Apoptosis Curation Manual
From GO Public
General guidelines for annotating to cell death-related terms
Rationale and aim
The motivation behind the apoptosis GO project was originally to clean up and expand the node under GO:0006915 “apoptosis”, which contained poorly structured and connected terms that did not reflect the current scientific knowledge on such an important area of biomedical research. Once the project started, it became apparent that its scope needed broadening, and some cell death-related terms above apoptosis have also been edited. In the present document we aim to provide a curation guide, starting from the most general terms directly under “cell death”, and walking our way down the ontology, hopefully providing a “decision tree” for GO curators. In particular, this document is meant as a useful tool to possibly re-house annotations that are currently placed under very generic terms in the new ontology (i.e. “apoptotic process”, “regulation of apoptotic process” and “induction of apoptosis”). There are still many direct manual annotations to these broad terms (1591, 691 and 861 respectively), and a good proportion could be made more specific and therefore gain in information content.
This draft does not aim to be fully exhaustive, and only the most relevant changes to the ontology made during the apoptosis project will be discussed. Regulatory terms are not discussed unless required. In several cases, we have added definition comments to cell-death related terms, as a ‘permanent’ guideline for curators. If you think that more terms would benefit from having such comments, or that their definitions are unclear, please let us know – your suggestions will be very welcome.
The ontology work consisted in changes and additions to Biological Process (BP) terms under the “cell death” branch, and to Molecular Function (MF) terms that have a relationship with apoptosis (caspase-related terms). Because some of the changes and new definitions challenge and even contradict the old ontology, we highly recommend reading through this document before using the new terms.
Below we will describe the most important branches of the cell death node as represented in the figure below. GO terms are linked to corresponding entries in QuickGO, and some annotation examples are provided (in green).
Types of cell death other than apoptosis
Current scientific knowledge of different types of cell death is not as rich as it is for apoptosis, and it is sometimes difficult to ascertain that the process we want to curate is indeed a distinctive type of cell death. However, some are sufficiently characterized, and there are terms available for them to annotate with, as follows.
GO:0070265 necrotic cell death - and children
Contrary to apoptosis and other types of cell death, necrotic cell death is mainly characterized by the loss of integrity of the plasma membrane and the loss of intracellular contents. However, in some instances it can occur under controlled conditions. In this case, the process should be annotated to the new term “programmed necrotic cell death”, or to its more specific child “necroptosis”, as follows.
GO:0097300 programmed necrotic cell death
This new term was required to update the placement of existing “necroptosis” (GO:0070266, see below); it is a child of both “necrotic cell death” and “programmed cell death”. Its definition is the same as the old one for “necroptosis”: “A necrotic cell death process that results from the activation of endogenous cellular processes, such as signaling involving death domain receptors and Toll-like receptors.” The definition of “necroptosis” has been made more specific:
“A type of programmed necrotic cell death characterized by activation of receptor interaction protein kinase 1 and/or 3 (RIP1 and/or RIP3).” Necroptosis is the focus of several recent papers, and it can occur via formation of a protein complex called ripoptosome (see GO:0097342).
- Annotation example:
See PMID:21052097, annotations to Birc2 (Q62210) and Birc3 (O08863).
GO:0012501 programmed cell death - and children
As discussed above, there are instances of programmed cell death that escape the definition of apoptosis. These may be annotated to the general term GO:0012501 “programmed cell death” or to one of its children. Also, the apoptotic process as defined in GO:0006915 does not occur in some taxa, such as plants. Where evidence is shown for a controlled (programmed) cell death modality in such species, the general term GO:0012501 “programmed cell death” may be used, and some specific children also exist: GO:0097468 “programmed cell death in response to reactive oxygen species” (with descendants GO:0010421 “hydrogen peroxide-mediated programmed cell death” and GO:0010343 “singlet oxygen-mediated programmed cell death”), and GO:0009626 “plant-type hypersensitive response” and its children, which are descendants of GO:0034050 “host programmed cell death induced by symbiont”.
