nature immunology • volume 2 no 2 • february 2001 • http://immunol.nature.com

CORRESPONDENCE

136

In a recent article, Dorfman et al.1 investigatedwhether basal T cell receptor (TCR)ζ chainphosphorylation depends on TCR interactionwith self–major histocompatibility complex(MHC) molecules. Although these data areinteresting in the context of peripheral T cellsurvival, the authors concluded differently andended up stating that lymphocyte survival isnot dependent on self-MHC expression. Thiswas because in their experiments donor CD4+

T cell numbers declined in both normal andclass II MHC–deficient (Aβ-KO) recipients.These conclusions are in marked contrast tothose of previous reports2–7, which addressedthe same question but in different experimen-tal systems. However, the inability ofDorfman et al.1 to determine an influence ofperipheral self-MHC on CD4+ T cell survivalmight be explained by the inappropriateexperimental system they used.

Peripheral T cell pool sizes are limited by, as

yet, unknown factors (ref. 8 andreferences therein). Dorfman etal. compared lymphocyte sur-vival after transfer into normaland Aβ-KO mice. This is notappropriate because in the lattercase donor CD4+ T cells do nothave to compete with recipientcells, whereas in the formercase they have to compete witha full preformed pool of hostCD4+ cells. In addition, the con-tinuous thymic output of CD4+

T cells in normal mice (but notAβ-KO mice) will graduallyreplace peripheral CD4+ T cells,

including those derived from the donor, as isthe case for CD8+ T cells9. For these reasons,normal mice cannot be used as controls forexperiments with Aβ-KO mice. It is like com-paring apples to oranges.

The authors observed proliferation of donorCD4+ T cells in Aβ-KO recipients. Proliferationceased once the mice were infused with anti-bodies to class II MHC, which indicated thepresence of residual class II MHC on host- ordonor-derived cells. Regardless of the cell type,the antibody-blocking studies were performedto exclude the contact of CD4+ T cells with(residual) self-MHC. Although the antibodieswere sufficient to block proliferation of donorCD4+ T cells in Aβ-KO recipients, we do notknow whether the dose was adequate to blockthe contact to self-MHC that is required forCD4+ T cell survival. This is because prolifera-tion could well require stronger stimuli thanthose needed for survival10 and, thus, would

result in a slower decline of CD4+ T cells. Here,it is irrelevant that the authors observed animmediate down-regulation of TCRζ chainphosphorylation when blocking self-MHCbecause we do not know how much phosphory-lation is required for survival. The differingkinetics between T cell survival in the absenceof self-MHC (t1/2 of approximately threeweeks2,5,7, J. Kirberg, unpublished data) anddiminished TCRζ chain phosphorylation (with-in 36 h), as seen by Dorfman et al.1, might sug-gest that this is actually the case.

In summary, the similar decline of CD4+ Tcells in the different mice may occur for verydifferent reasons in the experiments byDorfman et al.1: competition and replacementin normal recipients and death due to theabsence of TCR contact with self-MHC mol-ecules in Aβ-KO recipients. We find no basisto interpret the data as Dorfman et al.1 have,questioning the concept of TCR “tickling” byself-MHC as a prerequisite for peripheral survival. Finally, basic TCRζ chain phospho-rylation might still mediate survival down-stream of the TCR “tickling” but this cannotbe concluded from the data either.

1. Dorfman, J. R., Stefanová, I.,Yasutomo, K. & Germain, R. N. NatureImmunol. 1, 329–335 (2000).

2. Takeda, S., Rodewald, H. R.,Arakawa, H., Bluethmann, H. & Shimizu,T. Immunity 5, 217–228 (1996).

3. Kirberg, J., Berns,A. & von Boehmer, H. J. Exp. Med. 186,1269–1275 (1997).

4. Brocker,T. J. Exp. Med. 186, 1223–1232 (1997).5. Rooke, R.,Waltzinger, C., Benoist, C. & Mathis, D. Immunity 7,

123–134 (1997).6. Beutner, U. & MacDonald, H. R. Int. Immunol. 10, 305–310 (1998).7. Witherden, D. et al. J. Exp. Med. 191, 355–364 (2000).8. Berzins, S. P., Boyd, R. L. & Miller, J. F. J. Exp. Med. 187, 1839–1848

(1998).9. Tanchot, C. & Rocha, B. J. Exp. Med. 186, 1099–10106 (1997).10. Seddon, B., Legname, G.,Tomlinson, P. & Zamoyska, R. Science 290,

127–131 (2000).

