O. Mandleboim, Lautenberg Center for General and Tumor Immunology, The Hebrew University, The BioMedical Research Institute, Hadassah Medical School, Jerusalem 91120, Israel. Email: firstname.lastname@example.org Senior author: Ofer Mandelboim
NKG2D is one of the best characterized activating receptors and is expressed on natural killer cells and on various T-cell subsets. This receptor recognizes several different ligands that are induced by cellular stresses. In this review, we described the mechanisms controlling the expression of NKG2D ligands, with the emphasis on post-transcriptional and post-translational regulation.
For many years natural killer (NK) cells were thought to be controlled only by inhibitory mechanisms. Killing of target cells was considered to be a default, resulting from the absence of inhibitory signals. The discovery of activating receptors expressed by NK cells added another layer of complexity to our understanding of how NK killing is controlled, and thus the regulation of activity of NK cells is now considered in terms of a balance between inhibitory and activating signals.
The activating receptor NKG2D is a potent stimulatory immunoreceptor that is expressed on NK cells, NKT cells, γδ+ T cells and CD8+ T cells.1 Engagement of NKG2D by its ligands leads either to the direct activation of killing and cytokine secretion by NK cells or to a costimulation of cytotoxic T-lymphocyte (CTL) cytotoxicity. Because of this potent killing ability, surface expression of NKG2D ligands (NKG2DL) must be tightly regulated so that the immune response is not triggered inappropriately. In this review, we will concentrate on the possible mechanisms controlling NKG2DL expression.
NKG2D binds to a variety of major histocompatibility complex (MHC) class I-like proteins. In humans these proteins are divided into two families: the MHC class-I polypeptide-related chain (MIC) protein family that contains MICA and MICB; and the cytomegalovirus UL16-binding proteins (ULBP) family, which consists of five members (ULBP1–4 and REAT1G). In mice, there are five retinoic acid early transcript 1 (RAE1) proteins, H60 and MULT-1.1 Although all NKG2DL can bind and activate the NKG2D receptor, a comparison between the different ULBPs and MIC proteins shows no significant homology (< 20%). However, the crystal structure of NKG2D, together with ULBP3 or MICA, has revealed that the receptor–ligand interface is remarkably similar, despite the use of completely different residues for binding.2,3
NKG2DL are expressed on diseased or stressed cells, and numerous stress pathways lead to the up-regulation of these ligands. The first insight into the mechanism of NKG2DL regulation, demonstrating the concept of cellular stress detection by NKG2D, came from the finding that the MICA and MICB promoters contain heat shock elements similar to those found in heat shock protein 70 (HSP70) and that indeed heat shock induces MICA expression.4 Similarly, NKG2DL expression is also up-regulated in response to oxidative stress,5,6 genotoxic stress7 and viral infection.8
However, the exact molecular details explaining how NKG2DL expression is induced are still fragmentary, especially when considering the large number of diverse NKG2DL. We know quite a lot about the induction of NKG2DL transcription. Transcription of MICA and MICB was shown to be induced by proliferation, heat shock, viral infection and oxidative stress. Heat shock-induced binding of HSF-1 was shown to induce MICA and MICB transcription.6 A regulatory pathway that is chronically active in many tumour cells – the DNA-damage response – activates the transcription of genes encoding mouse NKG2DL and ULBP1–ULBP3 in humans.7 This response is initiated by the protein kinases ATM or ATR after the detection of double-stranded DNA breaks or stalled replication forks. However, to date, the exact molecular events linking the DNA-damage pathway or the expression of oncogenes with the up-regulation of NKG2DL transcripts and cell-surface expression still remain elusive.9
By contrast, viral induction of MICA and MICB transcription involves trans activation. It was demonstrated that human cytomegalovirus (HCMV) immediate-early proteins are able to displace histone deacetylases, which induces the transcription of MICA and MICB messenger RNA (mRNA).6 It is likely that differential regulation of NKG2D ligands may enable the immune system to respond to a broad range of cellular emergency cues; however, these are still needs to be fully defined.
