Download Supplementary Information (doc 108K)

1
Supplementary Materials and Methods
Mouse models relevant for PDAC pathogenesis
Supplementary Figure 1 allocates the different models throughout PDAC pathogenesis.
(I) Caerulein pancreatitis: Pancreatitis preconditions for PDAC development and recapitulates
its early changes in cell differentiation: acinar to ductal cell metaplasia. Caerulein-treated
mice1, 2 develop either acute or chronic pancreatitis depending on the dose and frequency of
the injections; acinar to ductal metaplasia is transient or persistent, respectively. As such, the
acute pancreatitis model at 48h resembles the chronic pancreatitis one. Both models allow
studying pancreatitis without subsequent development of pancreatic cancer. Acute pancreatitis
was induced in adult CD1 mice by 8 one-hourly, intraperitoneal injections of sulphated
caerulein on 2 consecutive days (2µg caerulein/200µl injection volume, Sigma, Munich,
Germany)3. Chronic damage was induced in adult C57BL/6 mice by administering 5 hourly
injections (50 µg/kg; Sigma) 3 days/week over 4 weeks1.
(II) Acinar cell in-vitro model: Murine acinar cells grown in suspension have been shown to
dedifferentiate and to acquire an embryonic phenotype, consistent with acinar to ductal
metaplasia and copying the mechanisms activated in experimental pancreatitis models 4. Acinar
cell fractions from CD1 mice were obtained as described previously5. Cells were cultured in
suspension in RPMI 1640 glutamax medium (Gibco, Karlsruhe, Germany) supplemented with
10% fetal bovine serum, penicillin, streptomycin, geniticin sulfate and soybean trypsin
inhibitor4. Inhibition of Erk1/2 activation was obtained by supplying UO126 (10µM, Sigma) to
the medium.
(III) KPC mouse model: it reproduces faithfully the progressive histological changes occurring in
human PDAC through conditional activation of the K-Ras oncogene and a mutant p53. PDX-1Cre;Trp53 R172H;LSL-K-RasG12D mice were generated as explained previously6. PDX-1-Cre targets
all pancreatic precursors for mutant expression of p53 and K-Ras. Human PDAC development is
recapitulated with invasive and metastatic carcinoma from 9 months of age.
2
Time points for analysis in the pancreatitis models have been carefully chosen to ensure a
‘peak’ change as judged by histological analysis, based on published observations as well as our
own results. For induction of chronic pancreatitis, mice received intra-peritoneal injections of
caerulein (MP Biomedicals,Solon, OH) five times a day, twice a week for 10 weeks 7. Analysis
was done at the 10 week time point.
Acute pancreatitis was induced by a series of caerulein (C9026, Sigma-Aldrich, St. Louis, MO,
USA) injections on 2 consecutive days3. Pancreata were collected after 48 hours and 8 days.
For the acute pancreatitis, temporal expression changes have been analysed in detail by real
time RT-PCR analysis, as well as immunostainings (Supplementary Figure 8). At 48 hours, the
lumen of the acini is dilated and tubular structures form with mild inflammatory infiltrates. This
time point is characterized by a decreased expression of acinar specific transcription factors
(Ptf1a and Rbpjl) and pancreatic digestive enzymes (Elastase 1 (Ela1) and Carboxypeptidase A1
(CpA1)). A transient induction of ductal markers as Keratin 19 (Krt19) is noted. The pancreas has
almost completely regenerated at 8 days of analysis, where changes in gene expression are
much less pronounced and acini have become negative for the marker Keratin19.
Initial and final states are as follows: (i) For the in vitro model: isolated acinar cell fraction and
cultured cell fraction, (ii) For caerulein induced acute pancreatitis: normal mouse pancreas and
48 hours after start of caerulein injections, corresponding to the peak in pancreatitis. (iii) For
caerulein induced chronic pancreatitis: normal mouse pancreas and 10 weeks after start of
caerulein injections.
3
DNA and RNA isolation from human and mouse pancreatic tissues
High-molecular weight DNA was isolated from fresh-frozen normal, CP and PDAC tissue samples
using the QIAamp DNA Mini kit (Qiagen, Hilden, Germany). DNA from formalin-fixed paraffinembedded (FFPE) tissue was extracted using the QIAamp DNA FFPE tissue kit (Qiagen). PanIN
lesions from 5 patients were macrodissected with a needle from deparafinised FFPE sections
and dissolved in 90 µl lysis buffer. After overnight incubation with 10 µl proteinase K (Qiagen),
bisulfite conversion reaction was directly performed with the digestion mixture (cf.
