Autophagy, meaning “self-eating” in Greek, is an evolutionarily conserved cellular degradative mechanism in eukaryotes. Autophagy maintains cellular homeostasis by degrading unwanted cellular materials in the lysosomes; and thus eliminates harmful contents such as intracellular pathogens including bacteria, viruses, fungi, and protozoa, damaged or surplus organelles, and harmful protein aggregates. The role of autophagy is not limited to the elimination of cargo - it plays a regulatory role in other processes such as innate and adaptive immunity, inflammation, development, diseases such as neurodegenerative diseases, cancer, metabolic diseases, and cardiovascular diseases among others [1]. For his pioneering work on autophagy, Dr. Ohsumi Yoshinori won the Nobel Prize in 2016 which underscores the importance of autophagy in health and diseases. Various pre-clinical and clinical studies have explored whether manipulation of autophagy can be used for treating various diseases in animal models.

Three types of autophagy exist: macroautophagy, microautophagy and chaperone-mediated autophagy. Macroautophagy encapsulates the cargo for degradation in a double-membrane structure called autophagosome which fuses with the lysosome and delivers the contents for degradation. Microautophagy mediates direct capturing of the cargo by lysosomes without the formation of autophagosomes [6]. Chaperone-mediated autophagy involves selective degradation of the target proteins directly into the lysosome which is delivered by a chaperone [7]. Of these, the most well-studied macroautophagy (hereafter called autophagy) includes mitophagy (for mitochondria), ERphagy (for endoplasmic reticulum), pexophagy (for peroxisomes), nucleophagy (for nucleus), ribophagy (for ribosomes), aggrephagy (for aggregates) and xenophagy (for pathogens). Here, we discuss the process of autophagy and some of the common methods for monitoring autophagy in vitro and in vivo.

Autophagy is initiated with the formation of a double-membrane structure called autophagosome around the cellular material earmarked for degradation. Autophagosome then delivers its contents to the lysosome by fusing with the lysosomes to generate autolysosomes where the contents are degraded by the activity of proteolytic enzymes. Multiple proteins are involved in authophagy (Table 1) and new involved proteins are being identified - the COPII-cargo adaptor complex Lst1-Sec23 was found to be involved in the ER-phagy [8]. Autophagy is initiated by ULK1, Beclin1 and PI3K-III complexes that promote the formation of an isolation membrane [9]. The isolation membrane gets decorated with LC3 (a mammalian homolog of yeast Atg8) which serves as an anchor for recruiting the selected cargo. LC3 is first cleaved by ATG4 to yield LC3-I revealing the glycine residue on LC3. Two ubiquitin-like conjugation systems ATG7-ATG3 and ATG5-ATG12-ATG16 further facilitate the recruitment of LC3-I to the autophagosomal membrane by conjugating LC3-I to lipid PE molecules to generate LC3-II which is recruited to the inner and the outer membranes of the autophagosomes [10]. Proteins such as p62 act as adaptor receptors that dock the cargo to the autophagosome by binding to the cargo on one end and to LC3-II on the other and deliver the cargo to the autophagosomes [11]. Fusion of autophagosome with the lysosome yields autolysosomes and this process is mediated by a Soluble NSF Attachment Protein Receptor (SNARE) protein syntaxin 17 [12]. The inner membrane of the autolysosome is subsequently degraded which includes LC3-II bound to the inner membrane while LC3-II on the outer membrane is released and recycled by the activity of ATG4. The process of cargo selection and degradation is referred to as autophagy flux and is indicative of functional and complete autophagy. Some of the well-studied inducers and inhibitors of autophagy are listed in Table 2, 3, respectively.

Several "standard" methods of monitoring autophagy are discussed below and summarized in Table 4. Often several methods are combined together to detect autophagy. For example, Saito T et al used the conversion of LC3-I to LC3-II and p62 degradation in Western blots as indicators of autophagy [14].

Multiple approaches can be used to study the LC3 II turnover.