Parthanatos, netosis, entosis, pyroptosis and other types of cell death
Although it has been proposed to create new, specific terms for each one of these subtypes of cell death, the discussions with the experts of the APO-SYS consortium made it clear that research in the field for most of these terms is still quite preliminary and it is not clear whether they reflect real biological events or not. For the time being, no new terms have been created, but the only existing one (GO:0070269 pyroptosis) has not been obsoleted because of an increasing number of publications on pyroptosis in the last few years.
GO:0006915 apoptotic process
The term name has been changed from “apoptosis” to “apoptotic process”, to better reflect its composite nature. Two part_of children have been added to illustrate the division of this process in two distinctive phases (signaling and execution). This concept has been incorporated in the definition of GO:0006915. Direct annotations to ‘apoptotic process’ should be restricted to those instances where it is not possible to distinguish which phase is specifically involved; see below for some annotation examples. The phases and their children terms are discussed at the end of this section. As for direct is_a children terms of GO:0006915:
This term indicates a specific type of apoptosis induced by inadequate adherence of the cell to its substrate.
GO:0097285 cell-type specific apoptotic process - and children
This term was created as a generic parent to group terms indicating occurrences of apoptotic process in specific cell types. The process of apoptosis may not vary, but the cell type does. Please do not annotate to the parent term, and refer instead to its more granular children, or to GO:0006915 “apoptotic process”.
- Annotation example:
This came up while re-curating PMID 10770950, where the human protein O15162 had previously been annotated to GO:0006915 apoptotic process. The curator found that the annotation could be made more granular by moving it to GO:0001781 neutrophil apoptotic process. However, the previous annotation to 'apoptotic process' had already been ISSed to other species (dog, pig, macaque and platypus). While these are all mammals, it's hard to say if neutrophils, the way they're defined in GO and in the Cell Ontology, are also found in platypus ('any of the immature or mature forms of a granular leukocyte that in its mature form has a nucleus with three to five lobes connected by slender threads of chromatin, and cytoplasm containing fine inconspicuous granules and stainable by neutral dyes').
In this particular example, if the curator moves the human annotation from 'apoptotic process' to 'neutrophil apoptotic process', the ISSes will be lost (or cause problems anyway) because in Protein2GO they can't point to a different GO term. Fixing this in P2GO would not be trivial in the short run. But if the curator leaves the human annotation to 'apoptotic process' AND adds an annotation extension pointing to CL:0000775 neutrophil, the human annotation will still be made more specific, and the ISSes won't be lost. So we would suggest that, if you run across similar instances, you add the cell type ID as an annotation extension, rather than using a cell-type specific term or using/requesting a pre-composed cell-type specific term. Note that it's still ok to use (or request) cell-type-specific apoptosis terms if yours is a new annotation, or if a pre-existing annotation is not ISSed to other species, or if you don't use P2GO/annotation extensions.
Signaling and execution phases of apoptosis
The most important change in the apoptosis branch of the BP ontology is the creation of two part_of children of “apoptotic process”, to make a clear distinction between two phases: GO:0097190 “apoptotic signaling pathway” and GO:0097194 “execution phase of apoptosis”. During the signaling phase, only initiator / apical caspases (e.g., CASP2, CASP8, CASP9, and CASP10) are activated, and no breakdown of cellular structures occurs. The execution phase starts with the activation of effector / executive caspases (e.g., CASP3, CASP6, CASP7), and during its course cellular structures break down in a controlled fashion, ultimately leading to cell death and the formation of apoptotic bodies. The creation of parent terms depicting both phases allows for the correct classification of many children terms that describe more granular aspects of apoptosis.
GO:0097190 apoptotic signaling pathway - and children
Gene products involved in events that take place in order to activate initiator caspases can be annotated under this parent term or its corresponding regulatory terms. Most of the mitochondrial events related to apoptosis (discussed later) take place during this phase.
