Catch us if you can

Tumours have ways of evading the body's immune system. A surprising example involves a mechanism that at first sight would seem to have the opposite effect and improve immune responsiveness.

How can tumours develop when the immune system should instead attack and destroy them? Especially difficult to understand is the situation in which tumour cells display surface molecules that should identify them as abnormal, and so should specifically activate immune cells. This makes little sense from the tumour's point of view. These molecules should flag the tumour cells for destruction, unless they somehow give the tumour an advantage — a situation that should remind those of a certain age of the Dave Clark Five, and the immortal lines of one of this 1960s band's songs: “Here they come again, Catch us if you can.” On page 734 of this issue, Groh et al.1 offer insight into this puzzle. They describe a mechanism of tumour evasion that may also apply to the impairment of other immune defence systems, and which may well have clinical potential.

Immune 'effector' cells express surface receptors that are involved in activating the cell by transmitting signals to the cell interior2. In the case of specialized immune cells termed cytotoxic T lymphocytes, or CTLs, such signals may result in the killing of an adjacent target cell, a process called cytotoxicity2. For signalling to occur, the target or tumour cell must display molecules or ligands on its surface that are specifically recognized by the effector cell's activation receptors.

Cytotoxic T lymphocytes possess exquisitely specific activation receptors — T-cell receptors — that initiate signalling2. But CTL activation also requires other, co-stimulatory, receptors; otherwise, only partial activation or sometimes immune paralysis occurs. The recently characterized NKG2D receptor takes on this function when it is coupled to a signalling chain known as DAP10 (Fig. 1)3. Many ligands for NKG2D — MICA and MICB (collectively termed MIC) in humans, and RAE1 and H60 molecules in rodents — are induced by stress and do not occur in large amounts under normal circumstances4,5. For example, virus infection can result in the production of MIC, thereby increasing the susceptibility of infected cells to NKG2D-dependent killing by CTLs6. Although there are other NKG2D ligands that may be present in normal tissues, they are less well understood7. However, we do know that many tumours display stress-induced NKG2D ligands, and that the ligand-expressing tumours are more susceptible to NKG2D-dependent cytotoxicity by CTLs and related natural killer cells than are tumours that lack these ligands. Furthermore, the transfer of genes for NKG2D ligands into tumour cells that do not express these molecules results in enhanced killing.

Figure 1: Tumour evasion of the immune system.

a, NKG2D ligands such as MIC are often induced in cells by 'stress' — viral infection, for instance, or transformation of the cell into a tumour cell. The result should be that the cell becomes more susceptible to killing by cytotoxic T lymphocytes (CTLs): NKG2D probably works as a receptor that co-stimulates the T-cell receptor (not shown) through the DAP10 signalling chain. b, As Groh et al.1 demonstrate, there are soluble forms of MIC circulating in the blood of patients with MIC-expressing tumours which can affect NKG2D function by blocking recognition of membrane-bound MIC by CTLs and downregulating NKG2D expression. The result is that the CTL effectively becomes blind to the presence of the tumour.

A peculiar situation thus exists in the many malignancies — especially those of epithelial tissues (breast, ovary, colon, lung and prostate) — whose cells express MIC on their surface in vitro, often at high levels, and which therefore show increased susceptibility to cytotoxicity (Fig. 1a). This is seen even in freshly isolated tumour cells, so that artefacts of in vitro cell culture cannot explain MIC expression1. And freshly removed, carcinogen-induced mouse tumours express RAE1 and H60 (ref. 8). So why would tumours display molecules that increase their sensitivity to the immune system?

Groh et al.1 have come up with an answer. They show that cancer patients whose tumour cells express surface MIC (MIC+) also have soluble forms of MIC (sMIC) in their blood (Fig. 1b). Blood-serum extracts from these patients decrease NKG2D expression and function, reactions that are also seen when normal lymphocytes are exposed to recombinant sMIC or to MIC+ or RAE1+ tumours themselves1,9. Also, NKG2D expression on circulating lymphocytes is abnormal in patients with MIC+ tumours, suggesting that the function of most lymphocytes may be affected. Obviously, the way in which sMIC expression correlates with clinical outcome, and its potential for diagnosing and monitoring patients, require further investigation. Whether these observations on sMIC extend to other NKG2D ligands also remains to be seen.

