In vivo primary induction of virus-specific CTL by immunization with 9-mer synthetic peptides

Journal of Immunological Methods, 153 (1992) 193-2{gl

193

© 1992 ElsevierSciencePublishers B.V. All rights reserved 0022-1759/92/$05.00

JIM 06403

In vivo primary inducti,-Jn of virus-specific CTL by immunization

with 9-mer synthetic peptides X i a n z h e n g Z h o u , Louise Berg, U s s a m a M. Abdel Motal and Mikaei J o n d a l Department of Immunology, Karolinska Institute, Box 60400, 104 OI Stockholm, Sweden

(Received 14 February 1992,revised received7 April 1992,accepted 7 April 1992)

A primary cytotoxic T l~nphocyte (CTL) response in vivo requires antigen presentation by cytosolic processing and can not in general be obtained by vaccination with soluble proteins. In the present work we have found that vaccination of mice with pre-processed synthetic peptides, corresponding to endogenous 9-mers orodnce0 in influenza A virus-infected cells, resulted in strong primary CTL responses. The generated CTL efficiently killed virus-infected target cells with preference for viral strains having the identical amino acid sequences to the peptides used for immunization. The optimal conditions for a primary in vivo CTL response was obtained with 100/zg peptide dissolved in incomplete Freund's adjuvant and injected s.c. at the base of tail. Spleen cells which had been primed 7-10 days earlier were rcstimulated for 5 days in vitro, using an optimal low peptide concentration (0.05/~M) and tested against virus-infected and peptide-treated target cells. The peptide-induced CTL were major histocompatibility complex class ! restricted and CD8 positive. Key words: CytotoxicT lymphocyte:InfluenzaA virus: Syntheticpeptide; Immunization

Introduction

Cytotoxic T lymphocytes (CTL) recognize short peptides derived from viral proteins in the cytosol and presented by cell surface major histocompatibility complex (MHC) class I molecules (Townsend et al., 1985, 1986; Maryanski et ai., 1986; Bjorkman et al., 1987). Such virus-specific CTL are an important cffector mechanism in the clearance of viral infections. To generate virus-specific Correspondence to: M. Jondal, Department of Immunology, Karolinska Institute, Box 60400, 104 01 Stockholm,Sweden. Tel.: 46-8-7286687; Fax: 46-8-302258. Abbreciations: CTL, cytotoxicT lymphocytes;1FA, incomplete Freund's adjuvant: LCMV, lymphocyticchoriomeningitis virus.

CTL in vitro, in v~vo priming is generally required. Priming has been done with live viruses (Deres et al., 1989), chemically modified lipopeptides (Deres et al., 1989; Schild et al., 1991) and viral proteins incorporated into immunostimulating complexes (ISCOMS) (Takahashi et al., 1990). More recently it has also been demonstrated that in vivo priming can be achieved with unmodified synthetic peptides (AIchele et ai., 1990; G a e et al., 1991; Kast et al., 1991). These peptides were 15-16 aa long and vaccinated mice were protected against virus challenge with lymphocytic choriomeningitis virus (LCMV) (Schulz et al., 1991) and Sendal virus (Kast et al., 1991). As it has been found that endogenous MHC class I presented peptides from virus-infected cells are shorter, usually 8 - 9 aa long (R6tzschke et at.,

194

1990; F a l l et al., 1991), we have investigated the immunogenicity of such peptides derived from the nucleoprotein (NP) of influenza A virus. We have found that two 9-mer H-2D b binding peptides give a strong primary in vivo CTL response and defined optimal conditions for the generation of peptide- and virus-specific CTL in vitro.

ology, Karolinska lnstitite, Stockholm). Cells were grown in RPMI 1640-5% heat-inactivated fetal calf serum (FCS). For CTL preparation, complete medium was used containing RPMI 1640 plus 10% heat-inactivated FCS, 0.1 mM non-essential aa, 1 mM sodium pyruvate, 5 x 10-s M 2-mercaptoethanol, 2 mM L~glutamine, 50 ~g streptomycin/ml and 100 ~g penicillin/ml.

Materials and methods

hnmunization Mice were immunized by one s.c. injection of 100 p.g each free synthetic peptides dissolved in IFA at the base of tail or by one intravenous injection of 20 HAU of influenza A virus diluted in PBS and used between 1-2 weeks after the immunization.