Extrinsic and intrinsic apoptosis
The term “apoptosis signaling pathway” has the children GO:0097191 “extrinsic apoptotic signaling pathway” and GO:0097193 “intrinsic apoptotic signaling pathway”. The existence of cross-talk points between these two pathways often makes it difficult to distinguish between them, as both extrinsic and intrinsic apoptosis can be mediated by mitochondrial events. To distinguish between these two pathways, curators should take the initial point of the signaling pathway as a reference, as detailed below. The more generic term ‘apoptotic signaling pathway’ can be used to annotate gene products involved in apoptotic events happening downstream of the cross-talk point between the extrinsic and intrinsic pathways. Coming from the extrinsic apoptotic signaling pathway, the cross-talk starts when caspase-8 cleaves Bid and the truncated product induces mitochondrial outer membrane permeabilization. From this point on it is not possible to distinguish between extrinsic and intrinsic pathways.
Also, in instances where it is not known what kind of signal initiated the apoptotic process, the generic parent term ‘apoptotic signaling pathway’ should be used.
GO:0097191 extrinsic apoptotic signaling pathway
It starts with the reception of an external signal at the level of the cell surface (either by ligand binding or by ligand withdrawal) and can be mediated by mitochondrial events or not. Children terms refer to specific signals that activate the pathway, the best-known case being [GO:0008625] extrinsic apoptotic signaling pathway via death domain receptors. It is important to note the new term [GO:0097192] extrinsic apoptotic signaling pathway in absence of ligand. This particular term refers to the apoptotic process triggered by the withdrawal of a ligand, such as a growth factor, that is under non-apoptotic conditions bound to a dependence receptor. It is important to remember that apoptosis triggered by growth factor withdrawal has been traditionally considered an instance of intrinsic apoptosis. However, the detailed characterization of the molecular mechanisms underlying this particular type of cell death resulted in a more appropriate classification as a specific variety of extrinsic apoptosis.
- Annotation examples:
GO:2001240 negative regulation of extrinsic apoptotic signaling pathway in absence of ligand: PMID:10666185 shows that platelet-derived-chemokine platelet factor 4 (PF4) induces the differentiation of monocytes into macrophages, and PF4 stimulation prevented monocytes from undergoing spontaneous apoptosis.
Monocytes undergo apoptosis presumably because signals preventing them to do so are absent in the culture conditions the authors use. This paper most likely points to extrinsic apoptosis in the absence of ligand. Depending on the information available:
If the paper itself includes information to support this (e.g., the methods cell culture conditions may state serum withdrawal), annotate using IMP or IDA to GO:2001240 'negative regulation of extrinsic apoptotic signaling pathway in absence of ligand'.
If you can find previous reports proving that apoptosis is indeed induced by this mechanism, then make an annotation to GO:2001240 'negative regulation of extrinsic apoptotic signaling pathway in absence of ligand' using the new apoptosis GO-REF and IC codes.
If there is no information confirming ligand withdrawal, annotate just using the IDA or IMP codes, and restrict the annotation to 'neg. reg. of apoptotic process'.
A diagram of the extrinsic apoptotic pathway including examples of molecular players can be found in Figure 1 in PMID:21760595.
GO:0097193 intrinsic apoptotic signaling pathway
It starts with the reception of an intracellular signal, although the cause of the signal can be extracellular: e.g., UV light exposure can induce DNA damage that leads to intrinsic apoptosis. (Also various chemicals can induce DNA damage see link). The source of the damage comes from the outside, but the signal that starts the apoptotic process is the damage itself, which is intracellular. It is always mediated by mitochondrial events (and often referred to as “mitochondrial apoptosis pathway”). Note that not the mode of action of some chemicals is not known see link Children of this term refer to the particular type of signal that starts the apoptotic process.
- Annotation example:
See PMID:21771788, annotations to ZPR9 (ZNF622, Q969S3) and ASK1 (MAP3K5, Q99683).
A diagram of the intrinsic apoptotic pathway including examples of molecular players can be found in Figure 2 in PMID:21760595.
