1. Introduction
Despite progress in surgery and chemotherapy, ovarian cancer still proves lethal in 70% of cases, leading to the death of more than 14,000 women in the United States each year [
1]. One of the most important factors contributing to poor outcomes is the persistence of dormant, drug-resistant cancer cells after primary cytoreductive surgery and combination chemotherapy. Despite normalization of CA125 and negative imaging following primary treatment, “second look” exploratory surgery of the abdominal cavity can detect small, quiescent, poorly vascularized nodules of persistent ovarian cancer on the peritoneal surface in ~50% of patients. After positive second look operations, persistent ovarian cancer may take months or years to become clinically evident, consistent with tumor dormancy. In more than 80% of positive second looks, persistent ovarian cancer cells express DIRAS3 and are undergoing autophagy [
2].
DIRAS3 is a maternally imprinted gene that encodes a 26 kD GTPase with 50–60% homology to Ras, but with a 34 amino acid N-terminal extension that reverses Ras function. We found that DIRAS3 is downregulated in 60% of ovarian cancers, and was the most downregulated gene compared to normal ovarian epithelial cells [
3]. Previous work from our group established that loss of expression from the paternal allele can occur genetically through loss of heterozygosity, epigenetically by hypermethylation of both alleles, transcriptionally by loss of E2F1 and E2F4 promoter binding or post-transcriptionally by loss of the HER RNA binding protein [
4,
5,
6,
7,
8,
9,
10]. Re-expression of DIRAS3 blocks mutant K-Ras
G12V and H-Ras
G12V -induced transformation of 3T3 and MCF-7 cells, prevents cancer cell growth, inhibits motility, induces autophagy and establishes tumor dormancy in xenografts [
11,
12].
Autophagy is a highly conserved catabolic process in which organelles and long-lived intracellular proteins are engulfed by double membrane vesicles, termed autophagosomes. Lysosomes then fuse with autophagosomes, forming autophagolysosomes [
13]. After acidification of the autophagolysosomes, proteases and lipases are activated that hydrolyze proteins and lipids, releasing amino acids and fatty acids that can be catabolized to produce energy. Depending upon the cellular context, induction of autophagy can either sustain or eliminate metabolically active cancer cells. In the short term, energy released by autophagy can facilitate survival of cancer cells in nutrient deprived microenvironments with inadequate blood supply. Persistent autophagy can, however, eliminate cancer cells. Enhanced mammary tumorigenesis has been observed in heterozygous
BECN1+/− mice with reduced expression of an essential component required for autophagy [
14].
DIRAS3 is a critical regulator of autophagy, (1) inhibiting AKT signaling, downregulating mTOR and decreasing phosphorylation of ULK1 to induce autophagy, (2) displacing Bcl-2 from BECN1 to form the autophagy initiation complex, (3) inhibiting Ras/MAPK and PI3K, retaining FOXO3a in the nucleus to induce critical proteins that participate in autophagy (LC3, Atg4), and (4) facilitating fusion of autophagosomes and lysosomes through Rab7 [
2,
11,
15].
Re-expression of DIRAS3 induces tumor dormancy in human ovarian cancer xenografts in immune deficient mice [
11]. Our group has developed the first inducible xenograft model for tumor dormancy in human ovarian cancer using tet-inducible expression of DIRAS3, demonstrating that autophagy is required to sustain small avascular deposits of human ovarian cancer cells in a nutrient poor environment [
11]. Treatment of dormant DIRAS3 expressing human ovarian cancer cells with chloroquine, a functional inhibitor of autophagy, delays the outgrowth of dormant ovarian cancer cells after downregulating DIRAS3 expression in this xenograft model [
11]. This model appears to be clinically relevant in that it mimics observation at second look operations of persistent DIRAS3 positive cancer cells that are undergoing autophagy [
2]. Whether or not this is dependent upon selection of a subset of pre-existing DIRAS3 expressing/autophagy positive cancer cells in the primary tumor or an adaptive phenotype that is conferred by a nutrient deprived, poorly vascularized tumor microenvironment remains to be elucidated.
In this study we have sought to determine whether nutrient starvation or amino acid deprivation can enhance DIRAS3 expression and induce autophagy in human ovarian cancer cells, providing one mechanism by which DIRAS3 can be upregulated in positive second look tumor specimens. Inhibition of mTOR following amino acid deprivation results in dissociation of E2F1 and E2F4 from the DIRAS3 promoter, with subsequent proteosomal degradation of these transcription factors. Loss of E2F1 and E2F4 permits transcriptional upregulation of DIRAS3, a critical member of the autophagy pathway. Similarly, inhibition of mTOR with the pharmacological inhibitor, rapamycin, results in increased E2F1 and E2F4 degradation, transcriptional upregulation of DIRAS3, and DIRAS3-mediated autophagy. In both cases, the transcriptional upregulation of DIRAS3 results in a positive feedback loop in which increased expression of DIRAS3 further decreases PI3K signaling and downstream phosphorylation of mTOR, resulting in further induction of autophagy. These data provide insight into the mechanism by which dormant tumors found at “second look” may adapt to the tumor microenvironment by upregulating DIRAS3 and autophagy.
3. Discussion
Our findings document a requirement for DIRAS3 in establishing the autophagy induced by amino acid deprivation in the presence or absence of serum. Under nutrient poor conditions, DIRAS3 is upregulated transcriptionally by decreased binding of E2F1 and E2F4 and increased binding of CEBPα to the
DIRAS3 promoter. Both genetic and pharmacological inhibition of E2F1/4 family members enhance DIRAS3-mediated autophagy in ovarian cancer cells. While DIRAS3 levels can be regulated by promoter methylation and miRNAs [
4,
5,
6,
7,
8,
9,
10], acute amino acid deprivation fails to affect either of these mechanisms by which DIRAS3 is traditionally silenced.
Autophagy could be important for the persistence of dormant chemotherapy resistant ovarian cancer cells following primary surgery and chemotherapy. Previous work from our group reported that small deposits of tumor which remain in collagenous scars on the peritoneal cavity following primary chemotherapy and surgery, express DIRAS3 and undergo autophagy in more than 80% of cases, whereas primary cancers from the same patients express DIRAS3 and punctate LC3 in less than 20% of cases [
2]. While it is possible that DIRAS3 expressing autophagic cells have been selected during primary treatment, the small fibrotic tumor nodules found on the peritoneal surface have few tumor associated vessels and it is likely that cancer cells are nutrient deprived. Our results suggest that amino acid or serum deprivation can induce DIRAS3 regulated autophagy, consistent with adaptation of persistent, drug resistant cancer cells to nutrient poor conditions.
Although autophagy has long been documented as a normal cellular process that is essential for recycling of mitochondria, peroxisomes and long-lived proteins, its role in carcinogenesis, tumor progression and dormancy remains controversial [
29]. This is largely due to the opposing, context-dependent roles that have been reported leading to the proposal of enhancing or inhibiting autophagy as a therapeutic strategy in cancer (Reviewed in references [
30,
31]). In nutrient- deprived ovarian cancer cells found in scars on the peritoneal surface, autophagy could provide energy to sustain cancer cells and facilitate tumor dormancy in the absence of tumor vessels [
32,
33]. Observations with a DIRAS3–induced model for ovarian cancer dormancy are consistent with this possibility. In this model, upregulation of DIRAS3 arrests cancer cell growth, prevents angiogenesis, and induces both autophagy and tumor dormancy. Treatment of dormant ovarian cancer cells with chloroquine, a functional inhibitor of autophagy, markedly delays outgrowth of tumors when dormancy is abrogated and growth permitted by downregulation of DIRAS3. Conversely, an excess of autophagy can lead to cancer cell death. DIRAS3-induced autophagic cancer cell death could be one mechanism that maintains the balance of cancer cell proliferation and cancer cell death, resulting in the ‘dormant’ phenotype. This hypothesis is supported by the work of Qiu et al. [
34] where they reported that overexpression of DIRAS3 in gastric cancer cells inhibited metastasis in vivo, and correlated with increased autophagy [
34].
In either case, autophagy could provide a therapeutic target for eliminating dormant, drug resistant ovarian cancer cells that remain after primary surgery and chemotherapy. Inhibition of autophagy has been achieved clinically by treatment with hydroxychloroquine with encouraging results in colorectal and pancreatic cancer [
35,
36], although the pharmacokinetics hydroxychloroquine is at the limit of concentrations required for effective inhibition of autophagy. Dimeric chloroquines and quinocrines have been shown to be 100- to 1000-fold more potent than hydroxychloroquine [
37] and may provide additional benefit. Conversely, autophagic cancer cell death could be augmented by eliminating survival factors secreted by ovarian cancer cells or found in the tumor microenvironment including IGF, IL-8 and VEGF. Treatment of mice with dormant autophagic human ovarian cancer xenografts using antibodies against VEGF, IL-8 and the IGF receptor significantly prolonged survival and cured a majority of mice [
38].
Another novel observation is that the CEBPα transcription factor positively regulates DIRAS3 expression and is also required for starvation-induced DIRAS3-mediated autophagy. Recently, emerging evidence has changed the classical dogma that transcription factor-DNA binding does not occur when high levels of methylation at CpG dinucleotides (mCpG) are present in the cis-regulatory region except for a few proteins which contain a mCpG-binding domain (MBD) [
39,
40]. Previous studies reported that a select few transcription factors without MBDs were found to bind methylated DNA with high affinity for specific sequences, but recent reports using systematic approaches have revealed that the number of transcription factors capable of binding and activating transcription of methylated DNA is much larger [
39] and CEBPα is one of the factors on that list.
4. Materials and Methods
4.1. Antibodies and Reagents
Antibodies against LC3 (2775S), β-actin (4910L), pAktS473 (9271S), AKT (9272), mTOR (2972), p-mTORS2448 (2971), E2F1 (3742), p-P70S6KT389 (9205), P70S6K (2708), p-ULK1S757 (6888), ULK1 (4773) p-ERKT202/Y204 (4370), ERK (4695) were purchased from Cell Signaling Technology (Danvers, MA, USA). Antibodies against DIRAS3 (ARHI) were generated in our laboratory, and do not show cross reactivity with other members of the DIRAS family. P62/SQSTM1 antibody was purchased from MBL (Woburn, MA, USA). Anti-E2F4 (10923-1-AP) antibody was purchased from proteintech (Rosemont, IL, USA). Anti-CEBPα D-5 (sc-365318), anti-AKT (sc-5298) and anti-ERK (sc-514302) antibodies was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA).
siRNAs were purchased from Dharmacon (Lafayette, CO, USA) and used as suggested by the manufacturer. E2F1 (1869) E2F4 (1874) DIRAS3 (9077) Sestrin2 (83667) CEBPα (1050) Non-targeting control (1810).
4.2. Cell Lines
Human ovarian cancer cells, A2780, SKOv3, ES2 and OVCAR8 were grown in RPMI medium, supplemented with 10% FBS and 1% L-glutamine. Amino acid free media was prepared from RPMI powder (US Biological R8999-04A) supplemented with 10% dialyzed (MWCO 10k) FBS (26400-044). Full starvation media was generated with HBSS supplemented with 3% glucose. Cells were routinely checked for mycoplasma contamination.