Mice Inbred C57B6/J female mice (H-2 h) were used at an age of 8-12 weeks. I//ruse$

Influenza A virus strains A / P R / 8 / 3 4 (PR8) and A / N T / 6 0 / 6 8 (NT60) were a gift from Dr. A. Douglas (National Institute for Medical Research, London). Virus was grown in the allantoic cavity of embryonated chicken eggs and used as ailantoic fluid for target cell sensitization and preparation of infected stimulator cells. Peptides The peptide ASNENMETM (designated as pep 9(PR8)) and ASNENMDAM (designated as pep 9(NT60)) are from the NP366.374 of influenza A virus strains A / P R / 8 / 3 4 and A / N T / 00/68, respectively. Pep 9(PR8) was synthesized using an Applied Biosystems 430A peptide synthesizer (Applied Biosystems, Foster City, CA). Pep 9(NT60) was synthesized by the solid phase 'tea bag' (Houghten, 1985; S~illbcrg, 1991). All peptides were purified by reverse-phase high performance liquid chromatography (HPLC) and analysed for amino acid composition by plasma desorption mass spectrometry. Stock solutions of peptides were prepared in phosphate-buffered saline (PBS) and stored at -20°C.

In vitro generation of influenza A virus-specific CTL Immune spleen cells were prepared into cell suspension. The red cells were lysed using lysis buffer consisting of NH4Ci 8.29 g, KHCO 3 1.0 g, EDTA 0.0372 g per 1 liter destilled water (pH 7.4). 25 × 10~ responder spleen cells from in vivo primed mice were cocuitured with 25 × 106 irradiated (2000 rad) syngeneic spleen cells, either infected by influenza A virus or in the presence of 0.05, 0.5 and 5 ~M peptides in 50 ml tissue culture flask (Costar, Cambridge, MA) with 10 ml complete medium for 5 days at 37°C in humidified air with 5.3% CO 2.

toma) and L929 (H-2 k, fibroblast) were pur-

Cytotoxicity assay As previously described (Jondal, 1975), virusinfected or peptide-coated target cells were prepared by infecting 2 × 10~ cells with 320 HAU virus for 1.5 h or incubated with 50/zM peptides for 1.5 h at 37°C, respectively. After washing, the cells were labelled with 100-200/zCi of Na~lCrO4 for 1 h. The cells were washed twice, and 104 tar6et cells in 50 ~.i were added to 100 p.I of varying numbers of effector cells which had been washed three times prior to the assay in 150 ~1 of complete medium in 96-well V-bottomed plates

chased from ATCC (American Type Culture Collection). RMA (H-2b), a mouse iymphoma cell line r.:ansformed with Rauscher virus (Ljunggren and l~rre, 1985), was a gift from Drs. K. K~irre and 7,-I.-G. Ljunggren (Department of Tumor Bi-

and were incubated for 4 h at 37°C, 5.3~ CO,. After incubation, the 50/~1 supernatant was collected and the percentage of specific 5~Cr release was calculated by the formula: % specific release = (experimental-spontaneous)/(maximal-sponta-

Ceil lines and culture medium E L A (I-'I-~'b, t h y m o m a ) .

P g l 5 (H-'~ d. m a s t o c y -

neous) cpm. Spontaneous release was always less than 15%.

Cell separation Effector cells were separated using the Dynabeads system (Dynal, Oslo, Norway). Briefly, 107 effector T cells were incubated with anti-Lyt 2 (53.6.72) or anti-L3T4 (GK 1.5) antibodies for 30 rain at 4°C. Then the cells were washed twice and 4 x 10 ~ super-paramagnetic Dynabeads coupled with sheep anti-rat IgG antibody were added. After 60 rain incubation at 4°C, cells rosetted with antibody-coated Dynabeads were removed by applying a magnetic concentrator MPC-I (Dynal, Oslo, Norway) on the outer wall of the test tube. The cells remaining in solution were collected and used as negatively depleted effector cells in the 5tCr-release assay.