Other children of apoptotic signaling pathway
There are events that escape the classification for extrinsic or intrinsic pathway of apoptosis, so they need specific terms and these have been placed under the more generic parent “apoptotic signaling pathway”:
GO:0008626 granzyme-mediated apoptotic signaling pathway
Granzymes are proteolytic enzymes released from cytoplasmic granules produced by NK cells and T-cells. As such, they can activate effector caspases directly (prominently, caspase-3), by-passing the extrinsic and intrinsic pathways. They can also cleave proteins with a role in the intrinsic pathway, activating apoptosis through it.
GO:0008628 hormone-mediated apoptotic signaling pathway
This term may be used to annotate apoptotic events triggered by hormones. Note that different hormones may activate apoptosis via different signaling pathways. Because of this variability, annotation towards the specific instance of the signaling pathway that is activated must be made conservatively, specifying it only when sufficient evidence is provided about the mechanism of action of the given hormone. For example: in some instances, (the absence of) a hormone may be triggering apoptosis via 'extrinsic apoptotic signaling pathway in the absence of ligand'. If this is clearly shown in a paper, annotations based on that publication may be made to both 'hormone-mediated apoptotic signaling pathway' and 'extrinsic apoptotic signaling pathway in the absence of ligand'. But unless clear evidence is provided, annotations should be made simply to 'hormone-mediated apoptotic signaling pathway'.
GO:0097194 execution phase of apoptosis - and children
Gene products involved in events that take place after effector/executioner caspases have been activated and before the cell dies can be annotated to this term and to its following children:
GO:0006921 cellular component disassembly involved in execution phase of apoptosis - and children
This term and its children can be used to annotate for the breakdown of cellular structures during apoptosis.
GO:0070782 phosphatidylserine exposure on apoptotic cell surface
In normal cells, phosphatidylserine residues are found exclusively on the inner side of the cellular membrane. During apoptosis, phosphatidylserine is transported to the outer cell surface by scramblase proteins. This event acts as an "eat-me" signal for macrophages to dispose of the dying cell. When annotating to this term, curators should be aware of the following: 1) phosphatidylserine exposure on the cell surface can occur in circumstances other than apoptosis (for example, when blood platelets are activated, they expose phosphatidylserine to trigger the clotting system, see PMID:21107324). Do not annotate to GO:0070782 unless phosphatidylserine exposure is shown to be part of an apoptotic event. 2) Caution should be applied when a study quotes annexin V assays. The annexin A5 protein binds to phosphatidylserine-containing membrane surfaces, which are usually only present on the inner leaflet of the membrane. However, in cells undergoing apoptosis, phosphatidylserine becomes exposed on the outer leaflet of the membrane. A positive annexin assay can therefore be linked to apoptotic death, but it shouldn't be confused with molecular events strictly involved in the process of phosphatidylserine exposure.
GO:0008637 apoptotic mitochondrial changes - and children
Due to the importance of the role of the mitochondrial compartment in apoptosis, this specific term was created to reflect events affecting it. When current knowledge allows for it, children and descendants of this term are linked to either the signaling or the execution phase of apoptosis, via “part_of” relationship. The following children of GO:0008637 deserve a specific comment:
GO:0043653 mitochondrial fragmentation involved in apoptotic process
Although most of the processes described under ‘apoptotic mitochondrial changes’ take place during the signaling phase of apoptosis, this one cannot be confidently placed there. It is still controversial whether this process is involved in the signaling phase of apoptosis or not, so it was placed under the more generic apoptotic mitochondrial changes parent and it will not be linked to the signaling or the execution phase until further research clarifies the matter.
GO:0097345 mitochondrial outer membrane permeabilization (MOMP) - and children
This important mitochondrial event allows for the release of cytochrome c from the mitochondrial intermembrane space (see below) and the formation of the apoptosome, triggering the activation of caspase 9. It is therefore a key step in the apoptotic signaling pathway and it is sometimes monitored experimentally to detect apoptosis. Curators should be cautious when annotating to its regulatory terms, since pretty much any apoptosis-inducing event will trigger MOMP. Only events directly and specifically affecting this permeabilization should call for an annotation to MOMP or to its regulation terms.