The best established role for NKG2D as an antiviral defense mechanism was demonstrated in HCMV infection. Infection with HCMV leads to the up-regulation of NKG2DL transcripts, including MICA, MICB and the ULBPs.8 In order to cope with the NKG2D-mediated killing of infected cells, HCMV have developed several counter-attack mechanisms. The UL16 protein of HCMV was found to bind to MICB, ULBP1 and ULBP2, resulting in the retention and sequestration of these ligands, which consequently prevented their surface expression.10,11 Subsequent studies demonstrated that the viral UL142 protein down-regulates the expression of MICA protein.12 We demonstrated that a microRNA (miRNA) encoded by the HCMV (hcmv-miR-UL112), down-regulates MICB expression by targeting a specific site in the MICB 3′ untranslated region (3′-UTR) and that this down-regulation protects infected cells from NKG2D-dependent killing.13 These observations, together with our recent finding that other virally encoded miRNAs from other herpesviruses also target MICB,14 suggest strong selective pressure to develop various viral mechanisms aiming at reducing NKG2DL expression to avoid immune elimination.
The MIC genes are highly polymorphic: around 60 alleles of MICA and 25 alleles of MICB have so far been identified.1 The reasons why such polymorphisms exist are still unknown, but it has been postulated that such polymorphisms provide an advantage for the host in its fight against viruses.1 Indeed, two studies15,16 have shown that a common MICA allele, which lacks the cytoplasmic tail, is resistant to down-regulation during HCMV infection and it has been suggested that this human variant escaped from the HCMV UL142 targeting.
In tumours, NKG2DL have often been shown to be up-regulated.17,18 Up-regulation of NKG2DL during tumorigenesis may render the emerging tumour cells sensitive to NKG2D-dependent elimination. Indeed, expression of NKG2DL on otherwise NK-resistant tumour lines was shown to mediate the in vivo elimination of these tumours19 and it was also recently demonstrated that NKG2D plays a critical role in tumour immunosurveillance in vivo.20
Similarly to viruses, tumours have developed multiple ways to evade the NKG2D-mediated immune response. At advanced stages of tumour growth, sustained surface expression of NKG2D ligands and shedding of soluble MICA induces the internalization and degradation of NKG2D, thus promoting tumour immune evasion.21,22
NKG2DL also participate in the cross-talk between immune cells, which could regulate innate and adaptive immune responses. In macrophages, lipopolysaccharide (LPS) treatment up-regulates transcription of the ULBP1–3 genes and also enhances surface expression of the constitutively transcribed MICA gene.23 The increased expression of NKG2DL induces NK cell-mediated cytotoxicity, which could allow the elimination of overstimulated macrophages.23 NKG2DL are also up-regulated on dendritic cells by Toll-like receptor stimulation, and they might participate in the activation of T cells and NK cells.24,25 Interestingly in vitro, T-cell activation also results in the up-regulation of NKG2DL expression. Such up-regulation could facilitate an NKG2D-mediated immune cross-talk between activated T cells and other immune cells expressing the NKG2D receptor.26,27
Post-transcriptional regulation of NKG2DL expression
As stated above, the mechanisms controlling NKG2DL expression are not fully understood and we still do not know why different NKG2DL exist. A critical requirement for the expression of NKG2DL is that they will not be expressed on the surface of normal cells (to avoid autoimmunity) and that upon stress their up-regulation would be rapid. Surprisingly, numerous reports demonstrate that different cells and tissues express mRNA for NKG2DL but lack any expression of the corresponding proteins.4,23,28 These observations thus suggest that at least some of the NKG2DL are also regulated post-transcriptionally.
These above-mentioned discrepancies observed between the expression of mRNA transcripts of NKG2DL and expression of NKG2DL protein, together with the surprising finding that the site which is targeted by hcmv-miR-UL112 is conserved in the 3′-UTR sequences of various MICB alleles and that a similar site also exists in MICA 3′-UTR sequences, lead us to hypothesize that the expression of MICA and MICB are controlled by cellular miRNAs.
Indeed, we have identified a group of endogenous cellular miRNAs that control MICA and MICB expression by binding to MICA and MICB 3′-UTR in sites that overlap with the site that is bound by hcmv-miR-UL112.29 We have shown that, under normal conditions, these cellular miRNAs maintain MICA and MICB protein expression under a certain threshold. Furthermore, we demonstrated that during short-term stresses such as heat shock and viral infection, although the amount of miRNAs did not change markedly, elevation in MICA and MICB mRNA expression probably exceeded the miRNAs repression activity, resulting in the observed expression of MICA and MICB protein.29 We therefore suggested that these cellular miRNAs set up a threshold for MICA and MICB protein expression (Fig. 1a).
