Chapter 12 - Monitoring the Autophagy Pathway in

C H A P T E R

T W E LV E

Monitoring the Autophagy Pathway in Cancer Frank C. Dorsey,* Meredith A. Steeves,* Stephanie M. Prater,* ¨ter,† and John L. Cleveland* Thomas Schro Contents 1. Introduction 2. LC3: A Phenotypic and Functional Marker of Autophagy 2.1. Real-time imaging of GFP-LC3 2.2. High content analysis of GFP-LC3 vesiculation 2.3. Monitoring GFP-LC3 by flow cytometry 2.4. Luciferase LC3: A high-throughput method to monitor autophagic activity 3. Assessing the Role of Autophagy in Em-Myc-Driven Lymphoma 3.1. Hematopoietic cell isolation and transplantation 3.2. Assessing hematopoietic chimerism 4. Concluding Remarks and Future Perspectives Acknowledgments References

252 254 255 258 260 262 264 265 268 269 269 269

Abstract Autophagy is an ancient cell survival pathway that is induced by metabolic stress and that helps prevent bioenergetic failure. This pathway has emerged as a promising new target in cancer treatment, where agents that inhibit autophagic degradation have efficacy in preventing cancer and in treating resistant disease when combined with conventional chemotherapeutics, which generally activate the pathway. However, agents that specifically target the autophagy pathway are currently lacking, and monitoring the effects of therapeutics on the autophagy pathway raises several challenges. Here we review the potential roles of the autophagy pathway in tumor progression and in maintenance of the malignant state, and introduce novel methods that we have developed that allow one to monitor autophagic activity ex vivo and in vivo.

* {

Department of Cancer Biology, The Scripps Research Institute, Scripps-Florida, Jupiter, Florida, USA Translational Research Institute, The Scripps Research Institute, Scripps-Florida, Jupiter, Florida, USA

Methods in Enzymology, Volume 453 ISSN 0076-6879, DOI: 10.1016/S0076-6879(08)04012-3

#

2009 Elsevier Inc. All rights reserved.

251

252

Frank C. Dorsey et al.

1. Introduction Though originally cast as part of the cell death machinery, recent evidence suggests the autophagy pathway supports tumor cell survival during metabolic stress. This is significant because autophagy is induced in response to a wide array of chemotherapeutic agents. In fact, activation of this pathway is a hallmark of nearly all therapeutic interventions. The clinical relevance of autophagy to cancer prevention and therapeutics is quickly coming into focus. For example, lymphomas arising in Em-Myc transgenic mice, a model of human B cell lymphoma (Adams et al., 1985), are sensitized to therapy-induced cell death following knockdown of the essential autophagy regulator Atg5 or treatment with the drug chloroquine (CQ) (Amaravadi et al., 2007), which impairs lysosomal-mediated degradation of cargo delivered by autophagosomes (Carew et al., 2007; Maclean et al., 2008). Furthermore, CQ treatment alone markedly delays the onset of B cell lymphoma in Em-Myc transgenics, and the development of T cell lymphoma in Atm null mice (Maclean et al., 2008). In addition, when combined with FDA-approved therapeutics such as the histone deacetylase inhibitor suberoylanilide hydroxamic acid (SAHA), CQ demonstrates efficacy even in refractory, Gleevec-resistant chronic myelogenous leukemia (CML) (Carew et al., 2007). Finally, recent clinical trials have demonstrated that CQ in combination with radiation and chemotherapy doubles the mean survival of patients suffering from glioblastoma multiforme (Briceno et al., 2003; Sotelo et al., 2006). Despite mounting evidence that inhibition of the pathway enhances the efficacy of anti-cancer regimens, autophagy appears to have, paradoxically, tumor suppressor properties. For example, deletion of one allele of beclin 1 (Atg6), a component of the class III PI-3-kinase complex required for autophagy, is common in breast and ovarian cancers (Liang et al., 1999), and mice haploinsufficient for beclin 1 are tumor-prone (Qu et al., 2003; Yue et al., 2003). Further, beclin 1 heterozygosity also accelerates lymphomagenesis in Em-Myc mice (Dorsey and Cleveland, unpublished data), a characteristic of classic tumor suppressors such as Arf or p53 (Eischen et al., 1999; Alt et al., 2003). Finally, mice lacking Atg4c, one of four Atg4 protease family members required for the cleavage of the proform of LC3 (Atg8) that mediates autophagosome formation (Fig. 12.1A), are more sensitive to chemically induced fibrosarcoma (Marino et al., 2007). Therefore, in some scenarios autophagy functions to suppress tumorigenesis, underscoring the need for important management and monitoring of the pathway before, during, and following therapy. One potential explanation for the apparent functional dichotomy of this pathway is that, similar to apoptosis, autophagy may curb oncogene-induced

253

Monitoring the Autophagy Pathway in Cancer

A

Atg4 pro-LC3 LC3-I

FGMKLSV

LC3-II

Atg4 FG

LC3-I FG

LC3-II LC3-II

PE

Cytosol organelles

LC3-II FG PE

LC3-II

LC3-II LC3-II LC3-II LC3-II

Cytosol organelles

Cytosol organelles

Cytosol organelles LC3-II

Phagophore

B

1.8

LC3-II

LC3-II LC3-II

Lysosome

Autophagosome

2.4

3

4

h

Figure 12.1 Monitoring autophagic vesicle formation by real-time microscopy using GFP-LC3. (A) Schematic depicting the steps of autophagic vesicle formation and maturation. The proform of LC3, a ubiquitin-like molecule required for fusion of isolation membranes of the phagophore to form the autophagosome, is first cleaved by the cysteine-dependent protease Atg4. This cleavage exposes a C-terminal glycine residue generating LC3-I. LC3-I is then activated by an E1 (Atg7), and is then transferred to an E2 (Atg3) resulting in its conjugation to phosphatidylethanolamine (PE) generating LC3-II. Then, LC3-II associates with phagophore membranes initiating the formation and maturation of autophagic vesicles. LC3-II remains associated with both the inner and outer membranes of autophagosomes and as a consequence is delivered along with the inner vesicle to the lysosome for degradation. LC3-II on the outer membrane of the autophagosome diffuses into the lysosomal membrane upon autophagic vesicle fusion with the lysosome. Here, Atg4 can then cleave PE from LC3-II, regenerating LC3-I that can then be used to form new autophagosomes. (B) Individual images of GFP-LC3^ expressing mouse embryonic fibroblasts (MEFs) taken from a real-time microscopy analysis of autophagosome formation and maturation in response to 50 mM chloroquine (CQ). Note that GFP-LC3 autophagosomes form devoid of LysoTracker Red staining. After formation, these autophagosomes lose their GFP staining as they gradually accumulate LysoTracker Red suggesting that they are either fusing with lysosomes, or that they become acidic.