Results

Primary in rive induction of anti-influenza A virus-specific CTL with 9-mers As previously reported (Townsend et" al., 1985, 1986), CTL induced with live influenza A virus in H-2 b mice are mainly directed against the immunodominant NP365-380 epitope. We initially tried to immunize mice with peptide NP365-380 derived from both PR8 and NT60 strains to induce CTL responses according to a protocol described by Aichele et al. (1990). However, an insignificant level of killing activity against both peptide-coated and virus-infected cells was observed in spite of repeated boosting. Instead, when 100/zg of the shorter pep 9(PR8) was used for s.c. injection and spleen cells restimulated with 5 p.M pep 9(PR8) in vitro, a strong CTL response against pep 9(PR8)-coated R M A (not shown) and EL4 cells was elicited although virusinfected targ~g ¢¢11~ were not killed (Table I). W h e n a lower concentration of pep 9(PR8) was used

for

restimulation

in

vitro,

the

generated

CTL had a higher killing activity against pep 9(PRg)-coated target cells and virus-infected cells were also lysed (Table I). As shown in Table 1,

the optimald0~cfor immunizationwas 100/~g and for restimulation 0.05 / t g / m l (0.05 IzM). Immunization with 10 ttg and 1 ttg generated less

TABLE I CT£ ACTIVITY INDUCED BY DIFFERENT DOSES OF PEPTIDES FOR PRIMING IN VIVO AND RESTIMULATION IN VITRO Priming Restim- Untreated dose ulation 40:1" 20:1 (/.tg) dose (p,M)

EL4targetlysis(%) PRSpepggPR8)infected coated 40: I 20: I 40: I 20: I

I00

7.0 30.8 15.0 31.2 6.9 12.4

5.00 0.05 5.00 0,05 5.00 0.05

10 I

4.6 0.1 5.6 5.7 3.3 4.1

4.2 0.5 4.2 4.1 2.7 3.3

8.5 30.1 14.0 23.8 5.5 10.6

43.7 59.4 23.5 45.5 I 1.9 16.4

34.7 50.7 22.1 34.6 9.3 15.5

Effector:target ratio.

or even no CTL response. Peptide dissolved in complete Freund's adjuvant (CFA) or PBS gave no response. Intravenous injection of peptides did not generate a primary in vivo C r L response either (not shown). Another 9-mer derived from NT60 gave stronger anti-peptide and anti-viral CTL responses, as shown in Fig. 1.

60

m

c') CX

,~ 3O"

~o io. o

,

.

.

o--.o....., o . . . . 4z~l

~ l

.

o---o---o .

jo~l

Fig. !. CTL induced hy pep 9(NT60) more efficiently lysed

both virus-infected and F l a l i d c e t c d

pM pep 9(NT60) used for restimulation than with 5 FM and 0.5 pM. CTL wece generated by one s.c. injection of 100/~g pep 9(NT60). Immune spleen cells were restimulated with

syngeneic irradiated spleen cells in the presence of pep 9(NT60) (A), 5 p-M: (B), 0.5 ~tM and (C), 0.05 P-M for 5 days and tested against EIA untreated (o o), NT60-infected (o e), pep 9(NT60bcoated ( ra ra ).

196

Specificity of CTL induced by pep 9(PR8) and pep 9(NT60) in the light of these results we investigated the fine specificity oi' the generated CTL in relation to pep 9(PR8) and pep 9(NT60). CTL preferentially killed their corresponding peptide-coated and virus-infected cells (Fig. 2). CTL showed a higher cross-reactivity with peptide-coated cells than with virus-infected target cells.

in vitro restimulation wilh PRB

in vitro ratimulation with Ixp 9(Pl~)

I

CTL responses to live t:irus or peptides Mice were primed with live PR8 virus and immune spleen cells were restimulated with either virus-infected stimulator cells (Fig. 3A) or normal syngeneic spleen cells in the presence of 0.05 g M pep 9(PR8) (Fig. 3B). CTL primed in vivo with live virus and restimulated with low concentration of pep 9(PR8) had a remarkable ability to recognize both virus-infected and peptide-coated cells (Fig. 3B). It should be noted that the low amotmt of res~imulation peptide was sufficient to stimulate virus primed CTL in vitro and had the same potency as virus-infected cells did. When pep 9(PR8) was used for in vivo priming and virus-infected cells or pep 9(PR8) for restimulation in vitro (Figs. 3C and 3D), CTL responses were generated against both peptide-coated and virus-infected cells. Live virus had a higher efficiency to prime CTL than the free peptide.

