Mystery Rays from Outer Space

Meddling with things mankind is not meant to understand. Also, pictures of my kids

April 30th, 2008

HERVs: zombie target practice for immunity

Image credit: DPPhotoJournal

A couple weeks ago I was having a chat with a friend about cancer immunity (as one so often does) and he asked if the Holy Grail of cancer immunity would be to identify tumor antigens. Not at all. There are hundreds of tumor antigens known. (The journal Cancer Immunity hosts a database that lists many of the known ones.) The problem is if anything the opposite; there are too many antigens, and many are one-offs, unique to one or a handful of tumors and of no use to most patients. A better Holy Grail would be a single target that many tumors have in common.

Our genomes are littered with the withered corpses of ancient retroviruses. Everyone has them. These human endogenous retroviruses (HERVs) are defective, and their proteins are usually not expressed, or are expressed at low levels. Because they’re not normally expressed much, they don’t necessarily tolerize the immune system. At least hypothetically, if there are pathologic conditions in which HERVs become expressed, they might form targets for immunity.

As it happens, there may be several such conditions. It’s been suggested (though not, to my inexpert eye, all that convincingly) that HERVs might represent targets in autoimmunity. More usefully, Douglas Nixon’s group showed some evidence, last fall, that HIV infection upregulates HERVs, offering a target for CTL that (unlike HIV itself) isn’t constantly mutating.1 And it’s been suggested for quite a while that HERVs might be immunogenic in tumors.

HERV buddingFor example, over ten years ago it was shown that patients with certain kinds of tumors, which consistently show high-level HERV activation, often have antibody responses to HERVs.2 However, in general, antibodies are not particularly effective against tumors, and as far as I know, nothing much arose directly from the antibody findings.

On the other hand, T cells are (at least sometimes) more effective against tumors; and T cell immunity was linked to HERVs first (as far as I know) in 2002,3 with the observation that a melanoma tumor antigen was derived from a HERV. Some similar work has followed.4

So: HERVs are potential antigens; they are more or less immutable; they can be upregulated in some tumors; and they can trigger an immune response by antibodies and by T cells. These are interesting observations, but is this at all relevant for tumor treatment?

Renal cell carcinomaThe next step in answering that question came out recently, in J Clin Invest. 5 Here we see not just reactive T cells (that is, T cells specific for HERV peptides) but a potent immune response that actually cleared a metastatic tumor. The response was due to an allogeneic bone marrow transplant, and when they tracked down the target peptide for the immune response, it was directed against a HERV peptide:

The genes encoding this antigen were found to be derived from human endogenous retrovirus (HERV) type E and were expressed in RCC cell lines and fresh RCC tissue but not in normal kidney or other tissues.

It’s still far from clear how universal a target HERVs might be. This group identified a HERV target in one of their patients, but they treated 74 patients, saw at least partial responses in 29 of those patients, sought to identify targets in four of the responders, and found the HERV target in just one of the four. Some of the other targets were apparently the more standard mutated proteins, specific to the individual tumor.

This peptide target, by the way, is from a group E HERV; most of the previous work has focused on group K HERVs, which tend to be more active and are expressed to some extent in normal tissue. HERV-E generally are pretty quiescent, so if tumors do upregulate HERV-E, it would be a more specific target. The authors did check, and found that most of that particular type of tumor expressed HERV-E. Interestingly, this is the kind of tumor that is most likely to be responsive to immunotherapy:

A histological review of the RCC6 cell lines and fresh RCC tissues used in experiments presented in this article showed all to be clear-cell carcinomas, with more than half expressing HERV-E transcripts. Furthermore, limited preliminary data from an ongoing study of fresh tumors suggest that this HERV-E may have transcriptional activity limited to the clear-cell variant of kidney cancer (unpublished observations), which is intriguing given the track record for this tumor being the immunoresponsive subtype of RCC.

It would be a very useful discovery if this turns out to be a common antigen among these tumors. That said, there are some other known common tumor antigens — such as tyrosinase in melanomas — and immunization hasn’t proven a silver bullet in those yet. But it’s early days, still.


  1. Garrison, K. E., Jones, R. B., Meiklejohn, D. A., Anwar, N., Ndhlovu, L. C., Chapman, J. M., Erickson, A. L., Agrawal, A., Spotts, G., Hecht, F. M., Rakoff-Nahoum, S., Lenz, J., Ostrowski, M. A., and Nixon, D. F. (2007). T Cell Responses to Human Endogenous Retroviruses in HIV-1 Infection. PLoS Pathog 3, e165. []
  2. Boller, K., Janssen, O., Schuldes, H., Tonjes, R. R., and Kurth, R. (1997). Characterization of the antibody response specific for the human endogenous retrovirus HTDV/HERV-K. J Virol 71, 4581-4588.[]
  3. Schiavetti, F., Thonnard, J., Colau, D., Boon, T., and Coulie, P. G. (2002). A human endogenous retroviral sequence encoding an antigen recognized on melanoma by cytolytic T lymphocytes. Cancer Res 62, 5510-5516.[]
  4. Rakoff-Nahoum, S., Kuebler, P. J., Heymann, J. J., E Sheehy, M., M Ortiz, G., S Ogg, G., Barbour, J. D., Lenz, J., Steinfeld, A. D., and Nixon, D. F. (2006). Detection of T lymphocytes specific for human endogenous retrovirus K (HERV-K) in patients with seminoma. AIDS Res Hum Retroviruses 22, 52-56.[]
  5. Takahashi, Y., Harashima, N., Kajigaya, S., Yokoyama, H., Cherkasova, E., McCoy, J.P., Hanada, K., Mena, O., Kurlander, R., Abdul, T., Srinivasan, R., Lundqvist, A., Malinzak, E., Geller, N., Lerman, M.I., Childs, R.W. (2008). Regression of human kidney cancer following allogeneic stem cell transplantation is associated with recognition of an HERV-E antigen by T cells. Journal of Clinical Investigation DOI: 10.1172/JCI34409[]
  6. RCC: “Renal cell carcinoma.” IY[]
April 27th, 2008

Elementary Dr Watson

Foot-and-mouth disease virusWe’ve been promised that as genome sequencing becomes faster and simpler, we’ll start seeing practical dividends as well as parlour tricks like sequencing Watson’s genome. Some of the dividends are already paying out, as a paper in the latest PLoS Pathogens1 shows.

