Three Es of cancer immunity A couple of weeks ago I talked briefly about immunity to tumors, saying that “this is actually a long-running controversy that has gone back and forth over the years. There’s too much history to treat all in one post, and indeed there’s circumstantial evidence arguing that in fact immune systems are not major players in cancer resistance. … However, the pendulum swing at the moment has it that the immune system does represent a major barrier to cancer progression.

A few days after I wrote that, a paper came out (so far just online) that helps support one of the models for immunity in cancer prevention:

Koebel, C. M., Vermi, W., Swann, J. B., Zerafa, N., Rodig, S. J., Old, L. J., Smyth, M. J., and Schreiber, R. D. (2007). Adaptive immunity maintains occult cancer in an equilibrium state. Nature doi:10.1038/nature06309

Last time I avoided talking about much of the history of this question, but for this paper a little background is necessary — anyway, it’s interesting in itself. Rather than try to explain the back-story and the new paper all in one go I’ll split them up into separate posts. Today I’ll run through some of the history, and later this week I’ll talk about the new paper itself.

The notion that tumors can be controlled by the immune system is an old one, dating back to Paul Ehrlich’s proposal in the early 20th century.1 However, there really weren’t any tools, such as immune-deficient animals, to test the hypothesis until the second half of the century. In the 1950s, as the understanding of immunity became a little clearer, people (especially Sir Macfarlane Burnet and Lewis Thomas) revisited and refined the idea:2

It is by no means inconceivable that small accumulations of tumour cells may develop and because of their possession of new antigenic potentialities provoke an effective immunological reaction with regression of the tumour and no clinical hint of its existence.

Blogging on Peer-Reviewed ResearchHowever, once some tools did become available, experiments didn’t support this “immunosurveillance” concept. At the time, the best model of immunodeficiency was the nude mouse — a spontaneous mutation that results in athymic mice, that (mostly!) lack T cells. These mice are quite immunodeficient, but they do not, it turns out, develop more tumors than do wild-type mice.3

We now believe that nude mice (and especially the mouse strain Rygaard and Povlsen used, which are unusually susceptible to tumors) are not a great model for this question; for one thing, they still do have some T cells, especially as they get older (which is when tumors are most likely to be an issue, of course). Nevertheless, the studies were the best that could be done at the time, and  the results were fairly definitively negative.  In the 1980s and early 1990s, it was at best controversial whether immunosurveillance was a reality, or if the immune system had any effect at all on tumor development. Hanahan and Weinberg’s massively influential 2000 paper4 on the events required for tumorigenesis didn’t even mention the immune system as a problem for the developing tumor.5

Spontaneous tumors in RAG-/- mice But in mid 1990s and early 2000s the pendulum began to swing back. Some of the most important work came from Robert Schreiber’s lab, showing that immunodeficient animals actually are more susceptible to tumors. This started with relatively artificial systems,6 but led to the critical observation that immunodeficient mice — RAG knockout mice, which lack T cells almost completely, a more complete effect than in nude mice — are more prone to spontaneous tumors as they age7 (see the figure at right for an example).

Another critical finding, that I personally found very exciting at the time,8 showed that tumors that form in immunodeficient mice are much more immunogenic than those that form in wild-type mice. Normally, if you transplant a tumor from one wild-type mouse to another, the tumor is likely to “take”, and continue to grow: It is not rejected by the immune system.  If you transplant from a wild-type to an immunodeficient mouse, again it takes.  But if you transplant a tumor from an immnodeficient RAG knockout mouse to a wild-type mouse, this tumor will be rapidly rejected (and the rejection is immune-mediated).9 This strongly suggests that tumors normally develop ways to avoid immunity.  When there’s no immune system to deal with, tumors don’t have to develop immune evasion mechanisms, and then when these immune-naive tumors are suddenly transplanted to normal mice and exposed to the immune system in all its fury, the tumors have no way of coping with immune attack.

Meanwhile, work from many groups, probably especially Soldano Ferrone, Nick Restifo , and Steven Rosenberg‘s, 10 was showing how frequently and drastically tumors mutate in order to develop resistance to the immune system, which argues that the immune system must be a major selective force during tumor growth. These findings started well before the more direct demonstrations I’ve mentioned, but at first it seemed that this sort of tumor immune evasion was unusual. As more and more cases were described, though, it has gradually come clear that immune evasion is — if not universal — the rule rather than the exception.

Hanahan & Weinberg tumor progression By now, as a result of these sorts of experiments, it’s become pretty much accepted that the immune system is a huge barrier to tumor growth; a seventh hurdle to add to the six that Hanahan and Weinberg described. The understanding is that the interaction between tumors and immunity takes three forms — the three “E”s of cancer immunity: “Elimination”, “equilibrium”, and “escape”.

“Escape” is what we usually see with cancer, because “elimination” mostly occurs before the tumors are clinically detectable. As I said in my earlier post, “by the time we can detect a cancer, it’s already been selected to be immune resistant. The cancers that were susceptible to the immune system were killed off when they were just a little cluster of cells, long before there was anything we could identify.

“Equilibrium” is an intermediate stage between elimination and escape. The tumor has mutated enough, or grows fast enough, or is hidden well enough, that the immune system can’t eliminate it altogether; but at the same time, it hasn’t mutated enough to completely escape immune control. The system is in dynamic equilibrium; the tumor can’t expand, because it’s being killed off as fast as it grows (or its growth is restrained); but the immune system can’t quite kill off the last few cells.

This “equilibrium” stage has been hard to demonstrate in an experimental system, but there have been some strong circumstantial arguments for it. The most famous example may be tumor regrowth after a kidney transplant; the donor had been successfully treated 16 years before the transplant and was believed to be tumor-free, but in the recipient the tumor — now no longer checked by the donor’s immune system — grew out within a couple of year. The tumor and the donor’s immune system had presumably been in equilibrium for 16 years.11

The Koebel et al paper shows evidence for the equilibrium stage in an experimental system. I’ll talk about it later this week.

You may be wondering, by the way, if I knew all these references and dates off the top of my head. The answer is that of course I did. However, last week I just had a general outline of the story. The difference is that this week, I’m covering tumor immunity for the grad immunology class I teach, so I’ve been doing some reading.12 I got the story straightened out, and the original references, from a couple of excellent reviews from Robert Shreider’s group:
Dunn, G. P., Old, L. J., and Schreiber, R. D. (2004). The three Es of cancer immunoediting. Annu Rev Immunol 22, 329-360.

Dunn, G. P., Bruce, A. T., Ikeda, H., Old, L. J., and Schreiber, R. D. (2002). Cancer immunoediting: from immunosurveillance to tumor escape. Nat Immunol 3, 991-998.


  1. Ehrlich P. 1909. Ueber den jetzigen Stand der Karzinomforschung. Ned. Tijdschr. Geneeskd. 5 (Part 1): 273-90[]
  2. Burnet, F.M. Cancer-a biological approach. Brit. Med. J. 1, 841-847 (1957).[]
  3. Rygaard, J., and Povlsen, C. O. (1974). The mouse mutant nude does not develop spontaneous tumours. An argument against immunological surveillance. Acta Pathol Microbiol Scand [B] Microbiol Immunol 82, 99-106.[]
  4. Over 4000 citations and counting![]
  5. Hanahan, D., and Weinberg, R. A. (2000). The hallmarks of cancer. Cell 100, 57-70. []
  6. For example: Kaplan, D. H., Shankaran, V., Dighe, A. S., Stockert, E., Aguet, M., Old, L. J., and Schreiber, R. D. (1998). Demonstration of an interferon gamma-dependent tumor surveillance system in immunocompetent mice. Proc Natl Acad Sci U S A 95, 7556-7561. And Smyth, M. J., Thia, K. Y., Street, S. E., MacGregor, D., Godfrey, D. I., and Trapani, J. A. (2000). Perforin-mediated cytotoxicity is critical for surveillance of spontaneous lymphoma. J Exp Med 192, 755-760. []
  7. Shankaran, V., Ikeda, H., Bruce, A. T., White, J. M., Swanson, P. E., Old, L. J., and Schreiber, R. D. (2001). IFNgamma and lymphocytes prevent primary tumour development and shape tumour immunogenicity. Nature 410, 1107-1111. []
  8. Note to self: Blog about this one some time! Very cool classic paper[]
  9. Shankaran, V., Ikeda, H., Bruce, A. T., White, J. M., Swanson, P. E., Old, L. J., and Schreiber, R. D. (2001). IFNgamma and lymphocytes prevent primary tumour development and shape tumour immunogenicity. Nature 410, 1107-1111.[]
  10. This is just off the top of my head and I’m probably forgetting some. Sorry![]
  11. MacKie, R. M., Reid, R., and Junor, B. (2003). Fatal melanoma transferred in a donated kidney 16 years after melanoma surgery. N Engl J Med 348, 567-568.[]
  12. And the Koebel et al. paper came out at a really fortuitous time.[]