. 40:1

. 20:1

. I01

.

21~1

40:1 20:1 10:1 $:1

Effeclor : target ratio Fig. 3. CTL responses primed and restiraulated with live virus and peptides. Mice were primed in vivo with PR8 live virus (A, B) or pep 0(PR8) (C, D), and restiraulated with Pg8-infected spleen cells (A, C) or irradiated spleen cells in the presence of 0.05 tzM pep 9(PR8) (B, D). Target cells were EL4 untreated (o o), PR8-infected (e e), pep 9(PR8)-coated ( n [7).

Time course of CTL activity after immunization with pep 9(PR8)

In vivo peptide-induced cytotoxicity is mediated by CD8 + and MHC class l-restricted T cells

. 40:1

40:1 20:1 10:1 5:1

IO:l

~o. o----r~

o---o._o._o

:1 . . . . . . . .

-A

I

|

"[

with pep 9(PR8)

In order to determine the kinetics of CTL activity after immunization with pep 9(PR8), mice were primed once with pep 9(PR8). Spleen cells of immunized animals were restimulated in vitro 2 - 3 0 days after immunization with pep 9(PR8) and assayed for CTL activity. The CTL activity against pep 9(PR8)-coated R M A target cells reached a peak 7 days after the immunization and the activity gradually declined afterwards (Fig. 4). At the day 30 after priming, a lower level of CTL activity could still be detected.

03)

(A)

in vivo priming

Effeclor : target ratio Fig. 2. Specificity of CTL induced by pep 9(PR8) ( A ) and pep

9(NT00) (B). The effector CTL were tested against EL4 untreated (o o), PRg-infected (e s), NT-60-infected (ra 1:3). pep 9(PR8)-coated (11 I1), pep 9(NT60)-coated ( A A ).

As shown in Fig. 5a, pep 9(PR8)-induced CTL depleted for CD8 + T cells failed to lyse pep 9(PR8)-coated EL4 ceils. On the other hand, depletion of CD4 + T ce|ls did not affect the cytolytic activity against pep 9(PR8)-coated target

80'

RMA

70.

4030~)I0

cells. Thus, p e p 9(PR8)-specific C T L express the C D 4 - C D 8 + phenotype. In rive p e p t i d e - p r i m e d C T L o f C 5 7 B 6 / J (H2 b) origin were tested o n syngeneic EL4 (H-2h), ailogeneic P815 (H-2 d) and L929 (H-2 k) target cells infected with influenza P R 8 strain. As Fig. 5b shows, there is a clear restriction specificity for H-2 b target cells.

"

Discussion 2 7 IO 20 3O Days after priming in rive with pep 9(PRS) Fig. 4. Kinetics of CTL activity induced by one s.c. injection of pep 9(PR8). Mice were primed ~vith 100/~g pep o~PR8) in rive and restimulated in vitro with 5/~M pep 9(PR8) at day 2, 7, 10, 20 and 30 after priming. The generated CTL were tested against RMA untreated (o • o) and pep 9(PR8)coated (e e). Effector:target ratio, 60: I. a 50

o

. ~ ,

4~1 ~l I0:1 Effector : target ratio

Fig. 5. Peptide-induced cytotoxlcityis mediated by CD8 + and H-2Db restricted T cells, a: CTL were induced by pep 9(PR8). Equal numbers of CTL ~vere then depleted for CD4 + and CD8 + T cells using the Dynabeads system. The remaining cells were then tested for lytic activity against pep 9(PRS)coated ELA cells. Untreated CTL (o o). CD4 + depleted ( e - - - - e ) and CD8 + depleted (ra ~) were tested for lytic activity against pep 9(PR8)-coated EL4 cells, b: CTL induced by 10ep9(PR8) were tested against PRS-infected EL4 (~ O), PR8-infected P815 ( o ~ o ) and PR8infected L929 ( ra [] ).