Probably most of you remember the outbreaks of foot-and-mouth disease in Britain in 2001, and again last year. FMD is a virus that affects many hooved animals; it’s not usually fatal, but causes productivity loss. FMD outbreaks are economically devastating, because aside from the productivity loss many countries, that are free of the disease, will refuse to take meat or other agricultural products from outbreak areas. The goal of FMD management, then, is to keep it away, and if it ever hit, to contain it and slaughter all infected and potentially-infected animals.

The 2001 outbreak in Great Britain came from outside the country. The 2007 outbreak, though, was clearly from a local source: The FMD research lab in the Institute for Animal Health (IAH), Pirbright, Surrey. The latest paper discusses the epidemiology of that outbreak, and how they used whole-genome sequencing to track and predict sites of FMD.

Samuel & Knowles, 2001, Fig 2(This is timely, because the US is planning to move the sole American FMD research center, now on Plum Island, to the mainland. There’s obvious concern that the virus could escape from containment within research labs and infect neighboring animals, causing the first American FMD outbreak since 1929. I am not particularly knowledgeable about the field, but I have to think that, at best, the timing of the planned move is unfortunate.)

FMD is caused by a picornavirus, the same broad family as polio and cold viruses. Like those viruses, FMD mutates rapidly, traveling around as a quasispecies cloud. The clouds can be easily divided into 7 broad groups, and within the most common serotype (O) there are 8 distinct subgroups (see the map2 to the right [click for a larger version] for their geographical distribution).

The FMD genome is 8134 nucleotides long, and the sequence analysis that has been used for epidemiology like the 7 different topotypes has been based on no more than 8% of that length — the VP1 gene, usually. That’s enough to track high-level changes, because of FMD’s rapid mutation rate:2

the rate of evolution is approximately 1% per year …. If the concept of a constant evolutionary rate is accepted and there are no constraints on virus evolution then it would expected that new topotypes could arise in approximately 15 years. In reality, this extent of evolution probably takes much longer. For example, FMD viruses belonging to the Asia 1 serotype, first identified in samples from Pakistan in 1954 … have not yet exceeded 15% nucleotide difference …

But 8% of the genome is not nearly enough to track changes within a single epidemic, like the one in Surrey last year; it simply isn’t long enough to pick up the handful of variations. It was known in the previous outbreak, in 2001, that the information was there in the genome (“virus recovered from closely housed animals can differ by 1 to 2 nucleotides and is likely to pass through a “bottleneck” on passage between farms”).3 The issue was a practical, technological one — being able to sequence entire virus genomes quickly enough to pass back information to people in the field.

Cottam 2008 Fig 2By 2007, the technology was there. The people at the IAH were able to sequence genomes from viruses isolated in the outbreak with a fine enough comb to track changes throughout the spread, and fast enough pass information back to the field within 24-48 hours. Their sequencing confirmed that the virus was in fact a lab escapee, because it was almost identical to a couple of lab strains but was different from circulating viruses. 4

The 40-odd viral genomes yielded a fair bit of useful information (see the figure to the left for a summary). For example,

The small number of nucleotide substitutions observed between viruses from source and recipient IP suggests that there has been direct transmission without the involvement of other susceptible species, e.g. sheep or deer.

It’s obviously useful to know if there’s a wild-animal reservoir of disease, but an even more important insight came from this work as well.

the virus from IP3b was nine nucleotides different from the virus from IP1b … This is a high number of changes for a single farm-to-farm transmission … and we predicted that there were likely to be intermediate undetected infected premises between the first outbreaks in August and IP3b. … Serosurveillance of all sheep within 3 km of the September outbreaks revealed another infected premises (IP5), on which it was estimated that disease had been present for at least two, and possibly up to five weeks. As Figure 2B shows, IP5 is a likely link between the August and September outbreaks.

I would be interested in hearing from the people on the ground just how useful this information was — for example, were they impelled to search more for an intermediate source based on this information, or did they already suspect it from other, classical ways? But in any case, it’s clear that genomics is capable of pushing epidemiology a lot further in the future.


  1. Cottam, E.M., Wadsworth, J., Shaw, A.E., Rowlands, R.J., Goatley, L., Maan, S., Maan, N.S., Mertens, P.P., Ebert, K., Li, Y., Ryan, E.D., Juleff, N., Ferris, N.P., Wilesmith, J.W., Haydon, D.T., King, D.P., Paton, D.J., Knowles, N.J. (2008). Transmission Pathways of Foot-and-Mouth Disease Virus in the United Kingdom in 2007. PLoS Pathogens, 4(4), e1000050. DOI: 10.1371/journal.ppat.1000050[]
  2. Samuel, A. R., and Knowles, N. J. (2001). Foot-and-mouth disease type O viruses exhibit genetically and geographically distinct evolutionary lineages (topotypes). J Gen Virol 82, 609-621.[][]
  3. Cottam, E. M., Haydon, D. T., Paton, D. J., Gloster, J., Wilesmith, J. W., Ferris, N. P., Hutchings, G. H., and King, D. P. (2006). Molecular epidemiology of the foot-and-mouth disease virus outbreak in the United Kingdom in 2001. J Virol 80, 11274-11282.[]
  4. As far as I know, it’s not yet known how exactly the virus escaped from the IAH. I’ve read what seems to be informed speculation that it may have come from the drains, as decontamination systems designed to prevent that weren’t properly maintained; but I don’t know if that’s true, an educated guess, or mere rumor and guesswork.[]
April 23rd, 2008

What happens in the liver, stays in the liver

Virus-Cell Interaction; Joerg Schroeer; Art of Science
“Human cytomegalovirus infected human endothelial cells”
by Joerg Schroeer

There’s a famous picture in Field’s Virology1 showing how ectromelia (mousepox virus, a model for smallpox) infects a new host, spreads within the mouse, and then is transmitted to a new host. The figure is below2 (click for a larger version). Simplified, ectromelia initially infects the skin through small cuts; it replicates at the site, then spreads through blood and lymph to organs (spleen, liver) where it replicates further. The progeny virus from this replication then spreads again through the blood, this time back to the skin, where it replicates once again (now vastly amplified from the initial infection) to form the classic “pock” lesions, which shed virus that can infect a new victim.