transformation by compromising cell survival and/or impairing cell proliferation, in a fashion akin to Arf or p53 or the retinoblastoma (Rb) tumor suppressor (Lowe and Sherr, 2003; Sherr, 2006). Alternatively, the regulation of autophagy may be an intrinsic component of the tumor suppressor

254


response. For example, though activation of the Arf/p53 tumor suppressor pathway by oncogenes generally leads to apoptosis, both Arf and p53 activation can induce autophagy (Feng et al., 2005; Reef et al., 2006; Abida and Gu, 2008). Thus, autophagy may also function to restrict oncogenesis. Importantly, autophagy may also limit accumulation of DNA damage, as beclin 1þ/– cells have increased DNA damage, centrosome abnormalities, and can become aneuploid (Mathew et al., 2007). Another unique attribute of the autophagy pathway is that it can play a significant cytoprotective role in established tumors undergoing nutrient limitation and metabolic stress, where it provides essential building blocks (e.g., amino acids and fatty acids) and energy (ATP) sources (Degenhardt et al., 2006). Thus, while dampening the tumor-suppressive properties of the autophagy pathway is advantageous to tumor cells, some level of the pathway must be sustained to respond to acute or chronic metabolic stress. Accordingly, the loss of both alleles of beclin 1 has never been observed in tumors from either mice or men. Furthermore, autophagy also protects cancer cells from metabolic stress-induced necrosis (Mathew et al., 2007), which often accompanies rapid tumor growth and provokes an inflammatory response that appears to be required for tumor progression (Luo et al., 2004; Karin and Greten, 2005; Mathew et al., 2007). Collectively, these observations indicate that autophagy plays important roles in both tumor initiation and progression, and in creating and sustaining the proper tumor microenvironment. Thus, this pathway is a suitable target for anticancer therapeutics on two fronts. However, given its cancer cellintrinsic versus non–tumor cell autonomous effects, a critical consideration is proper management and monitoring of this pathway. Herein, we summarize new reagents and methods that our laboratory has developed to understand the role of autophagy in the different phases of tumor development, maintenance, and treatment. These methods will support efforts focused on defining the complicated interactions between autophagy and cancer, and those that seek to define the activities of specific small molecules as future preventative or therapeutic agents that target this pathway.

2. LC3: A Phenotypic and Functional Marker of Autophagy The formation of autophagic vesicles is dependent upon a unique ubiquitin-like modification system that results in the conjugation of phosphatidylethanolamine (PE) to the C-terminus of Atg4-cleaved LC3 (Ichimura et al., 2000; Kabeya et al., 2000; Kirisako et al., 2000) (Fig. 12.1A). Consequently, GFP fused to the N-terminus of LC3


255

(GFP-LC3) can be used as a phenotypic reporter of autophagic vesicle formation (Kabeya et al., 2000; Mizushima et al., 2003) (see also the chapter by Kimura et al., in this volume). Under normal growth conditions, GFPLC3 is diffuse throughout the cytoplasm and nucleus. Upon the induction of autophagy, GFP-LC3 becomes covalently linked to PE and associates with the initiating phagophore membranes, redistributing GFP-LC3 to newly formed punctate, double-membrane autophagosomes (Fig. 12.1A). GFP-LC3 is an effective phenotypic marker where one can monitor autophagosome formation by fluorescence microscopy, and quantifying GFP-LC3 puncta is an important measure of the induction of autophagy in a wide variety of systems (Mizushima et al., 2004). Microscopy analysis alone, however, fails to distinguish between the induction of autophagy and the inhibition of turnover of autophagic vesicles. Specifically, following their formation, GFP-LC3-labeled autophagosomes fuse with lysosomes, an event that delivers their cargo, including GFP-LC3 localized to the inner membrane of the autophagosome, for degradation (Fig. 12.1B). We have exploited these properties and have developed a flow cytometric-based GFP-LC3 functional assay (see also the chapter by Shvets and Elazar in volume 452). In addition, we have developed several luciferase-LC3 (Luc2p-LC3) constructs that allow measurements of autophagic activity in a high-throughput format. Below we provide the methods where one can: (1) Use GFP-LC3 as a phenotypic marker; and (2) Use GFPLC3 and Luc2p-LC3 fusion proteins as functional markers for monitoring the autophagy pathway. Importantly, these methods can be applied to both in vitro and in vivo models, allowing one to monitor autophagic activity in a wide array of biological systems.

2.1. Real-time imaging of GFP-LC3 We routinely use GFP-LC3 to monitor the formation of autophagosomes by real-time fluorescence microscopy. This allows one to temporally dissect and define the involvement of signaling pathways, autophagic effectors and inhibitors, and/or organelles in autophagic vesicle formation and maturation. Importantly, the overexpression of GFP-LC3 can result in the aggregation of GFP-LC3 independent of the autophagy pathway (Ciechomska and Tolkovsky, 2007; Kuma et al., 2007). To avoid nonspecific aggregation, we routinely transduce cells with MSCV-based retroviruses (Hawley et al., 1992) that allow one to express transgenes such as GFP-LC3 at more physiological levels and that also allow one to express a selectable marker such as the gene encoding puromycin resistance (Puro) through an internal ribosome entry site (IRES) that resides downstream of the transgene. The resulting MSCV-GFP-LC3-IRES-Puro retrovirus then allows one to generate stable cell lines expressing GFP-LC3 by growth in puromycincontaining medium. To further control for nonspecific aggregation of

