Measuring Autophagy in Stressed Cells

Abstract

Macro-autophagy is a major catabolic process in the cell used to degrade protein aggregates, dysfunctional organelles and intracellular pathogens that would otherwise become toxic. Autophagy also generates energy and metabolites for the cell through recycling of degraded autophagosomal cargo, which can be particularly important for cell viability under stress. The significance of changes in the rates of autophagic flux for cellular function and disease is being increasingly appreciated, and interest in measuring autophagy in different experimental systems is growing accordingly. Here, we describe key methodologies used in the field to measure autophagic flux, including monitoring LC3 processing by western blot, fluorescent cell staining, and flow cytometry, in addition to changes in the levels or posttranslational modifications of other autophagy markers, such as p62/Sqstm1 and the Atg5–Atg12 conjugate. We also describe what cellular stresses may be used to induce autophagy and how to control for changes in the rates of autophagic flux as opposed to inhibition of flux. Finally, we detail available techniques to monitor autophagy in vivo.

1 Introduction

Methodologies used to quantify autophagy in experimental systems are continually being developed as our knowledge of the role of this key cellular process in growth control increases. Autophagosomes were first identified by transmission electron microscopy (TEM) [1, 2], and TEM remains a rigorous method to characterize the presence of ongoing autophagy [3]. However, TEM does not allow for the quantitative measurement of autophagic flux, nor is it particularly suited for routine visualization of autophagosomes, for example in co-localization analysis. Experimental visualization of autophagosomes and measurement of autophagic flux has typically focused on the protein LC3B, one of the seven mammalian orthologs of yeast Atg8 [4–6]. LC3B is conjugated to phosphatidylethanolamine (PE) in autophagosome membranes; thus, fluorescently tagged exogenous LC3 or immunofluorescence for endogenous LC3 can be used to identify autophagosomes by fluorescence microscopy. In addition, LC3 conjugated to the inner surface of the autophagosome is delivered to the lysosomes and degraded, and the detection of LC3 conjugation and subsequent lysosomal degradation by western blot can provide quantitative measure of the bulk autophagic flux occurring in a cell population.

The elongation and closure of phagophore membranes to form double-membraned autophagosomes and the subsequent fusion of autophagosomes with lysosomes for substrate degradation is a dynamic process that is regulated at multiple steps, both upstream through stress response pathways that feed into the core autophagy machinery and through posttranslational regulation of these core components in response to stress [7–9]. Defects in autophagy arising early versus late in the pathway have differing effects on autophagosome formation and maturation. In particular, it must be emphasized that because conjugation of LC3 occurs during autophagosome formation, whereas degradation of conjugated LC3 occurs at the last step in the pathway after lysosomal fusion, an increase in conjugated LC3-II can be the result of either increased conjugation (increased flux) or reduced lysosomal degradation (reduced flux due to a block in the autophagy pathway after the LC3 conjugation step). Therefore, simply measuring the levels of unconjugated LC3B-I versus conjugated LC3B-II is not sufficient to determine the extent of functional autophagy in a cell population. Instead, it is critical to measure LC3B-I/II levels both with and without inhibitors of lysosomal degradation, such as bafilomycin A1 or pepstatin/e64d, for every experimental condition, and it is the amount of LC3B-II delivered to the lysosome (the increase in LC3B-II when lysosomal degradation is inhibited for each condition) that reflects the amount of ongoing autophagic flux. There are similar caveats to the interpretation of other experiments designed to measure autophagy [10]. Thus, published analysis of methodologies in the autophagy field has rightly focused on defining how to interpret such experiments [10] with one particularly thorough overview of autophagy methods in 2008 [11] that was updated and extended in 2012 [12]. In addition, the journal Autophagy regularly publishes reports of novel method development in its Toolbox section (http://www.landesbioscience.com/journals/autophagy/protocols/). With such resources available online, the goal of the current methods chapter is to describe experimental procedures used routinely that are widely accepted as valid approaches in the field, with particular emphasis on details that are not routinely described in published methods sections but are essential to the successful performance of such experiments.

2 Materials

2.1 Cell Culture

1. Hyclone defined fetal bovine serum, catalog #SH30071.03 (see Note 1).

2. Tetracycline-free serum (Invitrogen catalog #10438-026).

3. Glucose-free DMEM (Invitrogen catalog #11966-025).

4. EBSS (Invitrogen catalog #14155-063).

2.2 Autophagy Inducing Chemicals

1. 2-deoxyglucose (Sigma-Aldrich catalog #D3179).

2. Lithium chloride (Sigma-Aldrich catalog #203637).

3. Tunicamycin (Sigma-Aldrich catalog #T7765).

4. Thapsigargin (Sigma-Aldrich catalog #T9033).

5. Brefeldin A (Sigma-Aldrich catalog #B7651).

6. ABT737 (Santa Cruz catalog #sc-207242).

7. PP242 (Sigma-Aldrich catalog #P0037).

2.3 Autophagy Inhibiting Chemicals

1. Hydroxychloroquine (Sigma-Aldrich catalog #H0915).

2. BafilomycinA1 (Enzo life sciences catalog #BML-CM110-0100).

3. Pepstatin A (Sigma-Aldrich catalog #P5318).

4. Leupeptin (Sigma-Aldrich catalog #L9783).

5. E64d (Sigma-Aldrich catalog #E8640).

6. Ammonium chloride (Sigma-Aldrich catalog #254134).

2.4 Plasmids

1. Tandem mCherry/GFP-LC3 fusion expression plasmid (Addgene #22418).

2. shRNAs to Atg5 (Sigma-Aldrich custom, see text below), Atg7 (Sigma-Aldrich TRCN0000375444), Beclin1 (sigma), Ulk1 (Addgene #27633), Ulk2 (Addgene #27634), p62 (Sigma-Aldrich TRCN0000098619).

3. siRNAs to Atg5 (Dharmacon L-004374-00-0005), Atg12 (Thermo Scientific siGenome SMARTpool M-050953-01), Ulk1 (Thermo Scientific siGenome SMARTpool M040155-00).

2.5 Western Blot

1. PVDF (0.45 μm) (GE Healthcare catalog #45-000-931).

2. Nitrocellulose (0.2 μm) (GE Healthcare catalog #45-001-230).

3. Polyclonal rabbit anti-mouse immunoglobulins/HRP (DAKO catalog #P0260).

4. Polyclonal swine anti-rabbit immunoglobulins/HRP (DAKO catalog #P0217).

5. ECL film (GE Healthcare catalog #28906839).

2.6 Immunocytochemistry

1. Blocking buffer (DAKO cat #X909).

2. Anti-LC3 rabbit polyclonal antibody (Cell Signaling Cat #2775).

3. Secondary fluorophore-conjugated (Alexa 488, Alexa 647, or Rhodamine Red-X) anti-rabbit antibody (Invitrogen Cat #A-11008, A-21244, R-6394, respectively).

4. ProLong Gold with DAPI (Invitrogen Cat #P-36931).

5. Axiovert 200M wide-field fluorescence microscope (Zeiss) or Olympus DSU Spinning Disk confocal microscope (Olympus).

6. Openlab software (PerkinElmer) or ImageJ software (NIH).

7. Anti-Lamp2 antibody (Abcam ab13524).

8. MitoTracker (Invitrogen Cat #M-7512, M-7514, M-22426, etc.).

9. ER-Tracker (Invitrogen Cat #E-12353, E-34250, E-34251).

10. LysoTracker (Invitrogen Cat #L-12490, L-7528, L-12492, etc.).

11. 35-mm #1.5 glass-bottom dishes (MatTek P35G-1.5).

12. Olympus LCV110U VivaView microscope (Olympus).

2.7 Immunohistochemistry

1. Antigen retrieval buffer (DAKO catalog #S1699).

2. Mouse IgG blocking reagent (Vector Laboratories catalog #MKB-2213).

3. Biotinylated secondary antibody (Vector laboratories catalog #BA-2001).

4. Elite kit (Vector Laboratories catalog #PK-6100).

5. DAB kit (DAKO catalog #K3468).

3 Methods

3.1 Choice of Systems in Which to Examine the Functions of Autophagy

As with all significant questions in biology, the choice of experimental system is key to achieving a successful outcome. In this chapter, we focus on the measurement of autophagy in mammalian cells. The examination of autophagy in other model systems is not explored here. It is important to recognize that different mammalian cell lines and tissues exhibit varying degrees of basal autophagy and also distinct capacities to increase rates of autophagic flux in response to stresses. This diversity likely reflects the selection against certain tumor suppressive roles of autophagy in some human cancers and interestingly selection for pro-tumorigenic properties in others [13]. This variation is clearly illustrated in Fig. 1, where we examined the capacity of a panel of human breast cancer cell lines to undergo autophagic flux. The T47D, SKBR3, MCF7, MDA-MB-468, and HCC38 breast cancer cell lines exhibit marked autophagic flux, as determined by the increase in LC3B-II in cells treated with HCQ for 4 hours to inhibit lysosomal degradation compared to untreated samples (Fig. 1). By contrast, MDA-MB-231, HCC1938, and HS578T cell lines exhibit a smaller increase in LC3B-II levels following treatment with HCQ, indicating that the basal rate of autophagic flux is much lower in these cells (Fig. 1). Thus, before embarking on an examination of autophagy in your preferred cell line of a given tissue origin, or indeed in a specific tissue type or tumor type in vivo, it is important to assess the basal level of autophagic flux.

Fig. 1.

Measurement of autophagic flux in a panel of human breast cancer cell lines by LC3B western blot. Western blot analysis for LC3B was performed as described in Subheading 3.4 using extracts from T47D, SKBR3, MCF7, MDA-MB-468, HCC1937, HCC38, MDA-MB-231, and HS578T cells grown in the presence or absence of 50 μM hydroxychloroquine for 4 hours. Autophagic flux is detected in T47D, SKBR3, MCF7, MDA-MB468, HCC38 cells but not in HCC1937, MDA-MB-231, or HS578T

3.2 Stresses Used to Induce Autophagy in Cultured Cells

A number of physiologically relevant stresses can be used to induce autophagy in cultured cell lines. These stresses include amino acid deprivation, hypoxia, glucose deprivation, and serum starvation. In addition, there are a number of chemical agents that can be added to cells in culture to induce autophagy in a less physiological manner. These agents include the addition of lithium chloride or the addition of mTOR inhibitors, such as rapamycin and its more recent analogs. We have found, however, that these chemical agents typically do not have significant effects in cells that have high levels of basal autophagic flux.

The most common method for in vitro “starvation” of cells to increase autophagic flux is to culture them for several hours in Earle’s Balanced Salt solution (EBSS), which lacks essential amino acids. Autophagic flux increases within 2 h in cells grown in EBSS, but most cells cannot sustain viability in EBSS for more than 4–8 h and undergo cell death. Growth of autophagy-competent cells at 1 % oxygen (hypoxia) will also increase autophagic flux relatively rapidly, within 2–8 h depending on the cell line. We routinely make use of an enclosed hypoxia chamber developed by Coy Systems that is set to maintain oxygen levels in the chamber at 1 % oxygen through the controlled delivery of nitrogen and carbon dioxide to the chamber. Alternatively, cells can be grown in a sealed chamber flushed with a mixture of 95 % N₂ /5 % CO₂ for a set amount of time to reduce oxygen levels to 1 %, which must be determined empirically. Note that agents such as desferroxamine and dimethyloxallyl- glycine, which stabilize HIFα subunits, do not induce autophagic flux. Although most cultured cell lines are routinely grown in 4.5 g/L glucose, the complete removal of glucose through growth in glucose-free media will also induce increased autophagic flux, whereas the combination of glucose deprivation with hypoxia, which more closely resembles physiological ischemic stress, will induce autophagic flux but may also markedly reduce cell viability. Serum starvation for 12–24 h can also be used to induce autophagic flux in some cell types. Inhibition of glycolysis with 2-deoxyglucose (5 mM for 12–24 h) will also induce energetic stress and autophagy in cells.

The addition of 10 mM lithium chloride to cell cultures of autophagy-competent cells elevates autophagy within 12 h by inhibiting inositol monophosphatase at the endoplasmic reticulum, thereby leading to reduced levels of free inositol and inositol triphosphate [14] that are known to activate autophagy. Agents that induce ER stress, such as tunicamycin (2 μg/mL), thapsigargin (3 μM), and brefeldin A (20 μM) will also induce autophagy, principally of the ER [15]. BH3 mimetic peptides, such as ABT737, originally designed to inhibit the anti-apoptotic activity of Bcl-2-related molecules by mimicking the interaction of the BH3 domain of Bad with the hydrophobic groove of Bcl-2 [16], also disrupt the inhibitory interaction of Bcl-2 with Beclin1 [17], which contains a weak BH3 domain, to promote autophagy [18]. Although some groups have used the mTOR inhibitor rapamycin to induce autophagy in their experimental system, we have found this approach to be relatively ineffective for inducing autophagy in most of the tumor cell lines tested (see Note 2). However, if mTOR inhibition is the desired method of autophagy induction, the stronger ATP-competitive inhibitor of mTOR PP242 (50 nM, 24 h) may be used.

3.3 Approaches Used to Inhibit Autophagy Experimentally

To determine whether autophagy is playing a role in response to specific stresses, the ability to inhibit autophagy is important. Autophagy can be inhibited in cultured cell lines using either chemical inhibitors or, more frequently, genetic knockdown or targeting of specific/critical autophagy genes.

Chemical inhibition of autophagy is plagued with problems relating to the lack of specificity of most of these agents for autophagy inhibition. One of the most commonly used chemicals to inhibit autophagy in culture is 3-methyl adenine (3MA), which actively inhibits PI3 kinases, including the autophagy-essential protein Vps34. Wortmannin, a different PI3K inhibitor, is also used. However, a major caveat to the use of these inhibitors is that they inhibit not just class III PI3Ks, such as Vps34, but also class I PI3Ks, and thus their use is open to major criticism. Beyond their lack of specifi city for class III PI3Ks, there is growing evidence that 3MA in particular inhibits other cellular processes and has numerous autophagy-independent effects that make the use of this inhibitor highly undesirable as a means to block autophagy [10]. Similar issues arise when considering the use of microtubule (MT) inhibitors such as Taxol, which will inhibit autophagy by blocking the transport of autophagosomes to the lysosome but will also clearly inhibit other MT-dependent processes, such as mitosis and cellular trafficking in general. As previously mentioned, although the short term use of lysosomal acidification agents, such as bafilomycin A1 or chloroquine, is useful to determine whether autophagic flux is taking place, these drugs cannot be used experimentally for longer term experiments due to their effects on proteasomal degradation amongst other processes (see Note 3).

Thus, the key approach to inhibiting autophagy in cells in culture has become shRNA-mediated knockdown of key autophagy genes. Our lab has made use of both customized shRNAs and commercially available shRNAs to Atg5, Atg7 and p62/Sqstm1 to knockdown expression of these genes in mouse and human cell lines. We have most commonly sourced these shRNAs from Sigma-Aldrich, but functional shRNAs are commercially available from other companies as well. The commercially available shRNAs from Sigma-Aldrich that we have validated include lentiviral (pLKO)-expressed shRNA to Atg7 (TRCN0000375444) and p62 (TRCN0000098619). We were not able to validate many of the commercially available shRNAs to Atg5 as effectively knocking this gene down in either mouse or human cell lines. Although we did detect reduced expression (up to 50 %) of Atg5 with a few of these shRNAs, we failed to detect any appreciable effect on LC3 processing or autophagic flux in stable cell lines (see Note 4). Thus, we generated our own shRNA based on the successful use of siRNA to the same seed sequence (gca tta tcc aat tgg ttt) recognized in both mouse and human Atg5/ATG5. Although siRNAs can also be used to inhibit autophagy more transiently, and in particular, we have validated the use of siRNAs against Atg5, Atg12, and Beclin1 in tumor cells, it should be noted that transient transfection of siRNA to Atg proteins to inhibit autophagy is less ideal due to the inductive effect of the transfection protocol itself on autophagy and to the short time frame in which such experiments are performed (see Note 4).

The generation of stable cell lines overexpressing shRNAs to specific autophagy proteins can be performed as follows:

1. Infection of cells is performed using pLKO-shRNA/selectable marker expressing plasmids packaged into lentiviral particles. Alternatively, plasmids (expressing the shRNA and a selectable marker, such as G418 or puromycin resistance) are transfected into cells using Lipofectamine 2000 (Invitrogen) in Opti-MEM/ reduced serum medium (Invitrogen). Ideally, stable clones expressing at least one “scrambled” control shRNA and two different shRNAs to the Atg gene of interest should be generated. This is not always feasible due to issues in generating shRNAs to Atg proteins that actually knockdown these genes effectively to the point of inhibiting LC3 cleavage (see Note 4), as shown in Fig. 2.

2. At 48 h after infection/transfection, the selection agent (for example, G418 or puromycin) is added to kill cells that have failed to stably incorporate the viral genome or transfected plasmid into their genome and thus fail to confer drug resistance. Each cell line exhibits differential sensitivity to any single selectable agent, and thus a “kill curve” should be performed ahead of time to determine the optimal concentration of selectable agent.

3. After 1–3 weeks, single drug-resistant clones emerge that are ring cloned and expanded for further analysis. It is important to pick single clones and not use pooled clones due to likely effects of integration site on the expression of encoded sequences.

4. Clones stably expressing the shRNA of choice are tested by western blot for successful knockdown of the targeted gene. In addition, the effects of knockdown of specific autophagy genes on LC3 processing in response to autophagy-inducing stresses must be tested. This is particularly important because, as already mentioned, levels of certain autophagy regulators are not limiting in many cell types (see Note 4).

5. Stable clones of cells expressing shRNAs to the autophagy genes of interest should be maintained in the selectable agent used to generate them (e.g., G418 or puromycin). However, for experiments making use of stable knockdown cell lines generated in this manner, it is advisable to remove the agent during the time course of the experiment, as some agents, such as puromycin, are known to modulate autophagy.

A variety of inducible shRNA expressing vectors are also available, such as the pTripZ vectors (ThermoSci) that express shRNAs to key autophagy genes under the control of tetracycline-inducible elements. These vectors offer the inducible control of expression of the genes of interest and may be of particular interest in generating stable knockdown of Beclin1, which we have found to be particularly recalcitrant to effective stable knockdown in numerous cell lines tested. However, it should also be noted that tetracycline and its analogs, such as doxycycline, inhibit mitochondrial protein translation resulting in elevated mitophagy as a result of mitochondrial dysfunction and so may not be the ideal inducible agent, particularly if your studies are focused on analysis of mitophagy (see Note 5).

Fig. 2.

Assessing the efficacy of shRNA knockdown of autophagy proteins. After testing out several commercially available shRNAs, we commissioned generation of a custom hairpin targeting the seed sequence 5′-gca tta tcc aat tgg ttt-3′ that is conserved in both mouse and human Atg5/ATG5. This hairpin is highly effective at knocking down Atg5 (a) as is required to block autophagy and to elicit an inhibitory effect on LC3 processing (b) and to promote p62/Sqstim1 accumulation (c). As discussed in the text, less effective knockdown of Atg5 does not always result in inhibition of autophagy since the Atg5 protein is so highly expressed in many cell lines

In addition to gene knockdown through the stable expression of shRNAs, the increasing availability of autophagy gene knockouts in mice allows us to examine the effects of autophagy inhibition in cells derived from these mice. Of particular use is the derivation of mouse embryo fibroblasts (MEFs) from Atg5-null or Atg7-null mice using similar approaches to those published previously with other genetically engineered mouse models [19].

3.4 Measuring Changes in Autophagic Flux by Western Blot

The most straightforward and workhorse assay to measure autophagic flux is western blot/immunoblot for LC3B-I and II [10, 20, 21]. As previously mentioned, measurement of autophagic flux depends on the ability to distinguish between autophagosome formation and completion of autophagy through lysosome fusion and degradation. This process is assessed experimentally by culturing cells in the presence or absence of agents that block the late maturation steps. Commonly used agents include bafilomycin A1, hydroxychloroquine and NH₄Cl that alter lysosomal pH, thus resulting in a block to degradation and the accumulation of autophagosomes unable to fuse with lysosomes. Other agents include leupeptin (50 μM), pepstatin A (15 μM), and E64d (29 μM), which can be used in combination to inhibit lysosomal proteases directly for longer periods of time. We strongly advise against the use of 3-methyl adenine to inhibit autophagy for reasons described in Subheading 3.3.

1. Cells are cultured for 16–48 h in the presence or absence of the agent or stress that modulates autophagy (e.g., 1 % oxygen or 10 mM LiCl₂ as described in Subheading 3.2). Bafilomycin A1 (100 nM) or hydroxychloroquine (50 μM) is added to control cultures 4 h prior to harvesting cells for protein extraction (see Note 3).

2. Cells are washed twice in ice-cold PBS with protease and phosphatase inhibitors (0.5 mM PMSF, 1 g/mL leupeptin and aprotinin, 1 mM sodium orthovanadate), harvested by scraping with a plastic cell remover, and pelleted by centrifugation at 4,000 × g for 3 min at 4 °C.

3. Cell pellets are lysed for 15 min on ice, vortexing every 5 min, in an equal volume of an NP-40-based lysis buffer (150 mM NaCl, 50 mM Tris–HCl pH 7.4, 1 mM EDTA, 1 % NP-40).

4. 60 μg (cell lines) or 100 μg (tissues/tumors) of protein are separated on a 13 % SDS-PAGE gel to resolve LC3B-I and LC3B-II at 16 kDa and 14 kDa respectively.

5. Following electrophoresis, proteins are transferred to 0.45-μm activated PVDF membrane (see Note 6) using a wet transfer system. PVDF membrane is activated first in 100 % methanol for 1 min and then washed well in double-distilled water before being equilibrated in transfer buffer for 10 min.

6. After transfer, the membrane is cut at 26 kDa to avoid cross-reaction of the primary antibody with nonspecific bands at higher molecular weights. The membrane is then blocked for 1 h at room temperature in PBS/5 % milk. The portion of the membrane below 26 kDa is probed with primary antibody to LC3. The portion of the membrane above 26 kDa is probed with an antibody that detects a constitutively expressed protein (such as β-actin) as a control for gel loading.

7. We routinely use the anti-LC3B rabbit polyclonal antibody from Novus (cat #NB600-1384) at 1:3,000 to detect LC3B-I/ LC3B-II in extracts from human cell lines and at 1:500 to 1:1,000 for extracts from mouse cell lines/tissues. The membrane is incubated with primary antibody in PBS/5 % milk overnight at 4 °C.

8. No Tween is added to the primary antibody incubation solution, but PBS/0.05 % Tween is used to wash the membranes (three times) after primary and secondary antibody incubations.

9. Anti-rabbit secondary antibody conjugated to horseradish peroxidase (DAKO) is incubated at 1:5,000 for 2 h at room temperature, membranes are washed three times in PBS/0.05 % Tween, and detection is performed by enhanced chemiluminescence (ECL film; GE Healthcare).

10. Interpretation of LC3 western blots has been discussed at length [10, 12, 21]. Previously, the ratio of processed LC3B-II (the faster migrating form of LC3B) to LC3B-I was used as a measure of autophagic flux. However, the overall levels of LC3B (the combined amount of LC3B-I and LC3B-II) can vary due to altered expression levels, and it is now accepted that quantification of the increase in LC3B-II during inhibition of lysosomal degradation provides a better surrogate for quantitating autophagic flux. An example of such a blot is presented in Fig. 1.

In addition to western blotting for processed LC3, levels of the Atg5–Atg12 conjugate can be used as a measure of effectiveness in early stages of autophagy leading up to autophagosome formation. During autophagosome formation Atg12 is conjugated to Atg5 resulting in the appearance of the ~50 kDa Atg5–Atg12 conjugate as autophagy increases (see Note 7). Levels of p62/Sqstm1 are often also used as an indirect measure of functional autophagy because p62/Sqstm1 accumulates when autophagy is defective [22, 23] (Fig. 2c) although autophagy-independent regulation of p62/Sqstm1 makes this a less robust readout of autophagic flux (see Note 8). For these two key western blot approaches, the same extracts as prepared to measure LC3 levels can be used but we find that extraction with RIPA buffer (1 % sodium deoxycholate, 0.1 % SDS, 1 % Triton X-100, 10 mM Tris–HCl pH 8.0, and 0.14 M NaCl, plus 0.5 mM PMSF, 1 μg/mL leupeptin, 1 μg/mL aprotinin, 1 mM sodium orthovanadate to inhibit proteases/phosphatases) is more efficient for p62/Sqstm1 and other Atg proteins. Table 1 summarizes the main conditions used routinely in our lab.

Table 1.

Summary of western blot methods for key autophagy proteins

Antigen	Extraction buffer/ membrane	Source/Cat.# primary antibody	Block	Primary antibody conditions	Washes	Secondary antibodya/ other conditions
LC3/Atg8	NP-40 PVDF	Novus NB600-1384 Rabbit polyclonal	PBS/5 % milk Room temp. 1 h	1:3,000 (human); 1:1,000 (mouse) PBS/5 % milk No Tween. 4 °C overnight	PBS/0.05 % Tween; 10 min, twice	PBS/5 % milk 2 h; room temp.
p62/Sqstml	RIPA Nitrocellulose	Progen CP62-C Guinea pig monoclonal	PBS/5 % milk Room temp. 1 h	1:1,000 to 1:5,000 PBS/5 % milk No Tween . 4 °C overnight	PBS/0.05 % Tween; 10 min, twice	PBS/5 % milk 2 h; room temp.
Atg5	RIPA Nitrocellulose	Novus NB110- 53818 Rabbit polyclonal	PBS/5 % milk Room temp. 1 h	1:500 PBS/5 % milk No Tween. 4 °C overnight	PBS/0.05 % Tween; 10 min, twice	PBS/5 % milk No Tween. 4 °C overnight.
Atg 12	RIPA PVDF	Cell signaling #4180 Rabbit monoclonal	0.1 TBST/5 % milk Room temp. 1 h	1:500–1:1,000 TBS/0.1 % Tween/5 % BSA; 4 °C overnight	TBS/0.1 % Tween; 10 min, twice	TBS/5 % milk 2 h; room temp.
Atg7 (human)	RIPA PVDF	Epitomics 2054-1 Rabbit Polyclonal	TBS/5 % milk Room temp. 1 h	1:1,000 TBS/5 % milk No Tween. 4 °C overnight	TBS/0.1 % Tween; 10 min, twice	TBS/5 % milk 2 h; room temp.
Atg7 (mouse)	RIPA PVDF	Sigma A2856 Rabbit Polyclonal	PBS/5 % milk Room temp. 1 h	1:1,000 PBS/5 % BSA No Tween. 4 °C overnight	PBS/0.05 % Tween; 10 min, twice	PBS/5 % milk 2 h; room temp.
Beclin1 Atg6	RIPA Nitrocellulose	Cell signaling #3738 Rabbit polyclonal	PBS/5 % milk Room temp. 1 h	1:1,000 PBS/5 % milk No Tween. 4 °C overnight	PBS/0.05 % Tween; 10 min, twice	PBS/5 % milk 2 h; room temp.
Ulk1	NP40 Nitrocellulose	Sigma A7481 Rabbit polyclonal	0.1 TBST/5 % milk Room temp. 1 h	1:1,000 TBS/0.1% Tween, 5 % milk; 4 °C overnight	TBS/0.1 % Tween 10 min, twice	TBS/0.1 % Tween 5 % milk; 2 h; room temp.

