medicine, Pittsburgh, PA 15261, USA
Received 29 April 2005; received in revised form 20 August 2005; accepted 20 November 2005
Abstract
Circumstantial evidence demonstrating a role for circadian rhythms in cancer has been presented but there is little direct
molecular evidence to support this idea in human cancer. Herein, we report a significant similarity between fly genes with strong
circadian rhythms and human genes under expressed in pancreatic cancer. The list of genes includes both circadian regulator genes,
such as period 1 and DEC1, and downstream effectors, such as ubiquitin specific protease 30. This observation may indicate that
the pancreas peripheral clock is disrupted in pancreatic cancer and are consistent with the recent proposals that circadian genes act
as tumor suppressors.
q 2006 Published by Elsevier Ireland Ltd.
Keywords: Circadian rhythms; Pancreatic cancer; Tumor suppressor; Period 1
The central regulator of circadian rhythms is the Disruption of circadian rhythms increases the
suprachiasmic nucleus (SCN) within the brain. The susceptibility to cancer in humans and animals.
clock regulates the biological rhythms of the organism Pinealectomy and exposure of rodents to constant
by entraining the peripheral tissues through neurohor- light accelerates the proliferation of breast epithelial
monal control mechanisms linked to factors from the stem-cells and increases the formation of mammary and
pineal gland [1]. Pinealectomy disrupts the circadian liver tumors [3,4]. Ablation of the suprachiasmic
nucleus in mice accelerates growth of implanted
rhythms of the pancreas, demonstrating the master
osteosarcoma and pancreatic adenocarcinoma tumor
clock controls some aspects of the pancreas. However,
cells [5]. Women whose circadian rhythms are
recent evidence indicates that the pancreas also
disrupted by working night shifts or by jet travel have
contains a peripheral clock, which enables it to
an increased risk for breast cancer and the risk increases
overwrite SCN control and respond to the environment
with the number of years or number of hours per week
independent of the SCN [2].
that were worked at night [6,7]. Maintenance of normal
circadian rhythms was found to be a significant
*
Corresponding author. Tel.: C1 412 359 4002; fax: C1 412 359
predictor of better survival for breast and colorectal
3239.
cancer patients [8,9]. Salivary cortisol levels, which are
E-mail addresses: [e-mail] (K.L. Pogue-Geile),
[e-mail] (K.L. Pogue-Geile). controlled by the circadian clock, were found to be a
0304-3835/$ - see front matter q 2006 Published by Elsevier Ireland Ltd.
doi:10.1016/j.canlet.2005.11.049
+ model ARTICLE IN PRESS
2 K.L. Pogue-Geile et al. / Cancer Letters xx (2006) 1 3
statistically significant predictor of survival time of additional potential transcriptional regulator of PER1,
breast cancer patients. Colon cancer patients who cAMP responsive element binding protein 3-like 1
maintained a regular pattern of rest and activity (CREB3L1) was also significantly underexpressed [10].
rhythms had a fivefold higher survival time than those The mouse Per1 promoter is activated by CREB proteins
who had chaotic circadian rhythms. via binding to cAMP responsive elements (CREs) [12].
Cancer of the pancreas represents one of the most The CREB3L1 protein contains CRE binding domains
aggressive of all cancers with death usually occurring making it a potential positive regulator of PER1.
within a few months of diagnosis. To determine which Underexpression of CREB3L1 is consistent with the
genes were important in the development and pro- underexpression of PER1 in pancreatic cancer. It
gression of pancreatic cancer, we previously developed remains to be determined whether these transcriptional
a custom pancreas gene enriched microarray termed regulators are responsible for the observed under-
Pittsburgh Pancreas Gene Enriched ARray-PittPEAR. expression of PER1 mRNA in pancreatic cancer tissues.
