Carcinogenesis, Vol. 23, No. 12, 2063-2080,
December 2002
© 2002 Oxford University Press
CARCINOGENESIS |
Development and molecular characterization of HCT-116 cell lines resistant to the tumor promoter and multiple stress-inducer, deoxycholate
1 Departments of Microbiology and Immunology,
2 Internal Medicine, College of Medicine, and
3 Arizona Cancer Center, University of Arizona, 85724–5049 and
4 Tucson Veterans Affairs Medical Center, Section of Hematology/Oncology, Tucson, AZ 85723, USA
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Evidence from live cell bioassays shows that the flat mucosa from patients with colon cancer exhibits resistance to bile salt-induced apoptosis. Three independent cell lines derived from the colonic epithelial cell line HCT-116 were selected for resistance to bile salt-induced apoptosis. These cell lines were developed as tissue culture models of apoptosis resistance. Selection was carried out for resistance to apoptosis induced by sodium deoxycholate (NaDOC), the bile salt found in highest concentrations in human fecal water. Cultures of HCT-116 cells were serially passaged in the presence of increasing concentrations of NaDOC. The resulting apoptosis resistant cells were able to grow at concentrations of NaDOC (0.5 mM) that cause apoptosis in a few hours in unselected HCT-116 cells. These cells were then analyzed for changes in gene expression. Observations from cDNA microarray, 2-D gel electrophoresis/MALDI-mass spectroscopy, and confocal microscopy of immunofluorescently stained preparations indicated underexpression or overexpression of numerous genes at either the protein or mRNA level. Genes that may play a role in apoptosis and early stage carcinogenesis have been identified as upregulated in these cell lines, including Grp78, Bcl-2, NF-
B(p50), NF-B(p65), thioredoxin peroxidase (peroxiredoxin) 2, peroxiredoxin 4, maspin, guanylate cyclase activating protein-1, PKC
, EGFR, Ras family members, PKA, PI(4,5)K, TRAF2 and BIRC1 (IAP protein). Under-expressed mRNAs included BNIP3, caspase-6, caspase-3 and serine protease 11. NF-B was constitutively activated in all three resistant cell lines, and was responsible, in part, for the observed apoptosis resistance, determined using antisense oligonucleotide strategies. Molecular and cellular analyses of these resistant cell lines has suggested potential mechanisms by which apoptosis resistance may develop in the colonic epithelium in response to high concentrations of hydrophobic bile acids that are associated with a Western-style diet. These analyses provide the rationale for the development of hypothesis-driven intermediate biomarkers to assess colon cancer risk on an individual basis.
Abbreviations: BH4, tetrahydrobiopterin; BNIP3, Bcl2/adenovirus EIB 19 kD-interacting protein 3; CAMK2D, calcium/calmodulin-dependent protein kinase II delta; DHAP, dihydroxyacetone phosphate; DOC, deoxycholate; EGFR, epidermal growth factor receptor; ER, endoplasmic reticulum; GAP, glyceraldehyde 3-phosphate; GC, guanylate cyclase; Grp78, 78 kD-glucose-regulated protein; IAP, inhibitor of apoptosis protein; IEF, isoelectric focusing; IKK-ß, IB-kinase-ß; IP3, inositol triphosphate; MALDI-MS, matrix assisted laser desorption ionization mass spectroscopy; MAPK, mitogen-activated protein kinase; MEK, MAPK kinase; MEKK, MEK kinase; NIK, NF-B-inducing kinase; NLS, nuclear localization signal; NaDOC, sodium deoxycholate; NO, nitric oxide; NOS2, inducible NO synthase; ONOO-, peroxynitrite; PDTC, pyrrolidine dithiocarbamate; PKC, protein kinase C-zeta; PKG, cGMP-activated protein kinase; PN-1, protease nexin-1; QDPR, quinoid dihydropteridine reductase; SERPIN, serine protease inhibitor; TEM, transmission electron microscopy; TPI, triose phosphate isomerase; Trx, thioredoxin; TPx, Trx peroxidase (peroxiredoxin); TR, Trx reductase; TRAF, tumor necrosis factor receptor-associated factor; TTFA, thenoyl trifluoroacetone.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Bile acids are natural detergents synthesized in the liver and stored in the gall bladder. High levels of certain bile acids, however, are known to promote G.I. cancer (1–4), including colon cancer (1,5,6). The bile acids that promote colon cancer are secondary bile acids that have been deconjugated and dehydroxylated, resulting in an increase in hydrophobicity. Conversion of primary bile salts to their secondary bile acid counterparts is catalyzed by the bacterial enzyme 7
-dehydroxylase which removes a hydroxyl group from the 7 position of the steroid nucleus. Cholic acid and chenodeoxycholic acid, respectively, are converted to deoxycholic acid (DOC) and lithocholic acid, the secondary bile acids found in greatest concentration in human fecal water. These secondary bile acids have been implicated in animal models of colon carcinogenesis as promoters of colon cancer (6), although the exact mechanism of tumor promotion by bile acids is unclear. The bile acid present in the highest concentration in the human colon and feces is DOC, and it is found at concentrations up to 0.78 mM in individuals consuming a high fat diet (7). Our laboratory previously reported that the sodium salt of DOC (NaDOC), at high physiological concentrations, induces apoptosis in colonic epithelial cells of the flat mucosa from normal subjects (8–11). We also found that resistance to NaDOC-induced apoptosis occurs in the normal appearing flat mucosa of patients with colon cancer (9–11). Decreased ability to undergo apoptosis is a risk factor in colon carcinogenesis (12–17) since apoptosis resistance creates a permissive environment for increasing genomic instability (e.g. aneuploidy, point mutations, loss of heterozygosity)(18–20), which can result in cancer. It has been calculated that sporadically arising polyps in the colon survive with nearly 11 000 genomic alterations per cell (21). It is assumed that some of these genomic alterations affect the expression of apoptosis and/or survival genes which may explain the further increase in apoptosis resistance during polyp development (22–24). We have proposed that excessive, frequent exposure of an individual's colonic epithelial cells to cytotoxic bile acids could lead to the selection of an apoptosis resistant population of cells, making that individual at risk for colon cancer (8–10). In an apoptosis resistant cell population, replication of DNA with unrepaired damage can lead to mutations, some of which might contribute to carcinogenesis. The experiments described here were carried out to gain insight into (i) the mechanisms involved in the cells' response to repeated exposure to a hydrophobic bile acid, and (ii), more importantly, the potential mechanisms by which apoptosis resistance can arise. The identification of candidate genes that contribute to apoptosis resistance in vitro may assist in the development of practical biomarkers for early detection of colon cancer risk in situ. Selection was carried out for resistance to induction of apoptosis by NaDOC in three separate serially passaged cultures of HCT-116, a colonic epithelial cell line that was very sensitive to apoptosis at the start of the selection experiments. The resulting apoptosis resistant cell populations were then analyzed for changes in gene expression using a variety of molecular and cellular techniques. Alterations in specific survival and apoptotic pathways were identified, and novel candidates for biomarker development and cancer prevention were identified.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Cell lines, media and chemicals
HCT-116, a colon adenocarcinoma cell line [American Type Culture Collection (ATCC), Bethesda, MD; ATCC # CCL 247] and apoptosis resistant HCT-116 cell lines were maintained in DMEM supplemented with 10% fetal calf serum (Omega Scientific, Tarzana, CA), 1% MEM non-essential amino acids, 100 µg/ml streptomycin, 100 U/ml penicillin, and 3.44 mg/ml L-glutamine. Unless otherwise indicated, media components were from Gibco BRL Life Technologies (Grand Island, NY) and chemicals were from Sigma Chemical Co. (St Louis, MO).