^aWe use the relevant HRP-conjugated secondary antibody from DAKO at 1:5,000, unless otherwise stated

3.5 Intracellular Staining for Autophagosomes

In addition to measuring autophagic flux by western blot, we also routinely perform intracellular staining for endogenous LC3. In unstressed cells, LC3 is expressed ubiquitously and is detected in a diffuse pattern in both the nucleus and cytoplasm of cells by immunofluorescence. As a result of elevated autophagy, this staining pattern changes from largely diffuse to predominantly punctate and cytoplasmic (Fig. 3). Quantification of LC3-positive puncta has therefore also emerged as a key methodology to measure increases in autophagy. As for western blotting for LC3, experiments are performed in the presence or absence of bafilomycin A1 or chloroquine to distinguish between increased autophagosome formation and decreased or inhibited autophagosome-lysosome fusion.

Fig. 3.

Immunofluorescent staining for LC3 in cultured cells. Using the approach described in Subheading 3.5, we stained HCC38 cells (a basal-like human breast cancer cell line) for LC3 before (a) and after (b) inducing autophagy by growth in EBSS for 3 h, and in the presence of 100 nM bafilomycin A1 (c) to inhibit autophagic flux

To stain cells for endogenous LC3, the following protocol may be used:

1. At least 24 h prior to cell staining, cells are plated at an appropriate cell density on glass coverslips in 6-well plates.

2. Following treatment of cells as described in Subheading 3.4, cells are washed once in PBS and fixed in 4 % paraformaldehyde (in PBS) for 15 min at room temperature.

3. Cells are washed three times in PBS.

4. Cells are then permeabilized in 100 % cold methanol for 10 min at −20 °C and washed three times in PBS.

5. Cells are incubated in blocking buffer (cat #X909; DAKO or 2% FBS 1 % Goat Serum/PBS) for 30 min to 1 h at room temperature.

6. Staining for LC3 is performed overnight at 4 °C by incubating cells in primary antibody solution made up of a 1:200 dilution of anti-LC3 antibody (rabbit polyclonal from Cell Signaling Cat #2775) in PBS.

7. The following day, cells are washed three times in PBS (5 min per wash) with gentle rotation.

8. Cells are then incubated in secondary antibody solution consisting of a 1:500 dilution of Alexa 488, Alexa 647, or Rhodamine Red-X conjugated anti-rabbit secondary antibody (Invitrogen) for 2 h in the dark at room temperature.

9. Cells are washed twice in PBS (5 min per wash) and then once in double-distilled water.

10. Coverslips are mounted with Prolong gold with DAPI (Invitrogen). Slides should be stored (up to 1 week) at 4 °C in the dark until viewed under the scope.

11. Cells can be imaged using an Axiovert 200M wide-field fluorescence microscope (Zeiss), and image deconvolution is performed with Openlab software (PerkinElmer). Alternatively, confocal imaging can be performed on an Olympus DSU Spinning Disk confocal microscope.

Co-staining with other antibodies is frequently performed to determine whether a putative novel autophagic substrate colocalizes with LC3-positive autophagosomes as well as to examine the maturation state of autophagosomes in cells. For example, we have co-stained cells for LC3 and for Lamp2 (Abcam ab13524) to examine when autophagosomes overlap with lysosomes as a measure of functional autophagy. Similar approaches can be used to examine co-localization of mitochondria (for example, using BNIP3 as a mitochondrial marker as shown in Fig. 4) with LC3-positive autophagosomes, merging of endosomes (for example using Rab5, Rab7, and Rab11 as endosomal markers) with LC3-positive autophagosomes and generally to interrogate the cargo or pathways interacting with LC3-positive autophagosomes.

Fig. 4.

Co-staining for LC3 and markers of autophagosomal cargo. As described in the text, intracellular staining for LC3B can be combined with staining for other markers, such as BNIP3 shown below. BNIP3 is an outer mitochondrial protein and molecular adaptor that targets mitochondria for degradation at the autophagosome [40]. Here we show overlap between punctate LC3 and BNIP3 induced by exposure of HCC38 human cancer cells to 1 % oxygen (hypoxia), consistent with BNIP3 playing a role in promoting mitochondrial turnover at the autophagosome as part of the adaptive response to hypoxia

Primary antibody (generated in mouse, rat, or species other than rabbit) to one of these additional markers is added at the same time as the anti-LC3 primary rabbit antibody, and an anti-mouse secondary antibody conjugated to a differently colored fluorophore is used to detect the mouse primary. Of note, we frequently use Triton, saponin, or other detergents to permeabilize cells for staining for most antibodies employed in the lab, but LC3 staining does not work with these other fixation techniques; thus, for co-staining approaches, it is best to stick to methanol permeabilization. Most other antibodies will stain well enough using methanol permeabilization to allow effective co-staining with LC3.

In addition to staining with other antibodies, cells can be co-stained with cell-permeable organelle-specific dyes such as MitoTracker, ER-Tracker, and LysoTracker alongside staining for endogenous LC3. These dyes are made up at stock concentrations of 100 μM and added to cell cultures at a 1:500 dilution (working concentration of 200 nM). The dyes can be added to LC3-stained cultures after the completion of step 8 above, incubated in the dark for 15 min at room temperature and washed twice in PBS before mounting in ProGold™ as described in step 10.

Prior to the development of antibodies that work well for endogenous LC3 staining, overexpression of exogenous GFP-LC3 was used to quantify autophagosome formation in cells in response to specific stresses, but we do not recommend using transient expression of GFP-LC3 to quantify autophagosome formation and autophagic flux due difficulties in interpretation and the high likelihood of experimental artifacts. Such issues include the effects of the heterogeneity of transfection efficiency and GFP-LC3 expression levels between cells, the induction of autophagy by both the transfection protocol and overexpression of GFP-LC3 and finally by the propensity of GFP-LC3 to associate with protein aggregates already present in cells as well as to self-aggregate when overexpressed, leading to the appearance of punctate fluorescence that does not represent autophagosomal structures [24].

We use transient GFP-LC3 overexpression for more qualitative approaches, for example, to examine interactions of LC3 with other molecules, but it is important to confirm findings by a second method that does not involve GFP-LC3, and we generally do not use GFP-LC3 overexpression by itself in assays requiring reliable quantification. If it is necessary to use GFP-LC3, it is preferable to generate single cell clones (as described in Subheading 3.3 for the generation of stable knockdown cell lines) stably expressing a set level of fluorescent-tagged LC3, thus ensuring that the cell population exhibits homogeneous expression of GFP-LC3 and eliminating the autophagy-inducing effect of transfection, although the caveat remains that the overexpression of GFP-LC3 itself may have an autophagy-inducing effect.

Despite these caveats to the use of GFP-LC3 as a single probe to examine autophagy in a quantitative manner, the development of tandem probes in which both red and green fluorophores are both fused to the amino terminus of LC3 [25] has permitted an internally controlled/ratiometric approach to quantifying the late maturation steps in autophagy. The principle relies on differential quenching of green fluorescence by the acid pH in the lysosome compared to red fluorescence that is largely undimmed by acid pH [6]. In addition to the mRFP-GFP-LC3 fusion reported by Kimura and colleagues, others have now utilized a mCherry- GFP-LC3 fusion equivalent that is commercially available (Addgene). The principle is to transfect cells with the plasmid as described above and then to quantify changes in the ratio of yellow to red puncta by time-lapse microscopy in response to specific stresses being tested for their effect on autophagic flux. A similar concept involves the use of the coral-derived fluorescent protein Keima that fluoresces at a different wavelength as a function of pH [26]. Thus, the ratio of fluorescence emission at 543 nm (preferred at acidic pH) to that at 458 nM (preferred at neutral pH) can be used as a measure of autophagic flux from autophagosome to lysosome/autophagolysosome [26]. The use of a single protein such as Keima over tandem red-green-LC3 fusions to monitor autophagic flux has certain advantages, including reduced effects due to intramolecular fluorescence transfer or selfaggregation of LC3.

In addition to intracellular staining of autophagosomes in fixed cells, we have also made use of time-lapse video microscopy to image movement of autophagosomes and changes in their number in response to specific stress. Cells stably expressing GFP-LC3 or dsRed-LC3 or mCherry-GFP-LC3 are seeded in 35-mm #1.5 glass bottom dishes (MatTek) for 24 h prior to imaging and timelapse DIC and fluorescence imaging is performed with an Olympus LCV110U VivaView microscope system maintained at 37 °C and 5 % CO₂. However, we have found that visualization and quantification of co-localization of autophagosomes with other subcellular structures (for example, mitochondria stained with MitoTracker deep red) in live cells stably expressing fluorescent-tagged LC3 requires the resolution of confocal microscopy, which we have achieved after growth of cells on glass bottom dishes as described above utilizing an Olympus DSU live-cell optimized spinning disk confocal microscope.

3.6 Immunohistochemical Staining of Tissues and Tumors for Markers of Autophagy

The ability to monitor autophagic flux in tissues and tumors in vivo remains difficult, and at present, the surrogates most commonly used include immunohistochemistry for LC3B and the autophagy substrate p62/Sqstm1 [27–33]. Punctate LC3 staining (Fig. 5) and lower p62 staining suggest ongoing autophagic flux, whereas diffuse LC3 staining and increased p62 suggest reduced or absent autophagic flux.