It’s generally accepted that this is a common pattern of pathogenesis for many of the viruses that go systemic; not necessarily all viruses, because certainly some remain localized or only spread through, say, direct contact, but for something that spreads through the entire host, it should be a reasonably accurate model.

Ectromelia pathogenesis (Fields)Human and mouse cytomegaloviruses certainly spread throughout the entire host, and infect many cell types within the body — endothelial cells, lymphoid cells in the spleen and elsewhere, liver, and probably other tissues as well. However, Ulrich Koszinowski’s group now suggests that in spite of this, it isn’t following the ectromelia pattern; replication within some of the organs (liver) is a dead end, that doesn’t help disseminate the virus. 3

Koszinowski has a habit of constructing very cool systems for analyzing his pet virus; sometimes so fancy that I wonder if he makes them a little baroque just because he can (and I know4 that he has occasionally been bitten by his elaborate systems). Here he used a Cre/lox recombination system, with the flox in the CMV and the Cre in the mouse under cell-specific promoters, so that the virus genome gets modified only when it replicates in the particular organ. Don’t worry about the details, the point is that the virus is tagged as soon as it replicates in a particular organ, so you can look at viruses throughout the whole mouse and identify whether their ancestors ever replicated in one particular organ. You can also work out timing of replication, and a few other things.

The unexpected bottom line is that what happens in the liver, stays in the liver. There is more MCMV in the liver than anywhere else in the body, but it’s a dead end; once it’s in the liver it doesn’t spread to other organs.5 Instead, the relatively small amount of virus that replicates in endothelial cells (and perhaps in the spleen) seems to be a major source for further spread within the body and for transmission.

The results challenge the concept that organs that produce the bulk of infectious virus during acute infection necessarily also play a major role in dissemination.

Surprisingly, to me anyway, even suppression of the adaptive immune system didn’t change this; the virus still hung out in the liver. (The door is still open for innate immune restriction of spread.)

This shows how little we really know about what goes on in authentic viral infections. So much of our understanding of virology is based on tissue culture, but it’s harder than it looks to extrapolate a simple in vitro observation to the complicated interactions within the body, and it’s dangerous to extrapolate from one virus to another.

Why does CMV replicate in the liver if it’s a dead end? Is this simply a matter of indifference to the virus (replicate anywhere you can and hope you’re in the right spot to spread) or is it doing something specific to modulate the host in some way? My bias is that this is something the virus is doing for a reason, but I don’t know what.


  1. It’s adapted from a figure in Fenner’s Viral Pathogenesis, which I haven’t read:
    Fenner F, Buller RM. Mousepox. In: Nathanson N, Ahmed R, Gonzalez-Scarano F, et al., eds. Viral Pathogenesis. Philadelphia: Lippincott-Raven; 1997:535-553.
    and is based on old research from Fenner:
    Fenner, F. (1949). Mouse-pox; infectious ectromelia of mice; a review. J. Immunol. 63, 341-373.[]
  2. I think this qualifies as fair use[]
  3. Sacher, T., Podlech, J., Mohr, C. A., Jordan, S., Ruzsics, Z., Reddehase, M. J., and Koszinowski, U. H. (2008). The major virus-producing cell type during murine cytomegalovirus infection, the hepatocyte, is not the source of virus dissemination in the host. Cell Host Microbe 3, 263-272.[]
  4. That is, “I have heard rumors that … “[]
  5. Even though the virus in the liver is actually perfectly competent for replication in other tissues, if taken from the liver and used to infect other mice.[]
April 20th, 2008

Immune evasion of CD1

MycobacteriaCD1 is a fascinating molecule, but it hasn’t traditionally been associated with antiviral protection. Viruses, however, seem to disagree.

CD1 is (actually, CD1 are, since these are a family of related molecules) members of the MHC class I family, with many of the traditional MHC class I features — binding to β2-microglobulin, a “groove” made of two alpha-helices on top of a beta-pleated sheet ( (in classical MHC, the peptide-binding groove: the “bun” of the peptide’s “hot dog”).

I previously wrote a field guide to the MHC family that shows these features across a wide range of the MHC family, but here are some comparisons of CD1d (PDB number 2GAZ) to classical MHC (HLA-A2; PDB number 2GTW). Here I’m showing the “heavy chain” of the complex in red, the ligand that fits in the binding groove is green, and β2-microglobulin in blue. (The “top view” is looking “down” at the molecule more or less as a T cell would “see” it; the ribbon view lets you see the ligand interactions a little more easily; the “ligand only” view shows the thing that goes in the groove with all the MHC or CD1 resides removed.) (Click on a molecule for a larger version.)

Side view Top view Top view (ribbon) Ligand only
CD1d CD1d, side view CD1d, top view CD1d, top view CD1d ligand
HLA-A2 HLA-A2, side view HLA-A2, top view HLA-A2, top view HLA-A2 ligand

The similarities are pretty obvious, but it’s the difference that makes CD1 particularly interesting. Classical MHC molecules present peptide ligands; CD1d presents, instead, very hydrophobic molecules: lipids, glycolipids, and lipopeptides. These sorts of things are typically found in mycobacteria (as with phosphatidylinositol mannoside, shown in the images here), and I’ve thought of CD1 as an example of how our physiology has been shaped by pathogens — this whole branch of the immune system, devoted to detection and elimination of tuberculosis and leprosy.

My view started to slip in around 2000, with Frank Chisari’s observation that NKT cells may be involved in control of hepatitis B virus in his transgenic mouse model.1 (NKT cells are the main lymphocytes that recognize CD1 molecules.) I’ve talked about Chiasri’s HBV mouse model before — it’s so artificial that I always am hesitant to extrapolate from it. That said, his findings in that model have all (as far as I know) held up in more natural systems, and the NKT observation is no exception; several other groups have seen similar things. 2

Raftery et al., 2008 Fig. 10What really confirmed to me that CD1 can be antiviral, though, was the virus’s side of the story. Viruses employ an arsenal of anti-immune molecules, presumably targeting whichever immune components that are especially dangerous to the particular virus. Over the past few years, there’s been an increasing number of sightings of viruses that block CD1-mediated presentation. The first (that I know of) was HIV,3 and since then vaccinia virus4 and herpes simplex5 have also been shown to block CD1-mediated antigen presentation. The latest addition to the list is human cytomegalovirus.6 These viruses (HIV, poxviruses, and herpesviruses) are particularly good at blocking classical MHC class I presentation as well; I don’t know if this dual blockade is typical, or if people have mainly looked in those viruses most renowned for immune evasion — in other words, maybe we’re seeing this double action because people are looking under the streetlamps.

It’s interesting that HSV and HCMV (though not HIV, which blocks both classical MHC class I and CD1 with the same protein, nef) have apparently developed separate systems to block CD1 and classical MHC. The molecules responsible for their CD1 blockade are not yet identified, but they don’t seem to be the same as the ones that block MHC class I. If CD1 blockade is the main function of these genes (and not a side-effect of blocking some other aspect of immunity, say), the implication is that CD1 is an important-enough player in controlling these viruses that they have had to maintain distinct pathways to escape from it.

I wonder what it’s doing.


  1. Kakimi, K., Guidotti, L. G., Koezuka, Y., and Chisari, F. V. (2000). Natural killer T cell activation inhibits hepatitis B virus replication in vivo. J Exp Med 192, 921-930.[]
  2. For example, Grubor-Bauk, B., Simmons, A., Mayrhofer, G., and Speck, P. G. (2003). Impaired clearance of herpes simplex virus type 1 from mice lacking CD1d or NKT cells expressing the semivariant V alpha 14-J alpha 281 TCR. J Immunol 170, 1430-1434. []
  3. Shinya, E., Owaki, A., Shimizu, M., Takeuchi, J., Kawashima, T., Hidaka, C., Satomi, M., Watari, E., Sugita, M., and Takahashi, H. (2004). Endogenously expressed HIV-1 nef down-regulates antigen-presenting molecules, not only class I MHC but also CD1a, in immature dendritic cells. Virology 326, 79-89.[]
  4. Webb, T. J., Litavecz, R. A., Khan, M. A., Du, W., Gervay-Hague, J., Renukaradhya, G. J., and Brutkiewicz, R. R. (2006). Inhibition of CD1d1-mediated antigen presentation by the vaccinia virus B1R and H5R molecules. Eur J Immunol 36, 2595-2600.[]
  5. Sanchez, D. J., Gumperz, J. E., and Ganem, D. (2005). Regulation of CD1d expression and function by a herpesvirus infection. J Clin Invest 115, 1369-1378.
    and
    Yuan, W., Dasgupta, A., and Cresswell, P. (2006). Herpes simplex virus evades natural killer T cell recognition by suppressing CD1d recycling. Nat Immunol 7, 835-842.[]
  6. Raftery, M.J., Hitzler, M., Winau, F., Giese, T., Plachter, B., Kaufmann, S.H., Schonrich, G. (2008). Inhibition of CD1 Antigen Presentation by Human Cytomegalovirus. Journal of Virology, 82(9), 4308-4319. DOI: 10.1128/JVI.01447-07[]
April 16th, 2008

How does alum adjuvant work?

Witch (F. Landerer after M. Schmidt)My first foray into research, when I started grad school, was to work on a vaccine against bovine adenovirus type 3.1 From a technical viewpoint, it was a great introduction to research; I got to do classical virology, animal work, tissue culture, protein purification, all kinds of immunology, and so on. Still, I ended up unenthusiastic about vaccinology (though not about vaccines, which I regard as one of the most important benefits to civilization in history). The problem was that so much of vaccine research seemed purely empirical, with no solid theory underlying it. The way to design a new vaccine was just to try a whole bunch of different things (sometimes applying rules of thumb) and see what worked.

To some extent that’s still true — there’s a great deal we still don’t understand about the immune system, and predicting how to drive a safe, protective response still has a lot of guesswork to it. 2 At the time, though, the most frustrating aspect of the field for me was adjuvants.

Adjuvants are factors that you include with your antigen, in order to drive a potent immune response. In general, the more pure an antigen is, the worse the immune response to it. Adjuvants allow you to provide a clean, defined antigen, without losing the immunogenicity of the filthy natural antigen.

The problem was that no one knew how adjuvants worked. They just … worked. There were a myriad of choices (for animals; in the US and Canada there’s only one adjuvant, alum, that’s licensed for humans), and they all mostly worked, and sometimes one worked better and sometimes another worked better, or differently; but there was no understanding of how, or why. Sometimes toe of newt was the best choice, and sometimes you were better off with eye of toad, and it depended on the phase of the moon and on which malign vapours were influencing your system.

Had I but known, of course, just about the time I switched out of vaccines and into other aspects of viral immunity, Charlie Janeway was offering up a grand unified theory of adjuvants3 which has for the most part proven triumphantly true. (I talked about it here.) He suggested that the immune system normally initiates a response when it recognizes conserved features of pathogens, and that adjuvants work because they mimic these conserved pathogen-associated molecular patterns. (Polly Matzinger also proposed a related model, in which immune responses start because cells are damaged — the danger hypothesis.) Since then, many of the pathogen-associated patterns have been identified, and many of the pattern receptors have been identified; adjuvants are no longer magic, they’re science.

AlumWell, all except for one: Alum, the most important one of all (because it’s the main adjuvant used for human vaccines). We had no real idea how that works, because it looks nothing like any plausible pathogen pattern.

A paper in J Exp Med4 now argues that alum’s adjuvant activity comes from uric acid. As it happens, this is less related to Janeway’s hypothesis, and is closer to Matzinger’s. Uric acid is released by dying or damaged cells, and is a powerful natural adjuvant5 — it’s an indicator to the immune system that cells are being damaged in the vicinity, meaning that there is “danger” nearby.