Development of resistant cell lines
Early passage HCT-116 cells were split and seeded into four flasks. The cells in flask A were serially passaged as an untreated control along with the cells in the other three flasks. After 48 h, the other three flasks (B, C and D) were treated with sodium deoxycholate (NaDOC). B cells were initially treated with 0.02 mM NaDOC, C cells with 0.1 mM NaDOC, and D cells with 0.2 mM NaDOC. Treatment lasted for 48 h, at which time the media plus NaDOC was removed and fresh media added. Remaining attached cells were allowed to grow until they reached ~80% confluency, at which point they were split and passaged. Cells of the A line were passaged weekly, while the NaDOC treated cells took longer to recover from the drug treatment and so were passaged less frequently. When the cells in a particular flask could survive a 48-h treatment with NaDOC, the concentration used to treat them in the following passage was increased. This was continued until all three lines were determined to be resistant to 0.5 mM NaDOC.
Evaluation of resistance to DOC-induced apoptosis
Initially, apoptosis resistance was evaluated using morphological criteria for the presence of apoptotic cells (8). Cytospin preparations were made, stained with Giemsa and 200 cells counted, as previously described (25–28). Later cultures were screened using the MTT assay.
Ultrastructural evaluation by transmission electron microscopy (TEM)
Untreated A cells and all three resistant lines B, C and D were fixed overnight in 3% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.2). The cells were pelleted, post-fixed in 2% osmium tetroxide, dehydrated in a graded series of ethanols, and embedded in epoxy resin. Cells were then viewed in a Philips transmission electron microscope.
Preparation of cells for confocal microscopy
Parental HCT-116 cells, sensitive cell line A, and all three resistant lines (B, C and D) were seeded directly onto coverslips. When the cells on coverslips had reached ~80% confluency, the coverslips were removed and the cells were fixed in 4% methanol-free formalin (Ted Pella, Inc., Redding, CA) in 1xPBS for 20 min. After fixation, the cells were permeabilized at –20°C in 100% methanol for 6 min. Slides were air dried, and stored at –20°C until immunostaining.
(i) Immunofluorescence procedures for activated NF-B
Cells on coverslips were stained with a monoclonal antibody to an epitope that spans the nuclear localization signal (NLS) sequence of the p65 subunit of NF-B (Boehringer Mannheim, Indianapolis, IN) (Kaltschmidt et al., 1995). The NLS sequence in its inactive state is bound by the inhibitory IB proteins which, upon NF-B activation, are degraded. Positive staining with this antibody therefore indicates the presence of activated NF-B. The antibody was used at a dilution of 1:100 in 1% bovine serum albumin in PBS. Treatment with a biotinylated goat anti-mouse secondary antibody (Vector Laboratories Inc., Burlingame, CA) (dilution 1:100) subsequently tagged with Cy-5 conjugated streptavidin (Vector Laboratories) (dilution 1:100) was used as the detection system. Nuclei were identified using YOYO-1 staining after RNase digestion as previously described (26–28).
(ii) Immunofluorescence procedures for NF-B p50 and NF-B p65 subunits
Cells on coverslips were stained with polyclonal antibodies against NF-B p50 (dilution 1:40) and p65 (dilution 1:20) subunits (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Proteins were detected using a biotinylated goat anti-rabbit antibody (dilution 1:100) followed by Cy-5 conjugated streptavidin (dilution 1:100) detection system. Nuclei were visualized as described in (i) above.
(iii) Immunofluorescence procedures for Grp78
A polyclonal antibody that was affinity purified and recognized an amino acid sequence at the amino terminus of the 78 kDa glucose-regulated protein (Grp78) was obtained from Santa Cruz Biotechnology, Inc. The antibody was used at a dilution of 1:50. Grp78 was detected using the biotin/Cy-5 streptavidin detection system [see (ii)above] and nuclei were visualized as described in (i) above.
(iv) Immunofluorescence procedures for Bcl-2
A monoclonal antibody to Bcl-2 (Santa Cruz Biotechnology, Inc.) was used at a dilution of 1:50. Bcl-2 was detected using the biotin/Cy-5 streptavidin detection system and nuclei were visualized as described in (i) above.
(v) Co-localization of Bcl-2 with mitochondria
The dye, Mitotracker Red ® (CMXRos) [Molecular Probes (Eugene, OR)], which is taken up into mitochondria as a function of the mitochondrial membrane potential (
mt) (29), was used to mark the mitochondria for co-localization studies. CMXRos was added to the cell suspension at a final concentration of 100 nM, and incubated for 30 min at 37°C, as previously described (30).
cDNA microarray analysis
The Oligotex mRNA Direct Kit (Qiagen, Valencia, CA) and protocol was used to isolate poly A+ mRNA from each of the resistant lines B, C and D, as well as the sensitive line A. Fluorescent first strand cDNA from each resistant line and the sensitive line A was made from 4 µg of poly A+ RNA using the Micromax Direct cDNA Microarray System (NEN Life Sciences, Boston, MA) following manufacturer's protocols, and all other components of the reaction were obtained from Gibco BRL. Lines B, C and D were Cy3 (Amersham, Piscataway, NJ) labeled while line A was Cy5 (Amersham, Piscataway, NJ) labeled. Three reactions were performed for sensitive cell line A, and one reaction each for the resistant lines. Each Cy5 labeled sensitive A line cDNA reaction was combined with one of the Cy3 labeled cDNAs obtained from the resistant lines using the Qiaquick PCR Purification Kit (Qiagen, Valencia, CA) following manufacturer's protocols. After elution from the purification column, the probe was lyophilized to dryness, and resuspended in 15 µl hybridization buffer, denatured by boiling for 2.5 min, and added to a microarray. Co-hybridization of the mixture was done on purified PCR products of human cDNA inserts robotically spotted onto chemically activated glass microscope slides. One microarray each was used for the A/B combination, the A/C combination and A/D combination. A coverslip (22x22 mm) was applied and the array was placed in a hybridization chamber (catalog number HYB-03, GeneMachines, San Carlos, CA) at 62°C for 18 h. Following hybridization, slides were washed by placing them into 50 ml conical tubes containing 2xSSC, 0.1% SDS for 5 min, 0.06xSSC, 0.1% SDS for 5 min, and 0.06xSSC for 2 min all at room temperature. Slides were scanned for Cy3 and Cy5 fluorescence using an Axon GenePix 4000 microarray reader (Axon Instruments, Foster City, CA) and quantitated using GenePix software. Gene expression results were analyzed using GeneSpring (Silicon Genetics, Redwood City, CA) software. For statistical considerations, the mean fold-induction or reduction values of the three resistant cell lines were compared with that of the long-passage sensitive cell line. The difference in these gene expression values between the sensitive and resistant cell lines was considered significant at the 95% confidence level. Genes that showed at least a 1.5-fold induction or a 0.5-fold reduction in expression in at least two of the three resistant cell lines compared with the sensitive cell line were also considered biologically important.
Real time RT–PCR (reverse transcriptase–polymerase chain reaction)
The modulation of several genes was confirmed by an independent measure of mRNA levels. The genes selected, i.e. calcium/calmodulin-dependent protein kinase (CAMK2D) and Maspin (SERPIN5B), had a high signal-to-noise ratio in the apoptosis-sensitive control (HCT-1116A) cell line. Two hundred ng of RNA was reverse transcribed using reagents from the Applied Biosystems (ABI) (Foster City, CA) TaqMan® kit (p/n N808–0234) with the following reaction conditions: 25°C for 10 min, 48°C for 30 min and 95°C for 5 min on an MJ PTC-200 Peltier Thermal Cycler (Waltham, MA) according to the manufacturer's instructions. The resulting cDNA was amplified on an ABI 7000 Sequence Detection System using SYBR® Green real time detection. Primers were designed using ABI Prism® Primer ExpressTM version 2.0 software from ABI (p/n 4329442) and then ordered through Bio·Synthesis (Lewisville, TX). Primers sequences were: CAMK2D (Unigene Hs.111460), forward 5'-CCACCTGCACCAGGTTCAC-3', reverse 5'-ATGCCCCCTTTCCAAGCT-3'; Maspin (Unigene Hs.55279), forward 5'-CTGACAACAGTGTGAACGAC-3', reverse 5'-CAAGCCTTGGGATCAATCATCT-3'. The reaction contained 4 ng of cDNA, 11 µl of PCR H2O, 12.5 µl 2X SYBR® Green Master Mix (ABI p/n 4309155), and 0.5 µl of primer mix at 25 pmol/µl. ABI's universal PCR reaction conditions were used according to the manufacturer's instructions and are as follows: Stage 1: 50°C for 2 min, Stage 2: 95°C for 10 min, and Stage 3 repeated for 40 cycles: 95°C for 15 s then 60°C for 1 min. Products were quantitated using the Ct method as described by the manufacturer. Products were size separated on a 2% agarose gel to confirm correct amplicon length. For CAMK2D delta and maspin, relative mRNA levels using cDNA microarray were reasonably consistent with relative mRNA levels determined using real time RT–PCR, as shown in Table I
.
|
2D gel electrophoresis and MALDI-MS
Protein was isolated from each of the resistant lines, from the sensitive line A, as well as from the parent HCT-116 cells according to the protocol suggested by Kendrick Labs (Madison, WI). Reagents for isolating protein samples were also provided by Kendrick Labs. Two-dimensional electrophoresis was performed according to the method of O'Farrell (31) by Dr Nancy Kendrick as follows: isoelectric focusing (IEF) was carried out in glass tubes of inner diameter 2.0 mm using 2.0% pH3.5–10 ampholines (Amersham Pharmacia Biotech, Piscataway, NJ) for 9600 volt-h. Fifty ng of an IEF internal standard, tropomyosin, was added to each sample. This protein migrates as a doublet with the lower polypeptide spot of MW 33 000 and pI 5.2. After equilibration for 10 min in buffer, each tube gel was sealed to the top of a stacking gel which is on top of a 10% acrylamide slab gel (0.75 mm thick). SDS slab gel electrophoresis was carried out for 4 h at 12.5 mA/gel. The following proteins (Sigma Chemical, St. Louis, MO) were added as molecular weight standards to a well in the agarose which sealed the tube gel to the slab gel: myosin (220 000), phosphorylase A (94 000), catalase (60 000), actin (43 000), carbonic anhydrase (29 000) and lysozyme (14 000). These standards appear as bands on the basic edge (right side) of the special silver-stained 10% acrylamide slab gel compatible with mass spectroscopy (32). Gels were dried between sheets of cellophane with the acid edge to the left.