We have found that immunization with two D b binding 9-mer synthetic p e p t i d e s corresponding to endogenously processed p e p t i d e s in virus-infected cells generate primary C T L responses which are M H C restricted, m e d i a t e d by C D $ + T cells and are virus-specific. Peptides were injected at the base of tail, emulsified in I F A and primary C T L responses were found in local lymph nodes and spleen at day 10. This is a simple immunization p r o c e d u r e which d o e s not generate a C T L response with whole protein material as p r o t e i r antigens must have access to the cytosol for proteolytic cleavage and further processing along the class I antigen presentation pathway (Moore et al., 1988). Only if the protein is pres e n t e d in ISCOMS (Takahashi et al., 1990, H e e g et al., 1991) or liposomes (Lopes and Chain, 1992), it enters the cytosol and induces C T L responses. Earlier work with 14-16 aa long peptides has also g e n e r a t e d primary C T L responses, either with modified lipopeptides, o r with free peptides by a similar immunization schedule as ours (Der e s e t al., 1989; Kast et al., 1991; Schild et al., 1991). We have optimized conditions for the induction o f virus-specific C T L responses by p e p tide vaccination (Table II) and e x t e n d e d these earlier findings to show that also short 9-mer peptides with the optimal length for strong D h binding, are immunogenic. In fact, in our hands, 16 aa long peptide variants from the same epitope as the 9-mer peptides, were not immunogenic in vivo, suggesting to us that an optimal peptide's length is important for their immunogeniciq,. We believe that the reason why the synthetic 9-mer peptides have such high immunogenicity in

198 TABLE 11 CONDITIONS FOR THE INDUCTION OF VIRUSSPECIFIC CTL BY PEPTIDE VACCINATION (I) Optimal length of peptide, similar to endogenouslyproduccd peptides in vires-infectedcells (2) Vaccination with 100/.tg in IFA by s.c. route at the base of tail (3) Restimulation ia vitro of spleen cells at day 10, using an optimal low dose (0.05 p.M) for 5 days

vivo is that they are 'pre-processed' to have an optimal fit into the antigen binding groove of empt~ D h molecules expressed on the surface of antigen presenting cells (Ljunggren et al., 1990; De Bruijn et al., 1991; R6tzschke and Falk, 1991). The emulsification in IFA may permit a slow release from the injection site to generate a continuously high membrane expression of class Ipeptide complexes. The optimal peptide's length has been found to be important for the long term stable expression of such complexes (Cerundolo et al., 1991). For in vitro restimulation of CTL, a lower poptide concentration was found to be better than higher concentrations. One possible explanation for this may be that high peptide concentrations induce a state of anergy in the CTL population, whereas low concentration allows restimulation initially, whereafter most peptide is consumed and the restimulated CTL left in an active state. Two more important parameters to be considered are the injection site and the adjuvant used. We injected peptides at the base of tail, which is a commonly used site for immunization, however, other sites may also be used. Schirrmacher et al. (1991) have found that the intra-pinna (intrapinna injection) is an optimal site to generate a primary CTL response against low immunogenic syngeneic tumors. In addition, we found that IFA was better than CFA for peptide immunizations in agreement with other reports (Alchele et al., 1990; Kast et al., 1991). This may be explained in terms of peptide stability. The generation of a ~trong local inflammation process may lead to the

proteolytic degradation of peptides and thus reduce immunogenicity. Also, with CFA, there may be a component of antigenic competition which negates the immunogenicity of the peptides. The fact that 9-mer peptides, which only bind to MHC class 1 in H-2 b mice, do not raise any helper T cell activity or antibody production (not shown), and generate CTL responses, demonstrates the autonomous character of this response. More recently it has also been reported that CTL can be generated without helper T cell activation in CD4 'knockout' mice (Rahemtulla et al., 1991). We found a peak of in vitro restimulated CTL at day 7-10, and at day 30 a lower response after the priming. The duration of the peptide generated CTL response is presently unclear, nevertheless published data suggests that the CD4 ÷ helper T cells are needed to generate a long term memory (Sanders et al., 1988; Grey et al., 1991). What could be the use of short 9-mer peptides as immunogens? In vaccines against infectious agents, a mixture of peptides that bind broadly to MHC class I may amplify CTL responses, espe~ ciaily if combined with helper peptides. Given that peptide vaccination can be optimized to include generation of memory, it may also be useful in the generation of tumor vaccines. For instance, if a certain tumor expresses MHC class I molecules, the precise dominating bound peptides can be isolated and sequenced using immunoprecipiration, acid elution, HPLC separation (R6tzschke et al., 1990, 1991; Falk et al., 1991) and Edman degradation. An individualized anti-tumor vaccine may then be constructed on the basis of the most stably expressed peptides. Although these ideas are presently hypothetical, suitable mouse tumor model systems are available for experimental testing (R6tzschke and Falk, 1991; Van den Eynde, 1991; Boon, 1992).