256


GFP, we also employ a MSCV retroviral construct that directs the independent expression of GFP and LC3, using the IRES to express GFP (MSCV-LC3-IRES-GFP). Here one selects for transduced cells simply by sorting for GFPþ cells using a fluorescence-activated cell sorter (FACS). Using real-time microscopy we have demonstrated that, although CQ inhibits autophagy by disrupting lysosomal-mediated degradation of autophagosomes (Maclean et al., 2008), it actually induces the rapid (within four hours) formation of large, morphologically distinct autophagosomes (Fig. 12.1B). Sometime after their formation, these autophagosomes accumulate the pH sensitive dye LysoTracker Red indicating that they ultimately become acidic and then summarily lose GFP fluorescence (Fig. 12.1B). Although both bafilomycin A1 (Baf-A1) and CQ both inhibit lysosomalmediated degradation of autophagosomes, Baf-A1 treatment is characterized by a much slower accumulation of small GFP-LC3 puncta and an absence of any LysoTracker fluorescence, indicating that it is much more effective than CQ at raising the pH of the lysosome (data not shown). These observations underscore the importance of real-time microscopy in the study of autophagy, as steady state analysis of autophagic activity in response to these agents, for example, using long-lived protein degradation assays (see the chapter by Bauvy et al., in volume 452) or by monitoring LC3-II (i.e., PE-conjugated LC3, Fig. 12.1A) formation by Western blot analyses would not distinguish these potentially important phenotypic differences. For these analyses, we use an Olympus DSU spinning disc confocal microscope (http://www.olympusamerica.com/seg_section/product.asp? product=1009) equipped with a stage enclosed in a Weather Station incubator. Unlike traditional scanning confocal microscopes, the DSU passes laser light through a spinning disc, which contains small slits that act as a virtual pinhole that simultaneously illuminates the entire field. Two advantages to this method are that (1) images can be taken rapidly, up to 15 frames per second; (2) samples are exposed to less light, reducing phototoxic effects seen with traditional confocal microscopy. Glass-bottomed culture dishes (MatTek, Fisher, #P35G.5-14-C) are used for these analyses. 1. The day before imaging, GFP-LC3 expressing cells are plated at an appropriate density (so they are 60%–80% confluent the following day) in DMEM medium supplemented with 25 mM HEPES buffer, pH 7.4, 10% fetal calf serum (FCS), 2 mM L-glutamine, 100 units penicillin, 100 mg/ml streptomycin, 1 mM pyruvate, and 0.1 mM nonessential amino acids (for mouse embryonic fibroblasts [MEFs] also add 55 mM b-mercaptoethanol) and are incubated in a 37 C incubator with 5% CO2. 300,000 cells are plated in a 35-mm dish using Hep2 or HeLa cells, or MEFs. Multiple devices have been developed to maintain proper CO2 concentration for pH stabilization, including an attached aerator, which bubbles 5% CO2 mixed with normal atmosphere through prewarmed


2.

3.

4.

5.

6.

257

water, thus delivering prewarmed, humidified CO2 to the cell culture dish. The addition of HEPES buffer helps maintain the correct pH. Additionally, when using small volumes of cell culture media, one can overlay the media with a thin layer of mineral oil, to prevent evaporation. Prior to imaging, cells are stained with LysoTracker Red (Invitrogen, #L-7528) to visualize the role of acidic lysosomal compartments in relation to autophagic vesicle formation and turnover. To stain the cells, LysoTracker is added to the medium at a concentration of 1 mM and incubated for 30–45 min in a 37 C 5% CO2 incubator. Cells are then washed twice to remove excess LysoTracker from the medium. For adherent cells, the media is simply exchanged twice with fresh pre-warmed media. If cells are nonadherent, the cells are centrifuged and are suspended in prewarmed media, centrifuged again, and resuspended in fresh prewarmed media and are replated. After LysoTracker staining, the plates are moved to the incubator on the microscope for 15–20 min to allow the plates to reach thermal equilibrium with the stage; this prevents thermal fluctuations, which can result in small movements in the plate and may cause cells to drift out of focus. After equilibration, initial images are collected to document the state of GFP-LC3 vesiculation before, for example, amino acid starvation or drug treatment. The most commonly used agents that affect the autophagy pathway used in our laboratory are chloroquine (20–50 mM; Sigma, #C6628), bafilomycin A1 (100 nM; Sigma, #11711), rapamycin (5 mg/ ml; Sigma, #R0395), and 3-methyladenine (3-MA, 5 mM; Sigma, #08592). As 3-methyladenine is difficult to solubilize, it should be dissolved in cell culture medium at 37 C just before treatment. For amino acid starvation, cells are rinsed with prewarmed Earl’s balanced salt solution (EBSS; Invitrogen, #14155048) supplemented with 25 mM HEPES buffer, pH 7.4, twice before incubating the cells in EBSS for the duration of the experiment. When using 3-MA to inhibit autophagy, cells are washed twice with prewarmed medium containing 5 mM 3-MA just prior to culture in EBSS, or in normal medium containing reagents that affect the autophagy pathway. To monitor vesiculation in response to different pharmacological agents, images are collected every 5 min for 4–6 h. To monitor vesicle formation and vesicle dynamics, images are typically acquired in the 15–30 ms range. Importantly, at the end of any acquisition, one should always move the objective to a field that has not been previously excited by the laser, to ensure that GFP-LC3 dynamics are not the result of excessive light exposure. Microscopes for real-time microscopy are now widespread and each manufacturer has their own individual software for image acquisition and analysis. We are currently using software developed by Intelligent Imaging Innovations (3i), which is standard for Olympus microscopes.

258


2.2. High content analysis of GFP-LC3 vesiculation An important new arena for the autophagy field is the discovery and development of specific small molecule inhibitors and activators of the autophagy pathway. A key component of such campaigns is the ability to analyze large numbers of compounds using GFP-LC3 phenotypic screens. We have developed a medium-throughput method to measure GFP-LC3 vesiculation in response to different pharmacological agents using high-content screening. Excitingly, these methods are also applicable for siRNA and cDNA screens to identify new players in the autophagy pathway. High content screening (HCS) is a method of screening fixed or living cells based on phenotype. It combines an automated microscope platform with image analysis software for the rapid acquisition and analysis of microscopy images in a microtiter plate format. Using this system, it is possible to accurately quantify GFP-LC3 vesiculation in a rapid and nonbiased manner. It is well accepted that amino acid starvation activates the autophagy pathway, yet using HCS we have demonstrated that amino acid starvation actually results in a slight reduction in GFP-LC3 vesiculation when compared to untreated cells (Figs. 12.2A–B). In contrast, CQ treatment resulted in the accumulation of a large number of autophagosomes at the same 6-h time point. Importantly, the accumulation of autophagic vesicles in response to CQ is autophagy gene-dependent (data not shown). CQ inhibits autophagic vesicle turnover by the lysosome, yet CQ also induces autophagic vesicle formation at early time points (Fig. 12.1B) (Maclean et al., 2008). Regardless of the exact mechanism, CQ-induced vesiculation can be used to identify new agents that inhibit autophagic vesiculation and new genes that induce or suppress the autophagy pathway. To measure autophagic vesiculation in response to CQ treatment by HCS, we use the following protocol: 1. We first generate stable cell lines using MSCV-GFP-LC3B-I-Puro or MSCV-LC3B-I-GFP retroviruses (see above) and then plate cells in DMEM medium (as previously) at 25,000 cells per well in a Packard View 96-well plate (PerkinElmer viewplate, #6005182). 2. The following day, the medium is replaced with pre-warmed complete medium, EBSS (amino acid [AA] starvation media), or medium containing 50 mM CQ and incubated for 6 h at 37 C with 5% CO2. 3. Cells are then washed twice with PBS and fixed with 4% paraformaldehyde for 10 min. 4. The cells are then washed twice with PBS and nuclei are stained with Hoechst 33442 (5 mg/ml; Invitrogen, #H3570) for 20 min and again washed twice with PBS. 5. Images are acquired on the InCell 1000 instrument (GE Healthcare) using a Q505LP dichroic mirror and corresponding UV/FITC filter sets with a 20x objective. Typically, we collect 8 independent images per well, which, for a 96-well plate, is a total of 768 images.

