Add immune complex

[gocentral]

In Immune complex clearance GO:0002434, immune complex is defined as "clusters of antibodies bound to antigen, to which complement may also be fixed"

Can this be defined in cellular component. Would like term without complement bound as well, please.

Reported by: alanr

Original Ticket: geneontology/ontology-requests/7140

[gocentral]

Alan,

The problem with making immune complex a cellular component is that immune complexes are by nature extremely heterogeneous. The immunoglobulins in a complex may be of any isotype, in any number with identical or different specificities. The antigen(s) being bound may be protein, carbohydrate, nucleic acid, or more complicated mixture of moieties, comprising one or more epitopes, and varying in size from individual molecules to whole bacteria.

Thus, immune complexes are not readily defined as protein complexes in the GO. A second issue is the similarity of the term 'immune complex' to the existing 'immunoglobulin complex' terms. It is likely that some users of the GO, including GO annotators, may confuse the terms. 'Immune complex' is of course an immunology term and widely used, whereas 'immunoglobulin complex' is basically a GO creation, since we like to add the word 'complex' to terms that name protein complexes in the GO even when the biological community does not.

Given these caveats, I am curious what use or need you have of an 'immune complex' and an 'immune complex lacking fixed complement' terms, and whether you are content that immune complexes do not fit the standard of protein complexes in the GO and are therefore simply a rather loosely defined 'extracellular part'.

Thanks,

Alex

Original comment by: addiehl

[gocentral]

As I understand it, the point is to make the biological process and cellular component branches consistent -- ideally, if we refer to a component or set of components in a BP term, there should also be a CC term for the component(s).

If it doesn't make biological sense to have a CC term to group the complexes formed by antibodies binding to antigens, that's fine with me, but it will annoy the purists. Perhaps we should consider renaming GO:002434, or adding a "systematic synonym" ? From an ontology point of view, I would also consider including a CC term such as "antibody-antigen complex" (is_a macromolecular complex, since not all antigens are proteins) -- but only if it won't look stupid to an immunologist.

m

Original comment by: mah11

[gocentral]

The use case is that there are many assays that detect such complexes. The complex seems no more or less loosely defined than immunoglobulin complex or protein binding. This is needed for the representation of molecular immunology and for OBI assays.

Original comment by: alanr

[gocentral]

Alan and Midori,

My term suggestions are below. I'm not entirely happy about the is_a to macromolecular complex, since the minimal immune complex would be a single immunoglobulin bound to a single small molecule. However in the majority of cases an immune complex would satisfy the definition of macromolecular complex, so I could be swayed pretty easily. Certainly the immune complexes that are of most interest in autoimmune disease are macromolecular complexes.

I would also like to change the definition of immune complex to "A process directed at removing immune complexes from the body. Immune complexes are clusters of immunoglobulins bound to antigen, to which complement may also be fixed, and which may precipitate or remain in solution." so that it more closely matches the CC defs proposed below.

[Term] id: GO:NewTer1 name: immune complex def: "A cluster of immunoglobulins bound to antigen, to which complement may also be fixed, and which may precipitate or remain in solution. The immunoglobulins in a complex may be of any isotype, in any number with identical or different specificities. The antigen(s) being bound may be protein, carbohydrate, nucleic acid, or more complicated mixture of moieties, comprising one or more epitopes, and varying in size from individual molecules to whole bacteria." [GOC:add, ISBN:068340007X "Stedman's Medical Dictionary"] comment: Note that immune complexes are by nature extremely heterogeneous. Also an immune complex is not a type of protein complex per se as defined by the GO, as the structure has no defined set of subunits. synonym: "antibody-antigen complex" EXACT is_a: GO:0044421 ! extracellular region part intersection_of: has_part GO:0042571 ! immunoglobulin complex, circulating

[Term] id: GO:NewTer2 name: immune complex lacking fixed complement def: "A cluster of immunoglobulins bound to antigen, to which complement has not been fixed, and which may precipitate or remain in solution. The immunoglobulins in a complex may be of any isotype, in any number with identical or different specificities. The antigen(s) being bound may be protein, carbohydrate, nucleic acid, or more complicated mixture of moieties, comprising one or more epitopes, and varying in size from individual molecules to whole bacteria." [GOC:add, ISBN:068340007X "Stedman's Medical Dictionary"] comment: Note that immune complexes are by nature extremely heterogeneous. Also an immune complex is not a type of protein complex per se as defined by the GO, as the structure has no defined set of subunits. synonym: "antibody-antigen complex" BROAD is_a: GO:NewTer1 ! immune complex

I am still very concerned about the potential confusion of these terms with the "immunoglobulin complex" terms, not only by curators, but by tools such as OBOL. I guess we'll have to see what happens in practice.

Thanks,

Alex

Original comment by: addiehl

[gocentral]

1) Are you sure the defs should be related as such? As I understand the situation, first we have the antibody-antigen complex, after which complement components bind. So there would be a derives_from relationship between the antibody-antigen complex and the complex+complement. Some processes happen only for the complement-bound complex. So perhaps (you figure the names - I'm trying to get the idea across)

immune-complex antibody-antigen complex antibody-antigen-complement complex

+ antibody-antigen-complement complex derives_from antibody-antigen complex

2) Is the immunoglobulin itself a complex? If so is that not sufficient to make it+small molecule a complex?

3) I'm not sure how the "defined subunit" criteria works. For example, consider NF-kappaB complex, which has a variety of forms. Or SCF ubiquitin ligase complex which is polymorphic on the F-box.

Original comment by: alanr

[gocentral]

A few comments / questions:

- I am not sure about the "may precipitate or remain in solution" part of the definition, as this implies that it is in solution in the first place, and e.g. in the case of a viral antigen integrated into a cell surface I don't think that is true.

- I am not that concerned about confusion of immune complex and immunoglobulin complex (there will be confusion, but it is not that much worse than what we have in other places - people will have to read definitions).

- Is 'a cluster' meant to be equivalent to 'one or more'?

Original comment by: bjoernpeters

[gocentral]

Hi Alex,

Could you let me know whether you'd recommend any changes to the terms you suggested? I'll add them, as given below or with any needed changes, and then close this item. (I can also include comments that may help a little in avoiding immune/immunoglobulin confusion.)

I'll change the assigned_to and do the work when I hear back from you.

midori

Original comment by: mah11

[gocentral]

• assigned_to: nobody --> aledie

Original comment by: mah11

[gocentral]

Midori,

The truth is, I don't think we should have these terms at all as protein complex terms in the GO, as you cannot define them as have particular subunits. Sure, they have some varying number of immunoglobulins of varying isotype and may or may not have varying amounts of complement proteins as well, and the antigen is in many cases one of the universe of proteins, although not necessarily, so these are not well-defined protein complexes at all.

If the OBI folks no longer need these terms, then we should skip them. Let's see if they still want them.

Thanks,

Alex

Original comment by: addiehl

[gocentral]

I don't feel strongly about having these in GO. up to Alan.

Original comment by: bjoernpeters

[gocentral]

• assigned_to: aledie --> nobody
• status: open --> closed

Original comment by: mah11