Quantitative DNA methylation analysis). mRNA from fresh frozen pancreatic tissues was
purified using the MagNA Pure LC System (Roche Diagnostics, Mannheim, Germany). The All
prep DNA/RNA kit (Qiagen) was used to isolate DNA and RNA from the mouse models. The All
prep DNA/RNA kit (Qiagen) and the GenElute Mammalian Total RNA miniprep kit (Sigma) were
used to isolate DNA and RNA from the acinar cell in-vitro model, respectively.
Cell lines
Human pancreatic cancer cell lines (Primary tumor cell lines: Colo-357, PANC-1; Lymph-node
metastasis cell line: MIA PaCa-2) were cultured in RPMI medium supplemented with 10% fetal
calf serum and 1% penicillin/streptomycin at 37°C in a humidified 5% CO2 atmosphere.
Autenthicity of the cell lines was checked by the Deutsche Sammlung von Mikroorganismen und
Zellkulturen GmbH (Braunschweig, Germany). Normal pancreatic ductal cell line (HPDE6-E6E7)8
was kindly provided by Dr. Ming-Sound Tsao (Ontario Cancer Institute, Toronto, Canada) and
was grown in Keratinocyte-SFM medium supplemented with 0.2 ng/ml Epidermal Growth
Factor and 30 µg/ml Bovine Pituitary Extract (Life Technologies, Darmstadt, Germany).
5-aza-2’-deoxycytidine (5-aza-dC) treatment in cell lines
Human pancreatic cancer cell lines (Colo-357, PANC-1, MIA PaCa-2) and the normal ductal cells
HPDE6-E6E7 were incubated for 48 hours with 0.5 µM 5-aza-2’-deoxycytidine (Sigma), a
cytosine nucleoside analog which is incorporated into DNA during replication and forms stable
complexes with DNA methyltransferases, resulting in a passive global demethylation of the
4
DNA. Medium containing fresh 5-aza-dC was renewed every 24 hours. Genomic DNA and RNA
were isolated using the the AllPrep DNA/RNA Mini kit (Qiagen).
CpG island microarray analysis
The 244K human CpG island microarrays (Agilent) were composed of about 199.000 probes
covering about 27.800 CpG islands. Three selection strategies were followed to identify
hypermethylated CpG islands:
(i) TOP 10.000: The top 10.000 probes (5% of all) with the highest log-intensity ratios (M-values)
were chosen. To minimize false-positives, only CpG islands with at least 2 positive neighboring
probes in the majority of the tumor arrays but in none of the normal or CP arrays were
considered.
(ii) M-threshold: An analysis based on statistical testing was done following similar principles. In
order to define the methylation status of a probe in the tested samples versus the reference, a
methylation threshold was calculated by fitting a log-normal mixture model with 2 components
(corresponding to hypermethylated and not hypermethylated compared to the reference pool).
Logistic regression models for individual oligonucleotides and corresponding likelihood ratio
tests were used to rank the oligonucleotides by their statistical significance. The CpG islands
were then ranked based on the best-scoring triplets of neighbouring oligonucleotides within
each island (with smallest mean of p-values). To test how strongly the CpG islands correlated
with tumor versus normal or CP status based on their methylation score, p-values were
estimated by sampling from the approximate null distributions of the CpG island scores (B=500
samples each). Candidate genes were selected within the islands having a multiple-testing
adjusted p-value lower than 0.05 (FDR control by Benjamini and Hochberg) and for which no
probe was methylated in any of the normal or CP samples. The statistical analysis was done
using the R statistical environment v2.11.
(iii) Matched-pair analysis: An additional Excel-based matched-pair analysis was performed for
3 array-pairs corresponding to matched tumor and adjacent normal tissues from the same
patients.
5
The list resulting from the union of the hypermethylated CpG islands obtained with these 3
strategies was considered for the next selection steps.
Candidate gene selection for MassARRAY analysis
Published data were used to further classify the candidate islands and to select the most
promising ones for deeper analysis:
(a) Only the islands located in the promoter region of a gene and known to be unmethylated in
normal pancreatic tissues (http://www.ensembl.org) were considered for further analysis.
(b) CpG islands associated with a gene known to be expressed in normal pancreas
(http://www.proteinatlas.org, http://biogps.gnf.org) were favored as well as the ones showing
down-regulation on the mRNA or protein level (http://www.pancreasexpression.org, 9, 10).
