Autophagy is an intracellular orderly recycling system to degrade unnecessary cytosolic proteins and damaged organelles. It is a dynamic cellular self-eating process that exists in all eukaryotic cells. Autophagy is dependent on lysosomal degradation but is distinct from endocytosis-mediated degradation of plasma-membrane proteins and extracellular components, including bacterial or viral pathogens and apoptotic cells during phagocytosis. Three major types of autophagy are recognized: macroautophagy, microautophagy, and chaperone-mediated autophagy (1, 2). Macroautophagy is responsible for bulky degradation of proteins and organelles of the cytoplasm by sequestering them into a
Autophagy is important for a cell to defend against intracellular and extracellular stresses and maintain metabolic homeostasis. Autophagy has been implicated in many physiological events, including starvation survival, organelle turnover and cellular renewal, cell growth, immunity, animal development, and ageing (5-7). Recent evidence indicates that alterations in autophagy occur frequently in many human diseases, such as neurodegenerative disorders, cancer, and cardiomyopathies (8-10). Although the roles of key proteins in mammalian autophagosome formation have been studied intensively (5, 11) since the discovery of ATG (autophagy-related) genes in yeast (12), their disease-related functions have yet to be elucidated. Currently, more than 230 proteins are recognized by the human autophagy database (from Luxembourg Institute of Health, http://autophagy.lu/index.html) as autophagy-involved human proteins directly or indirectly. A list of 56 compounds that act as an inducer or an inhibitor of autophagy is available, but direct targets of most compounds in the autophagy machinery are still being investigated (13). Searching for new pharmacological agents targeting the autophagy pathway, in addition to known activators and inhibitors of autophagy and evaluation of how each reagent affects autophagy, is critical for appropriate therapeutic application of each compound in different human diseases.
Methods for monitoring mammalian autophagy have been summarized and provided in several review articles (11, 14, 15). Immunoblot analysis of microtubule-associated protein light chain 3 (LC3), one of the mammalian homologs of Atg8 in yeast, shows that the number of autophagosomes increases in proportion to the band intensity of LC3-II, a phosphatidylethanolamine (PE)-conjugated LC3, migrating faster than LC3-I in SDS-PAGE (16). Biochemical analysis of the ratio of LC3-I to LC3-II is sometimes interpreted inappropriately because of degradation of LC3-II itself by autophagy. Detection of p62 protein, also called sequestosome 1, by immunoblot analysis is another useful biochemical method to monitor autophagy activation caused by selective degradation of p62 in the lysosome by autophagy (17, 18). The amount of p62 protein is inversely proportional to the magnitude of autophagy. Biochemical analysis of the ratio of LC3-I to LC3-II or the amount of p62 is limited for quantification of autophagy flux, because measurement of protein-band intensity can be inaccurate. Detection of both LC3 and p62 protein cannot provide information about the step at which a compound inhibits or activates during autophagy. Electron microscopy (EM) is useful for detecting different stages of the autophagy pathway. Early autophagosomes, autolysosomes, and autophagic degradation products are observed and identified as different structures (11, 14), indicating that EM provides qualitative information about autophagy. It is challenging to obtain quantitative information for the number of autophagosomes and autolysosomes. Application of EM to the massive analysis of autophagy has limitations because of the difficulty of sample preparation and observation. Optical imaging using wide-field microscopy, confocal microscopy, and super-resolution microscopy with proper imaging reporters or sensors has provided autophagy researchers with quantitative analytical tools to monitor autophagosomes and autolysosomes.
In this review, we describe confocal imaging methods with various fluorescent reporters to monitor autophagy flux and discuss their strengths and weaknesses. A brief quantification of imaging data for each method is described. We also provide background and general experimental procedures in a mammalian cell selected herein. Depending on the cell type of interest, stress conditions, and treated compounds, experimental procedures and interpretation of data will differ.