Antisense oligonucleotide techniques
Design of p65 antisense and missense oligonucleotides.
Two different 20-mer antisense oligonucleotide probes to the p65 subunit of NF-B were selected from 10 probes that targeted the ATG initiation codon and the region 3' of the ATG initiation codon (Figure 1) and an additional 10 probes that straddled the ATG codon (Figure 2), using the mRNA sequence of Ruben et al. (33)(GenBank accession # M62399). This selection was based on minimal secondary structure of the probes as determined using Eugene software [Daniben System, Inc. (Cincinnati, OH)]. The sequences of the two antisense probes, the missense probe and their secondary structural characteristics are shown in Table II. The antisense and missense probes were synthesized by Synthegen (Houston, TX) as phosphorothioates and labeled on the 5' end with 5(6)-carboxyfluorescein to visualize cellular uptake and localization and on the 3' end with acridine orange to assist cellular uptake and resist degradation by RNases.
|
|
|
Antisense protocols and quantitation of apoptotic cells
The apoptosis resistant cell line C (HCT-116RC) was used for antisense experiments. These cells, when exposed to 0.5 mM NaDOC, only undergo a low level of apoptosis (0–3%), indistinguishable from that of untreated control cells. Treatment of the unselected sensitive cell line with 0.5 mM NaDOC kills the vast majority of cells within 6–8 h. The resistant cells express high levels of the p50 and p65 subunits of NF-
B as well as activated NF-B (see Results). Various concentrations of the phosphorothioated antisense oligonucleotides were used to treat the HCT-116RC cells to determine cytotoxic levels and efficiency at down-regulating NF-B(p65) gene expression. After 24 or 48 h of treatment with the antisense oligonucleotide, cells were fixed and stained with an antibody to p65 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) to determine the amount of reduction in protein expression. A concentration of the most effective oligonucleotide was chosen that jointly optimized cell viability and caused a decrease in NF-B(p65) protein levels. Cells were then pretreated for 24 h with this concentration of oligonucleotide prior to NaDOC treatment. An apoptotic index was then determined using morphologic criteria, as previously described (25–27,34). |
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Development of HCT-116 cell lines resistant to deoxycholate-induced apoptosis
For our selection experiments, we used an adherent epithelial cell line, HCT-116, originating from a colorectal carcinoma that has a mutator phenotype. We started the selection process using three different concentrations of NaDOC (0.02, 0.1 and 0.2 mM for lines B, C and D, respectively), anticipating that we would generate resistant clones with varying patterns of resistance. Resistant line B became resistant to 0.5 mM NaDOC after 17 passages of incrementally increasing treatment concentration (control cells were at 45 passages, representing 45 weeks of long term culture), C became resistant to 0.5 mM NaDOC after 18 passages (control at 46 passages), and D became resistant after 9 passages (control at 40 passages). Ultimately, we generated 3 cell lines (B, C and D) with resistance to NaDOC-induced apoptosis at 0.5 mM, a concentration that would cause the vast majority of cells in the sensitive line A as well as the parent line to undergo apoptosis within 6 h. The resistance of lines B, C and D is evidenced by their ability to grow for at least 48 h in the presence of 0.5 mM NaDOC. Their resistance was demonstrated to be stable after at least 4 weeks growth in the absence of DOC. The sensitive long-passage cell line is hereafter referred to as HCT-116SA and the three resistant cell lines B, C and D are designated as HCT-116RB, HCT-116RC and HCT-116RD, respectively.
Molecular and cellular characterization of resistant cell lines
(i) TEM shows increased secondary lysosomes and normal-appearing mitochondria
We first examined these cell lines for morphological differences. Since bile acids are known to induce oxidative stress (35), we expected that the cells exposed persistently to NaDOC might exhibit megamitochondria (28,36) and/or other features of persistent oxidative stress. Upon examination of each of the resistant lines and comparison with the sensitive line (Figure 3), we found large secondary lysosomes in all three resistant cell lines, but no megamitochondria. In addition, there were no differences in mitochondrial ultrastructure between the HCT-116SA and the resistant cells. This suggests that the cells may have adapted to coping with extensive damage by inducing the formation of mitochondria that are resistant to oxidative stress and by increasing lysosomal activity, as the latter organelle is involved in removal of damaged cellular components (37).
|
(ii) Selection for resistance to NaDOC increases levels of GRP78 and Bcl-2
Grp78 and Bcl-2 are localized in the endoplasmic reticulum (ER) and mitochondria, respectively, and have been reported to protect against apoptosis (38–40). Grp78 was specifically evaluated by immunofluorescence/confocal microscopy since we have previously reported that NaDOC activated the Grp78 promoter (41) indicating ER stress, and Grp78 may, therefore, protect against ER stress. Bcl-2, a classic anti-apoptotic protein that localizes to the ER and mitochondria, was specifically evaluated since we have previously shown that NaDOC induces megamitochondria formation and a reduction in the mitochondrial membrane potential (28), indicating mitochondrial stress. Bcl-2 may, therefore, protect against both the ER and mitochondrial stress pathways. Using immunofluorescent detection and confocal microscopy, it was found that HCT-116RC cells exhibited dramatically increased levels of Grp78, HCT-116RD cells exhibited a smaller increase, and HCT-116RB cells exhibited little if any increase (Figure 4
). All three resistant cell lines showed increased levels of Bcl-2 with HCT-116RC cells showing the largest increase (Figure 5). HCT-116RC cells were then incubated with CMXRos (taken up by mitochondria based on the mitochondrial membrane potential), fixed and stained for Bcl-2. In addition to a cytosolic localization, Bcl-2 was found to be co-localized to mitochondria using confocal microscopy (Figure 6).
