Antigen recognition by human γδ T cells: one step closer to knowing

Sensing of self and non‐self phosphoantigens by human Vγ9/Vδ2 T cells in the context of the butyrophilin family members BTN2A1 and BTN3A1.

Immunology & Cell Biology 2020; 98: 351-354; doi: 10.1111/imcb.12334 The mammalian immune system is characterized by a magnificent division of labor between lymphocyte subsets that act in concert to sense and fight the widest possible range of threats to health and life. Broadly speaking, classical CD4 + T cells screen extracellular spaces for the presence of foreign structures that have been captured, processed and bound to major histocompatibility complex (MHC) class II by antigen-presenting cells. CD8 + T cells are in turn wired to inspect host tissues for the expression of abnormal proteins displayed in the context of MHC class I during viral infection or malignancy. Natural killer cells complement these responses by surveying the body for loss of functional MHC molecules on infected or transformed host cells.
cd T cells pride themselves in having so far escaped the scientists' desire to assign them a clear role within conventional schemes of immunology, not least because of their apparent lack of a general restricting element akin to MHC. Rather, evidence is mounting that in many cases these "unconventional" T cells may instead detect metabolic changes in host cells and survey tissue integrity by monitoring expression of stressrelated self-antigens and nonpeptide molecules in the absence of orthodox antigen presentation. 1,2 Two elegant studies by Rigau et al. 3 and Karunakaran et al. 4 now make an important contribution toward our understanding of how a subset of human cd T cells responds to such nonpeptide compounds.
In the vertiginous field of unconventional T-cell biology, human Vc9/Vd2 T cells are strangers in a strange land. Vc9/Vd2 T cells typically comprise 0.5%-5% of circulating blood T cells and respond readily to both self-and non-self-metabolites, often referred to as "phosphoantigens." 2 By far the most potent of these is (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate (HMB-PP), a microbial precursor of isopentenyl pyrophosphate (IPP), the ubiquitous building block of all higher isoprenoids. HMB-PP is produced by the majority of Gram-negative bacteria and bad microbes including Mycobacterium tuberculosis, Clostridium difficile and Listeria monocytogenes, as well as by malaria parasites and Toxoplasma gondii.
Humans do not produce HMB-PP, instead generating IPP via the mevalonate pathway, one of the commercially most attractive drug targets in the body. Inhibition of bhydroxy b-methylglutaryl-coenzyme A reductase by statins abrogates downstream generation of IPP, and consequently, the synthesis of higher isoprenoids, thereby lowering cholesterol levels in individuals at risk of cardiovascular disease. Further along in the mevalonate pathway, inhibition of farnesyl pyrophosphate synthetase can be achieved with aminobisphosphonate drugs such as zoledronate, which is used in the clinic to halt excessive bone resorption in individuals with osteoporosis, multiple myeloma and bone metastasis, but also leads to the elevation of levels of upstream metabolites, including IPP, within zoledronate-treated cells. These zoledronate-induced metabolic troubles of a mevalonate pathway out of control are sensed by Vc9/ Vd2 T cells and can be counteracted using statins.
The mysterious ways underlying the Vc9/Vd2 T-cell responsiveness to HMB-PP, IPP and zoledronate have long puzzled scientists. 5  illustrious family of only a handful of confirmed ligands for human cd T cells that comprise diverse molecules such as endothelial protein C receptor, annexin A2 and, most recently, also the MHC-related molecule MR1. 2,6 The clear implication of BTN2A1 in human Vc9/Vd2 T-cell responses is a sort of homecoming for cd T-cell aficionados, because the related protein BTN3A1 had already been shown to be required for Vc9/Vd2 T-cell responses to phosphoantigens. 7 Other butyrophilin family members with cd T-cell regulatory activity include BTNL3/BTNL8 (human Vc4 + T cells), Btnl1/Btnl6 (mouse Vc7 + T cells) and Skint1/Skint2 (mouse Vc5 + T cells). 8 Of note, HMB-PP and IPP were previously shown to bind to the intracellular B30.2 domain of BTN3A1, 9 yet it has remained unclear whether and how this binding might translate to recognition of the extracellular portion of BTN3A1 via the Vc9/ Vd2 T-cell receptor (TCR). Additional factors have been postulated to be involved in this response, including F1-ATPase, periplakin, RhoB and ABCA1, yet none of these acts as direct ligand for the Vc9/Vd2 TCR. 2 First and foremost, the papers by Rigau et al. 3 and Karunakaran et al. 4 are a striking testimony to the power of modern techniques that have allowed scientists to propel forward Figure 1. Vc9/Vd2 T-cell response to microbial antigens and self-phosphoantigens-the little things that give it away. BTN2A1 binds directly to the Vc9 chain of the cd TCR, possibly as homodimer, while also interacting closely with BTN3A1, which may additionally form homodimers and/ or associate with BTN3A2 or BTN3A3. Whether these contacts occur simultaneously, or involve sequential interactions at another time, another place remains to be resolved as indicated by the question marks in the figure. In particular, the role of BTN2A2 is unclear at present. Phosphoantigens may reach the B30.2 domain of BTN3A1 via various pathways: intracellularly upon changes in the metabolic flux through the mevalonate pathway, for example, after zoledronate treatment, or upon release from pathogens in the cytosol or in phagocytic vesicles. Exogenous HMB-PP and IPP may be shuttled directly into the cytosol, or be taken up by pinocytosis or endocytosis. Phosphoantigen binding to BTN3A1 may take place in the cell surface membrane or, far away so close, in a putative loading compartment inside the cell, and may be aided by other BTN3 isoforms and cofactors. Of note, baseline responses to BTN2A1 in the absence of phosphoantigens are still detectable, and the precise ligand specificity of the CDR3 regions of the Vc9/Vd2 TCR remains unknown. BTN2A1, butyrophilin 2A1; HMB-PP, (E)-4-hydroxy-3-methylbut-2-enyl pyrophosphate; Ig, immunoglobulin; IPP, isopentenyl pyrophosphate; TCR, T-cell receptor.
like a bullet the blue sky concepts that could not easily be tested with older methodologies. Rigau et al. employed a genome-wide CRISPR/ Cas9 (clustered regularly interspaced short palindromic repeats/CRISPRassociated protein 9) knockout screen of LM-MEL-62 melanoma cells that stained brightly with a soluble Vc9/Vd2 TCR tetramer, and identified BTN2A1 as the most significant guide RNA responsible for tetramer reactivity. 3  Of note, both groups report that BTN2A1 binds within seconds to Vc9/Vd2 as well as Vc9/Vd1 TCRs, and provide evidence for a direct interaction between BTN2A1 and the Vc9 chain while ruling out contacts with the Vd2 chain. Carefully targeted mutation of the Vc9 chain combined with structural modeling mapped a series of key amino acids involved in this interaction, in particular R20, E70 and H85. 3 Vice versa, targeted mutation of R65, R124, Y126 and E135 in the BTN2A1 protein completely abrogated binding to the Vc9/Vd2 TCR. 4 So how does the discovery of BTN2A1 as a Vc9/Vd2 ligand square with the previous implication of BTN3A1 in phosphoantigen sensing? Well, sometimes you can't make it on your own. Both molecules appear to be essential but not sufficient on their own, and only coexpression of BTN2A1 and BTN3A1 elicits Vc9/Vd2 T-cell responses to HMB-PP or zoledronate. And despite both molecules possessing intracellular B30.2 domains, phosphoantigens appear to only bind to the B30.2 domain of BTN3A1 but not to that of BTN2A1. 3 Using confocal microscopy and F€ orster resonance energy transfer analyses, Rigau et al. in fact demonstrate that BTN2A1 and BTN3A1 colocalize at the cell surface and that the two molecules associate intracellularly and extracellularly. 3 These results are complemented by crosslinking and immunoprecipitation studies by Karunakaran et al. using a 16-A spacer, indicating a close and possibly direct association of BTN2A1 with BTN3A1 on the cell surface. 4 The model emerging from these findings suggests that BTN2A1 exclusively touches the germlineencoded HV4 region of the Vc9 chain, with no involvement of the Vd2 chain, whereas phosphoantigen specificity of the TCR is conferred by BTN3A1 recognition via both the CDR3 region of the Vc9 chain and the CDR2 region of the Vd2 chain. 3,4 But is that all? The potential contribution of the other members of the BTN2/3 family, and whether target recognition occurs with or without them, remains elusive. Previous reports suggested that BTN3A2 and/or BTN3A3 may help facilitate surface expression of BTN3A1 and form heterodimers. Here, BTN3A2 was not required for the phosphoantigen-dependent activation of Vc9/Vd2 T cells by