Upon induction of autophagy, LC3-I is lipidated to become LC3-II which migrates faster than its parent molecule on an SDS-PAGE gel. LC3I and LC3II migrate as ~16 and ~14Kda proteins respectively. Thus, increase in LC3-II level has been considered as a gold standard for demonstrating autophagic induction. However, as LC3-II is subsequently degraded by the activity of proteases in the autolysosomes (a process referred to as autophagy flux), the increase in LC3-II by a test compound may not necessarily reflect autophagy activation but may also suggest a block in autophagy flux [18, 19]. To distinguish between these two possible outcomes, one uses inhibitors of lysosome-dependent degradation such as Bafilomycin A1, E64d, pepstatin A, and leupeptin that either prevent the fusion of autophagosomes with autolysosome or block the proteolytic activity of the degradative enzymes in the autolysosomes. For example, Ling Q et al treated Arabidopsis protoplasts with 1-10 uM E64 from Melford for 2 hours to inhibit autophagy [17]. Using these inhibitors, the following outcomes may be possible when testing a reagent for its ability to induce autophagy.

• The treatment causes an increased amount of LC3-II in the presence or absence of the inhibitors when compared to the untreated controls. This would suggest that autophagy flux is increased in this scenario.
• The treatment causes no change in LC3-II expression with or without the flux inhibitors in comparison to the untreated samples. This suggests that autophagy is inhibited under these conditions.
• The treatment causes increased LC3-II in the absence but not in the presence of the inhibitors. This suggests that under this condition, autophagy flux is suppressed.

Finally, as LC3 antibodies have an overall greater tendency for LC3-II compared to LC3-I, it is not recommended quantifying the LC3-II/LC3-I ratio, although it is still commonly used [20]. Instead, one should measure the LC3-II/actin or other housekeeping gene ratio when quantifying immunoblots [21]. Autophagy induction must also be validated by using genetic inhibitors such as siRNAs against autophagy genes such as Atg5, Atg7, or Beclin1 and/or by using pharmacological inhibitors such as 3-methyladenine (3 MA), or conditional knockout of genes like ATG7 [22]. Some of these inhibitors are discussed in Table 3. While testing the ability of a test reagent to induce autophagy, one should also use known autophagy inducers and inhibitors as positive and negative controls respectively in their studies. For instance, Tamaki Y et al used bafilomycin at 0.1 μM in HEK293A cell culture to demonstrate the involvement of autophagy–lysosome activity in intrabody-induced TDP-43 protein degradation [23].

LC3-II can also be monitored by fluorescence microscopy and can be distinguished from the parent LC3-I molecule due to its ability to become membrane bound via its lipidated moiety. Thus, while LC3-I is cytosolic, membrane-bound LC3-II represents autophagosomes that are visible as dots or “puncta”. Detection of puncta can be accomplished either by using antibodies against endogenous LC3 or by using GFP-LC3 expressing plasmid [27, 36]. While an increase in the number of GFP-LC3 puncta was considered to be indicative of an increase in autophagy in the past, now, the same criteria of using inhibitors such as bafilomycin A1 should be used to distinguish between the activation of autophagy and the suppression of autophagy flux, when using LC3 puncta to quantitate the autophagosomes.

A more advanced method for measuring autophagy flux by LC3 puncta is to use tandem labeled fluorescent LC3 expression vectors such as mRFP-GFP-LC3 [27] or mCherry-EGFP-LC3 plasmids [37]. EGFP-LC3 plasmid was initially developed as one of earlier probes to monitor autophagosome formation in the cell [36]. However, it presents significant limitations as GFP is low pH-sensitive and is quenched in the autolysosomes. Thus, information pertaining to the autophagic flux is not possible with the use of GFP-LC3 unless inhibitors such as bafilomycin A1 are used. RFP or cherry molecule (red) on the other hand is more stable in low pH autolysosomes and thus red puncta would reflect only autolysosomes while the yellow puncta (GFP and RFP signal) would reflect autophagosomes thereby providing a detailed analysis of autophagy flux. For example, Vodnala SK et al used an eGFP –mCherry-LC3b fusion reporter system to evaluate the impact of extracellular potassium concentration on autophagic flux in live T cells, using an autophagy-incompetent construct containing a Gly120 → Ala substitution (G120A) in LC3b as a negative control [20].

Another plasmid that has been recently developed to include an internal control is GFP-LC3-RFP-LC3ΔG. Upon cleavage of this protein by Atg4 in the cells, two distinct populations of LC3 are released, GFP-LC3 that decorates the autophagosomes and RFP-LC3ΔG that remains [38] in the cytosol and serves as an internal control for the quantity of GFP-LC3 before it is degraded by the autolysosomes. The outcome measure used in this assay is the ratio of GFP/RFP that is decreased upon an increase in autophagy.

Despite the certain advantages of LC3 puncta method, the galectin puncta formation assay is believed to be better for analyzing lysosomal membrane permeabilization due to its high sensitivity [24]. This technique employs the process of galectin translocation to impaired lysosomes, which is followed by a switch from homogenous to punctate staining distribution.

Detection of autophagosomes can also be monitored by using Flow cytometry. While microscopy is a great technique to visualize the autophagosomes in a small population of cells at a given time, flow cytometry offers a high-throughput analysis of the autophagosomes. Moreover, flow cytometry can also be used to detect autophagosomes in non-adherent cells such as blood monocytes. The outcome measured by flow cytometry is a decrease in GFP-LC3 fluorescence which is based on the principle that GFP-LC3 is degraded in the autolysosome and thus a decrease in GFP signal would indicate autophagy flux [39]. Flow cytometry can also be used with dyes such as CYTO-ID which are selectively accumulated in the autophagic vacuoles. CYTO-ID can also be detected by fluorescence microscopy [4].

Tandem labeled fluorescent LC3 expression vectors can also be analyzed with flow cytometry, concurrently with the fluorescent microscopy analysis as discussed above. Vodnala SK et al used the proportion of mCherry⁺GFP^– cells in flow cytometry as a parameter for autophagic flux [20].

P62 is an adaptor receptor that recruits cargo to the autophagosome. P62 also binds to the LC3 on the autophagosomes. P62 is degraded along with the designated cargo in the autolysosomes [11]. Thus, p62 degradation is considered a biomarker for autophagy activity. However, caution must be taken as p62 is also degraded by autophagy-independent pathways such as proteasomes. Thus, to confirm autophagy-mediated p62 degradation, one must use lysosomal inhibitors such as bafilomycin A1 to demonstrate that p62 is no longer degraded in the presence of this inhibitor. E Gelpi et al proposed to use nuclear p62 staining to identify Alzheimer type II astrocytes in metabolic/hepatic encephalopathy [40].

Traditionally this method involves labeling long-lived proteins using radiolabeled amino acids such as ³⁵S-methionine ³H-leucine, ¹⁴C-leucine, and ¹⁴C-valine [41] before inducing autophagy and monitoring the release of radioactivity as a measure of degradation of long-lived proteins upon induction of autophagy [29]. Again, one should use autophagy inhibitors such as bafilomycin A1 to demonstrate that the degradation of long-lived protein is mediated by autophagy and not by proteasomes. A non-radioactive method of detecting long-lived protein turnover is by using L-azidohomoalanine (AHA) through click-chemistry [27, 42].

Electron microscopy remains the best method for visualizing autophagosomes at the ultrastructural level. The nature of the specific cargo can also be detected using this technique. In addition, immunogold labeling can be utilized to identify special cargo and proteins within the autophagosomes [31]. While TEM remains by far the best technique to visualize autophagosomes, it presents significant challenges. The sample preparation is very critical when processing for EM and may require specific methods and expertise to maintain the integrity of the organelles. TEM is also a time-consuming technique and proper identification of the autophagic structure is difficult. Improper sample processing may also lead to the appearance of artifacts that may resemble autophagosomes [43]. Also, significant improvement in the visualizing of autophagosome ultrastructure has recently been achieved by implementing new fixation methods, such as ferrocyanide-reduced osmium and aldehyde/OsO4 mixture for detecting omegasome structures [30].

Increase in the mRNA and protein expression of several autophagy core genes and proteins such as Atg7, Beclin1, Atg5, LC3 often accompany the induction of autophagy [32, 44-46]. These outcomes are measured by real-time qPCR or northern blotting for analyzing gene expression and by western blotting using specific antibodies to analyse the protein expression. However, changes in mRNA and protein levels of autophagy players may not always be observed upon induction of autophagy and should not be used as a sole criterion to demonstrate the activation of autophagy and must be accompanied by other assays demonstrating functional autophagy.

Keima is a coral-derived fluorescent protein with a bimodal excitation spectrum depending on the pH. Keima exhibits excitation at 440 (appears as green) and 586 nm (appears as red) in neutral and acidic environments, respectively. Keima exhibits a low 550/438 excitation ratio in the cytosol under basal conditions. Upon induction of autophagy, an increase in the punctate structures with a higher 550/438 ratio is observed which represent autolysosomes [33]. Keima provides advantages over the RFP-GFP-LC3 system as it does not depend on the LC3 lipidation system and is useful in detecting autophagy that is independent of Atg5 conjugation system. However, Keima cannot be used in the fixed cells as it requires the lysosomal acidity to function. Keima has been successfully used for demonstrating mitophagy.