Autophagy consists of several dynamic steps but sequential measurement of each step is difficult (11, 14, 15). In mammalian cells, autophagy begins with a phagophore (an isolation membrane) (Fig. 1, step 1), which expands to engulf and sequester the cargo in a double-membraned autophagosome, which is loaded with protein aggregates and various organelles (Fig. 1, step 2). The autophagosome fuses with an endosome to form an amphisome (an acidic late autophagosome) (19). In the next step, the autophagosome fuses with a lysosome (Fig. 1, step 3) to form an autolysosome and promote degradation of the inner membrane of the autophagosome and the loaded materials by lysosomal hydrolases in the autolysosome (Fig. 1, step 4). The autolysosome is ready for the next cycle once degraded monomeric units are exported to the cytosol (Fig. 1, step 5) (20, 21). Some lysosomes are formed through the initiation of proto-lysosomal tubules and vesicles that originated from the autolysosomes and subsequent maturation with the supply of hydrolases (22).
Formation of phagophores requires class III phosphoinositide 3-kinase Vps34 activation by means of a macromolecular complex, including Vps34, Beclin-1 (mammalian Atg6), Atg14, and Vps15 (23-25). Vps34 generates phosphatidylinositol(3)phosphate (PtdIns(3)P; PI(3)P), which is critical for recruitment of effectors such as double FYVE-containing protein 1 (DFCP1) and WD-repeat domain phosphoinositide-interacting (WIPI) family proteins (5, 26), thereby promoting development of the isolation membrane or autophagosome (27). Imaging of fluorescent protein with PI3P-binding domain (two FYVEs of DFCP1, WIPI, or others) is useful for monitoring phagophore formation (26, 28, 29).
Autophagosome formation results from the sequestration of cytoplasmic compartments by the phagophore. At the late stage of phagophore formation, conjugation of the Atg8 family proteins―microtubule-associated protein 1 LC3, GABARAP, and GATE16―to the PE are produced by Atg7 and Atg3, and their conjugated Atg proteins are integrated into the membrane of autophagosomes (12, 30). LC3 without PE and with PE are described as LC3-I and LC3-II, respectively. LC3-II is known to function as a receptor for a selective substrate, p62/sequestosome 1, on the autophagosomal inner membrane (31). Imaging of fluorescent LC3-II is used to monitor autophagosomes or autolysosomes.
Autophagosomes containing cytosolic materials fuse with lysosomes that are responsible for massive cellular degradation to form autolysosomes (21). Autolysosome formation requires active microtubule organization during the fusion process (32). This process can be monitored by the analysis of colocalization of fluorescent LC3-II with lysosome markers, such as lysosomal-associated membrane protein 1 (LAMP1) and LAMP2 proteins. Inhibition of the fusion decreases the colocalization level of LC3 and LAMP1/2 proteins.
Lysosomal hydrolases, which are active at an acidic
GFP-LC3 is used commonly as an imaging marker for autophagosomes (31) but not for autolysosomes, because GFP possesses a pKa of 6.0 and is acid-labile (36). The mCherry protein is very acid-stable because it has a pKa of <4.5 span="" class="xref">37). Imaging of the mCherry-GFP-LC3 protein shows the fluorescence difference between autophagosomes and autolysosomes in a cell (18). The mCherry-GFP-LC3 proteins seen in neutral vesicles display both green and red fluorescence, but in acidic vesicles display red fluorescence only (Fig. 2A). In HepG2 cells expressing mCherry-GFP-LC3, nutrient starvation elicits an increase in the number of autophagosomes and autolysosomes (Fig. 2B). Quantification of the number of spots (yellow or red) shows the total number of autophagosomes and autolysosomes and their ratio (Fig. 2C). Analysis of the spot number ratio of autolysosomes (mCherry+GFP-) to autophagosomes (mCherry+GFP+) is useful for estimation of the transition from autophagosome to autolysosome. Treatment with compound X is suggested to inhibit the fusion of autophagosomes with lysosomes or to increase
HepG2 cells were cultured in DMEM supplemented with 10% fetal bovine serum (Gibco) and penicillin-streptomycin (Hyclone) at 37°C in a humidified atmosphere of 5% CO2. Cells were cultured on a 12-well plate containing cover slips (diameter, 18 mm,) coated with poly-L-lysine for both live-cell imaging and fixed-cell staining. Cells were infected with mCherry-GFP-LC3 adenovirus overnight (16-20 hours; infection with the virus for more than 20 h might be toxic to cells). Cells expressing mCherry-GFP-LC3 were incubated under selected conditions for 8 h (control, standard media; starvation, Earle's balanced salt solution [EBSS]). Cells were treated with 0.25 μM rapamycin for 8 h for autophagy induction. Cells were fixed for 10 min with 4% formaldehyde in phosphate-buffered saline (PBS), exposed for 30 min at room temperature to PBS containing 5% horse serum (Gibco-BRL) and 0.1% Triton X-100. The cells were also stained with 4’6’-diamidino-2-phenylindole (0.2 μg/ml) to detect DNA. We did confocal microscopy as described previously (40) with minor modifications and acquired the images using a 60× Plan Apochromat VC objective, NA 1.40, by illuminating with a 488-nm multi-Ar laser (for excitation of GFP fluorochrome) or with a 561-nm diode-pumped solid-state laser (for excitation of mCherry fluorochrome) with a microscope (A1R, Nikon or LSM880 Airy, Zeiss) equipped with a galvano detector. Images were processed with NIS-Elements AR 3.0 software (from Nikon) or Zen software (from Zeiss). For live-cell imaging, cells were maintained in an incubation chamber (Chamlide TC; Live Cell Instrument) at 37°C under an atmosphere of 5% CO2 in air.
For spot count, we analyzed each z-stack image to avoid miscounting overlapping autophagosome and autolysosome spots. For measurements with NIS or Zen software, we deleted background signals for the same value in all images to measure only the exact spots. Spots remaining on the deleted background can be measured. As a result, spots that have values greater than a certain fluorescence intensity were quantified. Spots with both mCherry (red) and GFP (green) were counted as autophagosomes and spots with only mCherry (red) were counted as autolysosomes.
The mCherry-positive and GFP-positive spots indicate not only autophagosomes but also autolysosomes that have a neutral
PtdIns(3)P is associated with endosomal functions and is mainly localized in the early endosomes (41). Accumulation of PtdIns(3)P by Vps34, a Class III PI(3)K, is critical for the initiation of autophagy (23, 24). PtdIns(3)P-binding proteins, including Atg18 (WD-repeat domain phosphoinositide-interacting 2, WIPI2) and DFCP1, are recruited to autophagic membranes and regulate LC3 lipidation, thereby promoting phagophore formation (27, 29). WD repeat and FYVE domain-containing 3 (WDFY3), known as autophagy-linked FYVE, is associated with autophagic membranes as a PtdIns(3)P-binding protein (42). Imaging of mCherry-WDFY (a PtdIns(3)P sensor) and GFP-LC3 is useful for monitoring early stages of autophagy (Fig. 2E). Phagophore-containing PtdIns(3)P is shown in red, early autophagosomes with both PtdIns(3)P and GFP-LC3 are shown in yellow, and late autophagosomes with only GFP-LC3 are shown in green. Selective Vps34 inhibitors such as SAR405 (43) inhibit early autophagy by restraining phagophore formation (Fig. 2E).
To image PtdIns(3)P during phagophore formation, we used the expression plasmid containing cDNA that encodes a PtdIns(3)P-binding protein (or domain), such as DFCP1 (26), UVRAG (44), or WDFY (42). HepG2 cells were cultured on a 12-well plate containing cover slips (diameter, 18 mm) coated with poly-L-lysine and transfected with GFP-LC3 and mCherry-WDFY (PI3P reporter). Cells were treated with 500 nM SAR405 (VPS34 inhibitor) under starvation (EBSS) for 8 h. Co-expression of mCherry-WDFY and GFP-LC3 can distinguish between early autophagosomes (WDFY+LC3+) and mature autophagosomes (WDFY-LC3+).
We did quantification as described for imaging analysis with mCherry-GFP-LC3 reporter. We calculated the number of PtdIns(3)P spots per cell from WDFY positive (red) spots per cell. From analysis in cells expressing both mCherry-WDFY and GFP-LC3, early autophagosomes were quantified from WDFY+LC3+ spots (yellow) and late autophagosomes from WDFY-LC3+ spots (green).
Expression of fluorescent PtdIns(3)P-binding proteins, such as DFCP1, UVRAG, and WDFY, shows somewhat basal structures, like ER, that are not involved in autophagy depending on their initial locations. During induced autophagy, newly formed PtdIns(3)P spots should be counted as phagophores or early autophagosomes. In normal conditions (for example, nutrient-rich media), PtdIns(3) spots seem to be localized on the endosomes, the ER, and the Golgi.
It is difficult to monitor the process of fusion of autophagosomes with lysosomes by using fluorescence imaging. Mature autophagosomes require a specific SNARE Syntaxin 17 on their outer membranes to fuse with lysosomes (45). To distinguish autophagosomes before and after their fusion with lysosomes, we quantify the fusion process by measuring what portion of LC3 positive spots are colocalized with lysosomal markers. Immunofluorescence of LAMP1 or LAMP2 proteins in cells expressing mCherry-GFP-LC3 shows that autophagosomes appear as a yellow circular vesicle (mCherry+GFP+LC3) and autolysosomes appear as red spots (mCherry+GFP–LC3) inside a circular far-red vesicle (LAMP1/2+) (Fig. 2F).
HepG2 cells were cultured on a 12-well plate containing cover slips (diameter, 18 mm,) coated with poly-L-lysine for both live-cell imaging and fixed-cell staining. Cells were infected with mCherry-GFP-LC3 adenovirus overnight. For immunofluorescence staining of cells for LAMP1 and LAMP2, cells were fixed with 100% cold methanol and incubated with 5% horse serum (Gibco-BRL), 0.1% Triton X-100 in PBS for 30 min and then incubated for 30 min at room temperature with primary antibodies to LAMP1 (#9091, Cell Signaling) or LAMP2 in the same solution. After washing with PBS, cells were incubated for 30 min at room temperature with Alexa Fluor 633– conjugated goat secondary antibodies (Invitrogen) at a 1:1,000 dilution in PBS containing 5% horse serum and 0.1% Triton X-100. Cells were also stained with 0.2 μg/ml 4’6’-diamidino-2-phenylindole to detect DNA. We acquired confocal images by illuminating with a 488-nm, a 561-nm, or a 640-nm laser with a microscope (A1R, Nikon) equipped with a galvano detector, and processed images with NIS-Elements AR 3.0 software.
We did quantification as described for imaging analysis with mCherry-GFP-LC3 reporter, with some modifications. We used a colocalization module to measure the number of autolysosomes by quantifying overlapping spots from both mCherry+ GFP-LC3 and LAMP1/2+. The number of autophagosomes was measured by counting mCherry+GFP-LC3 spots.
Treatment with an inhibitor of lysosome function, such as Bafilomycin A (a V-ATPase inhibitor), induces accumulation of enlarged autolysosomes. Red spots (mCherry+GFP-LC3) inside a circular far-red vesicle (LAMP1/2+) are counted as autolysosomes.
At the final stage of autophagy, cytosolic materials in the autolysosome are degraded by lysosomal hydrolases. A defect in lysosome function increases the number of autolysosomes that are larger than normal (34, 46). The difference between inhibition of autophagosome–lysosome fusion and a defect in the autolysosomal degradation can be perceived, because the fusion of autophagosome with lysosome is independent of V-ATPase-mediated acidification in the lysosome (35, 46). Careful analysis of enlarged autolysosomes shows malfunction of the degradation process in the autolysosome (Fig. 2G). In cells expressing mCherry-GFP-LC3, the size of mCherry+GFP+ spots (autolysosomes) increases during treatment with Baf A1 (V-ATPase inhibitor).
HepG2 cells were infected with mCherry-GFP-LC3 adenovirus overnight. Infected cells were incubated with 100 nM Baf A1 for 8 h, fixed, and observed using confocal microscopy as described for imaging analysis with mCherry-GFP-LC3 reporter.