BTN2A1-BTN3A1
co-transfected hamster cells, but nevertheless enhanced the response moderately. 3 In addition, BTN2A2 appears to be capable of binding to the Vc9 chain with similar affinity to BTN2A1, with as yet unknown relevance. 4 To walk on, we will require further biochemical and structural work into the precise contribution of the different BTN2 and BTN3 isoforms, the role of the B30.2 domains and their discrimination between selfand non-self phosphoantigens, the vesicular trafficking pathways that shuttle phosphoantigens to the cytosol, and characterization of the intracellular and extracellular partner molecules involved (Figure 1).
The implications of these new findings are manifold and light the way for future investigations. For instance, it is tempting to speculate that BTN2A1 and/or BTN3A1 may play a role in the exit of positively selected mature Vc9/Vd2 T cells from the thymus. Moreover, given that BTN2A1-BTN3A1 complexes can be formed independently of phosphoantigens and that BTN2A1 elicits a relatively small but discernible baseline response on its own, 5 there is a light for the existence of a further, true antigen that is recognized by the Vc9/Vd2 TCR in a CDR3-mediated fashion but has remained invisible so far. Intriguingly, such a hypothetical ligand/cofactor must be conserved evolutionarily -BTN2A1-BTN3A1transduced mouse and hamster cells are targeted readily by human Vc9/ Vd2 TCRs, 3,4 and for the first time a nonprimate original of the species, the alpaca (Vicugna pacos), has been reported to also possess a phosphoantigen-reactive Vc9/Vd2 T-cell subset. 11 Finally, given the strong translational and applied interest in Vc9/Vd2 T cells, BTN2A1 is a highly relevant target candidate for the development of agonistic and/or antagonistic miracle drugs in the context of microbial infection, autoimmunity and tumor immunotherapy.
The seminal publications by Rigau et al. and Karunakaran et al. on BTN2A1 as a novel ligand for human Vc9/Vd2 T cells extend previous reports of similar c-chain-specific, superantigen-like interactions of other butyrophilin family members with cd T cells in humans and in mice, evoking Susumu Tonegawa's unforgettable prediction that cd T cells "may be involved in an entirely new aspect of immunity," 12 clearly distinct from classical CD4 + and CD8 + T cells and MHC-restricted adaptive immunity. Yet one caveat remains. We may be one step closer to knowing but with regard to understanding the antigen specificity of the cd TCR CDR3 region we still haven't found what we are looking for. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.