259


A

GFP-LC3

Identified vesicles

UTD

EBSS

CQ

Figure 12.2 High content analysis of GFP-LC3 vesiculation. (A) GFP-LC3-expressing Hep2 cells were either left untreated (UTD), were starved of amino acids by incubation in EBSS, or were treated with 50 mM CQ for 6 h. The panels at left are representative images of GFP-LC3 vesiculation that occurred under each condition.The images in the panels at right correspond to those shown at left that have been analyzed using the InCell-based algorithm and shows the autophagosomes identified in blue. (B) Quantification of GFP-LC3 vesiculation in untreated, EBSS-treated, or CQ-treated Hep2 cells. To accurately quantify vesiculation, 48 individual images for each treatment were collected, containing approximately 30 cells/image, and GFP-LC3-associated vesicles were quantified using the InCell developer toolbox algorithm.

Autophagosomes are quantified using the InCell Developer Toolbox software. First, numbers of nuclei are quantified by object segmentation using a Kernel size of 31 and a sensitivity of 1. Post-processing of images is performed using an erosion of 11 and sieve greater than 50-mm2. GFP-LC3positive autophagosomes are then quantified by object segmentation using a Kernel size of 5 and a sensitivity of 1. This algorithm accurately counts GFP positive vesicular structures in a nonbiased manner (Fig. 12.2A–B). Data shown in Fig. 12.2B are the number of autophagosomes per cell (represented by nuclei count), yet this software is also capable of providing a quantitative measure of both vesicle size, and fluorescent intensity represented by the fluorescence multiplied by vesicle area. These methods make it possible to quantitatively assess morphological differences in autophagic vesicles in response to a wide array of agents (data not shown). One must be careful in interpreting quantification of autophagic vesiculation as a measure of autophagy. For example, though reductions in

260


GFP-LC3 vesiculation following amino acid starvation could suggest that this inhibits autophagy, a wealth of data has demonstrated that this is not the case. Thus one should assess autophagic activity by multiple methods to determine the control of autophagy in response to specific stimuli (Klionsky et al., 2008). Indeed, the most likely explanation of these results is that GFPLC3-positive autophagic vesicles are turned over at a much more rapid rate following amino acid starvation, which results in a steady state decrease in GFP-LC3-associated vesicles. Thus, while monitoring autophagy with this type of phenotypic screen can provide valuable of data, it is necessary to assess autophagy-specific degradation when screening for activators and inhibitors of this pathway.

2.3. Monitoring GFP-LC3 by flow cytometry In addition to its use for microscopy analysis, GFP-LC3 can also be used to measure the turnover of autophagic vesicles by flow cytometry. Using primary nontransformed MEFs or Hep2 cells, we have demonstrated that activation of autophagy, for example, following amino acid deprivation, results in the rapid degradation of GFP-LC3, as documented by reductions in mean fluorescence measured by flow cytometry (Fig. 12.3A). Importantly, reduction in the mean fluorescence following the induction of autophagy is Atg7-dependent (Fig. 12.3B). In addition, using MSCV-LC3-IRES-GFP virus infected cells as a control, where GFP degradation is not linked to that of LC3, this method allows for the measure of autophagic function even when it is not possible to specifically silence an autophagy gene (Fig. 12.3A). Therefore, in addition to being a phenotypic marker for autophagy, GFPLC3 can also be used as a functional marker for autophagic degradation. Finally, these constructs can also be used to monitor autophagic activity in vivo, for example in precancerous versus malignant tumor cells engineered to express GFP-LC3, or in GFP-LC3-expressing tumor cells treated with therapeutic agents. For these experiments, we use the following protocol: 1. Hep-2 cells or primary MEFs are transduced with either MSCVGFPLC3-IRES-PURO or MSCV-LC3-IRES-GFP retroviruses, and are then plated in a 24-well plate at 50,000 cells/well and cultured overnight in DMEM media (see previously). 2. The following day, cells are either left untreated, or are treated with various agents that effect autophagy, such as EBSS, 5 mM 3-MA, 3-MA in EBSS, 50 mM CQ, or 100 nM Baf-A1, in 0.5 ml of media. Treatment is anywhere from 2–24 h, and earlier time points are generally better to avoid complications arising from cell death. 3. Cells are then harvested at selected intervals after first collecting the medium from each well into 1275 mm polystyrene tubes (BD Falcon,

261


A 100

100

% of Maximium

80

UTD

6 h EBSS

80

60

60

40

40

20

20 0

0 101

102

103

104

Hep2 GFP-LC3

105

101

102 103 104 105 Hep2 LC3-IRES-GFP

B

% of untreated

100

50

0 H:

0

6

10

0

6 10

Atg7 −/−

Atg7 +/+ MEFs

Figure 12.3 Monitoring autophagic degradation by flow cytometry. (A) Hep2 cells were transduced with either MSCV-GFP-LC3-IRES-Puro or MSCV-LC3-IRES-GFP retroviruses. These cells were either left untreated (blue) or were deprived of amino acids by incubating the cells in EBSS for 6 h (grey). GFP fluorescence was then quantified by flow cytometry. GFP-LC3 levels were reduced following amino acid deprivation, while GFP alone was unaffected. (B) To confirm that reductions in GFP-LC3 fluorescence were the result of the activation of autophagy, Atg7þ/þ and Atg7^/^ MEFs were transduced with MSCV-GFP-LC3-IRES-Puro retrovirus and then cultured in EBSS for 6 or 10 h. The reduction of GFP-LC3 mean fluorescence only occurred in Atg7þ/þ MEFs; therefore, the reduction in GFP-LC3 fluorescence provoked by amino acid deprivation is autophagy dependent.