Acknowledgements This work was supported by Swedish Cancer Society and the Astra Company in Sweden. We would like to thank Ms. Anna-Lena Hammai'~n for growing influenza A virus, Dr. Abdur Rehman Siddiqi for reading the manuscript.

199

References Alchele, P., Hengartner. H., Zinkernagel, R,M. and Schulz, M. (1990) Anfiviral cytotoxic T cell response induced by in vivo priming with a free synthetic peptide. J. Exp. Med. 171, 1815. Bjorkman, P.J., Saper, M.A., Samroui, B., Bennett, W.S., Strominger, J.L. and Wiley, D.C. (1987) Structure of human class I histocompatibility antigen. HLA-A2. Nature 329, 506. Boon, T. (1992) Towards a genetic analysis of tumor rejection antigens. Adv. Cancer Res. 148, i445. Cerondolo, V., EIliott, T.. EIvin. J., Bastin, J., Rammensee, H.-G. and Towasend, A. (1991) The binding affinity and dissociation rates of peptides for class I major histocompatibility complex molecules. Ear. J. Immunol. 21, 2069. De Bruijn, M.L.H., Schumacher, T.N.M., Nieland, J., Ploegh, H.L., Kast, W.M. and Melief, C.J.M. (1991) Peptide loading of empty major histocompatibility complex molecules on RMA-S cells allows the induction of primary cytotoxic T lymphocyte responses. Eur. J. lmmunol. 21, 2963. Deres, K., Schild, H., Wiesmiiller, K.-H., Jung, G. and Rammensee, H.-G. (1989) In vivo priming of virus-specific cytotoxic T lymphocytes with synthetic lipopeptide vaccine. Nature 342, 561. F a l l K., R6tzschke, O., Stevanovic, S., Jung, G. and Rammensee, H.-G. (1991) Allele-specific motifs revealed by sequencing of self-peptides eluted from MHC molecules. Nature 351,290. Gao, X.-M., Zheng, B., Liew, F.Y., Brett, S. and Tite, J. (1991) Priming of influenza virus-specific cytoioxic T lymphocytes vivo by short synthetic peptides. J. lmmunol, 147, 3268. Gray, D. and Matzinger, P. (1991) T cell memory is short-lived in the absence of antigen. J. Exp. Med. 174, 969. Heeg, K., Kuon, W. and Wagner, H, (1991) Vaccination of class I major histocompatibility complex (MHC)-restricted routine CD8 + cytotoxie T lymphocytes towards soluble antigens: immunostimulating-ovalbumin complexes enter the class ! MHC-restricted antigen pathways and allow sensitization against the immunodominant peptide. Eur. J. lmmunol. 21, 1523. Houghten, R.A. (1985) General method for the rapid solidphase synthesis of large numbers of peptides: Specificity of antigen-antibody interaction at the level of individual amino acids. Proc. Natl. Acad. Sci. USA 82, 5131. Jondal, M. and Pross, H. (1975) Surface markers on human B and T lymphocytes. V1, Cytotoxicity against cell lines as a functional markers of lymphocyte-suhpopoulations. Int. J. Cancer. 15, 596. Kast, W.M., Roux, L., Curren, J., Blom, H.J.J., Voordouw, A.C., Metoen, R.H., Kolakofsky, D. and Melief, C.J.M. (1991) Protection against lethal Sendal virus infection by in vivo priming of virus-specific c3,totoxic T lymphocytes with a free synthetic peptide. Proc. Natl. Acad. Sci. USA 88, 2283. Ljunggren, H.-G. and K/irre, K. (1985) Host resistance di-