|
|
|
(iii) Repeated exposure to NaDOC increases levels of activated NF-
B and protein levels of NF-B subunitsNF-
B is a transcription factor activated by changes in the cellular redox state. Previous work from our laboratory, using pharmacologic inhibition, has shown that activation of NF-B is protective against NaDOC-induced apoptosis (26). We found that all three resistant cell lines had increased levels of activated NF-B as evidenced by increased immunofluorescence using confocal microscopy (Figure 7). In order to determine if the increase in activated NF-B was due to increased protein levels of the individual subunits, we stained the cells on coverslips with polyclonal antibodies to the p50 or p65 subunits of NF-B. Examination by confocal microscopy showed that HCT-116RC and HCT-116RD cells had increased protein levels of p50 and p65 subunits (Figure 8 and Figure 9, respectively), indicating that in these two resistant cell lines, increased levels of the individual protein subunits may be responsible for the increase in activated NF-B. The third line, HCT-116RB, did not show an increase in protein levels of the individual subunits, suggesting that another mechanism for the increase in activated NF-B may be involved in this cell line.
|
|
|
(iv) Determination by cDNA microarray analysis of altered gene expression at the mRNA level in the apoptosis resistant cell lines
To identify other genes involved in apoptosis resistance, we performed a DNA microarray analysis that detects gene expression at the mRNA level. Analysis of the data for more than 5000 genes identified genes that were over- (n = 593) or underexpressed (n = 246) at the 95% confidence level in all three resistant cell lines, or modulated by at least a factor of 1.5 or 0.5 in at least two of the three resistant cell lines when compared with the sensitive HCT-116SA line. In Table III
, we have grouped relevant genes from the microarray results into nine distinct categories. These categories are based on the distinct cellular stresses induced by hydrophobic bile acids (26,27,41) and include the ability of bile acids to (i) perturb membranes in a ligand-independent manner, resulting in activation of receptor-type proteins (42–46); (ii) activate protein and lipid kinases (44,45,47–55); (iii) activate phospholipase A2 (56) and affect arachidonic acid metabolism (57–60); (iv) induce oxidative/nitrosative stress (28,34,41,61–65); (v) modulate detoxification pathways (66,67); (vi) induce growth arrest and DNA damage (26,41,68–72); (vii) activate transcription factors (26,41,73–77); (viii) induce apoptosis (8–11,26,27,34,62,78–84); (ix) damage mitochondria (28,64,65,85–87).
|
(v) 2-D gels and MALDI-MS identify overexpression of protective proteins
In order to identify overexpressing genes at the protein level, we performed 2-D gel electrophoresis on our three resistant lines as well as the sensitive HCT-116SA cell line. A representative gel is shown in Figure 10
. Of the 159 spots analyzed, six spots were identified as proteins that were overexpressed in each of the three resistant lines as compared with the sensitive HCT-116SA cell line and the parental line, and three spots were identified as proteins that were underexpressed in each of the three resistant lines. MALDI-MS was performed to identify these proteins, and of the six overexpressed proteins, three were identified as cofilin, maspin, and triose phosphate isomerase (Table IV). Two spots gave a poor spectrum and could not be identified, and the sixth spot gave a good spectrum but was not identified as a known protein. Two proteins underexpressed in all three resistant lines were identified as elongation factor 2 and heat shock protein 90 (Table IV), while a third spot could not be identified.
|
|
(vi) Role of the p65 subunit of NF-
B in the modulation of apoptosisAs with all broad molecular characterizations, many candidate proteins are identified that could potentially contribute to apoptosis resistance. We directly tested one prominent transcription factor, NF-
B, for its contribution to apoptosis resistance in this study. The rationale for this selection is based on previous work from our laboratory indicating that NaDOC activates an NF-B-regulated survival pathway (26) and that pretreatment of cells with pyrrolidine dithiocarbamate (PDTC), a potent inhibitor of NF-B activation (88), sensitized cells to NaDOC-induced apoptosis (26). These findings indicated that NF-B plays an important role in the protection of cells against cell death induced by NaDOC, a multiple stress inducer. In addition, in the present study, NF-B was constitutively activated in all three resistant cell lines, underscoring its potential importance to apoptosis resistance. The p65 subunit of NF-B was targeted for down-regulation using antisense strategies because of its prominent role as a survival factor in knockout mice studies (89).
For these studies, we used the apoptosis resistant cell line HCT-116RC. As described above, HCT-116RC cells contain constitutively increased levels of both the p65 subunit of NFnf-B and the activated form of NF-B, using confocal microscopy in conjunction with appropriate polyclonal and monoclonal antibodies. Two antisense oligonucleotides [one targeted to the ATG codon of p65 mRNA (antisense oligo #1) and one targeted to a sequence 3' of the ATG codon (antisense oligo # 9) (see Table II
, Figure 1)] were used to determine the role of the p65 subunit in NaDOC-induced apoptosis. HCT-116RC cells were treated with 2 µM antisense oligonucleotide #1 or 2 µM antisense oligonucleotide #9 for 24 h and then treated with 0.5 mM NaDOC. Cells that received no treatments and cells that were pretreated with oligonucleotide probes and received no NaDOC treatment resulted in <2% apoptosis after 8 h or 24 h of incubation (Figure 11). Treatment with NaDOC alone increased apoptosis to 15% of cells by 24 h, measured by morphologic criteria (Figure 11). Pretreatment with either antisense oligonucleotide, however, dramatically enhanced apoptosis in NaDOC-treated cells (Figure 11). The greatest sensitization to apoptosis (43%) occurred with antisense oligonucleotide #9, which had better secondary structural characteristics when compared with oligonucleotide #1 (Table II).
|
Confocal microscopy experiments were then performed to determine if antisense oliogonucleotide #9 down-regulated the level of the NF-
B(p65) protein. It was shown that antisense oligonucleotide #9 was effectively taken up into the nucleus and cytoplasm of HCT-116RC cells (Figure 12C) and that the p65 protein levels had markedly decreased levels in the antisense-treated cells (Figure 12A), compared with levels of p65 in cells treated with a missense oligonucleotide (Figure 12B). Moreover, we tested the missense oligonucleotide generated by scrambling the antisense oligonucleotide #9 message (Table II) for its effect on NaDOC-induced apoptosis in HCT-116RC cells. The missense oligonucleotide had no significant effect on NaDOC-induced apoptosis, whereas antisense oligonucleotide #9 sensitized the resistant cells to apoptosis (Figure 13). The antisense and missense oligonucleotide probes, alone, had no significant effect on apoptosis (Figure 13).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
We have successfully developed three independent cell lines (HCT-116RB, HCT-116RC and HCT-116RD) from the parental HCT-116 cell line that exhibit resistance to NaDOC-induced apoptosis. The experiments described here were carried out to gain insight into stable alterations in specific gene expression of colonic epithelial cells that are frequently exposed to increasing levels of bile acids that accompany a high fat/low fiber diet (Western-style diet). The in vitro results shed light on the alterations in gene expression that may be responsible for the observed apoptosis resistance in the epithelial cells of the flat mucosa of colon cancer patients that has been previously described by our group (8–11). From a practical standpoint, one or more of these altered proteins may have potential as a biomarker for early detection of colon cancer risk.