#352058) on ice. Wells are subsequently washed with 0.5 ml of PBS/ well; the wash is added to medium that was previously collected to avoid biasing the analysis to tightly adherent cells. 4. Fully adherent cells are trypsinized with 0.5 ml of 0.05% trypsin (Invitrogen, #25300-054) for 5 min at 37 C and are examined under

262


the microscope to ensure complete trypsinization. The trypsinized cells are then pipetted several times to achieve a single-cell suspension, which is then added to the previously harvested supernatant fractions and washed; trypsin is neutralized by the addition of 3 ml of cold wash buffer (PBS containing 1% FCS) and cells are centrifuged at 300g for 5 min. 5. The supernatant fraction is removed, and the cells are resuspended in approximately 100 ml of wash buffer. Cells are immediately analyzed on a FACSAria for forward versus side scatter and for GFP fluorescence with the collection of 10,000 live-gated events. Median fluorescence intensity (MFI) is then determined for GFP-expressing cells.

2.4. Luciferase LC3: A high-throughput method to monitor autophagic activity Monitoring GFP-LC3 mean fluorescence in response to amino acid starvation confirms findings that LC3 is delivered along with autophagosomes to the lysosome for degradation (Tanida et al., 2004). Therefore, we reasoned that a luciferase-LC3 fusion (Luc2p-LC3) would allow one to monitor and quantify autophagy-dependent degradation in a high-throughput manner. We chose to use the highly unstable Luc2p luciferase gene that has been engineered to contain the PEST domain from the ornithine decarboxylase gene (ODC), which has one of the shortest half-lives known for any protein (Fan and Wood, 2007). This PEST domain directs the rapid destruction of Luc2p in a proteasome-dependent manner, and the fusion of Luc2p to LC3 stabilizes this protein, effectively shuttling it into the autophagy pathway for degradation. LC3 conjugation to PE and subsequent incorporation into autophagosomes requires a C-terminal glycine residue (Fig. 12.1A). Luc2p alone is rapidly degraded in response to amino acid starvation and this is not blocked by 3-MA, which inhibits autophagy (Fig. 12.4A). By contrast, Luc2p-LC3 degradation is largely autophagy-dependent, as its turnover is largely inhibited by 3-MA (Fig. 12.4A). As a control for autophagy-specific degradation, we also generated a Luc2p-LC3G120A mutant that cannot be covalently modified by PE, thus uncoupling its degradation from the autophagy pathway. Accordingly, the rate of turnover of Luc2p-LC3G120A is protracted relative to that of Luc2p-LC3 following the removal of amino acids (Fig. 12.4A). To verify that turnover of Luc2p-LC3 is indeed mediated via the autophagy pathway, we also assessed its turnover in wild-type and Atg7/– MEFs transduced with MSCV-Luc2p-LC3-IRES-Puro retrovirus. Notably, turnover of Luc2p-LC3 induced by withdrawal of amino acids was totally abolished in Atg7/– MEFs (Fig. 12.4B). This collection of LC3 fusion proteins provides powerful tools that allow one to monitor and quantify autophagy. Importantly, these assays

263


A

EBSS

EBSS + 3-MA 100

% of untreated

% of untreated

100 80 60 40 20

Luc2P

80

Luc2P-LC3

60

Luc2P-LC3G120A

40 20 0

0 0

5

10

15

20

0

25

5

10

15

20

25

H

H

B % of untreated

100

50

0 UTD EBSS

Atg7+/+

UTD EBSS

Atg7-/-

Figure 12.4 Measuring autophagic activity using luciferase-LC3 fusion proteins. (A) Luciferase activity was measured in Hep2 cells transduced with MSCV-Luc2p-IRESGFP, MSCV-Luc2p-LC3-IRES-GFP, or MSCV-Luc2p-LC3G120A-IRES-GFP retroviruses.Transduced cells were starved of amino acids by incubation in EBSS alone or in combination with the autophagy inhibitor 3-MA. Interestingly, Luc2p alone is degraded in response to amino acid deprivation, yet its degradation was unaffected by the 3-MA. Luc2p-LC3 is degraded in response to amino acid deprivation, while the degradation of the Luc2p-LC3G120A mutant was markedly delayed. In contrast to Luc2p, the degradation of both Luc2p-LC3 and Luc2p-LC3G120Awas inhibited by 3-MA; thus, these two fusion proteins are degraded by a mechanism different than that of Luc2p. (B) To confirm that Luc2p-LC3 is an accurate reporter of autophagy, luciferase activity was measured in Atg7 wild-type and null MEFs that were transduced with MSCV-Luc2pLC3-IRES-GFP virus.Transduced cells were cultured in EBSS for 6 h.The reduction in luciferase activity in response to EBSS treatment is autophagy dependent, as Luc2pLC3-driven luciferase activity was not reduced in Atg7/ MEFs when cultured in EBSS.

also demonstrate suitable dynamic range and z-scores, which make them amenable to medium- or high-throughput screening for antagonists and agonists of the autophagy pathway. To perform luciferase-based screening using the Luc2p-LC3 and Luc2pLC3G120A constructs, it is necessary to generate stable cell lines, as these constructs failed to behave as accurate markers of autophagic activity in transient transfection experiments. Importantly, virtually all transfection reagents contain chloroquine, which affects the autophagy pathway in rather profound ways, and it is quite likely that the failure of transient transfection approaches is attributable to the presence of CQ.

264


1. Once stable cell lines are generated, cells are plated in a CostarTM white 96-well plate at 50,000 cells/well in 100 ml of DMEM medium (see previously). All treatments are done in triplicate, and the vehicles for all compounds under investigation, are included as controls. As a negative control, 6 wells are left untreated, and as a positive control, 6 wells are incubated with EBSS after prewashing twice with prewarmed EBSS. Additionally, we use 5 mM 3-MA to inhibit autophagic degradation in response to amino acid starvation. These wells are first washed twice with EBSS containing 5 mM 3-MA and then incubated in EBSS þ 3-MA for the indicated time. Usually, we follow luciferase activity with respect to time using one plate or set of plates per time point. To date, we have found that 6- or 10-h intervals give the best dynamic range between the Luc2p-LC3 and Luc2p-LC3G120A constructs (Fig. 12.4A). Firefly luciferase produces light through the oxidation of the substrate D-luciferin in the presence of ATP and Mg2þ. This reaction has the highest quantum yield of all bioluminescent reactions (Fan and Wood, 2007). 2. To measure luciferase activity, we use PerkinElmer’s Britelite plus reagent (catalog #6016767). For each 96-well plate, we thaw 10 ml of Britelite reagent and bring it to room temperature. An equal volume of reagent is then added to each well (100 ml). 3. The plate is incubated at room temperature for 2–4 min. 4. Plates are read using an Analyst GT plate reader (Molecular Devices). Data are analyzed by subtracting the RLU values for vehicle alone from experimental values and are then normalized to the untreated controls. As a result, data are expressed as percent of control luciferase activity.