In measuring autolysosome size, we measured mCherry+GFP- spots in control cells (without treatment). Because Baf A1 blocks degradation of mCherry-GFP-LC3 in autolysosomes by inhibiting lysosomal acidification, Baf A1-treated cells have mCherry+GFP+ spots, indicating both autophagosomes and autolysosomes. We analyzed the size of the spots by measuring the diameter or circumference of each spot with NIS-Elements AR 3.0 software.
Quantification of the size of autophagosomes or autolysosomes has not been the focus of autophagy researchers looking at autophagy flux analysis. This size analysis is useful to estimate degradation and lysosome reuse processes from autolysosomes.
Confocal imaging using fluorescent reporters together with immunostaining has allowed the identification of autophagy regulators. Here, we summarize current fluorescence imaging methods to analyze autophagy flux. Careful quantification of imaging data is required to reveal whether a candidate compound is an inhibitor or a promoter of autophagy flux (Table 1). The number of PtdIns(3)P spots indicates the initial stages of autophagy. The total number of autophagosomes and autolysosomes and their ratio imply several processes during the transition of early autophagosomes to autolysosomes. Colocalization of autophagosomes with lysosomes shows fusion of autophagosomes and autolysosomes. Analysis of the size of autolysosomes indicates degradation processes in autolysosomes. We review modified steps of autophagy with the use of five quantitative imaging analyses (Table 1). When treatment with an unknown compound or malfunction of a gene shows a phenotype that is categorized into a specific family, its effect on autophagy flux will be proposed.
Methods to analyze autophagy in mammalian cells using imaging have improved substantially with regard to development of imaging reporters and instrumentation. A variety of tools have been introduced and summarized in many review articles (11, 14, 15, 47). In this review, we propose that comprehensive understanding of confocal imaging data of PtdIns(3)P, mCherry-GFP-LC3, lysosome markers, and autolysosome size sheds light on new regulators of autophagy and modes of action of a candidate compound. Especially, accurate quantification of autophagic vesicles using fluorescence imaging requires the reliable, uniform expression of reporters among cells. Development of stable cells expressing reporters and production of viruses to carry reporters are suggested solutions. We anticipate improvements to imaging instrumentation (structured illumination microscopy, SIM; stochastic optical reconstruction microscopy, STORM; stimulated emission depletion microscopy, STED microscopy; confocal microscopy with Airyscan) with super-resolution below the optical diffraction limit of ∼250 nm and imaging analysis programs to process and quantify large datasets. We hope these improvements motivate autophagy researchers to screen and characterize pharmacologic agents targeting autophagy to treat human diseases such as cancer and neurodegenerative disorders.
Autophagy flux is a dynamic process consisting of several intermediate membraned structures, such as phagophores, autophagosomes, and autolysosomes. Accumulation of specific intermediates results from not only inhibition but also activation of autophagy (11, 48). Careful investigations to distinguish specific intermediate vesicles among them, as briefly summarized in this review, are required for identifying modulators of autophagy flux and elucidating modes of action of specific modulators.
This study was supported by grants from the National Research Foundation of Korea (2019R1A6C1010020 and 2018R1D1A1 A02049371) and from the Ministry of Food and Drug Safety in 2016 (16173MFDS009) in Korea.
The authors have no conflicting interests.
Expectation of inhibition or promotion of autophagy flux with imaging analysis
|Modified step of autophagy processes||Number of PI3P spots||Total number of autophagosomes & autolysosomes||Ratio of autolysosomes to autophagosomes||Size of autolysosomes||Colocalization of autophagosomes with lysosomes|
|Inhibition of formation of phagophores||+||+||ND||ND||ND|
|Inhibition of maturation of autophagosomes||++||+||ND||ND||ND|
|Inhibition of formation of autolysosomes||++||+++||+||++||+|
|Inhibition of degradation process in lysosomes||++||+++||+++||+++||+++|
|Promotion of autophagy||+++||+++||++||++||++|
Five different imaging methods are used to evaluate effects of specific compounds on autophagy flux. +: low, ++: average, +++: high, ND: non-determined.