1. Tissues are collected in 10 % neutral buffered formalin, processed through increasing ethanol concentration to xylene, and embedded in paraffin. Four-micron tissue sections from paraffi n blocks are cut onto glass microscope slides for staining.

2. Tissue sections are deparaffinized and rehydrated through xylenes and serial dilutions of ethanol to distilled water.

3. Tissue sections are incubated in antigen retrieval buffer (DAKO, S1699 or 10 mM sodium citrate, pH 6.0) and heated in a steamer at over 97 °C for 20 min or in the microwave at maximum power for 11 min. Slides are quickly transferred to cold water afterwards to ensure that epitope denaturation was maintained. Slides are then washed twice in TBS.

4. Endogenous peroxidases are blocked by incubation in 10 % methanol/3 % hydrogen peroxide for 30 min.

5. After further TBS washes (2 × 5 min), sections are blocked in 1.5 % normal horse serum for 30 min at room temperature.

6. For LC3, tissue sections are treated with mouse IgG blocking reagent (MKB-2213, Vector Laboratories) if staining mouse tissues or tumors (primary antibody is a mouse monoclonal). This step can be skipped if staining human tissues/tumors.

7. Tissue sections are incubated with primary antibody to LC3 (1:250) or p62/Sqstm1 (1:500) at 4 °C overnight (see below for catalog numbers).

8. Following TBS wash, tissue sections are incubated with the appropriate biotinylated secondary antibody (1:100, BA-2001, Vector laboratories) for 1 h at room temperature.

9. Antigen-antibody binding is detected by Elite kit (PK-6100, Vector Laboratories) and DAB (DAKO, K3468) system.

10. Tissue sections are briefly immersed in hematoxylin for counterstaining and are mounted in Permount with glass coverslips.

In addition to immunohistochemical staining for LC3 and p62/sqstm1, we have also successfully stained for Atg7 and Atg5 on mouse and human tissues/tumors, although the levels of these proteins, if present, are not necessarily indicative of autophagic flux. The protocols for staining for these antigens are similar to that described above for LC3. Table 2 summarizes the conditions used for each antibody that we have successfully tested in immunohistochemical procedures.

Fig. 5.

Immunohistochemical staining of mouse mammary tumors for punctate LC3B. Detecting punctate LC3 in tissues by immunohistochemistry was performed as described above using sections of d80 mammary tumors from MMTV-PyMT mice (a). Although the anti-LC3B antibody used (NanoTools) is not specific for processed LC3, it does appear to preferentially stain LC3 puncta in tissue sections, as seen most readily at high magnification (b)

Table 2.

Summary of immunohistochemical protocols for key autophagy proteins

Antigen	Source Primary antibody	Epitope retrieval	Primary Ab conditions	Washes
LC3	NanoTools 0231-100/LC3-5F10 Mouse monoclonal	Heat denaturation in citrate buffer	1:250 (mouse, tissues); 1:100 (human tissues)	TBS
p62/Sqstml	Enzo Life Sciences BML-PW9860-0100 Rabbit polyclonal	Heat denaturation in citrate buffer	1:500	TBS
Atg7 (Note 9)	Sigma-Aldrich HPA007639 Rabbit polyclonal	Heat denaturation in citrate buffer	1:250	TBS
Atg5	Novus Biologicals NB110-53818 Rabbit polyclonal	Heat denaturation in citrate buffer	1:250	TBS

^{Note 9}Several publications note the use of Santa Cruz anti-Atg7 antibody sc-32211 to detect Atg7 in mouse tissues. We have found that this antibody in several lot numbers now tested crossreacts with a nonspecific nuclear epitope (present in tissues/ tumors from Atg7 knockout mice) and is not useful in our hands for detecting mouse Atg7.

All of these antibodies work for both human and mouse tissues.

3.7 Detecting Interactions with Processed LC3

Due to its lipophilic nature, it is difficult to extract endogenous LC3B-II in complexes with interacting proteins. Therefore, to detect protein-protein interactions, we have typically relied on over-expression of GFP-LC3 followed by immunoprecipitation using anti-GFP antibodies and western blotting for the putative interacting protein. If the target protein is degraded by autophagy, it can be helpful to perform the immunoprecipitation in the presence of bafilomycin A1 to inhibit autophagosome fusion with lysosomes and lysosomal degradation, which will increase the number of LC3- substrate complexes present in the cell by preventing their lysosomal degradation. Given the caveats relating to the use of overexpression of GFP-LC3 detailed above, however, it is crucial to validate any detected protein-protein interactions with immunoprecipitation of endogenous proteins. Although challenging, co immunoprecipitation of endogenous LC3 with interacting proteins has been reported [34] using a rabbit anti-LC3B antibody (Cell Signaling #3638) and NP40 lysis buffer (250 mM NaCl, 50 mM Tris–HCl pH 7.4, 5 mM EDTA, 1 % NP-40). Alternatively, LC3 interactions have been detected using protein cross-linking methods and yeast two-hybrid screening [35].

Fig. 6.

Optimized detection of LC3 by western blot. We have tested two different extraction methods and three different membranes in order to optimize LC3 western blotting techniques. These westerns worked best using PVDF 0.45 μm and an NP-40 based extraction buffer, as described in Subheading 3.4 and Note 2