But there is another interesting aspect of these findings. Because NKG2D may be involved in host defence against many pathogens6,7,10,11,12, persistent overall downregulation of NKG2D may provide another explanation for cancer patients' predisposition to infections. So the findings of Groh et al. might explain not only tumour evasion but also other aspects of immune suppression in cancer patients. Furthermore, these findings suggest that a similar evasion strategy might exist for other receptors and ligands in the immune system.

Soluble MIC affects NKG2D function in at least two ways — by acting as a competitive mimic that blocks recognition of membrane-bound ligand, and as a suppressor that downregulates NKG2D expression (Fig. 1b)1. The mechanism of this second activity is unclear, but is reminiscent of the ligand-stimulated internalization of other lymphocyte-activation receptors by soluble ligands2. In either case, it is notable that NKG2D has very high affinities and structural plasticity for its ligands13,14. This may partly explain why sMIC has a marked effect on NKG2D even when it is virtually undetectable by immunoassay. Regardless of this, sMIC could be used as a therapeutic agent to inhibit immune cells when the immune system is overactive, as occurs in autoimmune diseases. These include conditions where MIC expression is enhanced, such as inflammatory bowel disease12, and perhaps even disorders where MIC itself is not involved — NKG2D is widely expressed on large numbers of immune cells and has many other ligands.

How does sMIC arise? A clue comes from the association of sMIC with MIC+ tumours, from which it seems that surface-bound MIC may give rise to sMIC. This could occur by specific enzymatic cleavage, as is seen in the processing of soluble forms of members of the tumour-necrosis-factor family2. On the other hand, alternatively spliced forms of MIC RNA transcripts for the soluble forms might be produced at the same time as transcripts for the membrane-bound forms. Whatever the answer, this question should spawn research into potential pharmaceutical targets. If the first mechanism is involved, blockade of the enzymatic event could prevent sMIC from being produced, and possibly increase production of the uncleaved membrane-bound forms that flag tumours for destruction by NKG2D-activated immune cells. We therefore have a new hint as to how to catch the tumour cells that have long eluded both the immune system and immunologists.


  1. 1

    Groh, V., Wu, J., Yee, C. & Spies, T. Nature 419, 734–738 (2002).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Paul, W. E. Fundamental Immunology (Lippincott-Raven, Philadelphia, 1999).

  3. 3

    Wu, J. et al. Science 285, 730–732 (1999).

    CAS  Article  Google Scholar 

  4. 4

    Cerwenka, A. et al. Immunity 12, 721–727 (2000).

    CAS  Article  Google Scholar 

  5. 5

    Diefenbach, A., Jamieson, A. M., Liu, S. D., Shastri, N. & Raulet, D. H. Nature Immunol. 1, 119–126 (2000).

    CAS  Article  Google Scholar 

  6. 6

    Groh, V. et al. Nature Immunol. 2, 255–260 (2001).

    ADS  CAS  Article  Google Scholar 

  7. 7

    Cosman, D. et al. Immunity 14, 123–133 (2001).

    CAS  Article  Google Scholar 

  8. 8

    Girardi, M. et al. Science 294, 605–609 (2001).

    ADS  CAS  Article  Google Scholar 

  9. 9

    Gilfillan, S., Ho, E. L., Cella, M., Yokoyama, W. M. & Colonna, M. Nature Immunol. (in the press).

  10. 10

    Krmpotic, A. et al. Nature Immunol. 3, 529–535 (2002).

    CAS  Article  Google Scholar 

  11. 11

    Das, H. et al. Immunity 15, 83–93 (2001).

    CAS  Article  Google Scholar 

  12. 12

    Tieng, V. et al. Proc. Natl Acad. Sci. USA 99, 2977–2982 (2002).

    ADS  CAS  Article  Google Scholar 

  13. 13

    Radaev, S., Rostro, B., Brooks, A. G., Colonna, M. & Sun, P. D. Immunity 15, 1039–1049 (2001).

    CAS  Article  Google Scholar 

  14. 14

    Strong, R. K. Mol. Immunol. 38, 1029–1037 (2002).

    CAS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Wayne M. Yokoyama.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Yokoyama, W. Catch us if you can. Nature 419, 679–680 (2002).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing