JC? Sort of. CTL assay question, and more.
John Ladasky
I just finished a presentation of the paper "Cytotoxic T cell
memory without antigen" (Lau LL et. al., Nature 369:348, 23 June 1994)
for a face-to-face journal club here at Stanford. The authors claim
that they have demonstrated the existence of a memory T cell population
specific for lymphocytic choriomeningitis virus (LCMV) that can be
adoptively transfered from one host to another. Previous studies have
failed to show long-term CTL memory. Some interesting questions were
left unanswered by our journal club discussion, and I thought I would
bring them here.
1) A technical quibble: the authors claim to have an in vitro CTL
stimulation assay in which "virgin T cells do not generate an anti-
viral response." This comment is unreferenced. It's important
that this is true, because they use limiting dilution assays to
measure the fraction of CTL precursors (presumably memory cells) in
their animals, and there's typically a 100-fold excess of "normal"
CD8+ cells in their assay. I read several papers referenced by this
one, which described CTL stimulation methodology. None of them
said anything about memory cells. Does anyone have any insights?
2) The authors work pretty hard to prove that they have not transfered
LCMV along with their T cells. Yet previous studies, in which whole
spleens were adoptively transfered into mice (thereby increasing the
chance of transfering a few antigen-presenting cells along with the
CTL's), failed to show long-term memory. Another difference in the
earlier studies is that they used a single protein antigen (see e.g.
Gray DD, Matzinger PJ, J. Exp. Med. 174:969 [1991]) rather than a
whole virus. In a review article (Matzinger PJ, Nature, same issue
as the Lau article, pp. 605), it's suggested that "The complex
structures of... LCMV... offer a greater probability of cross-reac-
tions with environmental antigens." Personally, I don't buy this
argument. There are only a couple of peptides from LCMV that are
strongly immunogenic.
When the bionet.immunology journal club was being formed, I
offered (by email) to do a paper in September. I picked this time be-
cause I knew I would have another journal club presentation. I thought
that I could save myself some effort, and get some extra insights, by
presenting the same song-and-dance in two different locations. Perhaps
other people who don't want to see this forum disappear could do some-
thing similar.
O.K., tag! You're it!
--
Unique ID : Ladasky, John Joseph Jr.
Title : BA Biochemistry, U.C. Berkeley, 1989
Location : Stanford University Dept. of Cell Biology, Fairchild D-105
Keywords : immunology, music, running, Green
Shiv A. Prasad (MicroBio)
>
> I just finished a presentation of the paper "Cytotoxic T cell
>memory without antigen" (Lau LL et. al., Nature 369:348, 23 June 1994)
>for a face-to-face journal club here at Stanford. The authors claim
>that they have demonstrated the existence of a memory T cell population
>specific for lymphocytic choriomeningitis virus (LCMV) that can be
>adoptively transfered from one host to another.
>
>1) A technical quibble: the authors claim to have an in vitro CTL
> stimulation assay in which "virgin T cells do not generate an anti-
> viral response." This comment is unreferenced. It's important
> that this is true, because they use limiting dilution assays to
> measure the fraction of CTL precursors (presumably memory cells) in
> their animals, and there's typically a 100-fold excess of "normal"
> CD8+ cells in their assay. I read several papers referenced by this
I can't find my copy of this paper, but here are my thoughts from what I
remember. The frequency of anti-LCMV CTL in a naive animal should be on
the order of 1:100,000 or so, and the frequency that Lau et al. found in
the recipient mice was significantly higher, and remained at a higher
frequency than one would expect in a naive mouse. That would argue against
the memory effect being due to naive anti-LCMV CTL.
>2) The authors work pretty hard to prove that they have not transfered
> LCMV along with their T cells. Yet previous studies, in which whole
> spleens were adoptively transfered into mice (thereby increasing the
> chance of transfering a few antigen-presenting cells along with the
> CTL's), failed to show long-term memory. Another difference in the
[...]
I think that Rafi Ahmed argues that this is a big difference between his
group's work and the Zinkernagel group. The different results may be a
result of the cell types transferred. Perhaps the transfer of APC with
the memory cells allows peripheral deletion of those memory cells.
> whole virus. In a review article (Matzinger PJ, Nature, same issue
> as the Lau article, pp. 605), it's suggested that "The complex
> structures of... LCMV... offer a greater probability of cross-reac-
> tions with environmental antigens." Personally, I don't buy this
> argument. There are only a couple of peptides from LCMV that are
> strongly immunogenic.
I don't buy that environmental x-reactivity either. In fact the accompanying
paper shows that memory persists when donor "memory" cells are transferred
into beta-2 microglobulin knockout mice, so there is less chance of TCR
occupancy on the donor cells, unless they present to themselves.
>
>Unique ID : Ladasky, John Joseph Jr.