I’m not entirely convinced that this is the whole, or even the main, story. The implication is that alum acts by damaging cells. The authors say that “alum has been shown to induce a considerable degree of necrosis”. That may be true with the intraperitoneal injection model they used, but alum-adjuvanted vaccines in people are more often given intramuscularly, and I don’t know that alum is all that nasty in that context.6 After all, the reason alum is approved for use in humans is that it is so innocuous. And simple experience says that while vaccines sting, you don’t expect any kind of large-scale necrosis in your injected arm afterward — no more than you’d get from a modest bruise, which isn’t enough to trigger the kind of adjuvant effects we see with alum. Perhaps there is a small release of uric acid effect, and alum somehow amplifies the effect (perhaps by facilitating uric acid crystallization, which is essential for its adjuvant activity). Or perhaps uricase is important in intraperitoneal inject, but is less so in more clinically-relevant injections. I don’t know.

Still, it’s nice to see that adjuvant activities, nowadays, can actually be tested within well-defined theoretical contexts. That’s just a huge advance since the days when I had to play with them.

Update: Since I wrote this article there have been several more papers on alum’s mechanism of action. As I guessed here, alum can in fact work in other ways, and though the mechanism described here may be part of its effect it seems very likely that alum mainly acts in different ways. For updates see these posts:
Alum, take 2: A better answer
Silicosis parallels alum
and
A general rule for (some) adjuvants


  1. York, I., and Thorsen, J. (1992). Evaluation of a subunit vaccine for bovine adenovirus type 3. American journal of veterinary research 53, 180-183.[]
  2. Partly, of course, because most of the easy targets for vaccines, where we could predict protective responses, are already out there, and now we’re working on the hard cases like malaria and HIV.[]
  3. Approaching the asymptote? Evolution and revolution in immunology. Janeway, C.A.Jr.. Cold Spring Harb. Symp. Quant. Biol. 54, 1-13 (1989) []
  4. Kool, M., Soullie, T., van Nimwegen, M., Willart, M.A., Muskens, F., Jung, S., Hoogsteden, H.C., Hammad, H., Lambrecht, B.N. (2008). Alum adjuvant boosts adaptive immunity by inducing uric acid and activating inflammatory dendritic cells. Journal of Experimental Medicine DOI: 10.1084/jem.20071087[]
  5. Shi, Y., Evans, J. E., and Rock, K. L. (2003). Molecular identification of a danger signal that alerts the immune system to dying cells. Nature 425, 516-521.[]
  6. A quick and superficial scan through PubMed doesn’t turn up much support for the statement either , for what that’s worth.[]
April 13th, 2008

MHC isn’t sexy after all

Mickey and Minnie Mouse Inbreeding is a bad thing, genetically, and almost all species have ways of avoiding it. One way of avoiding inbreeding is to recognize individuals who are related to you, and not mate with them. That’s not so difficult when you’re a big-brained, highly social animal like a wolf, or a human, who have lots of brain devoted to issues of who is who and where they stand in a group. It’s a little more challenging for mice, though.

How do mice distinguish individuals? How do they determine relatedness? For the past 30 years1, the answer has been MHC: Mice select mates that differ at the major histocompatibility complex.

When I talked about this last fall, in the context of MHC diversity, I was rather skeptical that mating preference was the major driver of MHC diversity, quoting Piertney and Oliver:2

A lack of repeatability of several studies, and an apparent plasticity in response across experiments, questioned the robustness of the data, and the general relevance of mate choice as a primary driver of MHC diversity.

It didn’t occur to me to question the fundamental observation that MHC is even involved in distinguishing relatedness.

MHC seemed like a logical candidate for distinguishing individuals and determining relatedness because of its great polymorphism: in an outbred population, the MHC is so variable that few individuals are identical across the region, and individuals with similar MHC are most likely related to some extent. 3 And in fact, mice clearly can distinguish differences in MHC type by smell. 4 However, that doesn’t mean that mice recognize different individuals, or determine relatedness, by the differences in MHC. A couple of papers from Jane Hurst’s group in the past year suggest that in fact they do not. 5,6

Hurst’s group tried to move away from the artificial situation of highly-inbred lab mice, using instead wild mice breeding in semi-natural conditions. They find that under these conditions, mice do (as expected) avoid breeding with close relatives. But this incest avoidance doesn’t correlate with MHC type. Instead, there was a strong correlation with MUP type.

What, you cry, is MUP? These are “major urinary proteins”, which are known to be highly polymorphic in wild mouse populations — though not in lab mice — and which are also known to be very important in scent marking. Indeed, the only known function of MUPs is in scent marking. The lack of variability of MUPs in lab mice might have led to the use of MHC as markers instead in those studies, but in Hurst’s study MHC didn’t contribute to incest avoidance:

By contrast, MUP sharing had a strong and highly significant effect on the likelihood of successful mating (Table 1: model 3, p = 0.005; Figure S1). Specifically, there was no deficit when only one MUP haplotype was shared, but there were many fewer matings between mice that shared both MUP haplotypes (complete match) than expected under random mating conditions (Table 1: model 4, p < 0.002). … Mice thus avoid mating when shared MUP type reliably indicates very close relatedness.

Rodentia:Johnson's household book of nature, containing full & interesting descriptions of the animal kingdom. (New York : Johnson, c1880) Craig, Hugh, Editor.Incidentally, this is consistent with a recent paper from Peter Overath and Hans-Georg Rammensee.7 They looked for influences on urine odor in mice (try writing to your Mom and tell her that’s what you’re doing for your living, by the way, and see how long it takes before she starts talking about your cousin the investment banker) and didn’t find any influence of MHC:

… within the limits of the ensemble of components analysed, the results do not support the notion that functional MHC class I molecules influence the urinary volatile composition.

(However, there are non-volatile as well as volatile components to urine odor, so this isn’t definitive.)

MUPs are highly polymorphic in wild domestic mice, but are non-polymorphic (actually, basically non-existent) in humans. (In fact, MUPs are non-polymorphic even in Mus macedonicus, a mouse species closely related to M. musculus domesticus, but a species that doesn’t need as careful management of increeding because individuals normally disperse more. ) That means that MUPs can’t be a universal mechanism for inbreeding avoidance, so the work on MHC-linked mate choice in other species might still be valid. However, I still think the work on MHC and mate selection in humans is mostly pretty crappy unconvincing. Since the work in humans leans heavily on the assumption that MHC is important in mate selection in mice, that work can be looked at with an even more jaundiced eye now, I think.