3. Assessing the Role of Autophagy in Em-Myc-Driven Lymphoma The Em-Myc transgenic mouse express increased levels of c-Myc in all B lymphoid cells by virtue of the Em immunoglobulin heavy chain enhancer (Adams et al., 1985), a scenario that recapitulates MYC activation by the t (8;14) translocation of Burkitt lymphoma. Em-Myc mice develop rapid, clonal, pre B/immature B-cell lymphoma, with a mean survival of approximately 120 days (Eischen et al., 1999). Similarly, beclin 1þ/– mice are also tumor prone and can develop various types of mature B-cell lymphoma, but with a mean latency >13 months (Qu et al., 2003; Yue et al., 2003). Finally, beclin 1 heterozygosity accelerates the rate of lymphoma development in Em-Myc mice (Dorsey and Cleveland, unpublished data), yet the loss of the remaining beclin 1 allele in either mice or men has not been reported. Collectively, these data suggest that beclin 1 functions as a haploinsufficient


265

tumor suppressor, but that the complete loss of the autophagy pathway may be selected against during tumorigenesis. Studies using immortal, Atg5-null cells transformed with the E1A and Ras oncogenes have suggested that autophagy delays tumorigenesis by restricting necrosis, which induces an inflammatory response that promotes tumor progression (Degenhardt et al., 2006). However, these experiments fail to address the role of autophagy in regulating the initial phases of tumor development en route to transformation. Importantly, it is not currently known whether autophagy is necessary, at some level, for transformation and tumor development. These studies are complicated by the fact that beclin 1 knockout mice are embryonic lethals (E7.5–8.5), whereas Atg5 and Atg7 knockouts die soon after birth. Furthermore, the difference in the survival between beclin 1–/– versus Atg5- and Atg7-null mice suggests that beclin 1 may have other functions in development unrelated to autophagy. To address whether autophagy plays an in vivo role in tumor initiation and development, and to circumvent issues regarding lethality, we are using a transplant model, where one reconstitutes the hematopoietic system of lethally irradiated mice with hematopoietic cells derived from the fetal livers of Em-Myc transgenic mice that are either wild type, heterozygous, or null for Atg7 (Fig. 12.5A). Generation of hematopoietic chimeras is a wellestablished and highly valuable tool in cancer research. Here we outline the method that we are currently employing to address these questions.

3.1. Hematopoietic cell isolation and transplantation During vertebrate embryogenesis the fetal liver is the major site of hematopoiesis, from midgestation to the time of birth: This allows mouse hematopoietic stem cells to be easily isolated from embryonic day E11.5–E18.5 and used for transplant studies into lethally irradiated recipients (Schmitt et al., 2002). To isolate fetal livers, timed matings are established to generate embryos with known gestation dates. To eliminate the need to look for transient vaginal plugs that indicate mating has occurred, mating pairs are left together for 48 h over 2 successive nights, after which males are separated from females. Therefore, embryos isolated on the fifteenth or sixteenth day after the initial day of mating will either be E14.5 or E15.5. 1. Pregnant females are humanely euthanized using CO2, followed by cervical dislocation. 2. The abdomen is soaked in 70% ethanol, and surgically opened to expose the reproductive organs. 3. The uterine horns containing the embryos are collected, placed in a 50-ml tube containing 30 ml of cold 2% FCS/PBS, and kept on ice.

266 A


CD45.1 congenic recipients

C D45.2 donors

Eµ-Myc ; A tg7+/−

2× 500 Rad

Harvest fetal livers

X

Latency period

E13.5−15.5

Atg7+/−

B

Lymphoma onset

(Host = CD45.1; Donor = CD45.2)

100

% total PBMCs

80

60

T cells (CD3+) B cells (CD19+)

40

20

0 Host

Donor

Atg7+/+

Host

Donor

Atg7+/−

Host

Donor

Atg7−/−

Figure 12.5 Hematopoietic reconstitution using fetal liver transplantation. (A) Hematopoietic reconstitution using embryonic fetal livers. Timed-matings of CD45.2 Em-Myc/Atg7þ/^ and Atg7þ/^ animals were performed and embryonic day 13.5^ 15.5 fetal livers were collected and single cell suspensions were isolated. CD45.1recipients were exposed to 500 Rads of full body irradiation twice 6^8 h apart. After the second round of irradiation, 5 million fetal liver-derived hematopoietic cells were delivered to the congenic recipients via retro-orbital injection and observed for engraftment and subsequent lymphoma development. (B) Assessment of engraftment of donor Atg7þ/þ, Atg7þ/^, and Atg7^/^ lymphoid cells 6 weeks after transplant. Total peripheral blood mononuclear cells (PBMCs) were isolated, stained with antibodies for either CD45.1 (host) or CD45.2 (donor) along with CD3 (T-cells) and CD19 (B-cells), and analyzed by flow cytometry. Engraftment with all 3 donor genotypes occurred, but levels of engraftment were significantly reduced in mice receiving Atg7^/^ hematopoietic cells.

4. Isolated uterine horns are then taken to a sterile tissue culture hood, removed from the 50-ml tube, and placed in a 10-cm dish on ice containing cold 20 ml of 2% FCS/PBS. 5. Embryos are separated from one another with sterile scissors by cutting directly through the uterine horn between each embryo. 6. Individual embryos are then transferred to their own 10-cm dish where they are subsequently dissected away from both the uterine horn and amniotic sac, which are discarded. 7. The fetal liver is then removed with two curved forceps. Briefly, one forcep is used to puncture the embryo just in front of the spine, and the


8. 9. 10. 11.

12.

13. 14. 15. 16.

17.