Such post-transcriptional regulation of miRNAs has several possible advantages. First, MICA and MICB are already transcribed and thus, upon stress, a rapid induction of the expression of NKG2DL could be achieved. Second, this type of regulation can potentially prevent the undesired effects in which small fluctuations in the amounts of MICA and MICB mRNA will lead to NKG2DL expression and self-destruction and ensure that only ‘real’ stress will result in recognition and elimination by immune cells. Similar, miRNA-mediated threshold activity functions in several non-immunological scenarios were previously suggested.30,31
We predict that similar miRNA-based regulation may exist for other NKG2D ligands. Precise sequencing of the 3′-UTR of all NKG2DL is warranted to facilitate the discovery of cellular, and probably also of viral, miRNAs that target additional NKG2DL. Interestingly, under certain conditions the amount of cellular miRNAs could be manipulated to affect NKG2DL expression, because treatment with interferon-γ (IFN-γ) was reported to decrease the expression of MICA on melanoma through the up-regulation of the MICA targeting miRNAs.32 Although the biological significance of this phenomenon is still unclear, these results suggest that in different scenarios, the expression of MICA and MICB proteins could be directly affected by changes in the levels of their regulating miRNAs.
Interestingly, it was demonstrated that miRNAs could also, in certain situations, enhance the expression of their target genes. It was demonstrated that miRNAs which normally suppress the expression of target transcripts can instead activate the translation of the same target genes at different stages of the cell cycle or during stress (such as serum starvation).33,34 It is therefore possible that under normal conditions NKG2DL expression is silenced by miRNAs, whereas during stress the binding of the same miRNAs could actually activate NKG2DL expression. Such a switch in the function of miRNAs can immediately affect NKG2DL expression by enhancing the translation of the existing pool of mRNAs without any need for de novo synthesis and can therefore enable a fast and accurate elevation of NKG2DL expression upon stress (Fig. 1b). Whether this type of mechanism really affects NKG2DL expression, and which type of stresses have the ability to switch miRNAs function, is an issue that still needs further investigation.
In addition, it is well established that the 3′-UTR of mRNA plays an important role in regulating the stability, localization and translation of mRNA.35 Mammalian genes use alternative cleavage and polyadenylation to generate multiple mRNA isoforms differing in the length of their 3′-UTRs. Recently, it was shown that increased proliferation is associated with widespread reductions in the length of the 3′-UTR, which consequently results in widespread reductions in the 3′-UTR-based regulatory capacity of mRNAs.36 It is possible that also in the case of NKG2DL expression, highly proliferating cells, such as tumour cells, express a shorter version of MICA and MICB (and possibly also of other NKG2DL) 3′-UTRs, therefore releasing these proteins from their interactions with miRNAs and possibly other RNA-binding proteins, which enables the expression of NKG2DL (Fig. 1c). In agreement with this hypothesis, the expression of MICA and of MICB was markedly increased in proliferating cells compared with quiescent cells.5,37
These suggested mechanisms – the threshold, the switch in miRNA function, and the shortening of 3′-UTRs (Fig. 1) – are not mutually exclusive. It is possible that, under certain conditions, some of these mechanisms act together, providing an efficient miRNA-based mechanism to control NKG2DL expression.
A post-translational regulation mechanism was recently demonstrated for the murine NKG2D ligand, MULT-1. It was shown that under normal conditions, Mult1 protein undergoes ubiquitination on lysines in its cytoplasmic tail and consequently lysosomal degradation, preventing its surface expression.38 In response to stress such as heat shock or ultraviolet (UV) irradiation, Mult1 degradation and ubiquitination was reduced, providing a post-translational mechanism for stress-mediated cellular control of NKG2DL expression. Among the human NKG2DL, RAET1G and the MIC proteins are transmembrane proteins containing lysine residues within their cytoplasmic tail and thus these proteins might also be regulated by ubiquitin-mediated post-translational regulation.38
NKG2DL might represent one of the best examples of how danger is sensed by the immune system, leading to immune elimination of diseased cells. These ligands are expressed in various tissues and stresses and their expression is controlled, by different mechanisms, at the transcriptional, post-transcriptional and post-translational levels. On the one hand, this multi level regulation probably prevents the expression of NKG2DL on the surface of normal cells and on the other hand it might enable a rapid elevation of NKG2DL protein expression upon cellular stress. Together, it generates a complex and sensitive warning system that alerts the immune system of possible danger through the use of only one receptor, NKG2D.
The authors have no conflicts of interest to disclose.