  1. Yamazaki, K., Boyse, E. A., Mike, V., Thaler, H. T., Mathieson, B. J., Abbott, J., Boyse, J., Zayas, Z. A., and Thomas, L. (1976). Control of mating preferences in mice by genes in the major histocompatibility complex. J Exp Med 144, 1324-1335[]
  2. Piertney, S. B., and Oliver, M. K. (2006). The evolutionary ecology of the major histocompatibility complex. Heredity 96, 7-21.[]
  3. A review is here: Adv Genet. 2007;59:129-45. Genetic basis for MHC-dependent mate choice. Yamazaki K, Beauchamp GK.[]
  4. For example, Carroll, L.S., Penn, D.J., and Potts, W.K. (2002). Discrimination of MHC-derived odors by untrained mice is consistent with divergence in peptide-binding region residues. Proc. Natl. Acad. Sci. USA 99, 2187–2192.[]
  5. Sherborne, A., Thom, M., Paterson, S., Jury, F., Ollier, W., Stockley, P., Beynon, R., Hurst, J. (2007). The Genetic Basis of Inbreeding Avoidance in House Mice. Current Biology, 17(23), 2061-2066. DOI: 10.1016/j.cub.2007.10.041[]
  6. The Genetic Basis of Inbreeding Avoidance in House Mice
    Amy L. Sherborne, Michael D. Thom, Steve Paterson, Francine Jury, William E.R.
    Ollier, Paula Stockley, Robert J. Beynon, and Jane L. Hurst. Curr Biol. 2007 December 04; 17(23): 2061–2066. doi: 10.1016/j.cub.2007.10.041 []
  7. Röck F, Hadeler K-P, Rammensee H-G, Overath P (2007) Quantitative Analysis of Mouse Urine Volatiles: In Search of MHC-Dependent Differences. PLoS ONE 2(5): e429. doi:10.1371/journal.pone.0000429.[]
April 10th, 2008

NK cells do protect against cancer

NK cells killing a tumor cellNatural killer cells were originally identified as cells that spontaneously killed cancer cells. It’s been a bit surprising, then, that there’s been relatively little direct evidence that NK cells protect against spontaneous cancer.

For example, there was the study I talked about some time ago, looking at tumors in equilibrium with the immune system. There, the authors treated mice with a  carcinogen, waited until tumors stopped arising, and immunosuppressed the tumor-free survivors. 1 In most of these apparently cancer-free animals, immune suppression resulted in tumors appearing within a few weeks. This showed that the immune system can control tumors.

The interesting part (at least, the interesting part as far as NK cells are concerned) was that treatment with anti-NKG2D did not let tumors grow out; whereas shutting down T cells (with anti-CD4/anti-CD8) or interferon (with anti-IFN) did let the tumors reappear. NKG2D is an important receptor for NK cells, so the implication was that NK cells were not keeping the tumor in check, but T cells were.

Now, however, a similar set of experiments has shown that’s not necessarily true — NK cells probably are important in controlling some tumors. The paper is
Guerra, N., Tan, Y. X., Joncker, N. T., Choy, A., Gallardo, F., Xiong, N., Knoblaugh, S., Cado, D., Greenberg, N. R., and Raulet, D. H. (2008). NKG2D-Deficient Mice Are Defective in Tumor Surveillance in Models of Spontaneous Malignancy. Immunity 28, 571-580. DOI: 10.1016/j.immuni.2008.02.016

David Raulet’s group has been one of the leaders in NK cell research, and as far as I know they are the first to have made a knockout mouse lacking NKG2D. The mice are normal and happy and actually have normal numbers of NK cells that are functional. NK cells notoriously have many ligands, which is one of the reasons it took longer to figure out NK receptor/ligand interactions than for T cells, and presumably there are enough alternatives to NKG2D that NK cells can get whatever signals they need during development. However, of course, none of the ligands for NKG2D triggered NK cells.

NK cell killing a tumor cellRather than wait for truly spontaneous tumors (which are rare enough even in mice that you need very large numbers of mice to figure out what’s going on) they crossed the NKG2D -/-mice with a couple of transgenic lines that are highly cancer-prone. They also tried treating the mice with carcinogens, as was done in the Koebel et al study I mentioned earlier.

The carcinogen-treated NKG2D knockout mice got no more tumors than did wild-type mice — so that’s exactly consistent with the previous experiment. However, the “spontaneous” tumor transgenic models showed a big difference. The NKG2D knockouts had much earlier, more aggressive tumors than did the wild-type mice.

As well as evidence for cancer equilibrium, the Koebel et al paper showed evidence for immunoediting. 2 That is, tumors that grow in the presence of a healthy immune system are resistant to the immune response — they have been selected for immune invisibility or resistance. By comparison, tumors that grow in immune deficient mice are much more immunogenic — they haven’t had to develop immune resistance.

In one of the two transgenic systems, the NKG2D knockout mice showed the same thing: Their tumors were more likely to have NKG2D ligands than the wild-type mice. “These data suggest that NKG2D-dependent immunoselection (or editing) favors loss of NKG2D ligands on early-arising, aggressive tumors.”

But in the other tumor system no such evidence was seen; NKG2D ligands were just as prevalent:

There was no indication in this survey that expression of NKG2D ligands was selected against in Klrk1+/+ mice, despite the clear evidence that NKG2D-mediated surveillance is operative for these lymphomas. These data suggest that evasion of NKG2D-mediated surveillance by Eμ-myc-induced lymphomas occurs by mechanisms that do not depend on loss of NKG2D ligands.

So it seems that NK immune surveillance is much more complicated than the (already very complicated) T cell immune surveillance of tumors:

Taken together, these data suggest that the role of NKG2D-dependent surveillance differs in the three types of tumors studied here. In the case of early-arising prostate carcinomas in TRAMP mice, many of the tumors are eliminated, and the few that are not eliminated evade surveillance by extinguishing expression of NKG2D ligands. In the case of Eμ-myc lymphomas, it appears that the emerging tumors are mostly NKG2D sensitive, but a fraction of tumors escape NKG2D surveillance without losing NKG2D ligands. … The final category is represented by late-arising prostate carcinomas, which appear to be generally refractory to NKG2D-dependent surveillance.