267

other is used to pull back the skin on the abdomen. After the skin is pulled away, the fetal liver is removed from the abdomen by placing the curved forcep behind the fetal liver and gently moving it away from the rest of the embryo. 10 ml of cold 2% FCS/PBS are added to the 10-cm dish containing the isolated fetal liver and pipetted up and down several times to achieve a single-cell suspension. Unwanted connective tissue is removed by passing the cell suspension through a 70- to 100-mm filter. Cells are diluted 1:4 and counted using a hemocytometer. 20 ml of the cell suspension is added to 20 ml of a 2% acetic acid solution. Immediately after gentle mixing, 20 ml of PBS is added to neutralize the solution and 20 ml of 0.5% trypan blue (Invitrogen, #15250-061) is added as a viability dye. We typically recover 15–25 million hematopoietic cells per fetal liver with a viability of 70%–95%. In general, 5 million fetal liver-derived hematopoietic cells are transplanted per recipient, allowing for at least 2 transplants per embryo. Recipient mice are congenic for CD45.1, whereas donor hematopoietic cells express the CD45.2 alloantigen. This allotypic difference allows for later distinction between hematopoietic cells that are derived from the donor (CD45.2, mAB clone 104, BDBiosciences) and those remaining in the irradiated host (CD45.1, mAb clone A20, BDBiosciences) by specific antibody staining and FACS analyses. CD45.1 recipients are irradiated with a total of 1000 Rads (10 Gy) split into two equal doses separated by 3–5 h. Split-dose irradiation prevents much of the gut toxicity normally associated with lethal irradiation (Alpdogan et al., 2003). Once mice are irradiated, isolated fetal liver hematopoietic cells are centrifuged at 300g, at 4 C for 5 min. The supernatant fraction is removed and replaced with cold PBS containing 20 units of heparin per ml. Cells are resuspended in a volume to yield 5 million cells/100 ml. Irradiated recipients are then anesthetized using an IMPAC6 anesthesia machine from VetQuip (Pleasanton, CA) with settings at 2%–3% isoflurane and 2 liters/h of oxygen). Five million cells in 100 ml are then injected retro-orbitally, using a ½ inch, 27-gauge needle. Cells may also be administered via tail vein injection, but retro-orbital injections are simple to master and give excellent success rates with high levels of engraftment. Recipient mice are allowed to recover under observation until mobile. In some cases, a heat lamp may be used to assist in recovery and prevent hypothermia, but such use must be carefully monitored.

268


18. Mice are kept on 35 mg/ml BaytrilTM (enrofloxacin, Bayer Health Care, #08713254-186599)-treated water for 4 weeks posttransplant to prevent infection that can occur during hematopoietic reconstitution. 19. At 6–8 weeks, mice are bled retro-orbitally and peripheral blood is analyzed to determine basal levels of engraftment (section 3.2). With the above procedure, death due to failed engraftment is rare.

3.2. Assessing hematopoietic chimerism 1. At 6–8 weeks posttransplantation, blood is collected from the retroorbital sinus of lethally irradiated CD45.1 animals that were reconstituted with hematopoietic stem cells from CD45.2 donor fetal livers as described previously. Mice are anaesthetized until they are unconscious with isoflurane/oxygen and approximately 100 ml of peripheral blood is collected using heparinized capillary glass tubes (Fisher, #22-362-566) and placed in EDTA-coated collection vials (BD Biosciences, #367835). Similarly harvested peripheral blood from a C57Bl/6 (CD45.2) mouse is used as a control. 2. Then, 75 ml of whole blood is transferred to a 5-ml 1275-mm polystyrene tube. Red blood cells are lysed with 4 ml of Tris-ammonium chloride lysing buffer (144 mM NH4Cl, 17 mM Tris-HCl, pH 7.2), and cells are immediately centrifuged at 300g for 5 minutes, washed once with 1%FCS/PBS wash buffer, repelleted, then resuspended in 200 ml of 1%FCS/PBS and kept on ice. 3. Cell suspensions are split into 2 polystyrene tubes approximately 100 ml per tube, and blocked with 0.25 mg of BD Fc Block (BD Biosciences, #553141) for 5 min. 4. Cells are then immediately stained with a mixture of CD19-PE (0.1 mg; BD Biosciences, #553786), CD3e-PECy5.5 (0.1 mg; eBioscience, #35-0031) and either CD45.1-biotin (0.5 mg; BD Bioscience, #553774) or CD45.2-biotin (0.5 mg; BD Bioscience, #553771) for 30 minutes on ice. Unstained and single-color C57Bl/6 control tubes are generated using equal amounts of single antibodies for the purpose of setting the compensation of the flow cytometer. 5. Cells are then washed twice with 1%FCS/PBS and stained with 0.2 mg of streptavidin-APC-Alexa750 (Invitrogen, #SA1027) for 30 min on ice. 6. Cells are then washed twice with 1%FCS/PBS, and resuspended in 100 ml of 1%FCS/PBS. 7. Samples are then analyzed on a FACSCanto flow cytometer (BD Biosciences). Through gating, 20,000–50,000 live lymphocyte events are collected for subsequent analysis. Interestingly, we typically see a slight decline in engraftment when using Atg7þ/– fetal liver cells, and a further reduction in the complete absence of Atg7 (Fig. 12.5B).


269

Using this system recipient mice receiving Em-Myc transgenic stem cells will develop B cell lymphoma, usually within 3 months (Schmitt et al., 2002), and here it is thus possible to evaluate the effects of heterozygosity or total loss of autophagy components on Myc-driven lymphomagenesis. Further, by evaluating the hematopoietic phenotypes of control non-transgenic hematopoietic recipients one can also determine the effects of heterozygosity or loss of Atg7, or Atg5, on hematopoietic development and homeostasis.

4. Concluding Remarks and Future Perspectives Autophagy likely contributes to all aspects of cancer biology, including tumor initiation and progression, and in the maintenance of the malignant state. Cancer cells are, by their very nature, metabolically stressed. As they grow and expand, this problem becomes more acute with increasing hypoxia and nutrient deprivation, and autophagy is essential for surviving these stresses. Furthermore, autophagy also plays key roles in controlling the tumor microenvironment, the interaction of tumor cells with the immune system, and in the therapeutic response. Indeed, virtually all chemotherapeutic agents and radiotherapies induce such metabolic stress, and concomitant inhibition of autophagy potentiates the action of several known therapeutics. However, there are currently no small molecule inhibitors specific for the autophagy pathway. In addition to their use for understanding how autophagy contributes to the development and treatment of cancer, the methods presented herein provide key tools that allow for the development of screens for such small molecule autophagy antagonists and agonists. Finally, these methods can also be easily used to monitor and quantify the activity of the autophagy pathway in many other arenas of biology.

ACKNOWLEDGMENTS We thank the Lehman Brothers Foundation–Dorothy Rodbell Cohen Cancer Research Fund for providing funding for the development of assays to monitor the autophagy pathway. We also thank Scot Ouellette for helpful discussion and critique regarding these methods. In addition, we thank Juliana Conkright for technical support and the Cell-Based Screening facility at the Scripps Research Institute-Florida and Mark A. Hall for generation of an MSCV-gateway-I-GFP vector, which facilitated cloning of the MSCV-Luc2pLC3-IGFP construct. F.C.D. was also supported by a Ruth L. Kirschstein Fellowship from the National Cancer Institute, NIH and J.L.C. by NIH/NCI grant CA076379.

REFERENCES Abida, W. M., and Gu, W. (2008). p53-Dependent and p53-independent activation of autophagy by ARF. Cancer Res. 68(2), 352–357.