We identified 264 out of 5763 genes that were To determine if other circadian rhythm genes were
significantly differentially expressed in pancreatic underexpressed in pancreatic cancer, we searched
cancer [10]. PittPEAR for genes that were homologous to 22 fly
One of the five genes that were most strongly and head genes with strong circadian oscillating properties
consistently under expressed was the Drosophila [13]. We identified 30 human genes that were related to
homolog protein 1 (PER1) gene, one of eight known one of seven fly circadian genes that were present on
core circadian regulators. We verified the underexpres- our array (Table 1). The remaining 15 fly genes either
sion of PER1 with quantitative RT-PCR [10]. Interest- did not have a human homolog or the homolog was not
ingly, other transcriptional regulators of circadian genes present on PittPEAR.
including, PER1, were also significantly underex- Most of the PittPEAR human homologs of the fly
pressed. These genes included basic helix-loop-helix genes were close family members but were not the
domain containing, class B (BHLHB2 or Dec1) and specific homolog identified with NCBI s Homologene.
CREB3L1. BHLHB2 represses Bmal1/Clock-activation For example the human homolog for the fly cyp4d21
of PER1 transcription through the binding of E box gene is cytochrome P450, family 4, subfamily F,
elements in the PER1 promoter [11]. Binding of E box polypeptide 11 (CYP4F11). There were four different
elements is mediated by the basic helix-loop-helix cytochrome P450-related genes on our array but none
domain which is also present in other PER transcrip- of these were CYP4F11.
tional regulators, Clock (clk) and aryl hydrocarbon However, we noted that one of the four human
receptor nuclear translocator-like (BMAL1) [12]. An cytochrome P450-related genes, (platelet, cytochrome
Table 1
Genes similar to fly circadian genes are differentially expressed in pancreatic cancer
Fly genes with strong Specific human homolog No of fly related genes Significant differentially Cancer vs normal j5
circadian oscillation on PittPEAR expressed genes on PittPEAR expression value [14]
[13] [10]
Timeless Timeless homolog (Droso- 1 Timeless homolog (Drosophila) Not significant
phila)
Period None given on NCBI 1 Period homolog 1 (Drosophila) K13.132
period 1
Puckered Dual specificity phosphatase 5 Dual specificity phosphatase 5 K4.858
8
UDP-glycosyltrans- UDP glycosyltransferase 2 1 UDP glycosyltransferase 1 Not significant
ferase 35b family, polypeptide B15 family, polypeptide A9
Cyp4d21 Cytochrome P450, family 4, 4 Thromboxane a synthase 1 (pla- K7.775
subfamily F, polypeptide 11 telet, cytochrome P450, subfam-
ily V)
CG5798 Ubiquitin specific protease 8 13 Ubiquitin specific protease 30 K4.379
CG7149 Selenoprotein I 5 SelenoproteinX1 K5.472
Total genes 30 5
+ model ARTICLE IN PRESS
K.L. Pogue-Geile et al. / Cancer Letters xx (2006) 1 3 3
[4] S. van den Heiligenberg, P. Deprés-Brummer, H. Barbason,
P450, subfamily V) was significantly underexpressed in
B. Claustrat, M. Reynes, F. Lévi, The tumor promoting effect of
pancreatic cancer (Table 1). Furthermore, we identified
constant light exposure on diethylnitrosamine-induced hepato-
one significantly differentially expressed pancreatic
carcinogenesis in rats. Life Sci. 64 (1999) 2523 2534.
cancer gene for four other fly circadian genes. Analysis
[5] E. Filipski, V.M. King, X. Li, T.G. Granda, M.C. Mormont,
of the 5763 PittPEAR genes identified 179 genes as
X. Liu, et al., Host circadian clock as a control point in tumour
significantly under expressed in pancreatic cancer [10].
progression, J. Natl Cancer Inst. 94 (9) (2002) 690 697.
[6] S. Davis, D.K. Mirick, R.G. Stevens, Night shift work, light at
We observed a significant difference between the
night and risk of breast cancer, J. Natl Cancer Inst. 93 (2001)
observed number of underexpressed circadian pancrea-
1513 1515.
tic cancer genes (5) and the expected (0.93) based on
[7] V. Rafneeon, H. Tulinius, J.G. Jonasson, Hrafnkelsson, Risk
chi-square analysis (P!0.001). These findings support
of breast cancer in female flight attendants: a population-
the hypothesis that circadian rhythm-associated genes
based study (Iceland), J. Cancer Causes Control 12 (2001)
are preferentially underexpressed in pancreatic cancer
95 101.