Since apoptosis resistance is a stable characteristic of the cell lines HCT-116RB, HCT-116RC and HCT-116RD, persisting for at least a month after removal of NaDOC, we assume that the resistance phenotype is due to mutations and/or epigenetic changes (epimutations). Natural selection operates to favor any cell that has acquired a mutation/epimutation providing a growth advantage during NaDOC exposure. The evolution of the resistance phenotype in these cultured cell lines is the product of both mutation/epimutation and selection. Because mutations/epimutations occur randomly, the sequence of their occurrence will have differed between the three cell lines. Furthermore, since the initial exposures to NaDOC differed for the three lines, the initial selective pressures also differed. Consequently, the patterns of gene expression would be expected to differ among the three lines. Cell line HCT-116RD, which received the highest initial exposure to NaDOC, could only have survived if it had acquired a large early increase in resistance, whereas cell line B could have survived with a more gradual buildup of resistance mutations/epimutations. That considerable overlap in gene expression is actually found between the three independent lines is probably a reflection of the fact that the types of changes in gene expression which provide effective resistance to NaDOC is finite. Thus, similar changes, occurring independently, will be selected by repeated exposure to NaDOC.
Previous work from our laboratory suggested that NF-B is activated by deoxycholate (26) and plays an important role in protecting cells from the multiple stress inducer, NaDOC (41). We assessed the resistant cell lines, using immunofluorescence and confocal microscopy, for changes in protein levels of activated NF-B, as well as changes in levels of the individual subunits. All three resistant cell lines had increased protein levels of activated NF-B. The constitutively activated NF-B in all three resistant cell lines may have been selected for during the frequent stress endured by these cell lines during the selection process. In addition, all three strains had increased mRNA levels for the p65 subunit of NF-B (Table III vii). One possible explanation for the increase in activated NF-B is an increase in the protein levels of the p65 and p50 subunits of NF-B. Using antibodies to the individual subunits in conjunction with confocal microscopy, we found that resistant lines HCT-116RC and HCT-116RD, but not HCT-116RB, had increased protein levels of the p65 and p50 subunits in comparison with the sensitive cell line HCT-116SA. The increased levels of activated NF-B in HCT-116RB may be due to other factors that we did not assess. It may be that this cell line has decreased levels of NF-B inhibitory proteins, or, conversely, has increased kinase activity responsible for phosphorylating the inhibitory proteins resulting in IB degradation and the activation of NF-B. We also directly tested the anti-apoptotic function of the constitutively activated NF-B in one of the resistant cell lines (HCT-116RC) using antisense strategies, and found that down-regulation of the p65 subunit of NF-B dramatically reversed the apoptosis resistance. There are many NF-B downstream signaling pathways whose effectors may be responsible for protecting colon cells against apoptosis, including increased expression of anti-apoptotic proteins, such as Bcl-2 and Bcl-xL.
Figure 14 is a summary of established and suggested signal transduction pathways in which several of our key findings from this study are integrated based on plausible mechanisms of resistance to NaDOC-induced apoptosis. Space limitations do not permit a thorough discussion of all genes that were significantly overexpressed or underexpressed in the apoptosis resistant cell lines (Tables III and IV), or the multiple possible mechanisms associated with apoptosis resistance, in general (90). We have, therefore, focused these interacting pathways around the upstream activation of NF-B and downstream targets of NF-B, such as nitric oxide synthase (NOS)2 (inducible NOS) and NO (13,34) and Bcl-2 family members because of their key role in protecting cells against oxidative stress and/or a mitochondrial apoptosis pathway (28). The gene products shown in red in Figure 14 showed increased expression and those shown in blue showed decreased expression in the HCT-116 resistant cell lines at either the mRNA or protein levels.
|
NF-
B can be activated by kinases that were shown to be constitutively increased at the mRNA level in the apoptosis resistant cell lines. Upstream signaling proteins that may be responsible for constitutive NF-B activation in the HCT-116 resistant cell lines include EGFR (91–93), Ras GTPases (94–98), MAP kinase kinase (MEK) kinases (MEKK) (99), phosphatidylinositol kinases (93,100,101), PKA (102), PKC (PKC-zeta) (103–108) and NIK (NF-B-inducing kinase) (109–111) or NIK-like kinases (MINK) (Figure 14). Of the various PKC isoforms, it is interesting that the atypical isoform, PKC, with a known anti-apoptotic function (112) shows increased expression at the mRNA level in the resistant cell lines. Atypical PKC isoforms are known to associate with and directly phosphorylate IB-kinase-ß (IKK-ß) or IB-, leading to IB degradation and NF-B nuclear translocation (103–106), which provides a mechanism for its anti-apoptotic function. (NB. anti-apoptotic signaling through NF-B-independent kinase pathways is also possible and this possibility is indicated on the left side of Figure 14.) Although reduction–oxidation (redox) plays a critical role in NF-B activation (113,114), different antioxidants are variously selective for NF-B activation. Thioredoxin (Trx), for example, is a more potent antioxidant than either glutathione or N-acetylcysteine (115). The enzymes that serve to reduce oxidized thioredoxin were found to be up-regulated in the resistant HCT-116 cell lines in the present study. These include peroxiredoxin 2 [Trx peroxidase (TPx) 2], peroxiredoxin 4 (TPx 4 or AOE372) and Trx reductase (TR). The TPxs may, therefore, function as immediate regulators of peroxide-mediated activation of NF-B (116), and are shown in Figure 14 as being major players in the NF-B anti-apoptotic signaling pathway. The activation of NF-B in conjunction with a reduction in the level of peroxides in the cell may be responsible for the known anti-apoptotic functions of various members of the TPx family of antioxidant enzymes (117–120).