MOMP is initiated by the insertion of certain proteins into the mitochondrial membrane through a mechanism that is not yet fully understood. Gene products involved in this initiatory event can be annotated to the term GO:0001844 “protein insertion into mitochondrial membrane involved in apoptotic signaling pathway”.
- Annotation example:
See PMID:20850011, annotations to Drp1 (DNM1L, O00429).
GO:0001836 release of cytochrome c from mitochondria
Even more than in the previous case, annotation to this term, and especially to its corresponding regulatory terms, should be done with caution. Monitoring the release of cytochrome c is a common way to detect when apoptosis is occurring and it is widely used in the literature. Only when experimental evidence is provided that a gene product directly and specifically affects this release, an annotation should be made to this term or to its regulation terms.
- Annotation example:
See PMID:20850011, annotations to Drp1 (DNM1L, O00429).
GO:0032976 release of matrix enzymes from mitochondria
Annotation to this term should be done with extreme caution, since it is not clear whether this is a recognized process in the scientific community. In the only annotation that the term has (PMID:9843949) the authors perform a couple of assays in which they detect the release of matrix enzymes, but they consider them as evidence of induction of the “permeability transition pore”, a controversial structure whose physiological significance is still unclear.
Annotations to regulatory terms under GO:0006915 “apoptotic process”
Apoptosis is a fundamental cellular response to different types of injury or aggression. As such, experimental assays knocking-down or over-expressing particular gene products will often trigger it, or diminish the cell’s ability to enter apoptosis. It is then tempting to annotate these gene products with the “positive/negative regulation of apoptotic process”-type terms. However, apoptosis can occur in these cases as a side effect of the disruption of a major pathway; if so, the gene products may not really be acting as direct, specific regulators of apoptosis, and the annotation would be incorrect.
On the other hand, it is often difficult to ascertain if we are having a specific effect or not. Therefore, as a general rule of thumb, we would recommend avoiding annotating to regulatory terms for specific terms such as GO:0090201 “negative regulation of release of cytochrome c from mitochondria”, unless evidence for the specificity of the regulation is given. In this case, more general terms, such as GO:0006915 “apoptotic process”, allow for more flexibility in the annotation while still providing a link to the apoptotic process, and call for case-by-case decisions.
- Annotation example:
A situation where it was decided not to annotate to regulation of apoptotic process, since it was considered a side-effect of disrupting a key protein, is the knock-down of RPS19, a protein that has an essential role in ribosome biogenesis. The study reported that cells that had rps19 silenced displayed condensed DNA, a fragmented nucleus and cleavage of the caspase substrate, PARP. These are all indicative of apoptosis being triggered. Since the function of RPS19 in the assembly of the small ribosomal subunit has been well-studied it was considered that regulation of apoptosis is an indirect effect of the disruption of such an important process and as such an annotation was not made to regulation of apoptosis.
The word “caspase” was removed from all the term names related to this group of enzymes because it represents a gene product and because the distinction from other types of proteases is outside the scope of GO. Since the activity of caspases can be described as a “cysteine-type endopeptidase activity”, this wording replaces “caspase” wherever it was used in GO. See e.g. discussion in the following link:
There are terms such as GO:0006919 “activation of cysteine-type endopeptidase activity involved in apoptotic process” that specifically refer to a particular instance of the positive regulation of caspase activity. The aim of these terms is to capture the complexity of the caspase activity regulation process. Since caspases are synthesized as zymogens and then cleaved into active proteases, there is a need for terms that can capture both the activation process and the regulation of the enzymatic activity part. This way, proteins that directly perform or enable the cleavage of pro-caspases into their active, mature forms may be annotated using this term. If they increase the activity of an already active caspase or they activate it in an indirect fashion, they should be annotated using the parent term GO:0043280 “positive regulation of cysteine-type endopeptidase activity involved in apoptotic process”.
Only the positive regulation terms are mentioned here as guidance, but there are negative (inhibiting) ones as well when it applies.