270


Adams, J. M., Harris, A. W., et al. (1985). The c-myc oncogene driven by immunoglobulin enhancers induces lymphoid malignancy in transgenic mice. Nature 318(6046), 533–538. Alpdogan, O., Muriglan, S. J., et al. (2003). IL-7 enhances peripheral T cell reconstitution after allogeneic hematopoietic stem cell transplantation. J. Clin. Invest. 112(7), 1095–1107. Alt, J. R., Greiner, T. C., et al. (2003). Mdm2 haploinsufficiency profoundly inhibits Mycinduced lymphomagenesis. EMBO J. 22(6), 1442–1450. Amaravadi, R. K., Yu, D., et al. (2007). Autophagy inhibition enhances therapy-induced apoptosis in a Myc-induced model of lymphoma. J. Clin. Invest. 117(2), 326–336. Briceno, E., Reyes, S., et al. (2003). Therapy of glioblastoma multiforme improved by the anti-mutagenic chloroquine. Neurosurg. Focus 14(2), e3. Carew, J. S., Nawrocki, S. T., et al. (2007). Targeting autophagy augments the anticancer activity of the histone deacetylase inhibitor SAHA to overcome Bcr-Abl-mediated drug resistance. Blood 110(1), 313–322. Ciechomska, I. A., and Tolkovsky, A. M. (2007). Non-autophagic GFP-LC3 puncta induced by saponin and other detergents. Autophagy 3(6), 586–590. Degenhardt, K., Mathew, R., et al. (2006). Autophagy promotes tumor cell survival and restricts necrosis, inflammation, and tumorigenesis. Cancer Cell 10(1), 51–64. Eischen, C. M., Weber, J. D., et al. (1999). Disruption of the ARF-Mdm2-p53 tumor suppressor pathway in Myc-induced lymphomagenesis. Genes Dev. 13(20), 2658–2669. Fan, F., and Wood, K. V. (2007). Bioluminescent assays for high-throughput screening. Assay Drug Dev. Technol. 5(1), 127–136. Feng, Z., Zhang, H., et al. (2005). The coordinate regulation of the p53 and mTOR pathways in cells. Proc. Natl. Acad. Sci. USA 102(23), 8204–8209. Hawley, R. G., Fong, A. Z., et al. (1992). Transplantable myeloproliferative disease induced in mice by an interleukin 6 retrovirus. J. Exp. Med. 176(4), 1149–1163. Ichimura, Y., Kirisako, T., et al. (2000). A ubiquitin-like system mediates protein lipidation. Nature 408(6811), 488–492. Kabeya, Y., Mizushima, N., et al. (2000). LC3, a mammalian homologue of yeast Apg8p, is localized in autophagosome membranes after processing. EMBO J. 19(21), 5720–5728. Karin, M., and Greten, F. R. (2005). NF-kB: linking inflammation and immunity to cancer development and progression. Nat. Rev. Immunol. 5(10), 749–759. Kirisako, T., Ichimura, Y., et al. (2000). The reversible modification regulates the membrane-binding state of Apg8/Aut7 essential for autophagy and the cytoplasm to vacuole targeting pathway. J. Cell. Biol. 151(2), 263–276. Klionsky, D. J., Abeliovich, H., et al. (2008). Guidelines for the use and interpretation of assays for monitoring autophagy in higher eukaryotes. Autophagy 4(2), 151–175. Kuma, A., Matsui, M., et al. (2007). LC3, an autophagosome marker, can be incorporated into protein aggregates independent of autophagy: caution in the interpretation of LC3 localization. Autophagy 3(4), 323–328. Liang, X. H., Jackson, S., et al. (1999). Induction of autophagy and inhibition of tumorigenesis by beclin 1. Nature 402(6762), 672–676. Lowe, S. W., and Sherr, C. J. (2003). Tumor suppression by Ink4a-Arf: progress and puzzles. Curr. Opin. Genet. Dev. 13(1), 77–83. Luo, J. L., Maeda, S., et al. (2004). Inhibition of NF-kB in cancer cells converts inflammation-induced tumor growth mediated by TNFa to TRAIL-mediated tumor regression. Cancer Cell 6(3), 297–305. Maclean, K. H., Dorsey, F. C., et al. (2008). Targeting lysosomal degradation induces p53dependent cell death and prevents cancer in mouse models of lymphomagenesis. J. Clin. Invest. 118(1), 79–88.


271

Marino, G., Salvador-Montoliu, N., et al. (2007). Tissue-specific autophagy alterations and increased tumorigenesis in mice deficient in Atg4C/autophagin-3. J. Biol. Chem. 282 (25), 18573–18583. Mathew, R., Karantza-Wadsworth, V., et al. (2007). Role of autophagy in cancer. Nat. Rev. Cancer 7(12), 961–967. Mathew, R., Kongara, S., et al. (2007). Autophagy suppresses tumor progression by limiting chromosomal instability. Genes Dev. 21(11), 1367–1381. Mizushima, N., Kuma, A., et al. (2003). Mouse Apg16L, a novel WD-repeat protein, targets to the autophagic isolation membrane with the Apg12-Apg5 conjugate. J. Cell Sci. 116 (Pt 9), 1679–1688. Mizushima, N., Yamamoto, A., et al. (2004). In vivo analysis of autophagy in response to nutrient starvation using transgenic mice expressing a fluorescent autophagosome marker. Mol. Biol. Cell 15(3), 1101–1111. Qu, X., Yu, J., et al. (2003). Promotion of tumorigenesis by heterozygous disruption of the beclin 1 autophagy gene. J. Clin. Invest. 112(12), 1809–1820. Reef, S., Zalckvar, E., et al. (2006). A short mitochondrial form of p19ARF induces autophagy and caspase-independent cell death. Mol. Cell 22(4), 463–475. Schmitt, C. A., Fridman, J. S., et al. (2002). Dissecting p53 tumor suppressor functions in vivo. Cancer Cell 1(3), 289–298. Sherr, C.J (2006). Divorcing ARF and p53: an unsettled case. Nat. Rev. Cancer 6(9), 663–673. Sotelo, J., Briceno, E., et al. (2006). Adding chloroquine to conventional treatment for glioblastoma multiforme: A randomized, double-blind, placebo-controlled trial. Ann. Intern. Med. 144(5), 337–343. Tanida, I., Ueno, T., et al. (2004). LC3 conjugation system in mammalian autophagy. Int. J. Biochem. Cell Biol. 36(12), 2503–2518. Yue, Z., Jin, S., et al. (2003). Beclin 1, an autophagy gene essential for early embryonic development, is a haploinsufficient tumor suppressor. Proc. Natl. Acad. Sci. USA 100(25), 15077–15082.