(c) Genes located in regions potentially associated with Loss of Heterozygosity in PDAC
(http://www.intogen.org, 11-14) or being genetically altered in PDAC15 were considered as being
potential tumor suppressor genes, according to the Knudson two-hit hypothesis16.
(d) Hints towards a role of the gene in cancer were also considered.
The list of selected CpG islands is given in Supplementary Table 2.
mRNA expression analysis
cDNA was synthesized from 500 ng RNA using Superscript III reverse transcriptase (Invitrogen)
and oligo dT primers or random hexamers. Quantitative mRNA expression analysis (q-RT-PCR)
was performed in duplicates or triplicates on a LightCycler 480 (Roche Diagnostics) using the
Absolute qPCR SYBR Green Mix (Thermo Fisher Scientific, Ulm, Germany) as described
(http://www.roche-applied-science.com). Primer sequences and real-time PCR conditions are
given in Supplementary table 5. Primers were designed to cover all protein-coding isoforms of
the genes of interest. Mean normalized expression ratios were determined for each sample
using ACTB, GAPDH, and HPRT1 as reference genes for the human samples and ACTB and
GAPDH for the mouse samples. For the expression screen of 12 genes in the human samples
(Figure 1C), only HPRT1 and TBP were used as reference genes due to technical reasons.
6
Immunohistochemistry on Tissue Microarrays
Immunohistochemical analyses were performed as previously described 17. Briefly, tissue
microarrays were prepared from formalin-fixed, paraffin-embedded tissues, carrying a
duplicate of 205 PDAC, 27 CP and 14 normal pancreatic tissues. Clinical parameters, tumor
staging and gender of the patients are given in Supplementary Table 3. The
immunohistochemical stains were conducted with primary antibodies directed against ERK1/2
(1:100; Cell Signaling n°4696, Life Technologies), pERK1/2 (1:100; Abcam n°43765, Cambridge,
UK), PAK1 (1:100; Abcam n°58263), WNK2 (1:80; Abcam n°28852) and pPAK1 (1:50; Abcam
n°40795), using the avidin-biotin-complex-method. Antigen retrieval was performed by heat
pre-treatment in a citrate buffer (pH6). As previously reported17 for semi-quantitative analyses
of protein expression, staining intensities were scored as negative (score: 0), weakly positive
(score: 1), moderately positive (score: 2), or strongly positive (score: 3) by two observers (F.B.
and C.D.). Furthermore, the percentage of immunoreactive ductal cells was quantified. The
product of the scores of staining intensity and percentage of immunoreactive cells was
calculated based on the Remmele and Stegner scoring system as described in 18. To control the
specificity of the antibodies, the reactions were also conducted omitting the primary antibody
and immunoblotting of the antibodies were obtained.
Immunoblotting
Proteins were extracted from PANC1 cell lines overexpressing WNK2. Cells were harvested by
trypsinization, washed with PBS and lysed for 30 min on ice with a custom-made lysis buffer
(0.062 M Trizma-base pH 6.8 (Sigma), 2% SDS (Sigma), 10% Glycerol (Carl Roth, Karlsruhe,
Germany), 1% DTT (Sigma), 1% Benzonase (Novagen, Merck, Darmstadt, Germany), 2 Mini
protease inhibitor cocktail tablets (Roche), H2O). Lysis was stopped by heating at 95°C for 7 min.
After 5 min centrifugation, protein amounts present in the supernatant were quantified with
the Bicinchoninic acid (BCA) assay (Thermo Fisher Scientific). Sodium dodecyl sulfate
polyacrylamide gel electrophoresis (SDS-PAGE) was performed, using a 7.5% polyacrylamide gel
and by loading 10-30 µg proteins mixed with 5x loading dye (1 M Trizma-base, 10% SDS, 5% ßMercaptoethanol (Sigma), 50% Glycerol, Bromophenolblue (Sigma)). Blotting was carried out
7
on a polyvinylidene fluoride (PVDF) membrane (Invitrogen) using the Mini Trans Blot Cell
System (Bio-Rad Laboratories, Munich, Germany) with borate buffer. After blotting, blocking of
the membrane was done by 1 h incubation in 5% milk in TBS containing 0.1% Tween 20 (TBST,
Sigma). Incubation with the primary antibody diluted in 5% milk in TBST was then performed for
1 h or overnight at 4°C. The membrane was washed 4 times in TBST and was then incubated
with the secondary antibody for 1 h at room temperature. The following primary antibodies
were used: Rabbit polyclonal anti-WNK2 (dilution 1:500; Abcam n°28852), mouse monoclonal
anti-V5 tag antibody (1:5000; Life Technologies n°R960-25), mouse monoclonal anti-ERK1/2
antibody (1:2000; Cell Signaling n°4696), mouse monoclonal anti-pERK1/2 antibody (dilution
1:5000; Abcam n°43765), mouse monoclonal anti-PAK1 antibody (1:500; Abcam n°58263) and
rabbit monoclonal anti-pPAK1 antibody (1:10000; Abcam n°40795). Secondary antibodies were
purchased from Santa Cruz Biotechnologies. After 4 washing steps in TBST, the membrane was
incubated with 6 ml of Western Lightning Chemiluminescense Reagent Plus (Perkin-Elmer,
Waltham, USA) for 1 min and exposed to a Hyperfilm (Amersham Biosciences, Little Chalfont,
UK) which was subsequently developed on an AGFA Classic E.O.S developer (Siemens,
Erlangen). Mouse monoclonal Actin-B antibody was used as a loading control (1:20000; Santa
Cruz Biotechnologies n°47778, Heidelberg, Germany) and was detected on the same blot, after
stripping of the membrane.