The cargo sequestration analysis applies a cargo marker of total cytplasm to measure bulk autophagy [34]. A recent study assessed the sequestration of cytosolic protein lactate dehydrogenase in autophagic vacuoles of cells treated with inhibitors of intravacuolar degradation. The method was based on electrodisruption of the cell membrane followed by centrifugal sedimentation.

A recent study has presented a novel semiconductor-based method to estimate autophagy induced by nutrient starvation [35]. The authors have used a semiconductor-based field-effect transistor (FET) biosensor to monitor pH changes in starved HeLa cells. The detected positive shift correlated with the elevated numbers of hydrogen ions due to increased cellular respiration in autophagic cells.

Assays such an LC3 turnover and p62 degradation by western blot and immunohistochemistry can also be monitored in animal tissues [47, 48]. However, the optimal detection of LC3 and p62 in animal tissues is difficult and may depend on the source of tissue and the method of protein extraction. Some tissues have a higher level of LC3-I while others show a much higher level of LC3-II. Moreover, tissues have a heterogeneous population of cells which may present challenge of variability in detecting LC3-II. Moreover, autophagy flux is difficult to study in the whole animal and require the administration of flux inhibitors such as bafilomycin which may be toxic to the animals. Under such circumstances, ex vivo methods may be utilized where tissues are isolated and treated with the inhibitors and autophagy flux is determined by western blotting. Transgenic mice expressing GFP-LC3 have also been utilized to monitor autophagy in vivo [47]. Typically, cryosections are used for detecting GFP-LC3 positive autophagosomes in such animal models for testing autophagy induction. Advancement of this model is the development of mice which express RFP-GFP-LC3 [49, 50]. Moreover, cells derived from such animals can also be grown in vitro and used for further analysis [51]. Recently, mice expressing GFP-LC3-RFP-LC3ΔG have also been developed [38]. Such mice models would be very valuable in studying autophagy induction in vivo. Conventional and organ-specific autophagy-deficient mouse models have also been fundamental in advancing our knowledge of autophagy [52].

In addition to conventionally used laboratory methods, commercially available kits from various sources are also available that maybe used in conjunction with the methods described above. These kits often use certain dyes to label the autophagic vacuoles which then can be detected by fluorescence microscopy, flow cytometry, or a microplate reader. Some of these kits are:

• Autophagy Assay Kit (ab139484): Abcam
• Autophagy Assay Kit (MAK138-1KT):Sigma Aldrich
• CYTO-ID autophagy detection kit (ENZ-KIT175-0050): Enzo
• Autophagy Detection Reagent Pack (CF200097): Millipore

Of these, CYTO-ID detection kit has been used relatively commonly by several researchers to monitor autophagy in their studies [4, 53]. The dye used in this kit is a cationic amphiphilic tracer (CAT) dye that has been titrated to a dose that specifically labels autophagic vacuoles. Typically cells are first treated with a test reagent. Use of positive and negative controls such as rapamycin and chloroquine respectively are highly recommended for this assay. After incubation, the cells are washed and incubated in the dark with the CYTO-ID detection reagent and cells are then and analyzed by microscopy or flow cytometry. Both adherent and suspension cells can be analyzed using this reagent. Unlike conventional laboratory autophagy detection assays, these assays are less time-consuming. However, these assays should always be used in conjunction with more standard methods of monitoring autophagy rather than being used as a single method to detect autophagy.

Field of autophagy is dynamic and is ever evolving. New assays and techniques are constantly being developed to overcome existing caveats and to make progress in our current understanding of this field. Today, several methods exist that assist in the detection of autophagy in vitro and an in vivo. However, each method presents its own significant strengths and weaknesses. Thus, it is highly recommended that researchers use multiple methods when demonstrating autophagy in their experimental models and use appropriate positive and negative controls, wherever possible. In addition, researchers should also measure autophagy flux in their assays as a demonstration of a complete autophagic pathway. Finally, the autophagic assays should also be validated using genetic or pharmacological inhibitors of autophagy.