Recognition of microbiota-derived N-formyl methionine peptides by non-classical MHC class I, H2-M3-restricted CD8 T cells

A new study in journal Cell revealed that a specialized subset of CD8 T cells recognize conserved microbial derived N-formyl methionine peptides in context of non-classical MHC class I molecule H2-M3.

Using skin commensal Staphylococcus epidermidis (S. epidermidis) challenge model, the authors showed that recognition of this microbe by CD8+ T cells specifically relied on H2-M3 rather than other MHC molecules.

H2-M3 has been known to bind peptides that contain an N-formyl methionine (fMet), which is required to initiate protein translation in bacteria and mitochondria. Indeed, fMet peptide:H2-M3 tetramer (f-MIIINA:H2-M3) could stain a population of CD8 T cells from skin or lymph nodes. Interestingly, half of tetramer positive CD8 T cells were CD44^hi even in germ-free mice indicating cross-reactivity with other antigens or "ready-made" origin similar to thymus-derived Tregs.

To show biological significance of these CD8 T cell population recognizing commensals, the authors used H2-M3 KO mice and wound healing experiment. The authors claim that in the absence of H2-M3 wound healing was delayed, though data do not strongly support such assertion.

In summary, the authors described yet another T cell population recognizing products from bacteria. Because such recognition does not produce inflammation the authors suggest it could be involved in wound repair. But wound repair model provided very minimal support for such hypothesis. Basically, they missed the central point to make this paper relevant. So how it ended up in Cell? Ask editors. I could come up with couple of suggestions. When the paper's authors list includes Giorgio Trinchieri, John O’Shea and senior author, Yasmine Belkaid, all heads of big labs at NIH, it is quite difficult to say no.

posted by David Usharauli

Receptor-ligand specific labeling of immune cell interactions

This week journal Nature published new study in immunology that could be best described as a method paper. I personally don't understand the value of this paper been in Nature. Only positive characteristic I see in it is that experiments reported are done in a classical, cellular immunology "fashion" and very easy to follow and understand. Lets examine.

The new method, called LIPSTIC, which this study reported is about labeling receptor-ligand pair with naturally occurring enzyme, the Staphylococcus aureus transpeptidase sortase A (SrtA). SrtA appears to covalently transfers a substrate containing motif ‘LPXTG’ to a nearby oligoglycine (G5). Basically, receptor is genetically fused with SrtA and ligand is fused with G5 and when they interact, SrtA catalyses the transfer of the substrate onto the G5-tagged receptor. This transfer can be visualized by attaching to the substrate small labels such as biotin or a fluorophore.  

First the authors showed in vitro and ex vivo that the substrate transfer was specific to fused receptor-ligand pair and did not occur when enzyme was inactive or fused to unrelated receptor. Most of the reported experiments were done using CD40L/CD40 pair (CD40 signaling is relevant for DCs and CD8 T cell activation by CD40L expressing CD4 T cells).

Receptor-ligand specificity was maintained in vivo as well. Because CD40L upregulation on CD4 T cells depends on antigen-specific interactions, substrate transfer were restricted to those CD40+ DCs that were pulsed with cognate peptide (OVA).

However, this antigen-specific CD40L/CD40 interaction was maintained only for initial 10h-24h period. When OVA-specific T cells were left with DCs for longer period (48h) then even DCs pulsed with irrelevant peptide (LCMV peptide) got labeled with substrate. 

It is not clear what are the biological consequences of such non-specific T/DCs interactions. Are these two different DCs activated antigen or non-antigen specific manner somehow different with the regard of activation of CD8 T cells? We don't know. The problem with such model is that if activated CD4 T cells expressing CD40L can interact with CD40+ DCs and activate it (license it, to use polly matzinger's words) then we should expect that body should harbor only activated DCs because body constantly contains some number of CD40L+ activated CD4 T cells specific for all kind of antigens. It is possible high number of T cells transferred in these experiments created an artificial outcome.

In summary, this study showed new method how to label receptor-ligand pair in vivo. However, overall relevance of this method is not clear at present.

posted by David Usharauli

Microbiota-wide association studies and PD-1 immunotherapy

This week Science published 3 back-to-back studies with the findings that responses to PD-1 immunotherapy in cancer patients could be stratified based on presence of certain microbiota species (at least two of these studies were published couple of weeks back as first release papers). 

The trouble is that all three papers found different set of microbiota who they thought mediated responsiveness to PD-1 immunotherapy (dominated by Bifdobacterium, Akkermansia, or Faecalibacterium). 

In one study, it was actually a difference between set of beneficial microbiota versus nonbeneficial ones (not simply a single species), above certain ratio (>1.5), that determined responsiveness to PD-1 therapy.

Moreover, study of germ-free mice transplanted with opposite sets of microbiota (derived from 3 responders/nonresponders) were inconclusive because 1 set of microbiota from each group showed reverse effect with PD-1 immunotherapy. 

In summary, we have no clear understanding of these results. There is no way to predict if the same microbiota set would provide any benefit to a given patient. In general, presently microbiota research lacks direction and rules necessary to untangle its complexity. 

As far as I know, the model we have developed, SPIRAL, is the only one that provides a rational how to identify specific microbiota species. Right now it is just a guideline but when fully developed it will look similar to period table-like map that will make it easy to accurately pinpoint microbiota species relevant for antigen-specific immune response in any given individual.

posted by David Usharauli