BrdU incorporation assay and flow cytometry analysis
pcDNA3.1/V5–His–TOPO vectors containing WNK2 as well as an empty vector control19 were
transfected in equal molar amounts to PANC-1 cells with the TurboFect reagent (Thermo Fisher
Scientific). Stably overexpressing cells were selected with Geneticin (G418, GIBCO) at 1.5 mg/ml
for 3 weeks. Cell proliferation was determined by BrdU incorporation for 6h followed by flow
cytometry analysis, using the APC BrdU Flow kit (BD Pharmingen, San Jose, USA). BrdU assay
was performed in triplicates. FlowJo® software (Tree Star, Inc., Ashland, USA) was used to
identify the BrdU positive and negative populations. BrdU positive populations represent the
cells which went through the S-phase of the cell cycle.
8
Colony Formation Assay
Stable cell lines (PANC1 cells, overexpressing WNK2 or empty vector control) were seeded at
equal densities using three independent measurements with a CASY cell counter. These were
then grown in six well tissue culture plates for 10 days. Cells were fixed with 6% glutaraldehyde
and stained with 0.5% crystal violet according to Franken et al.20. Colonies containing >50 cells
were then counted, the different cell lines plating efficiencies calculated, and then expressed as
a percentage of empty vector controls.
Sulforhodamine B (SRB) Assay
Cell lines (PANC1 cells, overexpressing WNK2 or empty vector control) were first tested to find
a standard, optimal seeding density for each time point in 96 well plates over multiple day time
points. These densities were then used to seed cells at 4 day (4,000 cells/well) and 6 day (1,000
cells/well) time points. A standard curve for each cell line at each time point was also seeded on
the plate. Following the growth period, cells were fixed with 10% trichloroacetic acid, stained
with 0.5% sulforhodamine B (SRB), washed with 1% acetic acid and allowed to dry. 10mM Tris
base was added to the wells and the absorbance measured at 515nm. The standard curves
were then used to normalize the absorbance values and the proliferation differences between
cell lines were determined21.
9
References for Supplementary Materials and Methods
1. Strobel O, Dor Y, Stirman A, Trainor A, Fernández-del Castillo C, Warshaw AL, Thayer SP. Beta
cell transdifferentiation does not contribute to preneoplastic/metaplastic ductal lesions of the
pancreas by genetic lineage tracing in vivo. Proc. Natl. Acad. Sci. 2007;104:4419-24.
2. Niederau C, Ferrell L, Grendell J. Caerulein-induced acute necrotizing pancreatitis in mice:
protective effects of proglumide, benzotript, and secretin. Gastroenterology 1985;88:1192-204.
3. Jensen JN, Cameron E, Garay MVR, Starkey TW, Gianani R, Jensen J. Recapitulation of
elements of embryonic development in adult mouse pancreatic regeneration. Gastroenterology
2005;128:728-41.
4. Pinho AV, Rooman I, Reichert M, De Medts N, Bouwens L, Rustgi AK, Real FX. Adult pancreatic
acinar cells dedifferentiate to an embryonic progenitor phenotype with concomitant activation
of a senescence programme that is present in chronic pancreatitis. Gut 2011;60:958-66.
5. Means AL, Meszoely IM, Suzuki K, Miyamoto Y, Rustgi AK, Coffey RJ, Wright CVE, Stoffers DA,
Leach SD. Pancreatic epithelial plasticity mediated by acinar cell transdifferentiation and
generation of nestin-positive intermediates. Development 2005;132:3767-76.