But the bottom line is that NK cells do control some tumors. That’s not a surprise, because it’s pretty much been the assumption for a long time, but it’s reassuring to get evidence for it.


  1. 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 450, 903-907.[]
  2. Of course, this wasn’t the first evidence for immunoediting.[]
April 9th, 2008

Darwin

Darwin

I just got back from Toronto, where I visited the Darwin exhibit at the ROM;1 the picture above is what my kids look like after being sternly told to stop goofing around.2

Bonus pictures! while I stall on putting up a real post here:

Cap'n Matthew Alex the Albertosaurus
Cap’n Matthew navigates the HMS Beagle
through the stormy seas of the ROM
Alex the Albertosaurus


  1. Also made it to a ballgame to see my Red Sox lose to the Jays, and the Ontario Science Center for the kids[]
  2. Yeah, no pictures allowed at the Darwin exhibit. They didn’t shoot me, though.[]
April 7th, 2008

Non-cytotoxic cytotoxic T lymphocytes

Hepatitis (Wellcome)
Hepatitis

(Spending a few days in Toronto with my kids, so I can go to the Darwin exhibit at the Royal Ontario Museum.1 Depending on when I get back, my next blog post may be a little delayed.)

Even though cytotoxic T lymphocytes are called “cytotoxic”, it was no surprise when new techniques in the mid-90s suggested that CTL might have other strings to their bows. (I talked about this the other day.) Some experiments were also pointing the same way. For example, Frank Chisari’s group offered evidence that CTL might be able to control hepatitis B virus without actually being cytotoxic.

Hepatitis B virus was a particularly difficult case for study in general, because it’s pretty much a human-specific virus. (HBV is a pain in tissue culture. There are a couple of animal models, but they’re hardly convenient. I mean, ducks? Woodchucks?) From studies of the natural disease, it seemed pretty clear that natural HBV infection was often handled by CTL; people who control HBV show a strong CTL response, those in whom HBV progresses and becomes chronic usually do not. But testing the function of CTL in HBV infection wasn’t easy without an animal model.

So Chisari made an animal model. He made transgenic mice that express the HBV genome (that is, the mouse genome contained the HBV genome) under a liver-specific promoter.2 There you go, hepatitis B virus in the mouse liver. It can’t spread and infect new cells as would normally happen in humans, but on the other hand you’re expressing the virus in essentially all the liver cells anyway, so you don’t need to infect new cells.

Hepatitis B virus-transgenic mice

Of course, the “virus” is “self” under those conditions, meaning that the immune system doesn’t react to it. So Chisari’s group immunized other mice to raise HBV-specific CTL, and transferred the CTL into the transgenic mice.3 These HBV-specific CTL did two things: They apparently controlled the virus “infection”, and they damaged the liver.4 The presumption was that the two findings were the same effect, and the CTL were trying to eliminate virus by killing liver cells: “Our data show that antigen-specific immune effector mechanisms can destroy HBV-positive hepatocytes in vitro and in vivo … “.

Hepatitis viruses (Wellcome)
Hepatitis viruses

Six years later, though, they were suggesting the opposite, demonstrating that in fact CTL were controlling the “virus” through a non-cytolytic mechanism.5 (Actually, they showed very similar data a couple of years earlier than that,6 but the 1996 paper went further with the mechanisms and was overall more solid.) Essentially, they demonstrated that CTL were shutting down virus expression in the liver by releasing interferon and other cytokines. There is some liver damage, but it’s not necessarily because of direct cytotoxity; it’s a side-effect of the cytokines, not directly related to the viral clearance. (Recently, it’s been proposed that the CTL make a decision which route to go — cytotoxicity vs. cytokine-mediated control — based on the amount of virus antigen. 7 This might offer a way to manipulate this and drive the response to the most effective system in other infections, though it’s far from trivial to adjust amount of viral antigen.)

Non-cytotoxic control of HBV

So what? What difference does it make how the CTL are clearing virus, so long as they do clear it? Here are some reasons to care:

  • This way a few CTL can affect many target cells. There are some 1011 hepatocytes, and hepatitis viruses can infect a lot of them, very fast. There are different estimates for how long it takes for a CTL to deliver a lethal hit when it’s being cytotoxic, but let’s say something like 30 – 60 minutes per cell. That’s a long, long time for CTL to kill off all the infected cells. If all they have to do is release interferon, they can probably pick off many cells at once and move on. This is a faster way to control viruses, and it doesn’t take as many CTL. Guidotti et al transferred 5 x 106 HBV-specific CTL into the transgenic mice; within 24 hours, all of the virus had been shut down (remember. these are transgenic mice expressing “virus” in every liver cell).
  • Again: There are some 1011 hepatocytes, and hepatitis viruses can infect a lot of them, very fast. If CTL have to kill all the virus-infected cells, what’s left of the liver to do it’s usual liverly duties? If the infection can be shut down without killing the cells, you’ve saved your liver. Those 5 x 106 CTL shut down the virus without killing more than 10% of the liver cells.
  • Potentially, any inflammation in the liver can lead to protection. If you have HBV, and you’re infected by lymphocytic choriomeningitis, (well, first you’d likely be a mouse), the inflammation induced by the LCMV might shut down the HBV, as a handy side effect. This cross-protection from other viruses has actually been seen both in mice8 and — maybe — in humans as well.9

So the finding that CTL are capable of shutting down virus replication without having to kill the infected cells, fit very nicely with the new technology showing that CTL commonly have these mechanisms available.