Ian A. York
>memory without antigen" (Lau LL et. al., Nature 369:348, 23 June 1994)
>for a face-to-face journal club here at Stanford. The authors claim
>that they have demonstrated the existence of a memory T cell population
>adoptively transfered from one host to another. Previous studies have
Great choice for a JC paper, it covers a number of very interesting issues.
I thought that in case anyone is only peripherally interested in this, I would
try to post a little of the background and context for this paper. (Also, I
haven't yet re-read the papers - the background is as far as I've got.)
Be warned - this is not specifically my field; I'm merely an interested
bystander.
The question that's being addressed in this paper (and it's Nature neighbor) is
the nature of immunological memory. As we all know, exposure to an antigen
induces a number of acute-type responses (acute meaning days to a few weeks
after the initial exposure) including some of the obvious things such as anti-
body production/isotype switch, CTL lysis, etc. On subsequent exposure to the
same antigen, the immune response is much more rapid and much greater. "Sub-
sequent exposure" can mean very subsequent - i.e. years later. This is due to
immunological memory and happens because one of the things the acutely reacting
cells do is to form a population of memory cells.
When a memory response occurs months or years after an initial exposure
to antigen, is this because the same memory T cells are still around, persist-
ing from their intial exposure? - or is it because there has been an intermit-
tent, or continual, trickle of antigen in the intervening period? I had
assumed the former, without a very good reason for it. In the wild, of course,
the experiment is pretty tough to do right: are you immune to measles today
because you had that case of it twenty-five years ago when you were two, or
because you have been exposed to your neighbor's snotty-nosed brats shedding
their measles over you at intervals in the meantime? The problem is also more
complicated than that: even if there is no overt antigen hanging around, there
could be a very long-lived antigen depot - follicular dendritic cells are
frequently mentioned as possible depots. Although I feel uncomfortable
with the idea of cell-surface antigen persisting for months or years, this is
really a matter for my therapist and myself and should not influence the
hypothesis.
The paper that brought this to my attention (it was not the first to
discuss it) was from Zinkernagel's lab: Oehen et al, J Exp Med 176:1273-1281
(1992). They used lymphocytic choriomeningitis virus in mice, a paradigm of
a persistent virus, which induces a very long-lived memory response, and
asked what happened when the continual antigenic reminder - the LCMV - was
removed from the system. They addressed this first by using a different
immunogen - LCMV proteins in a vaccinia vector. Vaccinia doesn't replicate
well in mice and doesn't persist. They found that the vectors induced a short-
lived response, in which no memory was detectable after 30 days. Higher doses
of the vax vector gave longer lived responses (>60 d). However, by doing
adoptive transfer of immune spleen cells to naive recipients, they presumably
got rid of persistent antigen, and they found that the memory response in
the recipients was short lived. They argued that memory cells were an artifact
which were actually continually reactivated cells, constantly meeting new
antigen (which is at least partially consistent with the phenotype of mem-
ory cells: they show a number of activation-type markers.)
Arno Mullbacher (J Exp Med 179:317 - 321 (1994)) begged to differ.
Mice vaccinated with flu were used as donors into irradiated naive mice (I am
simplifying the experiments here, a little bit; he had more controls) and
he looked at the kinetics of the cytolytic response after infection of the
recipients; he argues that lysis 3 days after infection was diagnostic of
a memory reponse (which seems reasonable) and found that by this criterion
a memory response was detectable as much as 25 weeks after transfer - 42
weeks after the initial infection. However, Mullbacher was not really
anal about excluding possible persistent antigen from his transfers; either
(a very low level of) virus could have still been present (not likely,
but possible), or (more plausible) he could have transferred some of the
hypothetical antigen-depot cells along with his splenocytes: he selected
against MHC II expression, but who knows what these hypothetical cells
might have on their hypothetical surfaces.
The papers under discussion arise out of this. There are a couple of
other relevant papers people might be interested in glancing at - one
which I remember seeing but can't turn up - perhaps that issue of J Virol (if
that is where I saw it) is out for binding at the nearest library. This
was a study done on persistence of the immune response to vaccinia. If I
remember it right (and I might not) the population was one which had been
vaccinated for smallpox many years before, and then relatively isolated
from the rest of the world. Of course they hadn't been exposed to smallpox,
and it was unlikely that they had been exposed to vaccinia; but they had
detectable memory repsponses as much as 25 years later (This is all from
memory, and I didn't do much more than skim the article; if someone knows
better, please correct me) - anyway, if so, it argues that a long-lived
memory response doesn't require a continuous source of antigen; it doesn't
prove that antigen depots are not involved, but I really, really feel un-
comfortable with a cell presenting the same antigen for 25 years.
There is also an excellent review by Sprent and Tough: Science
265:1395-1399 (1994) [the Sept. 2 issue]: _Lymphocyte Life-Span and
Memory_ , which nicely discusses the issues in lymphocyte maturation and
lifespan; highly recommended.
Hope some others get involved in this discussion. When I've had a chance to
go through the papers more carefully, in the next couple days, I will try
to discuss John's specific questions.