It's important to note that changes in 'apoptotic effectors' (such as activation of caspase-3) do not always result in apoptotic programs. For example, while full-blown mitochondrial outer membrane permeabilization (MOMP) constitutes a point-of-no-return of intrinsic apoptosis, limited extents of MOMP (i.e., concerning a fraction of the mitochondrial pool) and the consequent (localized) activation of caspase-3 have been shown to participate in several cell death-unrelated programs such as the differentiation of megakaryocytes and granulocytes (see 'Pros and Cons...' in PMID:21760595). See PMID:18309324 for a review of apoptosis-unrelated functions of gene products that are commonly considered part of the 'apoptotic machinery'.
These review papers are recommended reading or reference for anyone undertaking curation or re-curation of apoptosis papers:
PMID:21760595 where cell death subroutines are accurately defined by biochemical features;
PMID:19373242 for interpretation of cell-death assays and related parameters, and aiding to distinguish between a simple readout and the actual occurrence of an apoptotic process or sub-process. For instance, caspase activation has often been (mis)interpreted as an unequivocal sign of apoptotic cell death, when it is in fact known that caspases also participate in many processes not linked to cell death;
PMID:21415859 contains useful information and figures on common apoptotic stimuli and their downstream mechanisms (see figures 5 and 6 in particular). More specific stimuli should be examined on a case-by-case basis. For instance, the mechanism of cell death induced by cyclosporin A is not yet fully understood and may vary depending on cell type or tissue (see e.g. fig. 6 in PMID:23347876 and fig. 2 in PMID:23295863.
Other annotation examples
In some cases, where detailed information is not provided in publications, it may be safer to adopt a more conservative approach and annotate to broad terms.
Example 1: PMID:10037816 the authors studied expression of a transfected gene and showed that it is unregulated when cells are induced to apoptose by external factors (staurosporine treatment). The curator debated about annotating to 'positive regulation of extrinsic apoptosis signaling pathway', but wasn't sure this was correct because it's not known where the gene product is regulating the apoptotic process. In fact, it is not clear how staurosporine induces apoptosis (see e.g. http://en.wikipedia.org/wiki/Staurosporine). Therefore, in the absence of further details, we suggested to annotate to the more general term 'positive regulation of apoptotic process'.
Example 2: PMID:10097103 the authors knock out a gene and show that retinal cells are more sensitive to light-induced apoptosis. The curator thought about annotating to 'negative regulation of intrinsic apoptosis signaling pathway' because presumably the light is causing some type of internal stress. But again, it is not sure where the regulation is taking place. The apoptosis signal might even triggered by a separate population of cells sending signals possibly mediating extrinsic apoptosis. In the absence of any evidence, we'd consider it safer to annotate to the broader 'negative regulation of apoptotic process'. In this particular paper, the annotation may gain in information by specifying the cell type involved.
Some commonly used chemicals and their mode of action on apoptosis
This section may be considered a work in progress. It lists drugs often quoted in experimental papers as inducing apoptosis, along with their mechanism of action. In some cases this is unclear or unknown. Curators may refer to the notions listed here to annotate to specific apoptotis terms. However, caution is recommended whenever a paper doesn't provide enough information, and when drug mode of action is unclear.
Curators are welcome (and encouraged) to add more examples as they come across them in papers.
- Chemicals that cause DNA damage (also referred to as genotoxic stress):
Use link for more details on topoisomerase inhibitors
Other DNA damage chemicals
Temozolomide interferes with DNA replication
Doxorubicin acts by intercalating DNA. Also known as hydroxydaunorubicin. Trade/brand names: Doxil, Adriamycin PFS, Adriamycin RDF, Rubex, Caelyx, Myocet.
Cisplatin (cisplatinum, cis-diamminedichloroplatinum(II), CDDP, Platin) binds DNA and causes crosslinking of DNA
- Chemicals whose mode of action to induce apoptosis is unknown/complex:
Cyclosporin A (CsA, Ciclosporin, cyclosporine)
Staurosporine The mechanism of how it mediates induction of apoptosis is not well understood. It has been found that one way in which staurosporine induces apoptosis is by activating caspase-3