  1. Also I promise it’s a complete coincidence that my Red Sox were playing the Blue Jays here last few days. The Jays gave them a whuppin’, but William and I had a fine time at the ballpark anyway.[]
  2. Chisari, F. V., Pinkert, C. A., Milich, D. R., Filippi, P., McLachlan, A., Palmiter, R. D., and Brinster, R. L. (1985). A transgenic mouse model of the chronic hepatitis B surface antigen carrier state. Science 230, 1157-1160.[]
  3. As you can see, the system is pretty elaborate, and I’ve never really felt all that comfortable with it — just too Rube Goldberg-ish for my liking — even though there’s nothing specific I can point to; and Chisari’s group is conscientious about their controls.[]
  4. Moriyama, T., Guilhot, S., Klopchin, K., Moss, B., Pinkert, C. A., Palmiter, R. D., Brinster, R. L., Kanagawa, O., and Chisari, F. V. (1990). Immunobiology and pathogenesis of hepatocellular injury in hepatitis B virus transgenic mice. Science 248, 361-364.[]
  5. Guidotti, L. G., Ishikawa, T., Hobbs, M. V., Matxke, B., Schreiber, R., and Chisari, F. V. (1996). Intracellular inactivation of the hepatitis B virus by cytotoxic T lymphocytes. Immunity 4, 25-36.[]
  6. Guidotti, L. G., Ando, K., Hobbs, M. V., Ishikawa, T., Runkel, L., Schreiber, R. D., and Chisari, F. V. (1994). Cytotoxic T lymphocytes inhibit hepatitis B virus gene expression by a noncytolytic mechanism in transgenic mice. Proc Natl Acad Sci U S A 91, 3764-3768.
    also
    Tsui, L. V., Guidotti, L. G., Ishikawa, T., and Chisari, F. V. (1995). Posttranscriptional clearance of hepatitis B virus RNA by cytotoxic T lymphocyte-activated hepatocytes. Proc Natl Acad Sci U S A 92, 12398-12402.[]
  7. Gehring, A. J., Sun, D., Kennedy, P. T., Nolte-‘t Hoen, E., Lim, S. G., Wasser, S., Selden, C., Maini, M. K., Davis, D. M., Nassal, M., and Bertoletti, A. (2007). The level of viral antigen presented by hepatocytes influences CD8 T-cell function. J Virol 81, 2940-2949.[]
  8. Guidotti, L. G., Borrow, P., Hobbs, M. V., Matzke, B., Gresser, I., Oldstone, M. B., and Chisari, F. V. (1996). Viral cross talk: intracellular inactivation of the hepatitis B virus during an unrelated viral infection of the liver. Proc Natl Acad Sci U S A 93, 4589-4594.[]
  9. Thio, C. L., Netski, D. M., Myung, J., Seaberg, E. C., and Thomas, D. L. (2004). Changes in hepatitis B virus DNA levels with acute HIV infection. Clin Infect Dis 38, 1024-1029.[]
April 2nd, 2008

Intravital microscopy: Set “Cool Factor” to “Extreme”

Breart et al Fig 3: Direct action of CTLs on individual tumor cells drives tumor regression
Breart et al, Fig. 3. Direct action of CTLs on
individual tumor cells drives tumor regression

Intravital microscopy — microscopic analysis, in real time, of processes within a living animal — has been used in immunology for maybe a decade now, but it hasn’t lost its cool factor yet. I don’t know that there have been any great intellectual breakthroughs arising from the work, but we have learned a fair bit about, say, interactions between T cells and their targets, and migration patterns, and so on. And of course there’s a huge help in visualization, which undoubtedly helps people understand what’s happening and, hopefully, develop other experimental approaches to test it.

Just as importantly, the “Awesome” factor of these things is absolutely off the scale. I pointed out Uli von Andrian’s collection of intravital videos the other day, and now the latest issue of Journal of Clinical Investigation has a paper from Phillipe Bousso’s group, showing 2-photon microscopy of cytotoxic T lymphocytes attacking a tumor.

The paper is:

Breart, B., Lemaître, F., Celli, S., Bousso, P. (2008). Two-photon imaging of intratumoral CD8+ T cell cytotoxic activity during adoptive T cell therapy in mice. Journal of Clinical Investigation, 118(4), 1390-1397. DOI: 10.1172/JCI34388

They used the EL4/EG7 tumor model in mice. These cells form solid tumors in C57BL/6 mice, and are not rejected by the immune system. The EG7 cells are derived from EL4; they have had a defined antigen introduced, and if you transfer activated T cells against the antigen, the tumor will be rejected. If you transfer naive T cells, and depend on them to be activated by the tumor, you’re out of luck; the tumor is not rejected. 1 They were able to watch all these things happening, in real time.

Here’s what happens with activated CTL (orange) around a tumor site (tumor cells in yellow/greenish). Watch the CTL zipping around merrily in areas where there are no tumor cells, and then screeching to a halt as they identify tumor antigen, engage their targets, and begin to kill:

(Embedded video! I’m so MySpace! I’m going to use tripple exclamation marks and mispell lot’s of words!!!)

The article is free access, I believe, so you should check it out for yourself; there are five videos to watch in the supplemental data. They show CTL engaging tumor cells and actively killing them, using indicators for cell death so they don’t have to guess what’s happening.

I think this is mainly a technical tour de force, and the amount of new information about tumor immunology is relatively small. But there are a couple things of interest. One is that naive T cells — the guys who do not reject the tumor — seem kind of indifferent to the whole thing. It’s not a question of the CTL entering the tumor, and then being turned off (which would have been my guess); rather, the naive cells never even entered the tumor in the first place:

Although CTL infiltration was quite variable in the different regions of the tumor (Figure 6A), EG7 patches were eliminated in CTL-rich areas, which was evidence that in vivo primed CTLs were not grossly impaired in their ability to kill target cells … Thus, the low level of CD8+ T cell infiltration, rather than a defect in the cytotoxic activity, appeared to be responsible for the inefficient response mounted by in vivo primed OT-I T cells.

Another surprising finding — which is so different from previous work in different systems that I’m hesitant to believe it — is the timing of cell killing. Previous studies (such as the von Andrian paper2 that produced this video) have suggested that CTL kill their targets in something under an hour; maybe 30 minutes or even less. Here. Bousso’s group find that the tumor cells take something like 6 hours to be killed. That’s such a large difference — and has such important implications for effectiveness of CTL killing — that, as I say, I’d like to see it confirmed before I take it to the bank.

Bousso’s web site has a bunch of other equally fascinating videos; check them all out.


  1. This is probably related to the ability of tumors to suppress immune responses, which I’ve talked about before.[]
  2. Mempel, T. R., Pittet, M. J., Khazaie, K., Weninger, W., Weissleder, R., von Boehmer, H., and von Andrian, U. H. (2006). Regulatory T cells reversibly suppress cytotoxic T cell function independent of effector differentiation. Immunity 25, 129-141.[]
|