Class II essential forCD4 survivalJÖRG KIRBERG1,HARALD VON BOEHMER2,THOMAS BROCKER3,HANS-REIMER RODEWALD4 AND SHUNICHI TAKEDA5

1Basel Institute for Immunology, Grenzacherstrasse 487, CH-4005 Basel,Switzerland. (kirberg@bii.ch) 2Dana-Farber Cancer Institute, 44 BinneyStreet, Boston, MA 02115, USA. 3University of Munich, Institute forImmunology, Goethestrasse 31, D-80336 Munich, Germany. 4UlmUniversity, Dept. for Immunology,Albert-Einstein-Allee 11, D-89081 Ulm,Germany. 5Kyoto University Medical School, Department of RadiationBiology,Yoshida Konoe, Sakyo-ku, JP-606-8501 Kyoto, Japan.

As Kirberg et al. and we1 point out, several pastreports2–8 concluded that self-MHC moleculerecognition is required for normal survival ofmature näive T cells. Believing this model to becorrect, we undertook our biochemical analysesin order to understand the signaling processesthat underlie this effect. Our survival experi-ments were conducted as controls that weremeant to verify that the CD4+ T cells we hadanalyzed showed the expected reduction inlifespan after MHC class II was eliminated

from their environment. Onlyupon obtaining the unexpectedresults reported in our paper didwe begin to question our under-lying assumptions on this issue.We then put a substantialamount of effort into consider-ing possible artifactual explana-tions, including the ones raisedby Kirberg et al.

Kirberg et al. maintain that our experimentalapproach led us to miss the survival effect ofMHC recognition they believe many groupshave documented. However, this perspectivedoes not incorporate new data reported afterpapers on “survival” first appeared, which dealwith the role of MHC recognition in the“homeostatic” proliferation of naïve T cells inlymphopenic hosts9–17. Naïve T cell populationsproliferate in T cell–deficient (that is, lym-phopenic or “empty”) environments. This pro-

liferation requires the presence of the selectingMHC class II molecule for naïve CD4+ cells andthe selecting MHC class I molecule for naïveCD8+ T cells. All the publications claiming thatnormal CD4+ T cell survival depends on MHCclass II recognition involve conditions in whichthe cells in the selecting MHC+ environment arelikely to proliferate much more extensively thanin the nonselecting or putatively MHC– host. Incontrast, the reports that show no effect of MHCexpression on CD4+ T cell survival for at least 1month1,18 use mice in which such proliferation isminimized. Thus, rather than assessing only therelative rates of cell death in nondividing cellpopulations, the experiments in papers that arguefor MHC-dependent survival actually comparethe combined effect of death and proliferation inMHC+ hosts to death alone in the MHC– hosts.

The data from many published experimentsby other laboratories are in accord with thisinterpretation: proliferation was seen in each

ResponseJEFFREY R.DORFMAN1, IRENA STEFANOVÁ1,KOJIYASUTOMO2

AND RONALD N.GERMAIN1

1Lymphocyte Biology Section, Laboratory of Immunology, Building 10 Room11N311,National Institutes of Health, Bethesda,MD 20892-1892,USA.(ronald_germain@nih.gov) 2Department of Pediatrics, School of Medicine,University of Tokushima,Kuramoto 3,Tokushima 770-8853, Japan.

©20

01 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://im

mu

no

l.nat

ure

.co

m© 2001 Nature Publishing Group http://immunol.nature.com