[8] S.E. Sephton, R.M. Sapolsky, H.C. Kraemer, D. Spiegel,
and support the hypothesis that circadian genes may act
Diurnal cortisol rhythm as a predictor of breast cancer survival,
as tumor suppressors. Interestingly, this list of genes
J. Natl Cancer Inst. 92 (2000) 994 1000.
includes both clock regulators such as PER1 and
[9] M.C. Mormont, J. Waterhouse, P. Bleuzen, S. Giacchetti,
downstream effectors of circadian rhythms such as
A. Jami, A. Bogdan, et al., Clin. Marked 24-h rest/activity
ubiquitin specific proteases, selenoproteins and glyco-
rhythms are associated with better quality of life, better
syltransferases (Table 1).
response, and longer survival in patients with metastatic
We have found that several circadian related genes colorectal cancer and good performance status, Cancer Res. 6
(2000) 3038 3045.
are underexpressed in pancreatic cancer, which may
[10] K.L. Pogue-Geile, J.A. Mackey, R.D. George, P.G. Wood,
indicate that pancreatic tumors have altered circadian
K.K.W. Lee, A.J. Moser, et al., A new microarray, enriched in
rhythms. Most late stage or fast growing tumors show
pancreas and pancreatic cancer cDNAs to identify genes
altered circadian rhythms so that cell proliferation of
relevant to pancreatic cancer, Cancer Genomics Proteomics 1
the tumors occurs more than once per day while normal
(2004) 371 386.
cells maintain normal circadian rhythms in which cell [11] S. Honma, T. Kawamoto, Y. Takagi, K. Fujimoto, F. Sato,
M. Noshiro, et al., Dec1 and Dec2 are regulators of the
division occurs once a day. Chronotherapy exploits
mammalian molecular clock, Nature 419 (6909) (2002) 841
these differences in cell cycle by delivering chemother-
844.
apeutic agents when tumor cells are dividing and
[12] Z. Travnickova-Bendova, N. Cermakian, S.M. Reppert,
normal cells are resting, thus maximizing the killing of
P. Sassone-Corsi, Bimodal regulation of mPeriod promoters
tumor cells and minimizing the killing of normal cells.
by CREB-dependent signaling and CLOCK/BMAL1 activity,
Several studies have reported that chronotherapy has Proc. Natl Acad. Sci. USA 99 (11) (2002) 7728 7733.
[13] Y. Lin, M. Han, B. Shimada, L. Wang, T.M. Gibler,
improved conventional therapy [15,16]. If pancreatic
A. Amarakone, et al., Influence of the period-dependent
tumor cells are asynchronous with normal pancreatic
circadian clock on diurnal, circadian, and aperiodic gene
cells, chronotherapy may be a reasonable treatment
expression in drosophila melanogaster, Proc. Natl Acad. Sci.
option but further work is required to test this
USA 99 (2002) 9562 9567.
possibility.
[14] S. Patel, J. Lyons-Weiler, caGEDA: a web application for the
integrated analysis of global gene expression patterns in cancer,
Appl. Bioinformatics 3 (2004) 49 62, http://bioinformatics.
References
upmc.edu/GE2/GEDA.html.
[15] P.A. Wood, W.J.M. Hrushesky, Critical review eukaryotic
[1] L. Fu, C.C. Lee, The circadian clock: pacemaker and tumour
circadian rhythms and cancer chemotherapy. Summary of
suppressor, Nat. Rev. Cancer 3 (2003) 350 361.
chronotherapy for human cancer patients, Gene Expr. 6 1996
[2] E. Muhlbauer, S. Wolgast, U. Finckh, D. Peschke, E. Peschke,
pp. 299 343.
Indication of circadian oscillations in the rat pancreas, Fed. Eur.
[16] G.C. Rivard, C. Infante-Rivard, M.F. Dress, J.M. Leclerc,
Biochem. Soc. Lett. 564 (1 2) (2004) 91 96.
J. Champagne, Circadian time-dependent response of childhood
[3] P.N. Shah, M.C. Mhatre, L.S. Kothari, Effect of melatonin on
lymphoblastic leukemia to chemotherapy: a long-term follow-
mammary carcinogenesis in intact and pinealectomized rats in
varying photoperiods. Cancer Res. 44 (1984) 3403 3407. up study of survival, Chronobiol. Int. 10 (1993) 201 204.