rected ~lcctively against H-2-deficient lymphoma variants; Analysis of the mechanism. J. Exp. Med. 162,1745. Ljunggren, H.-G., Stare. N.J., (3high, C., NeefIes. JJ., H6glund, P., Heemels, M.-T., Bastin, J., Schumacher, T.N.M., Town~nd. A.. K~ffe, K. and Plocgh, H.L (1990) Empty MHC cla~ I molecules come out in the cold. Nature 346, 47fi. Lopes, L.M. and Chain, B.M. (1992) Liposome-mediated delivery stimulates a cla~s I-restricted cylotoxic T cell response to .soluble antigen. Eur. J. immunol. 22, 287 Maryanski, J.L., Pala, P., Corradin, G., Jordan, B.R. and Ccrotfini, J.C. (1986) H-2 restricted cyloloxic T cells can recognize a synthetic HLA peptide. Nature 324, 578. Moore, M.W.. Carbone, F.R. and Bevan, MJ. (1988) Introduction of soluble protein into the class I pathway of antigen processing and presentation. Cell 54, 777. Rahemtulla, A., Fung-Leung, W.P., Schilham, M.W.. Kilndig, T.M., Sambhara. S.R.. Narendrun, A., Arabian, A., Wakeham, A., Paige, C.J., Zinkernagel, R.M., Miller, R.G. and Mak, T.W." (1991) Normal development and function of CD8 ÷ cells but markedly decreased helper cell activity in mice lacking CD4. Nature 353, 180. R6tz~hke, O. and F a l l K. (1991) Naturally-occurring pepfide antigens derived from the MHC class-l-restricted processing pathway, lmmunol. Today. 12, 447. R6tzschke, O., F a l l K., Dcres, H., Schild, H.. Norda, M., Metzger, J., Jung, G. and Rammensee, H.-G. (Ig00) Isolation and analysis of naturally processed viral peptides as recognized by cytotoxic T cells. Nature 348, 252. R6tzschke, O, F a l l K., Stevanovie, S., Jung, G., Walden, P. and Rammen~e, H.-G. (1991) Exact prediction of a natural T cell epitope. Eur. J. lmmunol. 21, 2891. Sanders, M.E., Makgoba, M.W. and Shaw, S. (1958) Human naive and memory T cells: reinterpretation of helper-inducer and suppress or inducer subsets. Immunol. Today 9, 203. Schild, H., Deres, K., Wiesmiiler, K.-H., Jung, G. and Rammensee, H,-G. (1991) Efficiency of peptides and lipopeptides for in vivo priming of virus-specific cytotoxic T cells. Eur. J. lmmunol. 21, 2649. Schirrmacher, V., Leidig, S. and Griesbach, A. (1991) In situ activation of syngeneic tumour-specific cytotoxic T lymphocytes: lntra-pinna immunization followed by restimulation in the peritoneal cavity. Cancer lmmunol. Immunother. 33, 299. Schulz, M., ZinkeroageL R.M. and Hengartner, H. (1991) Peptide-induced antiviral protection by cytotoxie T cells. Proc. Natl. Acad. Sci. USA 88, 991. S/illberg, M., Ruden. U., Magnius, L.O., Norrby, E. and Wahren, B. (1991) Rapid "tea bag" peptide synthesis using 9-fluorenylmethoxycarhoayl (Fmoc) protected amino acids applied for antigenic mapping of viral proteins, lmmue.ol. Left. 30, 59. Takahashi. H.. Takeshlta, T., Morein, B.. Putney, S., Getmain, R.N. and Berzofsky, J.A. (I,°90) Induction of CD8 + cytotoxic T cells by immunization with purified HIV-I envelope protein in ISCOMs. Nature 344, 873. Townsend. A.R.M., Gotch, F.M. and Davcy, J. (1985) Cyto-

200 toxic T cells recognize fragments of influenza nuclcoprorein. Cell 42, 457. Townsend. A.R.M., Rothbard, J., Gotch, F.M., Bahadur, G., Wraith, D. and McMichael, A.J. (1986) The epitopes of influenza nuclcoprotein recognized by cytotoxic T lymphocytes can be defined with short synthetic peptides. Cell 44, 959.

Van den Eynde, B., Lethe, B., Vanpel, A., Deplaen, E. and Boon, T. (1991) The gene coding for a major tumor rejection antigen of tumor P815 is identical to the normal gene of syngenic DBA/2 mice. J. Exp. Med. 173, 1373.