Three of the anti-apoptotic gene products [Bcl-2, IAP (inhibitor of apoptosis protein) and TRAF2 (tumor necrosis factor receptor-associated factor2)] whose expression was constitutvely increased in the apoptosis resistant cell lines (Figure 14) are known targets of NF-B gene transcription (121–125). Although the activity of NOS was shown to be protective in previous studies (33), NOS2 was not found to have increased expression at the mRNA level in the cDNA microarray analyses (protein analyses not performed). Although NOS2 is considered an inducible enzyme, we have previously reported that its constitutive expression can protect HepG2 cells against genotoxic, oxidative, xenobiotic and ER stress (126). NOS and its downstream NO-activated enzyme, guanylate cyclase (GC), and its target kinase, PKG (cGMP-activated protein kinase), may play a role in apoptosis resistance in the HCT-116 resistant cells, although functional studies need to be performed to determine a role for the NO/GC/cGMP/PKG signaling module in apoptosis resistance. The importance of NOS2, NO and GC to apoptosis resistance is supported by the constitutively increased expression of the recycling enzyme, quinoid dihydropteridine reductase (QDPR), the GC-activating protein-1 (GC activator 1A or GUCA1A), and the TPx/Trx/TR system in the resistant cell lines (see Figure 14). QDPR is an enzyme that recycles BH4 (tetrahydrobiopterin), an essential co-factor for NOS activity and the generation of NO (127). Schallreuter and Wood (128) have reported that the TPx/Trx/TR system is known to prevent the cytotoxic oxidation of BH4 under conditions of oxidative stress (129). GC activating protein-1 mRNA is increased in the resistant cell lines (Table III iv). It is a member of the neuronal calcium sensor family of Ca2+-binding proteins (130). Its high affinity for Ca2+ binding makes GC-activating protein an important sensor linking increased intracellular [Ca2+] to the generation of an important secondary messenger, cGMP.
An up-regulation of NO can have many biologic effects that result in apoptosis resistance (13), including S nitrosylation, inhibition of respiration and the inhibition of zinc finger proteins (131) that induce the expression of pro-apoptotic proteins, such as BNIP3 (Bcl2/adenovirus EIB 19 kDa-interacting protein 3) (132). BNIP3 expression is reduced in the resistant cell lines (Table III viii, Figure 14) as would be expected if NO has these effects. A consequence of an increase in NO is the formation of peroxynitrite (ONOO-), a potent toxicant. Selenoproteins, such as selenoprotein P, glutathione peroxidase 1, glutathione peroxidase 2 (gastrointestinal) and TR1, which are constitutively increased at the mRNA level in the resistant cells (see Table III iv) may protect against peroxynitrite-mediated oxidation (133–135).
The increased expression of five distinct metallothioneins in the resistant cell lines (Table III v) underscores their probable importance as cytoprotective proteins. We recently reported that the metallothionein promoter is activated by ursodeoxycholate (136), a hydrophilic bile acid that protects against the deleterious effects of cytotoxic bile acids such as DOC (137). This complements our finding here, and further suggests that metallothioneins protect cells against the cytotoxic effects of hydrophobic bile acids, such as NaDOC. The role of metallothioneins in cytoprotection is multi-factorial. Cai et al. (138) reported that metallothionein may react directly with peroxynitrite and prevent DNA damage. Zinc metallothionein may also be imported in mitochondria where it inhibits respiration (139). Since we have reported that the inhibition of respiration with rotenone and TTFA (thenoyl trifluoroacetone) inhibits NaDOC-induced apoptosis (28), the import of zinc-metallothionein to mitochondria may be a significant mechanism of bile salt-induced resistance. On the other hand, NO can mobilize Zn2+ from its stores in metallothioneins (140), with consequences for activation of zinc finger transcription factors (131) and/or inhibition of the Ca2+/Mg2+-activated endonuclease (141) and caspases (142).
Triose phosphate isomerase (TPI), an enzyme associated with glycolytic metabolism, was found to be increased in the resistant cell lines (Table IV) at the protein level using 2D-gel electrophoresis and MALDI-mass spectrometry. This enzyme may protect against oxidative stress initially induced by DOC. TPI catalyzes the reversible interconversion of dihydroxyacetone phosphate (DHAP) and glyceraldehyde 3 phosphate (GAP), transferring a proton from DHAP to GAP. GAP then reduces NAD+ to NADH, the main source of reducing power in the cell. Thus, the constitutively increased expression of TPI, selenoproteins, metallothioneins, glutathione synthetase and glutathione peroxidases in the resistant cell lines (see Tables III and IV), probably protects against the increased oxidative/nitrosative stress induced by frequent exposure to DOC during the selection process.
We have previously reported that DOC causes megamitochondrion formation in HT-29 colon cells within 4 h and induces apoptosis through a mitochondrial pathway (28). Oxidative stress is known to induce megamitochondria in studies performed by Karbowski et al. (36). Since bile acids induce oxidative stress, we anticipated that we might see megamitochondria in cells repeatedly exposed to bile acids in the present study. Careful analysis of each of the resistant cell lines indicated that there were no megamitochondria or any alterations in mitochondrial ultrastructure compared with the sensitive HCT-116SA cell line. Megamitochondria may be an early response to oxidative stress, and as these cells had been exposed to long term frequent treatments with NaDOC, megamitochondria induced in the parental cells may have been replaced by selection for more resistant mitochondria of normal size. We did observe large activated lysosomes at the ultrastructural level in all three cell lines, indicating a possible role of lysosomes in removing damaged organelles, including mitochondria. There are probably multiple mechanisms that protect mitochondria against damage in the resistant cells. The fact that many antioxidant enzymes are at increased constitutive levels, may protect, in part, against oxidative damage. Increased expression of 17 genes directly associated with mitochondria (Table III ix) was observed by cDNA microarray analysis. Of particular interest is the increased constitutive level of a subunit of the Fo complex of the H+ transporting ATP synthase (Table III ix). This may allow these frequently stressed cells to generate sufficient energy to protect against apoptosis. The interesting findings that BNIP3, a pro-apoptotic Bcl-2 family member (143) was constitutively reduced (see Table III viii) and Bcl-2 was constitutively increased (immunofluorescence results, Figure 5) in the resistant cell lines, are consistent with mechanisms of resistance that protect against mitochondrial damage (Figure 14). The fact that Bcl-2 localized to mitochondria, using confocal microscopy in the present study, is also consistent with a protection against mitochondrial damage and the inhibition of release of cytochrome c (144). The release of cytochrome c from mitochondria during apoptosis usually results in the activation of caspase-3 (145), one of the downstream effector caspases. The fact that the expression of caspase-3 and caspase-6 were constitutively reduced at the mRNA level in the resistant cell lines (Figure 14) indicates a possible mechanism of apoptosis resistance that reduces the availability of these molecules for activation after cytochrome c release (Figure 14).