4.3. Measurement of mRNA Expression
Cells were seeded in a 6-well plate at 1.5 × 105 cells/well, or following reverse transfection of siRNA at 1.0 × 105 cells/well, and treated as indicated. RNA was extracted with Aurum Total RNA Mini Kit (BioRad #7326820, Hercules, CA, USA). 200–400 ng of RNA was used in a SYBR green based quantitative PCR (One-Step RT-qPCR from RNA, BioRad #1725150, Hercules, CA, USA) to measure RNA levels. Relative expression was calculated by the 2−ΔΔCT method using glyceraldehyde-3-phosphatase dehydrogenase (GAPDH) as the reference gene. The experiments were repeated a minimum of three times, and samples were measured in technical duplicate.
Primers included (5′ → 3′):
GAPDH | Forward | ATGGAATCCATCACCATCTT |
Reverse | CGCCCCACTTGATTTTGG |
LC3 | Forward 356F | GAAGCAGCTTCCTGTTCTGG |
Reverse 588R | TCATCCCGAACGTCTCCTGG |
DIRAS3 | Forward | GTACCTGCCGACCATTGAAAA |
Reverse | GGGTTTCCTTCTTGGTCACTG |
E2F1 | Forward 734F | ACGTGACGTGTCAGGACCT |
Reverse 879R | GATCGGGCCTTGTTTGCTCTT |
E2F4 | Forward 202F | ATCGGGCTAATCGAGAAAAAGTC |
Reverse 350R | TGCTGGTCTAGTTCTTGCTCC |
4.4. Brightfield and Fluorescence Microscopy
Microscopy was performed using an Olympus IX71 microscope equipped with a DP72 camera and an XM10 camera at the specified magnifications (Shinjuku, Tokyo, Japan).
4.5. Transmission Electron Microscopy
Cells were seeded at (1.0–2.5 × 105 cells/well) in a 6-well plate and incubated for different intervals in amino acid-free medium. Samples were then fixed with Karnovsky’s fixative solution containing 3% glutaraldehyde plus 2% paraformaldehyde in 0.1 M cacodylate buffer at pH 7.3 and stored at 4 °C. After fixation, samples were submitted to the MD Anderson electron microscopy core facility for processing. Briefly, cells were washed in 0.1 M cacodylate buffer and treated with 0.1% Millipore-filtered buffered tannic acid, postfixed with 1% buffered osmium tetroxide for 30 min, and stained with 1% Millipore-filtered (0.2 μM) uranyl acetate. The samples were washed several times in water and then dehydrated in increasing concentrations of ethanol, infiltrated, and embedded in LX-112 medium. The samples were polymerized in a 60 °C oven for 2 days. Ultrathin sections were obtained using a Leica Ultracut microtome (Leica, Wetzlar, Germany), stained with uranyl acetate and lead citrate in a Leica EM Stainer, and examined in a JEM 1010 transmission electron microscope (JEOL USA Inc., Peabody, MA, USA) at an accelerating voltage of 80 kV. Digital images were obtained using AMT Imaging System (Advanced Microscopy Techniques Corp, Woburn, MA, USA).
4.6. Western Blotting
Cell lysates were prepared as indicated following incubation in lysis buffer (50 mM Hepes, pH 7.0, 150 mM NaCl, 1.5 mM MgCl2, 1 mM EGTA, 10 mM NaF, 10 mM sodium pyrophosphate, 10% glycerol, 1% Triton X-100) plus protease and phosphatase inhibitors (1 mM PMSF, 10 µg/mL leupeptin, 10 µg/mL aprotinin and 1 mM Na3VO4). Cells were lysed for 30 min on ice, and then centrifuged at 17,000× g for 30 min at 4 °C. The protein concentration was assessed using a bicinchoninic acid (BCA) protein assay (ThermoScientific, #23225, Waltham, MA, USA). Equal amounts of protein were separated by 8–16% SDS-PAGE, transferred to PVDF membranes and subjected to western blotting using an ECL chemiluminescence reagent (PerkinElmer, #NEL105001, Waltham, MA, USA).
4.7. siRNA Transfection
Cells were transfected with control or DIRAS family siRNAs (single and pooled oligos) using the Transfection #1 or #4 reagent (Dharmacon Research, #T-2001-01, #T-2004-01, Lafayette, CO, USA). Briefly, a mixture of siRNA (100 nM final concentration) and transfection reagents were incubated for 20 min at room temperature. This mixture was then added to cells and allowed to incubate for 48–72 h before cells were harvested for analysis.
4.8. Immunofluorescent Staining
Tumor cells (3 × 104) were seeded in chamber slides and treated as specified. Cells were fixed with 4% paraformaldehyde and permeabilized with 0.5% Triton X-100. Cells were washed with PBS, and blocked with 5% BSA/PBS followed by incubation with the primary antibody. After washing, cells were incubated with secondary antibodies conjugated with Alexa Fluor 488 or 594 (Molecular Probes, #A11017, #A11020, #A11070, #A11072), mounted and examined using a fluorescence microscope (Olympus IX71 microscope with an XM10 camera).
4.9. Quantitative Pyrosequencing
Methylation Analysis of the DIRAS3 CpG Islands. Genomic DNA was extracted from ovarian cancer cell lines following treatment with HBSS plus 3% glucose or amino acid free RPMI media for 0 h, 4 h and 24 h, using the QIAamp DNA Mini extraction kit. (Qiagen #51304, Lot 157030756, Hilden, Germany) 1ug of genomic DNA was treated with sodium bisulfite using the EZ DNA Methylation-Gold Kit (Zymo Research, Irvine, CA, USA) according to the manufacturer’s protocol. The samples were eluted in 40 μL of M-Elution Buffer, and 1 μL was used for each PCR reaction. Both bisulfite conversion and subsequent pyrosequencing analysis were done at the DNA Methylation Analysis Core, The University of Texas M.D. Anderson Cancer Center [
41].
PCR primers for pyrosequencing methylation analysis of the genomic area proximal to the transcription start site of DIRAS3 were designed using the Pyromark Assay Design SW 2.0 software (Qiagen, Hilden, Germany).
DIRAS3-1-F | GTAAGGGAGAAAGAAGTTAGA |
DIRAS3-1-R | biotin-TACTATCCTAACAAAACCCTC |
DIRAS3-1-SF2 | ATTTGGAAAAGGGATTGG |
Sequence to Analyze: YGGTGGGAGGYGTAGAGGGAAAAAGGAAYGATATAATYGGGTTTTTTAG
PCR size: 184bp
DIRAS3-2-F | GTTGGGTTAGTTTTTTATAGTTGGTT |
DIRAS3-2-R | biotin-AACCAAACAACCTAAAAAACAAATAC |
DIRAS3-2-SF2 | TTGGGGTGTTTAGTTGGTTG |
Sequence to Analyze: TYGYGGTAGTTTTTTYGAGTAGYGTATTTG
PCR size: 207bp
In brief, a sequencing primer was identified within 1 to 5 base pairs near the CpG sites of interest, with an annealing temperature of 40 ± 5 °C. After that, forward and reverse primers are identified uspstream and downstream to the sequencing primer. Optimal annealing temperatures for each of these primers were tested using gradient PCR. Controls for high methylation (SssI-treated DNA), low methylation (WGA-amplified DNA), partial methylation (S/W) and no-DNA template were included in each reaction. PCR reactions were performed in a total volume of 15 µl using ZymoTaq Polymerase kit reagent (Zymo Research, Irvine, CA, USA), and the entire volume was used for each pyrosequencing reaction as previously described (Estecio et al. 2007 [
41]). Briefly, PCR product purification was conducted with streptavidin-sepharose high-performance beads (GE Healthcare Life Sciences, Piscataway, NJ), and co-denaturation of the biotinylated PCR products and sequencing primer (3.6 pmol/reaction) was conducted following the PSQ96 sample preparation guide. Sequencing was performed on a PSQ HS 96 system (Biotage AB, Uppsala, Sweden) with the PyroMark Gold Q96 CDT Reagents (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. The degree of methylation was calculated using the PyroMark CpG SW 1.0 software (Biotage AB, Uppsala, Sweden).
4.10. Chromatin Immunoprecipitation (ChIP) Assay
A2780 ovarian cancer cells were seeded in 15 cm dishes at 80% confluency, and treated with amino acid free media for 0–4 h. An extra plate was used to determine cell number and viability following treatment, and an additional plate was prepared for chromatin immunoprecipitation. Using the protocol and reagents provided by Millipore (EZ-Magna ChIP A #17-408, Burlington, MA, USA), 1 × 107 cells were used for the assay. Sonication was performed on ice for 10 seconds at maximum intensity and resting for 2 min on ice between cycles for a total of 20–25 cycles. Once shearing of DNA was confirmed by agarose gel, lysate was mixed with 20 µL fully suspended Protein A magnetic beads, and 8 µg of antibody (IgG, E2F1, E2F4, CEBPα) per tube and allowed to rotate overnight at 4 °C. Immunoprecipitated DNA was purified using the spin columns provided and qRT-PCR was performed as described below. Each qPCR plate was performed in triplicate and GAPDH was used as a positive control for qPCR reaction. Analysis was performed to determine the relative DIRAS3 binding compared to the IgG negative control for each condition.
qPCR reagent assembly for 1 reaction:
ddH2O | 3.6 µL |
iTaqSYBR-Green master Mix (BioRad #172-5124) | 10 µL |
Primer Mix (10 µM working concentration F+R) | 0.4 µL |
DNA (20 µL diluted with 38 µL ddH2O) | 6.0 µL |
qPCR Parameters:
Initial Denaturation | 94 °C | 10 min |
Denature | 94 °C | 20 s |
Anneal and Extension | 60 °C | 1.0 min |
Repeat 60 cycles
Melt temperature curve
Primers included (5′ → 3′):
DIRAS3-0 | ChIP Forward 524F | TTTACCGGTCTTGCCACTAATG |
ChIP Reverse 341R | TCCAAAAGCAGTTTAATGCAGG |
DIRAS3-1 | ChIP Forward 76F | TTAAGAACCTTTTGCCTAGCC |
ChIP Reverse 227R | CGGTCCAACTGATTTTAGACGAA |
4.11. miRNA Quantitative Reverse Transcription-Polymerase Chain Reaction (qRT-PCR)
Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s suggested protocol. RNA was further purified with the miRNeasy Mini Kit from Qiagen (217004). RNA purity was assessed by measuring absorption at 260 nm (A260) and at 280 nm (A280) and the ratio of A260/A280 was determined. Those with a ratio of 1.9–2.1 were considered acceptable. qRT-PCR was performed to assess miR-221, miR-222, miR-222*, and miR-181 levels following amino acid deprivation or basal levels within the cell lines using the CFX Connect Real-Time PCR Detection system from BioRad (Hercules, CA, USA). Total RNA was reversely transcribed into complementary DNA (cDNA) with specific stem-loop RT primers and a cDNA synthesis kit (BioRad). Hsa-miR-221, 222, 222*, and 181a-c were purchased from ABI (Assay IDs 000524, 002276, 002097, 000480, 001098, 000482). Amplifications were carried out in triplicate on MicroAmp optical 96-well microliter plates (BioRad MLL9651). Thermal cycling conditions were as follows: 95 °C for 5 min, followed by 40 cycles of 95 °C for 10 s and 59 °C for 40 s. U6 RNA was used as an internal control in all qRT-PCR assays for miRNA. The ΔΔCT method was used to compare the relative expression levels at different timepoints following amino acid deprivation, and the final PCR results were expressed as the relative expression compared to individual control samples in each assay.
4.12. Statistics
All experiments were repeated independently at least two times and the data (bar graphs) expressed as mean ± s.d., unless noted otherwise. Statistical analysis was performed using Student’s t-test (two-sample assuming unequal variances). The criterion for statistical significance was taken as p < 0.05 (two-sided).