6. Hingorani SR, Petricoin Iii EF, Maitra A, Rajapakse V, King C, Jacobetz MA, Ross S, Conrads TP,
Veenstra TD, Hitt BA, Kawaguchi Y, Johann D, et al. Preinvasive and invasive ductal pancreatic
cancer and its early detection in the mouse. Cancer Cell 2003;4:437-50.
7. Neuschwander-Tetri BA, Talkad V, Otis Stephen F. Induced thrombospondin expression in the
mouse pancreas during pancreatic injury. The International Journal of Biochemistry & Cell
Biology 2006;38:102-09.
8. Ouyang H, Mou L, Luk C, Liu N, Karaskova J, Squire J, Tsao M-S. Immortal Human Pancreatic
Duct Epithelial Cell Lines with Near Normal Genotype and Phenotype. Am J Pathol.
2000;157:1623-31.
9. Chelala C, Hahn SA, Whiteman HJ, Barry S, Hariharan D, Radon TP, Lemoine NR, CrnogoracJurcevic T. Pancreatic Expression database: a generic model for the organization, integration and
mining of complex cancer datasets. BMC Genomics 2007;8:439.
10. Cutts RJ, Gadaleta E, Hahn SA, Crnogorac-Jurcevic T, Lemoine NR, Chelala C. The Pancreatic
Expression database: 2011 update. Nucleic Acids Research 2010.
11. Gundem G, Perez-Llamas C, Jene-Sanz A, Kedzierska A, Islam A, Deu-Pons J, Furney S, LopezBigas N. IntOGen: Integration and data-mining of multidimensional oncogenomic data. Nature
Methods 2010.
12. Karhu R, Mahlamäki E, Kallioniemi A. Pancreatic adenocarcinoma—Genetic portrait from
chromosomes to microarrays. Genes, Chromosomes and Cancer 2006;45:721-30.
13. Kowalski J, Morsberger L, Blackford A, Hawkins A, Yeo C, Hruban R, Griffin C. Chromosomal
abnormalities of adenocarcinoma of the pancreas: identifying early and late changes. Cancer
Genetics and Cytogenetics 2007;178:26-35.
14. Lin L-J, Asaoka Y, Tada M, Sanada M, Nannya Y, Tanaka Y, Tateishi K, Ohta M, Seto M,
Sasahira N, Tada M, Kawabe T, et al. Integrated Analysis of Copy Number Alterations and Loss of
Heterozygosity in Human Pancreatic Cancer Using a High-Resolution, Single Nucleotide
Polymorphism Array. Oncology 2008;75:102-12.
15. Jones S, Zhang X, Parsons DW, Lin JCH, Leary RJ, Angenendt P, Mankoo P, Carter H,
Kamiyama H, Jimeno A, Hong SM, Fu B, et al. Core Signaling Pathways in Human Pancreatic
Cancers Revealed by Global Genomic Analyses. Science 2008;321:1801-06.
16. Knudson AG. Antioncogenes and human cancer. Proc. Natl. Acad. Sci. 1993;90:10914-21.
10
17. Ceyhan GO, Demir IE, Rauch U, Bergmann F, Muller MW, Buchler MW, Friess H, Schafer K-H.
Pancreatic Neuropathy Results in [ldquo]Neural Remodeling[rdquo] and Altered Pancreatic
Innervation in Chronic Pancreatitis and Pancreatic Cancer. Am J Gastroenterol 2009;104:255565.
18. Goeppert B, Schmezer P, Dutruel C, Oakes C, Renner M, Breinig M, Warth A, Vogel MN,
Mittelbronn M, Mehrabi A, Gdynia G, Penzel R, et al. Down-regulation of tumor suppressor a
kinase anchor protein 12 in human hepatocarcinogenesis by epigenetic mechanisms.
Hepatology 2010;52:2023-33.
19. Hong C, Moorefield KS, Jun P, Aldape KD, Kharbanda S, Phillips HS, Costello JF. Epigenome
scans and cancer genome sequencing converge on WNK2, a kinase-independent suppressor of
cell growth. Proceedings of the National Academy of Sciences 2007;104:10974-79.
20. Franken N, Rodermond H, Stap J, Haveman J, van Bree C. Clonogenic assay of cells in vitro.
Nat Protoc 2006;1:2315-9.
21. Vichai V, Kirtikara K. Sulforhodamine B colorimetric assay for cytotoxicity screening. Nat
Protoc 2006;1:1112-6.