In addition to caspases (146–148), serine proteases have been determined to be important for bile acid-induced apoptosis in hepatocytes (78). Five members of the SERPIN (Serine protease inhibitor) family of proteins were found to be constitutively increased in the resistant cell lines [four assessed at the mRNA level (Table III viii) and one assessed at the protein level, i.e.maspin (Table IV)]. Of particular interest is maspin, which was determined to be overexpressed in all three of our resistant cell lines (Table IV), using 2-D gel electrophoresis/MALDI-MS. [NB. Maspin was not modulated at the mRNA level using both cDNA microarray analysis and real time RT–PCR (see Table I in Materials and methods section)]. Maspin may, therefore, be a novel anti-apoptotic protein in the colon. Several distinct functions have been proposed for maspin. (i) Maspin has been demonstrated to enhance surface adhesion by upregulation of cell adhesion molecules. This may affect resistance to stress. It has been shown that the ability of a cell to interact with extracellular matrix molecules and adjacent cells and tissues provides a sanctuary against apoptosis (149). (ii) The serpins inhibit proteinases and several serpins have been implicated in the regulation of cell death. The viral serpin protein, CrmA, prevents cytokine processing by inhibiting caspases, and protects against Fas-, TNF- and TRAIL-mediated apoptosis by inhibiting an unidentified proteinase specific to these pathways (150). (iii) An endogenous serpin, PN-1 (protease nexin-1), prevents the delivery of an apoptotic signal by inhibiting an extracellular proteinase from cleaving a cell surface receptor (150). Maspin may be acting in some fashion similar to CrmA, PN-1, or cell surface signaling causing apoptosis resistance. (iv) Increased expression of maspin occurs in response to DNA damage and the response is regulated by p53 (151). The maspin promoter is activated by binding of p53 directly to the p53 consensus-binding site present in the maspin promoter (151). If increased expression of maspin leads to growth arrest, it may be acting in a manner similar to the up-regulation of retinoblastoma protein in the inhibition of apoptosis (152–155) or may represent a novel anti-apoptotic molecule with a unique mechanism of action. Since the expression of p53 (Table III vii) and maspin (Table IV) were constitutively increased in the resistant cell lines, maspin may be one of the downstream anti-apoptotic molecules induced by the transcriptional activity of p53.
Although we have focused much of this discussion on the mitochondrial pathway of DOC-induced apoptosis (Figure 14), NaDOC is a multiple stress inducer (41) and also activates the Grp78 promoter (41), an indication of ER stress. Grp78 is a soluble non-glycosylated member of the heat shock family of stress-related proteins, which resides in the ER (156) and protects cells against ER stress. The chaperone function of Grp78 is involved in protein folding and Ca2+ binding, and maintains the balance in Ca2+ levels between the lumen of the ER and the cytoplasm. In the present study, we found increased expression of this protein in two of the three NaDOC resistant cell lines by immunofluorescence/confocal microscopy (Figure 4). The ER stress pathway is well documented to induce apoptosis through mechanisms that are less well understood (157,158) compared with those of the mitochondrial pathway (159,160). The outer membrane of the ER contains receptors for inositol trisphosphate (IP3). IP3 is formed during degradation of phosphatidylinositol bisphosphate in response to phospholipase C activation. IP3 receptor activation leads to the release of ER Ca2+ stores, and, once free cytosolic Ca2+ increases to nonphysiologic levels, Ca2+ dependent enzymes are activated. These enzymes include phospholipases, proteases, and endonucleases. It is believed that these enzymes may be ultimately responsible for carrying out the execution phase of apoptosis (161) through the ER stress pathway. This mechanism is plausible, since it is known that bile acids can increase intracellular [Ca2+] (162), induce calcium signals (163) and contribute to calcium-mediated hepatocyte apoptosis (164). Grp78 was also recently shown to inhibit caspase activation and caspase-mediated cell death, underscoring the important cytoprotective function of this chaperone protein (165). Thus, the constitutive increase of Grp78 at the protein level in our resistant cell lines probably contributes, in part, to the observed apoptosis resistance.
In summary, the in vitro development of these resistant colonic epithelial cell lines and their molecular and cellular characterization, has suggested potential mechanisms by which apoptosis resistance may develop in vivo in the colonic epithelium in response to a known tumor promoter in the colon, NaDOC. Several methods have been utilized to identify both novel proteins and proteins previously characterized as influencing apoptosis that may play a role in apoptosis resistance and early stage carcinogenesis. Our observations indicate that apoptosis resistance arises from alteration of expression of proteins acting through multiple pathways. Some of these modulated proteins may have potential as biomarkers of apoptosis resistance and increased cancer risk on an individual basis using targeted cDNA microarrays and/or immunohistochemical stains.
|
Notes |
|---|
5 To whom correspondence should be addressed Email: bernstein3{at}earthlink.net
[NB. An Excel file of all the data for the three resistant HCT-116 cell lines compared with the sensitive HCT-116 cell line is available as an Email attachment, upon request.]
|
Acknowledgments |
|---|
This work was supported in part by NIEHS grant #ES06694 (Experimental Pathology Core S